7th International Scientific Conference
on Defensive Technologies
Mileva Mari (1875  1948)
PROCEEDINGS
ISNB 9788681123829
Belgrade, 67 October 2016
MILITARY TECHNICAL INSTITUTE
Belgrade, Serbia
Publisher
The Military Technical Institute
Ratka Resanovia 1, 11030 Belgrade
Publisher's Representative
Col Assistant Prof. Zoran Raji, PhD (Eng)
Editor
Miodrag Lisov
Technical Editing
Dragan Kneevi
Liljana Kojiin
Printing
300 copies
CIP 
,
623.4/.7(082)(0.034.2)
66.017/.018:623(082)(0.034.2)
INTERNATIONAL Scientific Conference on
Defensive Technologies (7th ; 2016 ; Beograd)
Proceedings [Elektronski izvor] / 7th
International Scientific Conference on
Defensive Technologies, OTEH 2016, Belgrade,
0607 October 2016 ; organized by Military
Technical Institute, Belgrade ; [editor Miodrag
Lisov].  Belgrade : The Military
Technical Institute, 2016 (Beograd : The
Military Technical Institute).  1
elektronski optiki disk (CDROM) ; 12 cm
Sistemski zahtevi: Nisu navedeni.  Nasl. sa
naslovne strane dokumenta.  Tira 300. Bibliografija uz svaki rad.
ISBN 9788681123829
1. The Military Technical Institute
(Belgrade)
a)  b)

COBISS.SRID
7th INTERNATIONAL SCIENTIFIC CONFERENCE
ON DEFENSIVE TECHNOLOGIES
SUPPORTED BY
Ministry of Defence
www.mod.gov.rs
Organized by
MILITARY TECHNICAL INSTITUTE
1 Ratka Resanovia St., Belgrade 11000, SERBIA
www.vti.mod.gov.rs
ORGANIZING COMMITTEE
SECRETARIAT
Nenad Miloradovi, PhD, Assistant Minister for
Material Resources, Serbia, President
Major General Bojan Zrni, PhD, Head of Department
for Defence Technologies, Serbia
Major General Duan Stojanovi, Head of Department
for Planning and Development, Serbia
Brigadier General Slobodan Joksimovi, Head of
Department for Strategic Planning, Serbia
Major General Goran Zekovi, Head of the Military
Academy, Serbia
Major General Mladen Vuruna, PhD, Rector of the
University of Defence, Serbia
Vladimir Bumbairevi, PhD, Rector of the University
of Belgrade, Serbia
Branko Bugarski, PhD, Assistant Minister of the
Ministry of Education, Science and Technological
Development, Serbia
Radivoje Mitrovi, PhD, Dean of the Faculty of
Mechanical Engineering, Belgrade, Serbia
Zoran Jovanovi, PhD, Dean of the Faculty of Electrical
Engineering, Belgrade, Serbia
ore Janakovi, PhD, Dean of the Faculty of
Technology and Metallurgy, Belgrade, Serbia
Ivica Radovi, PhD, Dean of the Faculty of Security
Studies, Belgrade, Serbia
Mladen Bajagi, PhD, Dean of the Police Academy,
Belgrade, Serbia
Col. Zoran Raji, PhD, Director of the Military
Technical Institute, Serbia, Vice President
Col. Slobodan Ili, PhD, Director of the Technical Test
Centre, Serbia
Col. Stevan Radoji, PhD, Head of the Military
Geographical Institute, Serbia
Jugoslav Petkovi, JUGOIMPORT  SDPR, Belgrade,
Serbia
Mladen Petkovi, Director of "Kruik", Valjevo, Serbia
Zoran Stefanovi, Director of "Sloboda", aak, Serbia
Rado Milovanovi, Director of "Milan Blagojevi",
Luani, Serbia
Dobrosav Andri, Director of "Prvi Partizan", Uice,
Serbia
Stanoje Bioanin, Director of "Prva Iskranamenska",
Bari, Serbia
Milojko Brzakovi, Director of "Zastava oruje",
Kragujevac, Serbia
Marija Samardi, PhD, secretary
Mirjana Nikoli, MSc
Miodrag Ivanievi, MSc
Dragan Kneevi
Jelena Pavlovi
Liljana Kojiin
IV
SCIENTIFIC COMMITTEE
Miodrag Lisov, MSc, Military Technical Institute,
Serbia, President
Dragoljub Vuji, PhD, Military Technical Institute,
Serbia
Nafiz Alemdaroglu, PhD, Middle East Technical
University, Ankara, Turkey
Major General Nikola Gelao, Director of Military Centre
of Strategic Studies, Roma, Italy
Col. Zbyek Korecki, PhD, University of Defence, Brno,
Czech Republic
Evgeny Sudov, PhD, R&D Applied Logistic Centre,
Moscow, Russia
Stevan Berber, PhD, Auckland University, New Zealand
Constantin Rotaru, PhD, Henri Coanda Air Force
Academy, Brasov, Romania
Nenad Dodi, PhD, dSPACE GmbH, Paderborn,
Germany
Kamen Iliev, PhD, Bulgarian Academy of Sciences,
Centre for National Security and Defence, Sofia, Bulgaria
Col. Stoyan Balabanov, Ministry of Defence, Defence
Institute, Sofia, Bulgaria
Greihin Leonid Ivanovi, PhD, State College of
Aviation, Minsk, Belarus
Slobodan Stupar, PhD, Faculty of Mechanical
Engineering, Belgrade, Serbia
Col. Goran Diki, PhD, University of Defence, Serbia
Col. Boban orovi, PhD, University of Defence, Serbia
Col. Nenad Dimitrijevi, PhD, Military Academy, Serbia
Col. Miodrag Regodi, PhD, Military Academy, Serbia
Lt. Col. Dragan Trifkovi, PhD, Military Academy,
Serbia
Vlado urkovi, PhD, Military Academy, Serbia
Biljana Markovi, PhD, Faculty of Mechanical
Engineering, Sarajevo, Bosnia and Herzegovina
Branko Livada, PhD, Vlatacom Institute, Belgrade,
Serbia
Stevica Graovac, PhD, Faculty of Electrical Engineering,
Belgrade, Serbia
Col. Martin Macko, PhD, University of Defence, Brno,
Czech Republic
Col. Milenko Andri, PhD, Military Academy, Sebia
Col. Dejan Ivkovi, PhD, Military Technical Institute,
Serbia
George Dobre, PhD, University Politehnica of
Bucharest, Romania
Momilo Milinovi, PhD, Faculty of Mechanical
Engineering, Belgrade, Serbia
Dragutin Debeljkovi, PhD, Faculty of Mechanical
Engineering, Belgrade, Serbia
Slobodan Jaramaz, PhD, Faculty of Mechanical
Engineering, Belgrade, Serbia
Jovan Isakovi, PhD, High Engineering School of
Professional Studies, Belgrade, Serbia
Aleksa Zejak, PhD, RTRK Institute for Computer Based
Systems, Novi Sad, Serbia
Strain Posavljak, PhD, Faculty of Mechanical
Engineering, Banja Luka, Republic Srpska
Fadil Islamovi, PhD, Faculty of Mechanical
Engineering, Biha, Bosnia and Herzegovina
Tomaz Vuherer, PhD, University of Maribor, Slovenia
Silva Dobri, PhD, Military Medical Academy, Serbia
Elizabeta Ristanovi, PhD, Military Medical Academy,
Serbia
Zijah Burzi, PhD, Military Technical Institute, Serbia
Col. Ivan Pokrajac, PhD, Military Technical Institute,
Serbia
Mirko Kozi, PhD, Military Technical Institute, Serbia
Nikola Gligorijevi, PhD, Military Technical Institute,
Serbia
Vencislav Grabulov, PhD, IMS Institute, Serbia
Lt. Col. Ljubia Tomi, PhD, Technical Test Centre,
Serbia
Nenko Brklja, PhD, Technical Test Centre, Serbia
PREFACE
The Military Technical Institute, the first and the largest military scientificresearch institution in
Serbia with a 68 year long tradition, has been traditionally organizing the OTEH scientific
conference devoted to the defensive technologies. The Conference is sponsored by the Ministry of
Defense and it takes place every second year.
Its aim is to gather scientists and engineers, researchers and designers, manufactures and university
professors in order to exchange ideas and to develop new relationships.
The Seventh International Scientific Conference OTEH 2016 is scheduled as follows: lecture on the
occasion of Mileva Mari  Einstein, plenary session with two introductory lectures, working
sessions according to the Conference topics, exhibition of some actual exhibits of the weapons and
military equipment developed by the Military Technical Institute.
The papers which will be presented at the Conference have been classified into the following
thematic fields:
Aerodynamics and Flight Dynamics
Aircraft
Weapon Systems and Combat Vehicles
Ammunition and Energetic Materials
Integrated Sensor Systems and Robotic Systems
Telecommunication and Information Systems
Materials and Technologies and CBRN Protection
Quality, Standardization, Metrology, Maintenance and Exploitation.
The Proceedings contain 134 reviewed papers which have been submitted by the authors from 15
different countries. I would also like to stress that 24 papers are from abroad. The quality of papers
accepted for publication achieved very high standard. I expect stimulated discussion on many topics
that will be presented during two days of the Conference.
On behalf of the organizer I would like to thank all the authors and participants from abroad, as well
as from Serbia, for their contributions and efforts which made this Conference possible and
successful.
I would also like to thank the Ministry of Education and Science of the Republic of Serbia for its
financial support.
Finally, dear guests and participants of the Conference, I would like to wish you an enjoyable stay
in Belgrade and I am looking forward to see you again at the Eighth Conference.
Belgrade, October, 2016
Miodrag Lisov
Chairman of the Scientific Committee
VI
CONTENTS
OCCASIONAL LECTURE
3 MILEVA MARI EINSTEIN HER LIFE, WORK AND FATE, Velimir Abramovi
PLENARY LECTURES
7 IMPLEMENTATION OF INTEGRATED LOGISTIC SUPPORT TECHNOLOGIES:
FROM LOGISTIC SUPPORT ANALYSIS UP TO PERFORMANCE BASED
LOGISTICS, Evgeny V. Sudov
9 HISTORICAL DEVELOPMENT OF MODERN SMALL ARMS TECHNOLOGY:
OAK RIDGE NATIONAL LABORATORY PERSPECTIVE, Slobodan Raji
1. SECTION : AERODYNAMICS AND FLIGHT DYNAMICS
13 EFFECT OF BASE BLEED ON THE DRAG REDUCTION, Habib Belaidouni, Saa
ivkovi, Mirko Kozi, Marija Samardi, Boutemdjet Abdelwahid
19 DIVERGENCE ANALYSIS OF THIN COMPOSITE PLATES IN SUBSONIC AND
TRANSONIC FLOWS, Mirko Dinulovi, Aleksandar Grbovi, Danilo Petrainovi
24 AEROACOUSTIC ANALYSIS OF A JET NOZZLE, Toni Ivanov, Vasko Fotev,
Neboja Petrovi, Zorana Trivkovi, Dragan Komarov
29 NUMERICAL AND EXPERIMENTAL INVESTIGATION OF AERODYNAMIC
CHARACTERISTICS OF SPIN STABILIZED PROJECTILE, Damir D. Jerkovi,
Aleksandar V. Kari, Neboja Hristov, Slobodan S. Ili, Slobodan Savi
35 A HIGH SPEED TRAIN MODEL TESTING IN T32 WIND TUNNEL BY
INFRARED THERMOGRAPHY AND STANDARD METHODS, Slavica Risti,
Suzana Lini, Goran Ocokolji, Boko Rauo, Vojkan Luanin
41 AERODYNAMICS OF THE HIGH SPEED TRAIN BIOINSPIRED BY A
KINGFISHER, Suzana Lini, Boko Rauo, Mirko Kozi, Vojkan Luanin, Aleksandar
Bengin
47 OBSERVATIONS ON SOME TRANSONIC WIND TUNNEL TEST RESULTS OF A
STANDARD MODEL WITH A TTAIL, Dijana Damljanovi, ore Vukovi,
Aleksandar Viti, Jovan Isakovi, Goran Ocokolji
52 NUMERICAL AND EXPERIMENTAL ASSESSMENT OF TRANSONIC
TURBULENT FLOW AROUND ONERA M4 MODEL, Jelena Svorcan, Dijana
Damljanovi, Dragan Komarov, Slobodan Stupar, Neboja Petrovi
58 COMPUTATIONAL ANALYSIS OF HELICOPTER MAIN ROTOR BLADES IN
GROUND EFFECT, Zorana Trivkovi, Jelena Svorcan, Marija Balti, Dragan Komarov,
Vasko Fotev
64 SIMULATION OF ROLL AUTOPILOT OF A MISSILE WITH INTERCEPTORS,
Milan Ignjatovi, Milo Pavi, Slobodan Mandi, Bojan Pavkovi, Nataa Vlahovi
68 DESIGN OF THE MAIN PIVOT ON THE FORCED OSCILLATION APPARATUS
FOR THE WIND TUNNEL MEASUREMENTS, Marija Samardi, Dragan
Marinkovski, Duan uri, Zoran Raji, Abdelwahid Boutemedjet
VII
73 PRELIMINARY AERODYNAMIC COMPUTATION OF LONG ENDURANCE
UAV WING, Abdelwahid Boutemedjet, Marija Samardi, Zoran Raji
2. SECTION : AIRCRAFT
79 DEVELOPMENTS IN HEADUP DISPLAY TECHNOLOGY FOR BASIC AND
ADVANCED MILITARY TRAINING AIRCRAFT, Robert Wilsey Fraes
85 FLIGHT PERFORMANCE DETERMINATON OF THE PISTON ENGINE
AIRCRAFT SOVA, COMPUTER PROGRAM SOVAPERF, Nemanja Velimirovi,
Kosta Velimirovi
90 POSSIBLE APPROACHES TO EVALUATION OF TRAINING AIRCRAFTS USED
IN FLIGHT SCREENING, Slavia Vlai, Franc Hudomal, Aleksandar Kneevi
95 INTEGRATION OF TACTICAL  MEDIUM RANGE UAV AND CATAPULT
LAUNCH SYSTEM, Zoran Novakovi, Zoran Vasi, Ivana Ili, Nikola Medar, Dragan
Stevanovi
102 CONTRIBUTION TO THE MAINTENANCE OF Mi8 HELICOPTER IN THE
SERBIAN AIR FORCE, Zoran Ili, Boko Rauo, Miroslav Jovanovi, Ljubia Tomi,
Stevan Jovii, Radomir Janji, Nenko Brklja
108 ON THE EFFECTIVE SHEAR MODULUS OF COMPOSITE HONEYCOMB
SANDWICH PANELS, Lamine Rebhi, Mirko Dinulovi, Predrag Andri, Marjan Dodi,
Branimir Krsti
114 EFFICIENT COMPUTATION METHOD FOR FATIGUE LIFE ESTIMATION OF
AIRCRAFT STRUCTURAL COMPONENTS, Stevan Maksimovi, Mirjana uri,
Zoran Vasi, Ognjen Ognjanovi
119 ANALYSIS OF AIRCRAFT STRUCTURES CROSS SECTION, Bogdan S.
Bogdanovi, Dario A. Sinobad, Tonko A. Mihovilovi
125 STRESS CALCULATION OF NOSE GEAR SUPPORT WITH ASPECT OF
WELDING OF AEROSPACE STEEL 15CRMOV6, Aleksandar Petrovi, Bogdan S.
Bogdanovi, Aleksandar Stanaev
131 INFLUENCE OF PILOTS AVERAGE BODY MASS INCREASING ON BALANCE
OF LIGHT PISTON TRAINING AIRCRAFT, Zorica Sari, Zoran Vasi, Vojislav
Devi, Boris Glava
139 PROTOTYPE SOVA DEVELOPMENT: AIRCRAFT LYFE CYCLE EXTENSION,
Vanja Stefanovi, Marija Blai, Marina Ostoji, Tonko Mihovilovi, Dragan Ili
145 SOME ASPECTS OF THE DIFFERENT TYPES WIRELESS SENSORS
IMPLEMENTATION WITHIN AIRBORNE FLIGHT TEST CONFIGURATION,
Zoran Filipovi, Vladimir Kvrgi, Dragoljub Vuji
152 UAS  FROM MINI TO TACTICAL, Adi Cohen
3. SECTION : WEAPON SYSTEMS AND COMBAT VEHICLES
157 A PRELIMINARY DESIGN MODEL FOR EXPLOSIVELY FORMED
PROJECTILES, Mohammed Amine Boulahlib, Milo Markovi, Slobodan Jaramaz,
Momilo Milinovi, Mourad Bendjaballah
VIII
163 TENDENCIES OF DEVELOPMENT OF AMPHIBIOUS ASSETS IN ARMED
FORCES OF NATO COUNTRIES, Nenad Kovaevi, Nenad Dimitrijevi
168 ON ALGORITHM OF SYNCHRONIZED SWARMING AGAINST AN ACTIVE
THREAT SIMULATOR, Radomir Jankovi, Momilo Milinovi
173 DETERMINING PROJECTILE CONSUMPTION DURING INDIRECT MORTAR
FIRE, Aca Randjelovi, Vlado Djurkovi, Petar Repi
177 PROPELLER AND SHIP MAIN EGINE SELECTION IN CORRELATION WITH
OVERALL EFFICIENCY PROPULSION COEFFICIENT IMPROVEMENT, Jovo
Dautovi, Vojkan Madi, Sonja urkovi
182 OPTIMIZATION OF PLANETARY GEARS AND EFFECTS OF THE THINRIMED GEAR ON FILLET STRESS, Milo Sedak, Tatjana M. Lazovi Kapor, Boidar
Rosi
188 PROJECTION OF QUALITY A COMPLEX TECHNICAL SYSTEM, Ljubia Tani,
Petar Jovanovi, Samed Karovi
194 STRESS ANALYSIS OF INTEGRATED 12.7 MM MACHINE GUN MOUNT,
Aleksandar Kari, Duan Jovanovi, Damir Jerkovi, Neboja Hristov
199 STRATEGY IMPLEMENTATION OF DUALSEMIACTIVE RADAR HOMING
GUIDANCE WITH COUPLING OF TANDEM GUIDED AND LEADING MISSILE
OF AIR DEFENCE MISSILE SYSTEM ON REAL MANEUVERING TARGET,
Markovi Stojan, Milinovi Momilo, Nenad Sakan
205 EXPERIMENTAL INVESTIGATION OF OILS IN FOURSTROKE ENGINES,
Sreten Peri, Bogdan Nedi
211 OPTIMIZATION OF THE BOX SECTION OF THE SINGLEGIRDER BRIDGE
CRANE BY GRG ALGORITHM ACCORDING TO DOMESTIC STANDARDS AND
EUROCODES, Goran Pavlovi, Vladimir Kvrgi, Stefan Mitrovi, Mile Savkovi, Neboja
Zdravkovi
218 MATHEMATICAL MODELING DYNAMIC PERFORMANCE OF ARTILLERY
FIRE SUPPORT IN THE OFFENSIVE OPERATION, Damir Projovi, Zoran
Karavidi, Miroslav Ostoji
223 MODELING AND MULTIBODY SIMULATION OF LAND ROVER DEFENDER
110 RIDE AND HANDLING DYNAMICS, Nabil Khettou, Dragan Trifkovi, Slavko
Mudeka
231 PERSPECTIVES OF USE OF SWITCHED RELUCTANCE MOTORS IN COMBAT
VEHICLES, Radoslav Rusinov
4. SECTION : AMMUNITION AND ENERGETIC MATERIALS
237 PHYSICOCHEMICAL PROPERTIES AND THERMAL STABILITY OF
MICROCRYSTALLINE NITROCELLULOSE ISOLATED FROM WOOD FIBER,
Mohammed Amin Dali
243 A METHOD OF GUNPOWDER GRAIN SHAPE OPTIMIZATION, Stefan Jovanovi
249 COMPOSITE SOLID PROPELLANTS WITH OCTOGENE, Vesna Rodi, Marica
Bogosavljevi, Aleksandar Milojkovi, Saa Brzi
255 SOLVING TECHNICAL PROBLEMS WHILE WORKING WITH ORDNANCE
USING INNOVATION PRINCIPLES, Obrad abarkapa, Duan Raji, arija Markovi
IX
260 APPLYING OF NANOTECHNOLOGY IN PRODUCTION OF RIFLE
AMMUNITION, Mihailo Erevi, Veljko Petrovi, Branka Lukovi
266 DETERMINATION OF COMPATIBILITY OF DOUBLE BASE PROPELLANT
WITH POLYMER MATERIALS USING DIFFERENT TEST METHODS, Mirjana
Dimi, Bojana Fidanovski, Ljiljana Jelisavac, Slavia Stojiljkovi, Nataa Kariik
272 CHARACTERIZATION OF BEHIND ARMOR DEBRIS AFTER PERFORATION
OF STEEL PLATE BY ARMOR PIERCING PROJECTILE, Predrag Elek, Slobodan
Jaramaz, Dejan Mickovi, Miroslav orevi, Nenad Miloradovi
278 VISUALIZING THE THERMAL EFFECT OF THERMOBARIC EXPLOSIVES,
Uro Aneli, Danica Simi, Dragan Kneevi, Marko Devi
283 RELIABILITY OF SOLID ROCKET PROPELLANT GRAIN UNDER
SIMULTANEOUS ACTION OF MULTIPLE TYPES OF LOADS, Nikola Gligorijevi,
Saa ivkovi, Vesna Rodi, Saa Antonovi, Aleksandar Milojkovi, Bojan Pavkovi,
Zoran Novakovi
290 AN EXAMPLE OF PROPELLANT GRAIN STRUCTURAL ANALYSIS UNDER
THE THERMAL AND ACCELERATION LOADS, Saa Antonovi, Nikola
Gligorijevi, Aleksandar Milojkovi, Sredoje Suboti, Saa ivkovi, Bojan Pavkovi
297 TRANSFER OF GRANULATED PBX PRODUCTION TO THE INDUSTRIAL
SCALE, Slavica Terzi, Stanoje Bioanin, Aleksandar orevi, ivka Krsti, Biljana
Kostadinovi, Zoran Borkovi
304 EXPLOSIVE REACTIVE ARMOR ACTION AGAINST SHAPED CHARGE JET,
Dejan Mickovi, Slobodan Jaramaz, Predrag Elek, Nenad Miloradovi, Dragana Jaramaz,
Duan Mickovi
310 AMMUNITION SURPLUS  THREAT TO POSSESSORS DISPOSAL METHODS:
REVIEW OF DEMILITARIZATION TECHNOLOGIES, Bla Miheli
324 SHOCKWAVE OVERPRESSURE OF PROPELLANT GASES AROUND THE
MORTAR, Miodrag Lisov, Slobodan Jaramaz, Mirko Kozi, Novica Ristovi
5. SECTION : INTEGRATED SENSOR SYSTEMS AND ROBOTIC SYSTEMS
331 ACOUSTIC SOURCE LOCALIZATION USING A DISCRETE PROBABILITY
DENSITY METHOD FOR POSITION DETERMINATION, Ivan Pokrajac, Nadica
Kozi, Predrag Okiljevi, Miodrag Vraar, Brusin Radiana
336 STATISTICAL APPROACH IN DETECTION OF AN ACOUSTIC
BLAST WAVE, Miodrag Vraar, Ivan Pokrajac
340 ADAPTIVE TIME VARYING AUTOPILOT DESIGN, Nataa Vlahovi, Stevica
Graovac, Milo Pavi, Milan Ignjatovi
345 MATHEMATICAL MODEL FOR PARAMETER ANALYSIS OF PASSIVELY QSWITCHED Nd:YAG LASERS, Mirjana Nikoli, eljko Vukobrat
350 HFSW RADAR DESIGN: TACTICAL, TECHNOLOGICAL AND
ENVIRONMENTAL CHALLENGES, Dejan Nikoli, Bojan Doli, Nikola Tosi, Nikola
Leki, Vladimir D. Orli, Branislav M. Todorovi
356 EFFECTIVENESS OF ACTIVE VIBRATION CONTROL OF A FLEXIBLE BEAM
USING A DIFFERENT POSITION OF STRAIN GAGE SENSORS, Miroslav Jovanovi,
Aleksandar Simonovi, Neboja Luki, Nemanja Zori, Slobodan Stupar, Slobodan Ili
362 INFLUENCE OF GEOMETRICAL PARAMETERS ON PERFORMANCE OF
X
MEMS THERMOPILE BASED FLOW SENSOR, Danijela Randjelovi, Olga Jaki,
Mile M. Smiljani, Predrag Poljak, arko Lazi
367 MONITORING PHYSIOLOGICAL STATUS OF THE SOLDIER DURING
COMBAT MISSION VIA INTEGRATED MEDICAL SENSOR (HEART RATE,
OXYGEN SATURATION) SYSTEM, Oliver Mladenovski, Jugoslav Ackoski, Milan
Goci
371 ALUMINIUM TILES DEFECTS DETECTION BY EMPLOYING PULSED
THERMOGRAPHY METHOD WITH DIFFERENT THERMAL CAMERAS, Ljubia
Tomi, Vesna Damnjanovi, Goran Diki, Boban Bonduli, Bojan Milanovi, Rade Pavlovi
377 CHANNEL SELECTOR FOR OPTIMIZATION OF TEST AND CALLIBRATION
PROCEDURES OF ICTM PRESSURE SENSORS, Predrag Poljak, Milo Vorkapi,
Danijela Randjelovi
381 SECURITY SYSTEM IN MILITARY BASES WITH MATLAB ALGORITHM,
Tamara Gjonedva, Sofija Velinovska, Jugoslav Achkoski, Boban Temelkovski
385 IMAGING DETECTOR TECHNOLOGY: A SHORT INSIGHT IN HISTORY AND
FUTURE POSSIBILITIES, Branko Livada, Dragana Peri
391 IMAGE QUALITY PARAMETERS: A SHORT REVIEW AND APPLICABILITY
ANALYSIS, Jelena Koci, Ilija Popadi, Branko Livada
398 MULTISENSOR SYSTEM OPERATORS CONSOLE: TOWARDS STRUCTURAL
AND FUNCTIONAL OPTIMIZATION, Dragana Peri, Saa Vuji, Branko Livada
404 STATIONARY ONROAD OBSTACLES AVOIDANCE BASED ON COMPUTER
VISION PRINCIPLES, Mourad Bendjaballah, Stevica Graovac, Mohammed Amine
Boulahlib, Milo Markovi
411 GPS AIDED INS WITH GYRO COMPASSING FUNCTION, Ivana Trajkovski, Nada
Asanovi, Vladimir Vukmirica, Milan Miloevi
417 MODERNIZATION OF THE RADAR P12, Verica Marinkovi Nedelicki, Branislav
Pavi, Boris Mikovi, Mladen Mileusni, Predrag Petrovi, Aleksandar Lebl, Dragan
Borjan, Dejan Ivkovi, Dragan Nikoli
422 DISTRIBUTED TARGET TRACKING IN CAMERA NETWORKS USING AN
ADAPTIVE STRATEGY, Nemanja Ili, Khaled Obaid Al Ali, Milo S. Stankovi, Srdjan
S. Stankovi
428 AUTONOMOUS MOBILE ROBOT PATH PLANNING IN COMPLEX AND
DYNAMIC ENVIRONMENTS, Novak Zagradjanin, Stevica Graovac
434 SENSORLESS BRUSHED DC MOTOR SPEED CONTROL USING NATURAL
TRACKING CONTROL ALGORITHM, Milo Pavi, Milan Ignjatovi, Nataa
Vlahovi, Mirko Miljen
6. SECTION : TELECOMMUNICATION AND INFORMATION SYSTEMS
441 EVALUATION OF SELFORGANIZING UAV NETWORKS IN NS3, Nataa Maksi,
Milan Bjelica
446 CONCEPTUALIZING SIMULATION FOR LAWSONS MODEL OF COMMAND
AND CONTROL PROCESSES, Neboja Nikoli
451 GENERATING EFFECTIVE JAMMING AGAINST GLOBAL NAVIGATION
SYSTEMS, Sergei Kostromitsky, Aliaksandr Dyatko, Petr Shumski, Yury Rybak
XI
457 EFFICIENT POWER FLOW ALGORITHM, MODIFIED ALGORITHM NAHMAN
AND PERI, Branko Stojanovi, Milan Moskovljevi, Tomislav Raji
462 SOLID STATE LBAND HIGH POWER AMPLIFIER USING GAN HEMT
TECHNOLOGY, Zvonko Radosavljevi, Dejan Ivkovi, Dragan Nikoli
466 PERFORMANCE EVALUATION OF NONLINEAR OPTIMIZATION METHODS
FOR TOA LOCALIZATION TECHNIQUES, Maja Rosi, Mirjana Simi, Predrag
Pejovi
472 GPUBASED PREPROCESSING FOR SPECTRUM SEGMENTATION IN
DIRECTION FINDING, Marko Mii, Ivan Pokrajac, Nadica Kozi, Predrag Okiljevi
478 TECHNIQUES FOR INTELLIGENCE DATA GATHERING IN MOBILE
COMMUNICATIONS, Saa Stojkovi, Ivan Tot, Fejsov Nikola
481 AN IMPLEMENTATION OF MANET NETWORKS ON COMMAND POST
DURING MILITARY OPERATIONS, Vladimir Risti, Boban Z. Pavlovi, Saa Devetak
486 PRACTICAL IMPLEMENTATION OF DIGITAL DOWN CONVERSION FOR
WIDEBAND DIRECTION FINDER ON FPGA, Vuk Obradovi, Predrag Okiljevi,
Nadica Kozi, Dejan Ivkovi
494 STATISTICS OF RATIO OF TWO WEIBULL RANDOM VARIABLES WITH
DIFFERENT PARAMETERS, Ivica Marjanovi, Dejan Rani, Danijela Aleksi, Dejan
Mili, Mihajlo Stefanovi
500 SOFTWARE AND INFORMATIONAL SYSTEMS IN THE PRODUCTION OF
DTM25 OF THE MILITARY GEOGRAPHICAL INSTITUTE, Aleksandar Pavlovi,
Viktor Markovi, Ana Vuievi, Saa Bakra
7. SECTION : MATERIALS, TECHNOLOGIES AND CBRN PROTECTION
507 THERMAL STABILITY AND MAGNETIC PROPERTIES OF FE2O3
POLYMORPH, Violeta N. Nikoli, Marin Tadi, Vojislav Spasojevi
513 TECHNOLOGY FOR COMBATING BIOTERRORISM, Elizabeta Ristanovi
517 ESTIMATION OF SAFT AND PCSAFT EOS PARAMETERS FOR NHEPTANE
UNDER HIGH PRESSURE CONDITIONS, Jovana Ili, Mirko Stijepovi, Aleksandar
Gruji, Jasna Staji Troi, Gorica Ivani, Mirjana Kijavanin
522 THE APPLICATION OF IR THERMOGRAPHY FOR THE CRACKS DETECTION
IN THE COMPOSITE STRUCTURES USED IN AVIATION, Stevan Jovii, Ivana
Kosti, Zoran Ili, Ljubia Tomi, Aleksandar Kovaevi
525 THERMAL AND CAMOUFLAGE PROPERTIES OF ROSALIA ALPINA
LONGHORN BEETLE WITH STRUCTURAL COLORATION, Ivana Kosti, Danica
Pavlovi, Vladimir Lazovi, Darko Vasiljevi, Dejan Stojanovi, Dragan Kneevi, Ljubia
Tomi, Goran Diki, Dejan Panteli
530 ISOGEOMETRIC ANALYSIS OF FREE VIBRATION OF ELLIPTICAL
LAMINATED COMPOSITE PLATES USING THIRD ORDER SHEAR
DEFORMATION THEORY, Ognjen Pekovi, Slobodan Stupar, Aleksandar Simonovi,
Toni Ivanov
536 ON THE CORRELATION OF MICROHARDNESS WITH THE FILM ADHESION
FOR SOFT FILM ON HARD SUBSTRATE COMPOSITE SYSTEM, Jelena Lamovec,
Vesna Jovi, Ivana Mladenovi, Bogdan Popovi, Milo Vorkapi, Vesna Radojevi
541 A COMPARISON OF DIFFERENT CONVEX CORNER COMPENSATION
XII
STRUCTURES APPLICABLE IN ANISOTROPIC WET CHEMICAL ETCHING OF
{100} ORIENTED SILICON, Vesna Jovi, Jelena Lamovec, Mile Smiljani, arko Lazi,
Bogdan Popovi, Predrag Poljak
547 RADIOACESIUM137 IN THE ENVIRONMENT AND THE EFFECT OF
RADIATIONHYGIENE CERTIFICATION ON FOOD, Nataa Paji, Tatjana Markovi
550 SEPARATION OF THE CARBONDIOXIDE FROM THE GAS MIXTURE, Dragutin
Nedeljkovi, Lana Puti, Aleksandar Staji, Aleksandar Gruji, Jasna StajiTroi
556 ELECTRODEPOSITION OF METAL COATINGS FROM EUTECTIC
TYPE IONIC LIQUID, Mihael Buko, Jelena B. Bajat
561 IMPACT OF THE ALTERED TEXTURE OF THE ACTIVE FILLING OF THE
FILTER ON THE SORPTIVE CHARACTERISTICS WITH THE SPECIAL
REFERENCE TO THE EFFECIENCY OF FILTERING, Marina Ili, eljko Seni,
Vladimir Petrovi, Biljana Mihajlovi, Vukica Grkovi
567 INFLUENCE OF DAMAGED INJECTORS USED IN COMMON RAIL SYSTEMS
ON ECOLOGICAL AND ENEGRGY EFFICIENY, Dejan Jankovi, Mileta Ristivojevi,
Dimitrije Kosti
572 DEFECT DURING PRODUCTION OF STEEL CARTRIDGE CASE, Nada Ili,
Ljubica Radovi
577 FAILURE ANALYSIS OF THE STATOR BLADE, Jelena Marinkovi, Duan Vraari,
Ljubica Radovi, Ivo Blai
581 READOUT BEAM COUPLING STRATEGIES FOR PLASMONIC CHEMICAL OR
BIOLOGICAL SENSORS, Zoran Jaki, Mile M. Smiljani, arko Lazi, Dana
Vasiljevi Radovi, Marko Obradov, Dragan Tanaskovi, Olga Jaki
587 COMPLETE KINETIC PROFILING OF THE THREE NANOMOLAR
ACETYLCHOLINESTERASE INHIBITORS, Maja VitoroviTodorovi, Mirjana
Jakii, Sonja Bauk, Branko Drakuli
594 DEPENDANCE OF CBRN INSULATING MATERIALS PROTECTION TIME
UPON BUTYLRUBBER AND FLAME RETARDANT CONTENT, Vukica Grkovi,
Vladimir Petrovi, eljko Seni, Maja VitoroviTodorovi
598 FILTERING HALF MASKS USAGE FOR PROTECTION AGAINST AEROSOL
CONTAMINATION OF BIOLOGICAL AGENTS, Negovan Ivankovi, Duan Raji,
Radovan Karkali, Dejan Indji, Duan Jankovi, eljko Seni, Marina Ili
603 LASERS POSSIBILITIES IN BRASS SURFACE CLEANING, Bojana Radojkovi, Slavica
Risti, Suzana Poli, Bore Jegdi, Aleksandar Krmpot, Branislav Salati, Filip Vueti
609 EFFECT OF IFWS2 NANOPARTICLES ADDITION ON PHYSICALMECHANICAL AND RHEOLOGICAL PROPERTIES AND ON CHEMICAL
RESISTANCE OF POLYURETHANE PAINT, Dragana S. Lazi, Danica M. Simi,
Aleksandra D. Samolov
614 THERMAL ANALYSIS OF NANOCRYSTALLINE NIFE2O4 PHASE FORMATION
IN SOLID STATE REACTION, Vladan osovi, Aleksandar osovi, Toma ak,
Nadeda Talijan, Duko Mini, Dragana ivkovi
618 PRELIMINARY ANALYSIS OF THE POSSIBILITY OF PREPARING PVB/IFWS2
COMPOSITES. EFFECT OF NANOPARTICLES ADDITION ON THERMAL AND
RHEOLOGICAL BEHAVIOR OF PVB, Danica M. Simi, Duica B. Stojanovi, Mirjana
Dimi, Ljubica Totovski, Saa Brzi, Petar S. Uskokovi, Radoslav R. Aleksi
XIII
624 HIGH PERFORMANCE LIQUID CHROMATOGRAPHY DETERMINATION OF
2,4,6TRINITROTOLUENE IN WATER SOLUTION, Jovica Nei, Ljiljana Jelisavac,
Aleksandar Marinkovi, Slavia Stojiljkovi
630 MEASURING CLEANING CLASS OF OIL AFTER TRIBOLOGICAL TESTING,
Radomir Janji, Slobodan Mitrovi, Dragan Duni, Ivan Maui, Blaa Stojanovi, Milan
Bukvi, Zoran Ili
636 RECYCLING LITHIUM  ION BATTERY, Milan Bukvi, Radomir Janji, Blaa
Stojanovi
642 NUMERICAL CALCULATION OF JINTEGRAL USING FINITE ELEMENTS
METHOD, Bahrudin Hrnjica, Fadil Islamovi, Denana Gao, Esad Bajramovi
646 INFLUENCE OF DIFFERENT TYPES OF POLYMER IMPREGNATION ON
SPECTRAL REFLECTION OF TEXTILLE MATERIALS, Aleksandra Samolov, Milan
Kuli
649 QUALITY OF RECOVERED EXPLOSIVES OBTAINED FROM DELABORATED
MUNITIONS, Maja Matovi, Ljiljana Bundalo
654 THE STRENGTH INVESTIGATION OF SPECIFIC POLYMERIC COMPOSITE
ELEMENT/METALLIC ELEMENT JOINT REALIZED BY PINS, Slobodan
itakovi, Jovan Radulovi
659 QUALITATIVE AND QUANTITATIVE ASSESSMENT OF BOND STRENGTH OF
SOLID ROCKET PROPELLANT AND THERMOPLASTIC MATERIAL FOR
CARTRIDGE LOADED GRAIN, Jovan Radulovi
665 SYNTHESIS OF RE/PD HETEROGENEOUS CATALYSTS SUPPORTED ON HMS
USING SOLGEL METHOD FOLLOWED BY SUPERCRITICAL DRYING WITH
EXCESS SOLVENT, Dragana Proki Vidojevi, Sandra B. Glii, Aleksandar M. Orlovi
671 UNDERSTANDING PLASMA SPRAYING PROCESS AND APPLICATION IN
DEFENSE INDUSTRY, Bogdan Nedi, Marko Jankovi
678 THERMAL STABILITY AND MICROSTRUCTURAL CHANGES INDUCED BY
ANNEALING IN NANOCRYSTALLINE FE72CU1V4SI15B8 ALLOY, Radoslav Surla,
Milica Vasi, Neboja Mitrovi, Ljubica Radovi, Ljubica Totovski, Dragica Mini
682 LOW LEVEL TRITIUM DETERMINATION IN ENVIRONMENTAL SAMPLES
USING 1220 QUANTULUS, Nevena Zdjelarevi, Marija Leki, Nataa Lazarevi
685 OPTIMIZATION AND VIRTUAL QUALITY CONTROL OF A CASTING, Sreko
Manasijevi, Radomir Radia, Janez Pristravec, Velimir Komadini, Zoran Radosavljevi
8. SECTION : QUALITY, STANDARDIZATION, METROLOGY,
MAINTENANCE AND EXPLOITATION
695 RELIABILITY PREDICTION OF ELECTRONIC EQUIPMENT: PROBLEMS AND
EXPERIENCE, Slavko Pokorni
701 APPLICATION OF INNOVATION STANDARDS IN THE FIELD OF WEAPONRY
AND MILITARY EQUIPMENT, Duan Raji, Obrad abarkapa
705 SURFACE TEXTURE FILTRATION INTERNATIONAL STANDARDS AND
FILTRATIONS TECHNIQUE OVERVIEW, Srdjan ivkovi, Branka Lukovi, Veljko
Petrovi
XIV
710 SYSTEM FOR REMOTE MONITORING AND CONTROL OF HFOTH RADAR,
Bojan Doli, Dejan Nikoli, Nikola Tosi, Nikola Leki, Vladimir D. Orli, Branislav M.
Todorovi
715 MODEL OF IMPROVING MAINTENANCE OF TELECOMMUNICATION
DEVICES, Vojkan Radonji, Milenko iri, Branko Resimi, Ivan Milojevi
721 USAGE AN INFRARED THERMOGRAPHY FOR THE PROCESS CONDITIONBASED MAINTENANCE OF SHIPS SYSTEMS, Veselin Mrdak
727 VOLUMETRIC CALIBRATION FOR IMPROVING ACCURACY OF AFP/ATL
MACHINES, Samoil Samak, Igor Dimovski, Vladimir Dukovski, Mirjana Trompeska
733 DIAGNOSTIC APPROACH TO THE MAINTENANACE OF MARINE SYSTEMS,
Duan Cincar
739 MAINTENANCE OF HYBRID VEHICLES, Blaa Stojanovi, Milan Bukvi, Radomir
Janji
745 A NEW APPROACH TO CREATING AND MANAGING TECHNICAL
PUBLICATIONS FOR AIRCRAFT LASTA USING S1000D STANDARD, Branko
Dragi, Vojislav Devi, Miodrag Ivanievi
750 NEW ISSUE OF STANDARD AS/EN 9100:2016, EXPECTATION AND BENEFITS
FOR CUSTOMERS, Biljana Markovi
XV
OCCASIONAL LECTURE
MILEVA MARI EINSTEIN HER LIFE, WORK AND FATE
Prof. dr VELIMIR ABRAMOVI
Mileva Mari (18751948) a Serbian mathematician and physicist, was Einstein's colleague,
confidante and wife. 1896, at the age of twentyone, as the only woman in that year beginning
studies in the mathematical section, Mileva entered the Swiss Federal Polytechnic, Zurich, the same
date as Einstein, who was threeandahalf years younger than her. They married 1903. Doubtless,
after their marriage, Mileva subordinated her professional goals to Einstein's. There are strong
indications and evidences that her influence in Einstein's most creative time to come was enormous.
But, it is still on the way to properly research and determine to what extent mathematician Mileva
Mari Einstein truly contributed to Brownian motion, (Einstein's doctoral thesis, 1906),
Electrodynamics of moving bodies 1905, (while working on the subject, Einstein wrote to Mileva:
our work on relative motion), then, Photoeffect explanation 1905, and energy to matter
conversion formula (E=mc2) which Einstein described 1919 as the most important upshot of the
special theory of relativity). For the sake of scientific truth clarity we have to single out, recover
and evaluate Milevas original mathematical and physical ideas that underlined the Theoretical
Physics of XX century.
There is a philosophical saying that one who really does the work is rarely credited for it.
Overshadowed by the worldwide famous and selfcentered personality of her great husband Albert,
after divorce 1919, due to tiring and painful family obligations, Mileva never again could raise
enough strength and motivation to return to scientific problems by herself.
PLENARY LECTURES
IMPLEMENTATION OF INTEGRATED LOGISTIC SUPPORT
TECHNOLOGIES: FROM LOGISTIC SUPPORT ANALYSIS UP TO
PERFORMANCE BASED LOGISTICS
Dr. SCI. (TECH) EVGENY V. SUDOV
R&D Center Applied Logistics
Abstract:
Integrated Logistic support (ILS)  is a combination of modern management, engineering and
information technologies dedicated to development and support of system of exploitation of
machinebuilding products. The core of ILS is a logistic support analysis (LSA). This is a complex
of formal tasks around modelling of product and maintenance procedures needed to keep the
product in necessary condition. Such model allows to estimate expected availability of the product
and value of corresponding maintenance costs. During operation of product this model may be
tuned up on the base of operational statistics. Some parameters of exploitation system may be also
changed, such as maintenance tasks conditions and intervals, stratification of tasks between levels
of repair, material management algorithms and so on.
It's quite obvious, practical implementation all this sophisticated methods requires great efforts as
well as it requires methodical and normative support.
In 20062016 in Russian Federation are developed and implemented more than 30 National
standards in the area of ILS. At the same time a number of software tools for LSA and creation of
interactive maintenance / repair documentation were developed and implemented in different
branches of industry (aircraft, shipbuilding, defense).
All this accumulated practices and experience allows to get close to new model of relation between
Customer and Supplier, when a subject of the contract are not only services, but also the
achievement of certain indicators of readiness (availability). Such modern approach now is known
as Performance Based logistics.
HISTORICAL DEVELOPMENT OF MODERN SMALL ARMS
TECHNOLOGY: OAK RIDGE NATIONAL LABORATORY PERSPECTIVE
SLOBODAN RAJI
Oak Ridge National Laboratory, Oak Ridge TN, USA, rajics@ornl.gov
Abstract:
It can be argued that the development of modern small arms had its origins in 1944 Nazi Germany with the introduction
of the StG44. This first assault rifle was intended to replace both the bulky longrange battle rifle and the pistol caliber
submachine gun with a single weapon. This idea of an intermediate caliber mated with a large ammunition capacity and
full automatic capability revolutionized the individual combat weapon. For the next forty years virtually all individual
weapons were directly based on this gasoperated and auto loading design, including the two most ubiquitous assault
rifles of the west and east, the AR and AK platforms, respectively. Although a limited amount of research was ongoing,
very few original ideas were fielded during that period. The significant lack of advancement in small arms technology
during this time was in large contrast to the exploding new capabilities of modern air forces and navies.In the 1980s it
seemed that some advanced technology, available to the individual soldier, was finally being developed for actual
implementation. The Heckler & Koch G11 and the Barrett Firearms 82A1 provided decidedly new technology and
novel capabilities. However, even though the G11 was advanced in many ways, its high cost and final deployment near
what turned out to be the end of the cold war reduced it to a mere research and development demonstration. Although
the 82A1 was adopted by the United States (US) military due to its ability to extend the individual soldiers reach from
~1,000 m to ~2,500 m, its lack of advanced lightweight materials resulted in its adoption in relatively small numbers.
After another 25 years, only minor modifications to the AR and AK weapon systems have been made. The only major
advancement to individual soldier weapons over many decades has been the XM25 grenade launcher, a highly
sophisticated weapon and firecontrol system that can change battlefield tactics. Although very expensive, this system
is the first significant advancement in small arms technology since the AR and AK platforms were first developed,
about 60 and 70 years ago, respectively. Even though modern air forces would never consider using, even in their
reserve inventories, platforms that were nearly 70 years old, the worlds premier armies seem resolved and destined to
soon be using nearly centuryold small arms technology. This lack of small arms development and implementation
during the cold war is difficult to understand since this time period represented the largest arms race in the history of the
world. The typical answer given by some to explain this situation is the relatively large cost associated with regularly
replacing the entire small arms portfolio. However with a present global defense budget of approximately $2T, and the
US representing about 40 percent of that amount, it seems inexplicable based solely on cost. The US spends less than
one percent of its research and development budget on small arms. US casualty data from recent wars, Iraq and
Afghanistan, show that the probability of a US soldier expiring from a rifle bullet is much larger than that from an
explosive device. If the many billions of dollars spent on detecting and defeating improvised explosive devices was
justified, then the extremely small amounts spent on small arms development is particularly difficult to explain. Due to
the very advanced present technological status of sensors, materials, processing power, manufacturing, and battery
storage, it seems clear that a much higher level of small arms capability can easily be implemented. At a minimum it
should be a goal of a military to produce small arms that can have the projectile pointofimpact coincident with the
aimpoint (zeroed) under all condition and all times. Presently, weapons are zeroed at some known distance. The
problem is that at any other distance and environmental condition, the correct aimpoint has to be estimated by the skill
of the operator, a challenging task for a young and inexperienced soldier. The Oak Ridge National Laboratory has been
developing advanced technologies related to small arms applications for many years in response to the concerns stated
above. We have developed rapid manufacturing methods that can take advantage of exotic new materials and alloys.
We have also developed surface treatments that are lubricous or superhydrophobic and improve both maintenance and
reliability of existing small arms. We have also been able to apply the knowledge gained during our vast history of
sensor development to the small arms field and have developed a barrel deflection sensor with reticle compensation in
realtime. Very recently we have developed a projectile tracking approach that can terminally track large caliber sniper
bullets. Since small arms are often the weapon of choice for many terrorists, ORNL is now focusing on the development
of small arms technologies that will provide soldiers with weapons that possess a large technological overmatch
capability. With the new hot war of terrorism on the rise recently, hopefully the next 70 years will be more
technologically productive in terms of small arms development compared to the cold war.
SECTION I
Aerodynamics and Flight Dynamics
CHAIRMAN
Marija Samardi, PhD
Duan uri, PhD
EFFECT OF BASE BLEED ON THE DRAG REDUCTION
HABIB BELAIDOUNI
Military Academy, University of Defense, Belgrade, h_belaidouni@yahoo.fr
SAA IVKOVI
Military Technical Institute, Belgrade, sasavite@yahoo.com
MIRKO KOZI
Military Technical Institute, Belgrade, mkozic@open.telekom.rs
MARIJA SAMARDI
Military Technical Institute, Belgrade, majasam@ptt.rs
ABDELWAHID BOUTEMEDJET
Military Academy, University of Defense, Belgrade, abdelwahed1954@gmail.com
Abstract: One of the most important aerodynamic performance characteristics for the artillery projectiles shell is the
total drag. The total drag of projectile can be divided into three components consisted of pressure drag, viscous drag,
and the base drag. The base drag is major contributor to the total drag, for that reason its important to have its good
estimation in the preliminary design stage of a projectile. Projectiles with base bleed use concept of reducing base
drag, by injection of gas generated by burning of composite propellant, into the base area. In the paper, the internal
ballistic calculation of existing base bleed configuration is presented. Using CFD computations various turbulence
models were tested. The numerical results were compared and validated with semi empirical theory. Obtained
numerical results with the most appropriate turbulence model served in further CFD calculation of aerodynamic drag
of the projectile with and without base bleed effect.
Keywords: Base drag, Base bleed, Computational fluid dynamics.
In recirculation region the point along the axis of
symmetry, where the streamwise velocity diminish, is so
called shear layer reattachment point. As the shear layer
reattaches, the flow is forced to turn along the axis of
symmetry, causing the formation of a reattachment shock.
1. INTRODUCTION
Aerodynamic bodies such as projectiles, missiles and
rockets, generally, undergo significant deterioration of
flight performance by the drag. For these types of flight
bodies, especially the drag in the base region has the most
significant contribution to total drag. At transonic speeds,
for example, base drag constitutes a major portion up to
50% of the total drag for typical projectiles at Mach 0.9
[1]. Therefore the base drag should be considered
separately from the other pressure drag components. For
this reason, the minimization of base drag has been an
important issue to date, and considerable effort has been
made to find suitable techniques for obtaining low base
drag shell design.
Picture 2 shows that injecting small amounts of gas into
the flow field behind the base of the projectile will split
the originally large recirculation zone into two parts. One
recirculation region remains at the symmetry axis, and the
other one is formed right behind the base corner [47]. As
the mass flow rate increases, the recirculation zone at the
axis is pushed further out, and the other one at the base
corner becomes larger. If the mass flow rate is increased
away, the recirculation region near the axis disappears,
and the base bleed follows a straight path.
During projectile flight, directly behind the base, the
reverse flow can be seen (Picture 1).
2. NUMERICAL MODEL
The large turning angle behind the base causes separation
and formation of reverse flow known as the recirculation
region or the separation bubble [23]. The size of the
recirculation determines the turning angle of the external
flow, and therefore the strength of the expansion waves.
A smaller recirculation region causes the flow sharp
turning, leading to a stronger expansion wave, and lower
pressures behind the base. Therefore, small separated
region causes larger base drag than large regions.
The objective of the present study is to estimate
numerically the coefficient of drag at different Mach
number and angles of attack for 122 mm supersonic
artillery shell.
Analysis was done numerically using the software Gambit
2.4 and ANSYS FLUENT [8].
13
OTEH2016
EFFECTOFBASEBLEEDONTHEBASEDRAGREDUCTION
Picture 1. Flow flied behind a projectile without base bleed [8]
Picture 2. Flow flied behind a projectile with base bleed [8]
The geometry of projectile and meshing were made using
GAMBIT 2.4. A computational domain of about 65000
cells structured mesh was constructed around the
projectile with base bleed. The computational domain
was set up with dimensions of 40 d, where d is the model
diameter, downwards from the base and 25 d upwards
14
OTEH2016
EFFECTOFBASEBLEEDONTHEBASEDRAGREDUCTION
Pressure far field boundary condition, as shown in Picture
3, has been used. Calculation was performed using twodimensional axisymmetric, density based solver with
various turbulence models and second order upwind
discretization of the flow and turbulence equations.
from the model axis, to ensure freestream conditions and
thus to obtain better convergence. The aerodynamic
coefficients were the criterion of convergence in all cases.
Several turbulence models were investigated:
The two equations realizable k turbulence model,
which solves transport equations for the turbulence
kinetic energy k, and its rate of dissipation ,
the three equations kkl turbulent viscosity model
solving turbulent kinetic energy, laminar kinetic energy
and specific dissipation rate,
and the last one was twoequation SST k model
which solving shearstress transport equations.
Grid in the whole numerical domain for 122 mm
projectile can be seen in Picture 4.
Picture 3. Computational grids [9]
Picture 4. Mesh of numerical domain for 122 mm projectile
3. INTERNAL BALLISTIC CALCULATION
The gas generator contains two identical solid propellant
grains. These two elements provide an internal
combustion surface consisting of two cylindrical surfaces
and four flat surfaces, which results in total decreasing
burning surface area.
The gas generator for the 122 mm base bleed projectile is
housed in the afterbody, whose cross is shown in cross in
Picture 5.
In experimental research [10], a series of measurements
of base bleed effect were performed for different flight
Mach numbers. Results are shown in Picture 6. Picture 6
shows reduction of the base pressure as a function of
dimensionless injection parameter I for different values of
flight Mach number. The parameter I is ratio of the bleed
mass flow rate and product of the base area with
freestream mass flow rate,
I=
m p
pV Abase
(1)
Conclusion was made that for most of the Mach numbers
there is an optimal value of I. This optimum value is
higher with decrease of the projectile flight Mach number.
According to the results of research [10], the semiempirical model was established, which allows to
determine optimum working regime of the gas generator,
in order to achieve reduction of the base drag coefficient.
Picture 5. Scheme of gas generator unit and propellant
grain
15
OTEH2016
EFFECTOFBASEBLEEDONTHEBASEDRAGREDUCTION
Picture 6. The experimental results representing of the effect of Mach number [10]
Picture 7. Required pressure vs. time curve to obtain desired optimum mass flow rate change.
Picture 8. Experimental chamber pressure vs. Time for gas generator 122 mm
16
OTEH2016
EFFECTOFBASEBLEEDONTHEBASEDRAGREDUCTION
The mass flow rate of combustion products bleeding
through the orifice by subsonic velocity is:
m p = I V Abase
(2)
where: V = M ( kRT ) 2
In dependency of projectiles trajectory altitude all
influential functions can be expressed in dependency from
time, using altitude y(t) and speed M(t) parameters: air
density (y) = (t) and temperature T(y) = T(t), Finally
optimal mass flow rate change in time can be determined
as:
m p ( t ) = I ( M ( t ) ) ( t ) kRT ( t )M ( t ) Abase
Picture 10. Effects of base bleed with m = 0.1231 kg/s
on drag coefficient
(3)
On the other hand required combustion products mass
flow rate exiting from gas generator, where is shown in
Picture 7, can be achieved by appropriate internal ballistic
design of gasgenerator (GG). In dependency of thermochemical characteristics of propellant, and using gas
dynamics relations, the products mass flow rate as a
function of the pressure in combustion chamber is
obtained as:
m p ( p ) = 2 ( p ) V2 ( p ) A2
5. CONCLUSION
In the paper, numerical computations were performed for
projectile 122 mm at the different values of the Mach
number. The addition of base bleed with base cavity also
was computed. A series of calculation drag coefficient
was run for projectile 122 mm with and without base
bleed.
The Mach number range computed was 0.8<M<2 with an
angle of attack = 0. For the case without base bleed
was found a best turbulence model. The turbulence model
k was selected for simulation projectile with gas
generator (base bleed).
(4)
Designed gas generator GG is tested in Military Technical
Institute Laboratory ''Technicum'' in Baric. According to
the experiments obtained curve of chamber pressure vs.
time shown in Picture 8 is accurate enough.
The influence of the base bleed flow effects on the drag
coefficient was calculated. The computed drag coefficient
of the projectile with and without base bleed showed a
reduction about 11%.
4. RESULTS
The zero yaw drag CX0 is compared first with remarkably
good agreement to semi empirical data when the k
turbulence model is utilized, Picture 9. The kkl does
not do quite as well. However, the twoequation SST was
implemented with no significant differences noted with
the semiempirical data [11].
This investigation will be continued through studying the
effect of base bleed on drag coefficient at different angles
of attack, the influence of the base bleed effects on the
lateral aerodynamic coefficients, and especially on the
dynamic derivatives and stability parameters.
A best model k turbulence model will be selected for
simulation projectile with base bleed because it shows
good agreement with CX0.
NOMENCLATURE
Picture 9. Zeroyaw drag coefficient vs. Mach number
The effect of mass flow rate on drag coefficient at
different flight regimes is shown in Picture 10. It shows
that there is a reduction of about 11% in drag coefficient
in both the subsonic and the supersonic flow.
17
m p
 mass flow rate of combustion products.
Abase
 base area of projectile.
A2
 orifice area.
V2
 velocity of combustion products at section 2.
 freestream velocity.
 density of combustion products at section 2.
 freestream density.
 gas constant.
 temperature.
 ratio of specific heats.
 angle of attack.
Cdb
 base drag coefficient.
CX0
 zero yaw drag coefficient.
 dimensionless injection parameter.
OTEH2016
EFFECTOFBASEBLEEDONTHEBASEDRAGREDUCTION
 Mach number.
pb
 base pressure.
MTI
Supersonic Stream, Aeronautical Research Council
(Great Britain), Reports and Memoranda No. 3224,
Dec. 1959.
[6] Bowman,J.E., Clayden,W.A.: Cylindrical Afterbodies
in Supersonic Flow with Gas Ejection, AIAA
Journal, 5, (6), 1967, 1524,1525.
Turbulent
[7] Valentine,D.T.,
Przirembel,C.E.G.:
Axisymmetric NearWake at Mach Four with Base
Injection, AIAA Journal, 8, (12), 1970, 2279, 2280.
[8] Ansys Inc., ANSYS FLUENT 14.0 and GAMBIT 2.1
licensed to VTI, 2010
[9] Lee,Y.K., Kim,H.D., Raghunathan,S.: A Study of
Base Drag Optimization Using Mass Bleed, 15th
Australasian Fluid Mechanics Conference, 1317
December 2004, Australia.
[10] Bowman,J.E., Clayden,W.A.: Reduction of base drag
by gas ejection, R.A.R.D.E report 4/64, 1969
[11] Jaramaz,S., Injac,M.: Methode of calculation range
of base bleed projectile (in Serbian), Military
Technical Institute, Belgrade, 1989.
 Military Technical Institute.
References
[1] Sahu,J. Nietubicz,C.J.: NavierStokes Computations
of Projectile Base Flow with and without Mass
Injection, AIAA Journal, 23(9), 1985, 13481355.
[2] Herein,J.L., Dutton,J.C.: Supersonic Base Flow
Experiments in the Near Wake of a Cylindrical
Afterbody, AIM Journal, 32, (1), 1994, 7783.
[3] Herrin,J.L., Dutton,J.C.: Supersonic NearWake
Afterbody Boattailing Effects on Axisymmetric
Bodies, Journal of Spacecraft and Rockets, 31,
(6),1994, 10211028.
Preliminary
[4] Cortright,E.M.,
Schroeder,A.H.:
Investigation of Effectiveness of Base Bleed in
Reducing Drag of BluntBase Bodies in Supersonic
Stream, NACA RM E51A26, March 1951.
[5] Reid,J., Hastings,R.C.: The Effect of a Central Jet on
the Base Pressure of a Cylindrical Afterbody in a
18
DIVERGENCE ANALYSIS OF THIN COMPOSITE PLATES IN SUBSONIC
AND TRANSONIC FLOWS
MIRKO DINULOVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, mdinulovic@mas.bg.ac.rs
ALEKSANDAR GRBOVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, agrbovic@mas.bg.ac.rs
DANILO PETRAINOVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, dpetrasinovic@mas.bg.ac.rs
Abstract: In the present paper the static aeroelastic phenomenon known as torsional divergence is investigated on rocket
stabilizers (fins) made of composite materials. Using analytic approach, the differential equation for torsional divergence of
composite trapezoidal stabilizer is derived. The equation obtained was in the form of second order differential equation with
variable coefficients. The solution to divergence equation was obtained using Galerkin's approach and the complete solution
procedure is presented. Required material elastic coefficients were calculated using micromechanics composite analysis for
lamina, and classical lamination theory (CLT) for complete stabilizer laminate layup.
It was found that Galerkin's approach can be successfully deployed in solving differential divergence equation and
divergence speed (VD) of composite stabilizer in subsonic airflow can be effectively calculated.
1. INTRODUCTION
Static aeroelasticity investigates problems that involve the
interaction between steadystate aerodynamic forces and
flexible structure deformations. Lifting surfaces (airplane
wing, tail, rocket stabilizer...) exposed to airflow may exhibit
instability that is known as torsional divergence. Torsional
divergence is static aeroelastic problem, and it results from
the fact that when the lifting surface is exposed to airflow an
aerodynamic moment is generated. This aerodynamic
moment is proportional to the square of flight speed. On the
other hand, the structure elastic stiffness is independent of the
flight speed since it is inherent characteristic of the structure
itself. Based on this, it is obvious that a critical airspeed may
exist at which the structure elastic stiffness is not sufficient to
keep the lifting surface in a deformed position (as a result of
aerodynamic moment). Above critical airspeed even a small
deformation of a lifting surface leads to a large angle of twists
(therefore large strains and stresses) which may cause lifting
surface failure.
Figure 1. The effect of structure flexibility
Structural stiffness is relatively low for flexible plates and
shells which represent the basic building block of modern
flight structures. These structures are therefore more
susceptible to instabilities even at low flow velocities. In
this case, divergence and flutter can easily occur at
relatively low flow velocity
During the design phase of the lifting surfaces, it must be
insured that the structure is capable of sustaining applied
loads, it remains stable (divergent free) within whole range
of the flight envelope [14].
In aerodynamics lifting surface can be represented by an
object which transforms input parameters: velocity (V) and
angle of attack () into aerodynamic force and aerodynamic
moment, whereas in the aeroelasticity the feedback loop
exists, where, is through the elasticity of the structure, the
value of the of the angle of attack () is transformed, by
addition of the structural angle of twist ( ) (Figure 1).
2. COMPOSITE FIN TORSIONAL
DIVERGENCE SPEED
Let us assume that the lifting surface has variable
aerodynamic, geometric and mechanical parameters along
the span as presented in the following figure (Figure 1).
19
OTEH2016
DIVERGENCEANALYSISOFTHINCOMPOSITEPLATESINSUBSONICANDTRANSONICFLOWS
dM T ( y ) 1
dC ( y )
+ V 2 z
dy
d
2
[ ( y ) + ( y )] e ( y ) l 2 ( y )
+ 1 V 2 Cmac ( y ) l 2 ( y )
2
n g ( y ) h ( y ) = 0
(5)
Using the relation between applied torque or moment of
torsion (MT), angle of twist () shear modulus (G) and the
torsion constant (J), the change of the angle of twist along
the fin span may be expressed as:
d = M T
dy GJ
Figure 2. Differential element of the fin and forces
distribution
Differentiating previous relation in respect to span
coordinate y one obtains:
Differential lifting aerodynamic force, denoted as dRz(y)
acts in the aerodynamic center of the differential element of
dM T
= d GJ d =
dy
dy
dy
2
d
d
GJ ( y ) 2 +
d ( GJ ( y ) )
dy dy
dy
the fin. The area of the differential element is l ( y ) dy .
Assuming the quasisteady flow around the stabilizer fin the
lifting aerodynamic force on the fin differential element is
therefore,
dRz ( y ) = 1 V 2
2
dC z ( y )
[ a ( y ) + ( y )] l ( y ) dy
l ( y ) dy l ( y )
(1)
+ d ( GJ ( y ) )
dy 2 dy
[GJ ( y )] d
(2)
element mass per nit span.
In the shear center, on one side of the differential element
the torsion moment M T ( y ) is applied, and on the other side
the torsion moment, with corresponding gradient is:
(3)
GJ = const, l = const , e = const, = const,
(9)
Equilibrium condition of the differential fin element, about
the shear center, requires that the moment equation is
satisfied in the following form:
MT ( y ) +
dCz
= const, Cmac = const, h = const, g = const.
d
dM T ( y )
dy M T ( y ) +
dy
+ dRz ( y ) e ( y ) l ( y ) +
(8)
This equation represents a linear nonhomogeneous secondorder equation with variable coefficients. The objective is to
solve the previous equation for the angle of twist () along
the fin span as a function of quasi steady flow speed (V) and
determine the extreme values of twisting angle which
correspond to divergence speed (VD). However, due to its
complexity, finding the analytical solution to previous
differential equation is not possible for the real lifting
surface geometries used nowadays. Only in cases where,
acts, where n is load factor, and g ( y ) is the differential
dM T ( y )
dy
dy
d + 1 V 2 dCz ( y ) l 2 ( y ) e ( y )
dy 2
d
2 dC z ( y )
2
1
= V
( y ) l ( y ) e ( y )
d
2
1 V 2 Cmac ( y ) l 2 ( y ) +
2
n g ( y) h( y)
Also in the center of gravity differential inertia force ng ( y )
MT ( y ) +
(7)
Therefore, the equilibrium condition of the differential fin
element about the shear center is:
The differential value of the aerodynamic moment is:
dM ac ( y ) = 1 V 2 Cmac ( y )
2
(6)
which represents the rectangular surface with constant
parameters along the span (y), finding the analytic solution
to the derived differential equation is possible, and it has
been investigated by many authors. The solution for the
divergence speed (VD) is:
(4)
+ dM ac ( y ) n g ( y ) h ( y ) dy = 0
VD =
Substituting the relations for the quasisteady differential
aerodynamic force and differential aerodynamic moment in
the previous equation, the equilibrium condition becomes:
20
2GJ 2
dC
z l 2 eb 2
d
(10)
OTEH2016
DIVERGENCEANALYSISOFTHINCOMPOSITEPLATESINSUBSONICANDTRANSONICFLOWS
However, for the more complex geometries and variable fin
parameters along the span, solution to the differential
equation in hand, has to be sought by deploying numerical
approach. In this paper the solution of the equation is found
by Galerkin method [5]. The Galerkin method is weighted
residual methods, where weight function functions are
2y
,
b
2
2y
2 ( y ) = y 1 ,....,
b
N
2y
N ( y ) = y 1
b
1 ( y ) = y 1
chosen from the basis functions w( x) {i ( x )}i =1 and it is
n
required that the following n equations hold true:
It is obvious that functions (eq. 17) satisfy required initial
conditions (eq.16)
( x ) ( L [u ( x )] + f ( x )) dx = 0 i = 1, 2,..., n
i
(10)
( b)
k ( 0 ) = 0 1 0
(11)
dC ( y )
B ( y, V ) = 1 V 2 z
( y ) l 2 ( y ) e ( y )
d
2
k 1
(19)
Substituting the assumed solution in the differential
equation on hand, one may obtain:
Where, by comparison one may conclude that in the
previous equation variable coefficients are:
dC ( y ) 2
l ( y ) e ( y )
A ( y, V ) = 1 V 2 z
d
(18)
2 (b 2)
2 (b 2)
2 (b 2)
1 b k b = 0
k ( b 2 ) = 1
2
F ( y ) d 2 + F ( y ) d +
dy
dy
F ( y ) = GJ ( y ) ,
= 0,
( )
The differential equation for the equilibrium condition of
the differential fin element about the shear center (eq. 8),
derived in the previous section of this paper, for
convenience can be expressed as:
F ( y ) = d ( GJ ( y ) )
dy
2y
+
b
k 1
2y
y k 1 2
b
b
k ( y ) = 1
3. SOLUTION OF THE DIVERGENCE SPEED
DIFFERENTIAL EQUATION BY
GALERKIN`S METHOD
A ( y, V ) = B ( y, V )
(17)
F ( y)
i =1
A ( y, V )
(12)
aii( y ) + F ( y )
a ( y) +
i i
i =1
a ( y ) = B ( y, V )
(18)
i i
i =1
Using Galerkin's method for error minimization of assumed
solution for the differential equation (eq. 8) it follows that:
(13)
b2
[ F ( y ) ( y ) + F ( y ) ( y ) + A ( y,V ) ( y )]
ai
i =1
j ( y ) dy =
(14)
1 V 2 Cmac ( y ) l 2 ( y ) + n g ( y ) h ( y )
2
(19)
b2
B ( y,V ) ( y ) dy
j
Assuming the solution in the form:
( y) =
With following notation:
a ( y )
i
(15)
b2
Aij (V ) =
i =1
[ F ( y ) ( y ) + F ( y ) ( y ) + A ( y,V ) ( y )]
i
(20)
j ( y ) dy
And functions i ( y ) are chosen to satisfy initial
conditions:
b2
i ( 0 ) = 0, i ( b 2 ) = 0
B j (V ) =
(16)
B ( y,V ) ( y ) dy
j
(21)
In General, there are many functions that satisfy these
conditions. In a particular problem following functions are
chosen:
Previous equations assume the form of linear equations:
N
A (V ) a = B (V )
ij
i =1
21
(22)
OTEH2016
DIVERGENCEANALYSISOFTHINCOMPOSITEPLATESINSUBSONICANDTRANSONICFLOWS
If the value for the determinant is not equal to zero new
flight speed V2 is assumed and coefficients and determinant
are recalculated. Based on new values it can be verified
whether the determinant has changed the sign suggesting
that the determinant zero value (condition given by eq. 28)
lies between the two guesstimate flight speeds.
Coefficients are functions of flight speed (V). The system of
equations can be expressed in the following form:
A11 (V ) a1 + A12 (V ) a2 + ....
+ A1N (V ) aN = B1 (V )
A21 (V ) a1 + A22 (V ) a2 + ....
+ A2 N (V ) aN = B2 (V )
(23)
4. NUMERICAL EXAMPLE
The solution to the lifting surface torsional divergence
problem, presented in previous section is illustrated on a
trapezoidal plate, with cross section in a form of an airfoil.
The lifting surface geometry and airfoil data is given in the
following picture.
AN 1 (V ) a1 + AN 2 (V ) a2 + ....
+ ANN (V ) aN = BN (V )
In order to find the solution to the above system of linear
equations Cramer's rule can be used. Forming the
determinants:
A11 A12
A21 A22
D (V ) =
AN 1 AN 2
A1N
A2 N
ANN
(24)
A11 A12
A21 A22
Di (V ) =
AN 1 AN 2
B1
B2
Bn
A1N
A2 N
ANN
Figure 2. Lifting surface geometry and control sections
locations
Table 1. NACA 006 airfoil data
The required coefficients can be determined as follows
(eq.25):
ai (V ) =
Di (V )
D (V )
(25)
It has to be emphasized that all parameters in previous
equations are functions of quasi steady flow speed (V).
The final solution can be written in the following form:
( y, V ) =
1
D (V )
a (V ) ( y ) =
i
i =1
D (V ) ( y )
i
(26)
i =1
Critical divergence speed is defined that the angle of twist
( ) at that particular speed becomes infinitely large
( y,VD ) .
0.00000
0.00947
0.01307
0.01777
0.02100
0.02341
0.02673
0.02869
0.02971
0.03001
0.02902
0.02647
0.02282
0.01832
0.01312
0.00724
0.00403
0.00063
Table 2. Fin material data
(27)
AS4 Carbon fiber /3501 6 Epoxy matrix (UD) 35016 Epoxy
It may be concluded that this is satisfied when:
D (V ) 0 .
NACA 006 airfoil data
0.00063
0.00000
0.00403
0.01250
0.00724
0.02500
0.01312
0.05000
0.01832
0.07500
0.02282
0.10000
0.02647
0.15000
0.02902
0.20000
0.03001
0.25000
0.02971
0.30000
0.02869
0.40000
0.02673
0.50000
0.02341
0.60000
0.02100
0.70000
0.01777
0.80000
0.01307
0.90000
0.00947
0.95000
0.00000
1.00000
1.00000
0.95000
0.90000
0.80000
0.70000
0.60000
0.50000
0.40000
0.30000
0.25000
0.20000
0.15000
0.10000
0.07500
0.05000
0.02500
0.01250
0.00000
(28)
Therefore, in an engineering application of method
presented, the critical divergence speed may be calculated
by assuming certain flight speed V1, and based on this value
coefficients Aij (V1 ) and determinant D (V1 ) are calculated.
22
Vf
0.63
1.58
E11
E22
G12
12
142
10.3
7.2
0.27
F1t
F1c
F2t
F2c
F6
2280
1440
57
228
71
DIVERGENCEANALYSISOFTHINCOMPOSITEPLATESINSUBSONICANDTRANSONICFLOWS
Table 3. Fin Geometric characteristics
5. CONCLUSION
J[m x Xcg
Es
a.c
e
1010] [m]
0 0.07624 11.3 0.0320 0.0279 0.0191 0.0088
25 0.06988
8.01 0.0294 0.0255 0.0175 0.0080
50 0.06353
5.45 0.0267 0.0232 0.0159 0.0073
75 0.05718
3.59 0.0240 0.0209 0.0143 0.0066
100 0.05083
2.25 0.0213 0.0186 0.0127 0.0059
section %
L[m]
In this paper the solution for torsional divergence speed is
presented. This type of static aeroelastic phenomenon can be
mathematically expressed with linear nonhomogeneous
secondorder differential equation with variable coefficients,
that was derived in section of this paper. Analytic solution
can not be found in the closed form for this type of problem
in cases when the geometry of the lifting surface is
relatively complex and material is not isotropic.
Table 4. Fin Geometric characteristic variables for
divergence calculation
L
J
Xcg
Es
a.c
e
OTEH2016
In order to solve torsional divergence equation for complex
lifting surface geometries and orthotropic materials
(composites), numerical approach using Galerkins method
was used. It was found that this solution method can be
successfully applied. Numerical example is presented and
using this method the divergence speed is calculated for the
lifting surface in the form of composite stabilizer fin.
setion span
section polar moment of inertia
location of sectional center of gravity
location of sectional shear center
aerodynamic center position
distance between shear center and aerodynamic center
expressed in percent of span length.
The solution found represents theoretical divergence speed
value (i.e the the speed at which the lifting surface becomes
unstable or divergent). However, the results obtained are in
practice nonrealistic (rotation of the tip section lifting
surface are two high). These results were expected since the
structural failure criteria were not used. Incorporating the
material failure criteria into the divergence speed calculation
method proposed is the course for the future work in this
area.
One of the required parameters in torsional divergence
equation, derived in previous section is polar moment of
inertia. For the fin (lifting surface) analyzed in this example
it is clear that the values for the polar moment of inertia
change along the span. Calculating moments of inertia for
all control sections (1 to 5, Table 3), the function that best
fits was expressed in the following form:
J = 5 1014 y 2 1.4 1011 y + 1.1 109
(29)
References
[1] Librescu,L., Maalawi,K.Y.: Aeroelastic design
optimization of thinwalled subsonic wingsagainst
divergence, ThinWalled Structures 47 (2009) 8997.
[2] Kirch,A., Clobes,M., Peil,U.: Aeroelastic divergence
and flutter: Critical comments on the regulations of EN
199114, Journal of Wind Engineering and Industrial
Aerodynamics, Volume 99, Issue 12, December 2011,
Pages 12211226.
[3] Mahran,M., Negm,H., ElSabbagh,A.: Aeroelastic
characteristics of tapered plate wings, Finite Elements
in Analysis and Design, Volume 94, February 2015,
Pages 2432.
[4] Kameyama,M., Fukunaga H.: Optimum design of
composite plate wings for aeroelastic characteristics
using lamination parameters Computers & Structures,
Volume 85, Issues 34, February 2007, Pages 213224.
[5] Feng,X., Lewis,T., Neilan,M.: Discontinuous Galerkin
finite element differential calculus and applications to
numerical solutions of linear and nonlinear partial,
differential equations, Journal of Computational and
Applied Mathematics, Volume 299, June 2016, Pages
6891.
Figure 3. Section rotation along fin span for 500 m/s
supersonic flow
Figure 4. Fin tip section rotation as a function of flow
speed
23
AEROACOUSTIC ANALYSIS OF A JET NOZZLE
TONI IVANOV
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, tivanov@mas.bg.ac.rs
VASKO FOTEV
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, vfotev@mas.bg.ac.rs
NEBOJA PETROVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, npetrovic@mas.bg.ac.rs
ZORANA TRIVKOVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, ztrivkovic@mas.bg.ac.rs
DRAGAN KOMAROV
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, dkomarov@mas.bg.ac.rs
Abstract: Computational fluid dynamics (CFD) based on the finite volume method were used in order to obtain the aerodynamic characteristics of a jet nozzle. After the transient CFD results were gained the Ffowcs WilliamsHawking formulation
for farfield noise modeling was used for aeroacoustic computation. A grid sensitivity was done and different turbulence
models were tested. The obtained results were then compared with publicly available experimental data after which a conclusion was derived.
Keywords: computational aeroacoustics, computational fluid dynamics, jet noise.
1. INTRODUCTION
Nowadays most researchers use inhouse codes for the large
eddy simulation (LES) or detached eddy simulation (DES)
of the flow field and the Ffowcs WilliamsHawking (FWH)
formulation for the farfield noise prediction [13]. Also
researchers have made significant efforts in order to implement the Unsteady Reynolds Averaged NavierStokes
(URANS) solvers with the use of different turbulence
viscosity models for predicting jet noise [47].
One of the most important issues in the air traffic industry
today is the aircraft noise. Many governments have limited
the allowable noise level in and around airports through
regulations and strict requirements for aircraft external noise
certification.
Jet noise is an issue in military aircraft as well. Military aircraft
have engines with very low bypass ratios and high exit temperatures and velocities due to which they have very loud noise
characteristic that can pose annoyance to communities near
military bases as well as a health threat to ground crews.
In this paper aeroacoustic analysis of the SMC000 nozzle
were done by the use of the two dimensional axisymmetric
URANS with the FWH methodology for the far field noise
prediction.
Although in recent years aircraft engines have been
significantly improved, the rise in air traffic mandates that the
noise generated by aircraft engines needs to be further lowered.
Main source of the overall aircraft noise is the jet noise produced by the jet engine. This is especially noticeable during takeoff when the engines work at maximal power.
2. MATHEMATICAL BACKGROUND
Most of the modern models for aeroacoustic jet noise prediction are based on the acoustic analogy of Lighthill [8,
9].This is a two step analogy that separates the sound generation induced by the fluid flow and the sound propagation
in an acoustic medium. In this analogy it is shown that
aerodynamic sound is a consequence of turbulence which
provides quadrupole source distribution in ideal gas at rest.
Hence the equations of the propagation of sound in uniform
medium at rest due to externally applied fluctuating stresses
are derived by the momentum and mass conservation
equations and are given as:
In order to better predict the generated jet noise researchers
have been using different approaches over the years.
Recently computational fluid dynamics has been extensively
used for noise prediction. One of the methods used for noise
calculations is the direct numerical simulation (DNS) of the
aeroacoustic noise. This method although most accurate
requires that the sound field is calculated all the way from
the sound source to the receiver. This makes it highly
computationally demanding and impractical for noise
estimates in the far field. Another approach is to use CFD
for the flow field and some acoustic analogy for the noise
prediction in the far field.
2Tij
2
2 2
=
0
xi x j
t 2
24
(1)
OTEH2016
AEROACOUSTICANALYSISOFAJETNOZZLE
where Tij is the instantaneous applied stress at any point
and is often called the Lighthillian acoustic tensor:
Tij = ui u j + pij a02 ij
2
1 p 2 p = 2 T H ( f )
{
}
2
xi x j ij
a0 t 2
(2)
{[ pij n j + ui ( un vn )] ( f )}
xi
The computation of the far sound can then be calculated
with the integral:
2
p ( x, t ) = 1
4 xi x j
x 1 y T y, t
ij
xy
a0
+ {[ 0 vn + ( un vn )] ( f )}
t
3
d y (3)
where the function f=0 denotes a mathematical surface used
to embed the exterior flow problem (f >1) in an unbounded
space. H(f) is the Heaviside function and (f) is the Dirac
function, ui, vi, un and vn are the fluid velocity and surface
velocity components in the xi direction and normal to the
surface f=0 respectively. p` is the sound pressure at the far
field.
Lighthill then used dimensional analysis to derive a law for
scaling the total acoustic power of the jet with the jet speed.
The acoustic power from isothermal subsonic jets is:
P = K 0u 8 a0 5 d 2
(4)
This equation can be integrated analytically under the assumption of free space flow and absence of obstacles
between the sound sources and the receivers. The solution is
consisted of surface integrals which contribute for the monopole, dipole and partially from quadrupole sources and
volume integrals which represent quadrupole sources outside the source surface.
where K is a proportionality constant called acoustic power
coefficient.
The acoustic power formulation given in Eq. 4 applies only
to stationary sources. Having in mind that quadrupoles are
convecting downstream the effect of moving sources is accounted for by a convection factor.
P = K 0 u 7.5 a0 5 d 2 C 5
Qi ni
dS
f = 0 4 x y
e
Lij n j
dS
xi f = 0 4 x y
e
2
Tij
dV
+
xi x j f > 0 4 x y
e
p ( x, t ) =
t
(5)
1/2
Where [ ]e denotes evaluation at emission time e. The source terms under the integral sign are Qi = (ui vi) + 0vi
and Lij = ui(uj vj) + pij. The third integral is evaluated in
the region outside the surface f >0 in order to account for
sources outside the FWH surface. Since volume integrals
become small when the source surface encloses the source
region they can be dropped. It should be mentioned that the
time t in Eq. (10) corresponds to the observer time, that is
the time when the observer at position x perceives the pressure fluctuations, and is different from the emission time e.
(7)
is defined as:
2 =
2f L2
a02
= const
(10)
(6)
where Mc is the convection Mach number, the angle from
the jet axis, and accounts for the decay time of the eddies.
Mc in stationary ambient may be approximately related to
the Mach number:
M c = 0.55M j
The convection factor is given as:
2
C ( M c , ) = (1 M c cos ) + 2 M c 2
(9)
Since there are three characteristic problem types in design
procedures, considering the observer and the ambient fluid,
the presence of both temporal and spatial derivatives can
make the FWH equation difficult for numerical implementation. In order to tackle this issue over the years different
formulations of the FWH analogy that are better suited for
numerical implementation were introduced [12].
(8)
where f represents the characteristic frequency and L the
length scale of the eddies.
Lighthills theory is not valid for supersonic Mach numbers.
In fact through observing experimental data it can be seen
that for M=1 to 1.5 there is a u6 dependency of the sound
power level. For M up to 2.5 this dependency is of the order
u4 and for M larger than 2.5 of the order u3.
3. NUMERICAL SETUP
In this paper the two inch SMC000 nozzle geometry was
considered (Fig. 1). This geometry has been extensively
tested over the years [1315]. The nozzle geometry consists
of a diameter base with six inch inlet (left of the arrow)
which is used for different nozzle cones and a cone with 2
inch outlet (right of the arrow). In this paper only the cone
part is modeled in order to reduce the number of cells in the
grid and since the inside flow is not of interest.
More suitable approach for high speed jets acoustic noise is
given by Ffowcs Williams [10].
The Lighthill analogy doesn't account for interaction with
surfaces but the sound is only a result of turbulence quadrupoles. This was solved by Ffowcs Williams and Hawkings
[11]. The FWH equation can be written as:
25
OTEH2016
AEROACOUSTICANALYSISOFAJETNOZZLE
quadrilateral grid with 627 cells in the longitudinal and 327
in the lateral direction was chosen for the rest of the computations (Fig.4). The dimensionless wall distance on the
nozzle walls was set to y+ < 1.
Figure 1. SMC000 nozzle geometry [13]
Since aeroacoustic computations are computationally very
demanding the SMC000 geometry is going to be represented via an axisymmetric model. Axisymmetric models are
not used very often for aeroacoustic calculations for they
cannot completely accurately resolve the turbulence which
is the main source of noise in jets but efforts in this direction
have been made since the computational time necessary is
significantly smaller than that of the three dimensional models [16]. In figure 2 the computational domain with the
FWH surface used is shown.
Figure 4. Computational grid
The boundary condition values that were used for the computations are given in Table 1 where NPR is the nozzle
pressure ratio and the Mj is the jet Mach number, T0 is the
ambient temperature and p0 the ambient pressure.
Table 1. Boundary condition values
NPR
po [Pa]
T0 [K]
1.86
100 000
300
Mj
0.98
The Ansys FLUENT 16.2 commercial CFD flow solver was
used for the computations. The pressure based axisymmetric
transient solver with the pressurevelocity coupled scheme
was used. Both the temporal and spatial discretization were
of the second order.
The time step size needed to be set as to be able to capture
the highest frequency of 100 kHz and according to the
equation t=1/(nf) where t is the time step size, f is the
desired frequency and n is the nth sample that is read in the
acoustic module.
Figure 2. Computational domain
In order to ensure that statistically stationary state solution
was achieved the values in few points were monitored and
the transient calculations were done for at least 5000 time
steps after which the acoustic data on the FWH surface was
gathered for at least another 5000 time steps.
Grid sensitivity study for the steady flow using the k turbulence viscosity was done. Three different grids sizes have
been considered. The coarsest grid had 45434 cells, the medium one 172904 and the finest grid had 363312 cells.
Comparison of the centerline velocities on the jet axis obtained by the different grids is given in Fig. 3.
4. RESULTS
The velocity decay along the jet centerline obtained with the
different turbulence models is given in Fig.5.
Figure 3. Centerline velocity decay obtained with the 3
different grid sizes and the k turbulence model
As it can be seen the medium and fine grid both give very
close results so the medium grid which is a structured
Figure 5. Centerline velocity decay obtained with different turbulence models
26
OTEH2016
AEROACOUSTICANALYSISOFAJETNOZZLE
It can be seen that the k turbulence model gives the most
accurate result in comparison with the experimental results
given in [13]. The SA turbulence model best predicts the
point of the start of the decay but largely overestimates the
decay rate.
In figure 6 the turbulent kinetic energy (TKE) obtained with
the k, k SST and the transitional SST turbulence models
is shown.
Figure 9. Observer positions
The sound pressure level (SPLLsp) is calculated with the
equation Lsp = 10(log(PSD/pref2)), where PSD is the power
spectral density and pref is reference acoustic pressure and
has value of 2x105 [Pa].
In figure 10 a comparison between the sound pressure levels
for the 90 degree observer position obtained by the different
turbulence models is shown.
Figure 6. TKE values along jet centerline
In figures 7 and 8 the axial velocity and the turbulent kinetic
energy contours are shown. The images were mirrored over
the x axis in order to show the complete flow field. These
values are in good agreement when compared to PIV experiments data shown in [5].
Figure 10. Sound pressure levels for 90 degree observer
by different turbulence models
The calculated maximum sound pressure level for all turbulence models is a little larger than the experimental values while
the peak frequency and the SPL in regards to frequency in general differs a lot from the experimental values. There is a rapid
decay in the sound pressure level as the frequency increases
which is completely different from experimental results.
Figure 7. Axial velocity [m/s] contours
In figure 11 the overall sound pressure levels (OASPL) at
different observer positions for the k and SA turbulence
models in comparison with experimental data are shown.
Figure 8. TKE [m2/s2] contours
The sound pressure levels were obtained for five different
observer locations on a 50 nozzle diameters (100 inch) radius from the nozzle exit. The observer positions in regards to
the nozzle exit are given in Fig 9.These positions were set in
such manner as to replicate the microphone positions in the
experimental investigations done in [13,14].
Figure 11. OASPL at different observer positions
27
OTEH2016
AEROACOUSTICANALYSISOFAJETNOZZLE
Simulation, 47th AIAA Aerospace Sciences Meeting, Orlando, Florida, AIAA Paper 200914, (2009), pp. 125
[6] Engel, R., C., Silva, C., R. and Deschamps, C., J.: Application of RANSbased method to predict acoustic
noise of Chevron nozzles, Applied Acoustics, 79 (2014),
pp. 153163
[7] Benderskey, L., A. and Lyubimov, D.,A.: Investigation
of Flow Parameters and Noise of Subsonic and Supersonic Jets Using RANS/LES High Resolution method,
29th Congress of the International Council of the Aeronautical Sciences, St. Petersburg, Russia, September
2014
[8] Lighthill, M., J.: On Sound Generated Aerodynamically. I. General Theory, Proceedings of the Royal Society of London Series A, Mathematical and Physical
Sciences, 211, 1107 (1952), pp. 564587
[9] Lighthill, M., J.: On Sound Generated Aerodynamically. II. Turbulence as a Source of Sound, Proceedings
of the Royal Society of London Series A, Mathematical
and Physical Sciences, 222, 1148 (1954), pp. 132
[10] Ffowcs Williams, J., E.: The Noise from Turbulence at
High Speed, Philosophical Transactions of the Royal
Society of London Series A, Mathematical and Physical
Sciences, 255, 1061 (1963), pp. 469503
[11] Ffowcs Williams, J., E. and Hawkings, D., L.: Sound
Generation by Turbulence and Surfaces in Arbitrary
Motion, Philosophical Transactions of the Royal Society of London, Series A. Mathematical and Physical
Sciences, 264, 1151 (1969), pp. 321342
[12] NajafiYazdi, A., Bres, G.,A. and Mongeau, L.: An
Acoustic Analogy Formulation for Moving Sources in
Uniformly Moving Media, Proceedings of the Royal
Society Series A, 467, (2011), pp. 144165
[13] Brown, C. and Bridges, J.: Small Hot Jet Acoustic Rig
Validation, NASA/TM2006214234 Report, April 2006
[14] Froening, L., V., et.al. : Experimental Investigation of
Turbulent Jet Flow From a Chevron Nozzle, 22nd International Congress of Mechanical Engineering COBEM
2013, Brazil, November 2013
[15] Bridges, C., A. and Brown, C.: Parametric Testing of
Chevrons on Single Flow Hot Jets, NASA/TM2004213107 Report, September 2004
[16] Koch, D., L., Brigdes, J. and Khavaran, A.: Flowfield
Comparisons From Three NavierStokes Solvers for an
Axisymmetric Flow Jet, NASA/TM2002211350 Report, February 2012
It is seen that the overall acoustic sound pressure level is
largely overestimated while the increase of the OASPL with
the decrease of the observer angle is slightly underestimated
by both the k and the SA model.
5. CONCLUSION
Two dimensional axisymmetric analysis of the aeroacoustic
noise generated by the SMC000 axisymmetric two inch
nozzle was done. It was shown that although the URANS
give good results for the velocity decay and the turbulence
kinetic energy there is a large discrepancy in the sound pressure level estimation. This is probably due to the fact that
the axisymmetric model is not able to fully resolve the turbulence especially the large scale turbulence at larger downstream distances. Also adding a buffer zone on the right
exit of the computational domain may improve the SPL
estimate.
Tide and Babu [5] did 3D URANS analysis of the SMC000
on a 30 degree pie section and were still not able to accurately predict the peak frequency.
It is imposed that a full three dimensional model should be
used for accurate computation of the peak frequency.
ACKNWOLEDGMENT
The research work is funded by Ministry of Science and
Technological Development of Republic of Serbia through
Technological Development Project No. 35035.
References
[1] G., A., Faranosov, et. al.: CABARET method on unstructured hexahedral grids for jet noise computation,
Computers & Fluids, 88 (2013), pp. 165179
[2] Morris, P., J., Du, Y. and Kara, K.: Jet Noise Simulations for Realistic Jet Nozzle Geometries, Procedia Engineering, 6 (2010), pp. 2837
[3] Uzun, A. and Hussaini, M., Y.: HighFidelity Numerical Simulations of a Round Nozzle Jet Flow, 16th
AIAA/CEAS Aeroacoustics Conference, AIAA Paper
20104016, (2010), pp. 116
[4] Kenzakowski., D., C. and Kannepalli, C.: Jet Simulation for Noise Prediction Using Advanced Turbulence
Modeling, 11th AIAA/CEAS Aeroacoustics Conference,
AIAA Paper 20053086, (2005), pp. 118
[5] Tide, P.S. and Babu, V.: Aerodynamic and Acoustic
Predictions from Chevron Nozzles Using URANS
28
NUMERICAL AND EXPERIMENTAL INVESTIGATION OF
AERODYNAMIC CHARACTERISTICS OF SPIN STABILIZED
PROJECTILE
DAMIR D. JERKOVI
University of Defense in Belgrade, Military Academy, damir.jerkovic@va.mod.gov.rs
ALEKSANDAR V. KARI
University of Defense in Belgrade, Military Academy, aleksandar.kari@va.mod.gov.rs
NEBOJA HRISTOV
University of Defense in Belgrade, Military Academy, nebojsahristov@gmail.com
SLOBODAN S. ILI
University of Defense in Belgrade, Military Academy, slobodan.ilic@vs.rs
SLOBODAN SAVI
University in Kragujevac, Faculty of Engineering Sciences, ssavic@kg.ac.rs
Abstract: The paper presents the numerical and experimental research and the analysis of the aerodynamic coefficients
of the spin stabilized projectile. The numerical prediction method of aerodynamic coefficients is performed with CFD
(Computational Fluid Dynamics) steady RANS (Reynolds Averaged Navier Stokes equation) method, four different
models of turbulence and three different types of mesh. The semiempirical methods are performed to predict the values
of aerodynamic coefficients and derivatives. Experimental investigation is performed through aerodynamic wind tunnel
tests and ballistic proving ground tests. The analyses of static aerodynamic coefficients are performed for subsonic,
transonic and supersonic flow for different numerical and experimental research. The experimental proving ground test
investigations are done using 3D ballistic radar for transonic and supersonic flight Mach numbers. The comparison of
the numerically predicted values of the aerodynamic characteristics is accomplished through the 6DoF flight model of
40 mm model of projectile in relation to the experimental results. The performed numerical techniques and methods on
the structured type of mesh coupled with SST k turbulence model are generated wide and qualitative aerodynamic
description for projectile flight dynamic modeling, according to acquired experimental flight test results on the proving
ground.
Keywords: aerodynamic coefficient, RANS, SST k, experimental aerodynamic measurement, ballistic radar
measurement
calculated values in relation to the test values. The
significance of the accurate prediction is that the
symmetric projectile with initial velocity as the main
energy resource, flies to target and the main influence is
the air drag, depending on the flow regimes according to
geometric parameters and boundary conditions.
1. INTRODUCTION
The accuracy and precision of a flight dynamic system
depends on the proper model and the experimental results.
The projectile as a flight dynamic system with the specific
geometric and dynamic characteristics has to save the
initial energy, during the flight through the atmosphere.
The optimal aerodynamic shape of the projectile provides
stable flight, decreasing drag and preserving velocity.
The projectile is assumed to be either a body of revolution
whose spin axis coincides with a principal axis of inertia,
or a finned missile with three or more identical fins
spaced symmetrically around the circumference of a body
of revolution. In addition to the requirements of
configuration and mass symmetry, the projectile is also
restricted to small yaw flight along its trajectory. In
conventional aircraft aerodynamics, the terms pitch or
angle of attack refer to the aircrafts nose pointing
above or below its flight path; the terms yaw or angle
of sideslip refer to the nose pointing to the left or right of
the flight path, [1].
The classic spin stabilized projectile, observed in the
research, as symmetric solid body, is consisted of the
front part, nose (ogive shape), the middle cylindrical part
(added rotating band) and the rear part, boattail
(truncated cone shape). The specific dimensions and
construction of the projectile determine the specific
physical effects of the air flow. The main task is to
determine influence of air flow to the projectile with
adequate aerodynamic flow model and to verify
29
NUMERICALANDEXPERIMENTALINVESTIGATIONOFAERODYNAMICCHARACTERISTICSOFSPINSTABILIZEDPROJECTILE
In the paper, the numerical research of the aerodynamic
coefficients of axial force is described. The numerical and
semiempirical prediction models of the aerodynamic
coefficients are provided by the different techniques and
methods, according to the flow regime, and took into
account the influence of aerodynamic parameters and
their interaction.
OTEH2016
3. THE EXPERIMENTAL RESEARCH
The series of wind tunnel tests of the projectile model of
40 mm are performed in the T38 wind tunnel [2,3] of the
Military Technical Institute in Belgrade (VTI). The series
of proving ground ballistic tests are performed in the
facility of the Technical Test Center of Serbian Armed
Force (TOC).
2. AERODYNAMIC MODEL OF THE
PROJECTILE
3.1. The wind tunnel tests
Projectile model of 40 mm mounted in the T38 test
section is shown in Picture 2.
The aerodynamic axial force opposes the forward velocity
of the projectile and that is the classical aerodynamic
force of exterior ballistics as the air resistance or
drag. The aerodynamic force acting on projectile in the
center of pressure is given by, [1,2],
X = q S Cx
(1)
where are,
q =
S=
V2
2
d2
4
, Dynamic Pressure,
, Reference Area of the Projectile,
C x , Aerodynamic Coefficient of Axial Force,
Picture 2. Projectile in the wind tunnel test section, [2]
, Air Density of the Free Stream and,
The T38 test facility of Military Technical Institute in
Belgrade is a blow down type pressurized wind tunnel
with a 1.5m x 1.5m square test section [2,3].
V , Free Stream Velocity.
The axial aerodynamic coefficients, representing
aerodynamic force depends on airflow parameters (Mach
number, Reynolds number), aerodynamic velocity and the
angle of attack.
The wind tunnel tests of the model are performed in the
Mach number range from 0,2 to 3,0. The angle of attack
was in the interval from 10 degrees to +10 degrees and
roll angle was 0 degrees. The instrumentation and data
acquisition system are used within VTI facility. The data
reduction is performed after each run, using the standard
T38APS software package in use with the windtunnel
facility. The reduction is done in several stages, [2,3]:
The component of axial aerodynamic force X, and drag
force D, are presented in Picture 1,
X Axial Aerodynamic Force,
D Drag Aerodynamic Force,
x
Data acquisition system interfacing and signals
normalization,
Determination of flow parameters in the test section
of the wind tunnel,
Determination of model position (orientation)
relative to test section and airflow,
Determination of nondimensional aerodynamic
coefficients of forces and moments.
Picture 1. The Aerodynamic Force on Projectile
The stagnation pressure p0 in the test section is measured
by a Mensor quartz bourdon tube absolute pressure
transducer pneumatically, connected to a Pitot probe in
the settling chamber of the wind tunnel. The range of the
used transducer was 7105 Pa. The difference between the
stagnation and static pressure (pst p0) in the test section is
measured in subsonic/transonic flow regime by a Mensor
quartz bourdon tube differential pressure transducer,
pneumatically connected to the p0 Pitot probe and to the
orifice on the test section sidewall. In transonic and
supersonic flow regimes, the absolute pressure transducer
The axial aerodynamic force coefficient is given by (2)
and depends on Mach number and the angle of attack ,
[1,2]. The force represents the main component of the
total aerodynamic force (Drag),
C x ( Ma ) = C x 0 ( Ma ) + C x 2 ( Ma )
(2)
The aerodynamic axial coefficient Cx ( Ma ) depends on
Mach number (e.g. Reynolds Number), according to the
geometry parameters of the projectile, [1,2].
30
NUMERICALANDEXPERIMENTALINVESTIGATIONOFAERODYNAMICCHARACTERISTICSOFSPINSTABILIZEDPROJECTILE
Frequency Modulated Continuous Wave (FMCW).
of same type and range are used. The range of the
transducers was 1,75105 Pa. The atmospheric pressure
patm is measured by a Mensor quartz bourdon tube
absolute pressure transducer, pneumatically connected to
the pressure port in the wind tunnel exhaust. The range of
the transducers was 1,75105 Pa. The stagnation
temperature T0 is measured by a custommade RTD probe
in the settling chamber of the wind tunnel. The pitching
and rolling angle of the model are measured by NPL
resolvers integrated in the model support mechanism. The
accuracy of the pitching angle reading was 0,05 degrees
and the accuracy of the rolling angle reading was 0,25
degrees, [2,3].
The performance of the radar system antenna:
The aerodynamic forces and moments acting on the
model are measured by ABLE 1.00 MKXXIIIA internal
sixcomponent strain gauge balance. The nominal load
range of the balance was 2800 N for normal, 620 N for
side force, 134 N for axial force, 145 Nm for pitching, 26
Nm for yawing moment and 17 Nm for rolling moment.
The accuracy was approximately 0.25% F.S. for each
component. The data acquisition system is consisted of a
Teledyne 64 channels front end controlled by a PC
computer. The frontend channels for flow parameters
transducers were set with 30 Hz, fourthorder low pass
Butterworth filters and appropriate amplification. The
data from all analog channels are digitalized by a 16bit
resolution A/D converter with the overall accuracy of the
acquisition system being about 0,05% to 0,1% F.S. of the
channel signal range. All channels are sampled with the
same 200 samples/s rate, [2,3].
X
q S
Monopulse Phased Array Multi Frequency Doppler
Radar MFDR Antenna,
Output Power: 120 Watts,
Antenna Gain: 40 dB,
Maximum Beam: 10x10,
Minimum Beam: 1x2,
Transmitter Type: Continuous Wave, synthesized
solid state,
Frequency: Xband, adjustable between 10400 and
10550 GHz,
Noise figure: 3dB,
Transmitter Type: Multiple Frequency CW, solid
state PLO, Operation Mode: CW, FMCW, MFCW,
RF
Bandwidth: 10 MHz
The values of the aerodynamic axial coefficient is
calculated on the basis of the negative acceleration, i.e.
retardation of the body in relation to the local coordinate
frame, bounded at the initial point of the flight, according
to the following equation,
Cx ( Mai , sr ) =
(3)
Vi Vi +1
4m
Vi + Vi +1 i , sr S ( xi +1 xi )
(5)
where are, m is mass of flight body projectile, S is crosssection area of projectile, Vi and Vi+1 are measured flight
velocities, xi and xi+1 are horizontal distances, i,sr is
average value of air density and Mai,sr is average value of
flight Mach number.
The axial aerodynamic coefficient in the body axis system
is calculated from the components of aerodynamic force
X, as,
Cx =
The measurement is consisted of the set proving ground
test of trajectory flight measurements at measured initial
conditions of flight and conditions of atmosphere, Picture
3. The results of the measurement are represented as timedependent values of position of the flight body, in polar
coordinates, with time resolution grade of 103 s.
Mach number Ma is calculated using the isentropic
relation:
1
2 p0
1
Ma =
1 pst
OTEH2016
(4)
3.2. The radar inflight tests
The experimental inflight tests are performed in the
facility of the TOC, with system of equipment for field
ballistic radar measurements and the instrumentation for
GPS and atmospheric measurements.
The system of equipment for field radar measurements is
consisted of the 3D ballistic radar, the acquisition system
and the support system. The radar measurement system is
based on Doppler principles, with monopulse phase
comparison and range measurements through integrating
range, Multi frequency ranging, Frequency modulated
ranging and the comparison of principles. The range
measurements are done by integrating velocity, according
to Multi Frequency Continuous Wave (MFCW) and
Picture 3. Trajectory Flight Measurement
The ballistic flight experimental test is consisted of ten
31
NUMERICALANDEXPERIMENTALINVESTIGATIONOFAERODYNAMICCHARACTERISTICSOFSPINSTABILIZEDPROJECTILE
measurements of 40 mm projectile at five different
elevation angles.
Momentum,
( ui u j ) =
x j
4. THE NUMERICAL AND SEMIEMPIRICAL RESEARCH OF
AERODYNAMIC CHARACTERISTICS
The research model is the model of the spin stabilized
projectile with following characteristics:
Reference diameter (caliber) 40 mm,
Total length ~ 5,2 caliber,
Nose length, ~ 3 caliber,
Boat tail length, ~ 0,5 caliber,
Center of gravity from nose, ~ 3,3 caliber.
+
xi x j
ui u j 2 ui
+
ij
+
x j xi 3 xi
(7)
( uiu j )
x j
where p is mean pressure, is mean density, is
molecular viscosity, ui and uj are mean velocities.
Reynolds stresses were given at term,
u u j
uiu j = t i +
x j xi
On the basis of the geometric and dynamic characteristics
of the research model and according to the performed
methods, the aerodynamic coefficients are derived. The
graphics of the characteristics of the aerodynamic
coefficients in relation to Mach number are given in the
paper. Mach numbers represents the characteristics of
flow field, i.e. the velocity of the projectile in relation to
the total atmosphere conditions.
2
u
k + t i ij ,
xi
3
(8)
where t is turbulent (eddy) viscosity, k is turbulent
kinetic energy. To correctly account for turbulence,
Reynolds stresses are modelled in order to achieve closure
of (7). The method of modelling employed utilizes the
Boussinesq hypothesis to relate the Reynolds stresses to
the mean velocity gradients within the flow.
The research deals with numerical simulation of static
aerodynamic coefficients. The governing equations are
given on the basis of Reynolds Averaged NavierStokes
equations, of the steady state flow. The derivatives of the
aerodynamic coefficients are improved in relation to the
measurements and results of the experiments. In this
chapter the results of the numerical calculation of the
aerodynamic coefficients are described, [4,5].
The numerical researches are provided with four
numerical results, presented in the paper as CFD1, CFD2,
CFD3 and CFD4, and described at Table 1.
Table 1. Numerical predictions
Mark
The research of aerodynamic data, presented in the paper,
is consisted of two aerodynamic predictions: the semiempirical aerodynamic predictions (ADP0), [2,3] and
numerical predictions with numerical software of
Computational Fluid Dynamics (CFD) incorporated into
Ansys Fluent software [4].
The prediction research model of body was 0,04 m
referent diameter and 5,2 referent diameter long. The
semiempirical aerodynamic predictions (ADP0) are
performed using aerodynamic prediction technique
presented at [2]. The results of the aerodynamic
prediction technique ADP0 are axial force aerodynamic
coefficients at zero yaw. The values of zeroyaw drag
coefficient are consisted from the results of the
components, according to the body sections and flow
characteristics.
Domain Mesh Type
Quad
mapped
Number
of Cells
Turbulence
Model
75103
2 equation
k RNG
CFD1
2D
CFD2
2D
Hybrid/ Tri19104
Quad
2 equations
k RNG
CFD3
2D
Hybrid/ Tri21104
Quad
3 equations
tkkl
CFD4
3D
Hexahedra
18105
2 equations
SST k
The set of numerical simulations CFD1, with oneequation turbulence model SpalartAlmaras, in 2D
numerical domain, consisted of about 75 000 quadrilateral
cells, is performed at three flow regimes (Boundary Layer
of ~0,025 d size and 1,032 aspect ratio).
The set of numerical simulations CFD2, with twoequations turbulence model RNG k, in 2D numerical
domain consisted of about 19104 triangle cells, is
performed at three flow regimes (Boundary Layer of
~0,015 d size and 1,2 aspect ratio).
The aerodynamic numerical prediction simulations of the
research model of projectile are performed using the CFD
code. The governing equation is based on Reynolds
Averaged Navier Stokes equations (RANS), given by
equations for the conservation of mass and momentum,
and presented in the following forms [4,5,6,7]:
Third set of numerical simulations CFD3 in 2D numerical
domain, consisted of about 21104 triangle cells, is
performed at three flow regimes, with threeequations
turbulence model tkkl (Boundary Layer of ~0,001 d
size and 1,2 aspect ratio).
Continuity,
( ui ) = 0 ,
xi
OTEH2016
(6)
The fourth set of numerical simulations CFD4 is
performed at three sonic regimes, in 3D numerical
32
OTEH2016
NUMERICALANDEXPERIMENTALINVESTIGATIONOFAERODYNAMICCHARACTERISTICSOFSPINSTABILIZEDPROJECTILE
The outer boundaries were set to the free stream
conditions at standard atmosphere for the total
temperature, T = 288 K and the total pressure p = 100 000
Pa. The inner boundary of the model was modeled as noslip, isothermal wall boundary.
domain, consisted of about 1.8 million hexahedral cells,
with twoequation turbulence model SST k (Boundary
Layer of ~0,0002 d size and 1,2 aspect ratio).
The applied turbulence models are, [4,5]:
SpalartAlmaras, one equation turbulence model
(CFD1),
RNG (ReNormalization Group) k, where
additional terms improve accuracy, and represents
turbulence dissipation rate (CFD2),
Transitional tkkl model, where kl represents
laminar kinetic energy (CFD3),
SST (shear stress transport) k, where the turbulent
viscosity is computed through solution of two
additional transport equations for the turbulent
kinetic energy k, and either the turbulence specific
dissipation rate, , (CFD4).
The criteria of convergence were constant values of
aerodynamic coefficients of axial force, within last 100
iterations and the residuals below 106, for CFD1, CFD2
and CFD3 simulations and bellow 104 for CFD4
simulations.
The results of computational fluid dynamic simulation
was obtained through sets of separated calculations for
different Mach numbers of three flow regimes and
different values of angle of attack in the range of 0 to 10
degrees.
The numerical discretization of the computational domain
around the model was designed with mentioned four types
of mesh (Table 1). The computational domain for 2D and
3D meshes is created with longitudinal length of 75 to 80
referent diameter of model and lateral width of about 25 40 referent diameter of model. The spatial discretization
schemes of the equations were second order upwind. The
computational domains are presented in Picture 4.
4. THE ANALYSIS OF RESULTS OF
AERODYNAMIC AXIAL COEFFICIENT
According to the performed research, results are presented
as dependencies of flow regimes, i.e. Mach number, and
also in relation to the angle of attack.
In Picture 5 was presented the axial aerodynamic
coefficient in relation to Mach number. The results of
CFD predictions are marked as CFD1 to CFD4. The
experimental results are: EXPA for aerodynamic wind
tunnel tests and EXPB for ballistic proving ground tests.
The semiempirical aerodynamic prediction results of
zeroyaw drag coefficient are marked as ADP0.
CFD1
CFD2
CFD3
CFD4
EXPA
EXPB
ADP0
0.6
0.5
x0
0.4
0.3
0.2
0.1
0
0
0.5
1.5
Ma
2.5
0.5
CFD4
EXPA
EXPB
ADP0
0.45
0.4
0.35
x0
0.3
0.25
0.2
0.15
0.1
0.05
0
0
0.5
1.5
Ma
2.5
Picture 5. Axial AD coefficient vs. Mach number
The CFD3 computational prediction in 2D domain is
enable fast and qualitative results through all flow
regimes. The CFD4 prediction is shown very good
results, in 3D domain, and enabled the analyses of the
coefficient in relation to the angle of attack. Further
research based on CFD4 prediction is shown good
Picture 4. The part of computational domains:
a) CFD1, b) CFD2, c) CFD3, d) CFD4
33
OTEH2016
NUMERICALANDEXPERIMENTALINVESTIGATIONOFAERODYNAMICCHARACTERISTICSOFSPINSTABILIZEDPROJECTILE
=4,2o
agreement of other static and dynamic coefficients with
experimental results.
0.4
0.35
Ma=0.2 CFD4
Ma=0.2 EXPA
Ma=0.5 CFD4
Ma=0.5 EXPA
160
EXPB
EXPB
140
SIM
Ma=0.7 CFD4
Ma=0.7 EXPA
120
100
y [m]
0.3
80
0.25
Cx0
60
0.2
40
0.15
20
0.1
0
0
500
1000
1500
0.05
0
0
Ma=0.8 CFD4
Ma=0.8 EXPA
0.6
Ma=0.9 CFD4
Ma=0.9 EXPA
5
Ma
Ma=0.95 CFD4
Ma=0.95 EXPA
Cx0
0.3
0.2
0.1
Ma=1.5 CFD4
Ma=1.5 EXPA
Ma=2.0 CFD4
Ma=2.0 EXPA
5
Ma
10
Ma=3.0 CFD4
Ma=3.0 EXPA
Ma=2.5 CFD4
Ma=2.5 EXPA
0.4
Cx0
4000
4500
5000
The trajectory inflight measurements are shown
agreement with simulated trajectory based on the CFD
aerodynamic results. The trajectory with all its elements,
as velocity and angle, are shown same trend and value
level.
0.5
0.3
References
0.2
[1] McCoy Robert L.: Modern Exterior Ballistics,
Schiffer Military History, Atglen PA, 1999.
[2] Jerkovi,D., Samardi,M.: The aerodynamic
characteristics determination of classic symmetric
projectile, pp. 275282, 5th International
Symposium about design in mechanical engineering,
KOD, Novi Sad, 2008.
[3] Jerkovi,D., Ilic,S., Kari,A.. Regodic,D.: The
research of influence of the aerodynamic coefficient
on the stability of the axissymmetric projectile, 4th
International Scientific Conference on Defensive
Technologies, OTEH 2011, Belgrade, Serbia, 67
October 2011.
[4] Ansys Inc., Ansys Fluent Theory Guide, Release
14.0, Canonsburg, PA, 2011.
[5] Masatsuka,K. (2013) I do like CFD, Vol. 1,
Governing Equations and Exact Solutions, CRADLE
and NIA CFD Seminar Series, p. 299.
[6] DeSpirito,J., Silton,S., Weinacht,P.: NavierStokes
Predictions of Dynamic Stability Derivatives:
Evaluate of SteadyState Methods, ARLTR4605,
US Army Research Laboratory, Maryland, 2008.
[7] Silton,S.I.: NavierStokes Computations for a
Spinning Projectile from Subsonic to Supersonic
Speeds. J. Spacecraft Rockets 2005, 42 (2), 223231.
0.1
0
0
3500
The numerical researches of the aerodynamic coefficient
are shown very good agreement with experimental results.
The 3D numerical research with three equations SSTk
is shown qualitative results and enabled analysis in the
relation to the angle of attack. Also, the numerical
research in 3D numerical domain is very convenient for
further research of the aerodynamic coefficient in relation
to the angular velocities (spin) and angle of attack,
separately and coupled.
0.4
3000
5. CONCLUSION
Ma=1.2 CFD4
Ma=1.2 EXPA
0.5
0
0
2500
x [m]
Picture 7. Trajectory in Vertical plane
10
Ma=1.1 CFD4
Ma=1.1 EXPA
Ma=1.0 CFD4
Ma=1.0 EXPA
2000
5
Ma
10
Picture 6. Axial AD coefficient vs. AOA
a) subsonic, b) transonic, c) supersonic regime
In Picture 6 are presented the results of axial aerodynamic
coefficient of numerical prediction RANS SSTk in 3D
numerical domain (CFD4), and the aerodynamic wind
tunnel tests EXPA, in relation to the angle of attack
(AOA) for different Mach numbers, of three groups of
flow regimes.
The differences between experimental and numerical
results are caused by the limitation of the mounting
measuring equipment on the model base, Picture 2.
The deviation of the results at the zero AOA are 3,6%,
and increases with the values of AOA. The deviations are
the smallest in the supersonic flow regime.
The agreement of the experimental field results of the
trajectory and calculated aerodynamic coefficient with
CFD4, are shown in Picture 7 (EXPB1 and EXPB2). The
trajectory of projectile (SIM) is simulated with 6DoF
model with the values of aerodynamic coefficient
obtained from CFD4.
34
A HIGH SPEED TRAIN MODEL TESTING IN T32 WIND TUNNEL BY
INFRARED THERMOGRAPHY AND STANDARD METHODS
SLAVICA RISTI
Institute Gosa, Belgrade, s1avce@yahoo.com
SUZANA LINI
Institute Gosa, Belgrade, sumonja@yahoo.com
GORAN OCOKOLJI
Military Technical Institute, Belgrade, ocokoljic.goran@gmail.com
BOKO RAUO
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, brasuo@mas.bg.ac.rs
VOJKAN LUANIN
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, vlucanin@mas.bg.ac.rs
Abstract: Experimental techniques in wind tunnel tests were always the stateofart technologies and provide high
quality results. In the resent experiments of the highspeed train model, run at the semiopen lowspeed wind tunnel of
the VTI, the infrared thermography (IRT) measurements were introduced. The intention was to observe the flow field
behavior in vicinity of the highspeed train model surface by a multidisciplinary approach. Reynolds analogy make
possible to correlate fluid dynamic to thermal field. By using highsensitivity thermographic systems, temperature
pattern on the surface of test model can be analyzed. Temperature distribution on the model surface is very complex,
varies in the time and space, and depends on many combined effects related with flow and model characteristics. IC
thermography can indirectly perform flow visualization in the boundary layer. Two standard techniques, aerodynamic
drag measurements and the flow visualization with a TiO2 emulsion were used, too. Experiments were performed for
three velocities. The results show good correlation and thus verified the use of IR thermography for the noninvasive
investigation of flow in the boundary layer in precision wind tunnel testing.
Keywords: wind tunnel, highspeed train model, visualization, thermography
When reviewing a development of the highspeed trains
[5], in last decades, it is noticeable that the aerodynamics
steps are in a focus [58]. The aerodynamic characteristics
of the highspeed trains were involving two approaches
windtunnel testing and numerical methods [9]. Although
the tendency is to cutoff the expenditure of the projects,
the windtunnel tests remained irreplaceable. Nowadays,
following of the trends is a necessity, furthermore, in a
field of the railway vehicles trends in aerodynamics meant
lowering of the power consumption and cheaper
transportation. On the other hand, a trend of application of
the stateofart technologies of the measurements needs
multidisciplinary research. At the VTI for a decades, upon
a significant experience in the fields of a nonaeronautical
windtunnel research [10], contemporary measurement
techniques, as here presented IRT, were applied. As
overlapping technique, the oil emulsion visualization, was
used for a justification of the CFD observations, as a
necessary step in aerodynamic practice [11].
1. INTRODUCTION
With the intention to contribute to the challenges in a field
of highspeed train aerodynamics, wind tunnel tests,
applying the IRT on an initial model of the highspeed
train, were performed in the wind tunnel T32. A task was
to observe the applicability of the thermography as a
testing technique in the wind tunnels at lowspeed
conditions. Furthermore, performed tests gave the insight
to the temperature distribution of the simplified highspeed train in the presence of the ground. The results are
used as the input parameters for the computational fluid
dynamic, CFD, observations at the bionic design process
of the highspeed train [1].
The wind tunnel facilities and testing techniques were
generally developed upon the highest standards with a
purpose of advanced researches of aeronautical and
missiles models and fluid phenomenon from subsonic to
hypersonic speeds [24]. Development in other areas
perceived and used benefits of the wind tunnel testing
and, by time, many of nonaeronautical tests were
developed from the buildings, over the racecars to the
highspeed trains [5].
The IRT, become a very helpful noninvasive temperature
measuring method for instance in the cases of: the object
surface temperature and insulation checks of the industrial
facilities [12]; the energy efficiency and building
35
OTEH2016
AHIGHSPEEDTRAINMODELTESTINGINT32WINDTUNNELBYINFRAREDTHERMOGRAPHYANDSTANDARDMETHODS
between the model surface and the surrounding air. A BL
over the flat plate is consisted of three zones: laminar
from the leading edge of the plate, further, transition and
afterwards turbulent regime. In accordance to the
Reynolds analogy [1618], the velocity boundary layer
(VBL) and corresponding thermal boundary layer (TBL)
have the similarly defined BL thicknesses. Both, for the
VBL and TBL the BL edges thickness are defined as
values near the free stream values (0.99V and 0.99T).
Reviewing the flow, the VBL exists right from the leading
edge, while the TBL a bit later, thus the flow over the
small length from the leading edge has the equal value to
the free stream temperature. A development of the VBL
and the TBL results, for both, in thickening (Picture 1a.),
from laminar VBL (parabolic type of the edge) to a
turbulent VBL (Picture 1b.), but the characters of velocity
and the temperature profiles differ (Picture 1c.).
Depending of the flow conditions, at a transition point,
flow starts transformation, from the laminar to the
turbulent flow. Inside the turbulent flow, a buffer layer
and viscous sublayer exist and grow along the turbulent
layer. However, temperature values, as well as the
thickness of the TBL differ to the flow regimes thus the
transition point might be detected by temperaturedifference measurements.
insulation of the museums, heritage buildings [13]; also a
material loading behavior as well a defect or crack
recognition. Nowadays, for wind tunneltesting, the IRT
is of the great support for insight of the fluid flow
convective heat transfer [24] and boundary layer
transition [14,15]. The main advantage of the IRT in the
windtunnel testing are the testbudget savings because
the IRT gives very fast, transient and associative data
upon which advanced, and reciprocally high time/cost,
experiments of the boundary layer are employed.
2. HEAT TRANSFER IN THE BOUNDARY
LAYER
The air wetting the smooth flat plate model surface in twodimensional case, for the simplicity, in a general case of an
incompressible flow is considered. The energy equation
leads to an energy balance that is subjected to the internal
energy e ( e = cvT ), the conduction of heat (caused by
collision of the molecules of the fluid), the convection of
heat by a flow of fluid (caused by the flow streaming over
the surface in contact) and the heat sourced by a friction [1719]. The convection in flow is recognized in two types as:
natural, when the flow was powered by the density
difference caused by the heat transfer between the fluid and
the model, and forced convection that is powered by the
mechanical source of fluid motion.
The conduction is generally described by the Fouriers
Law where the coefficient of heat transfer between the
model and a fluid in contact, qconduction, is given in (1).
( )
qconduction = k T
n
n=0
(1)
(a)
where k denotes a thermal conductivity and ( T n )n = 0 is
a magnitude of the temperature gradient in direction
normal to the model surface [17,18]
The forced heat convection involves the air motion and
the heat conduction, is described by the Newtons Law of
cooling [17] defining a convective heat transfer, qconvection,
as in (2)
qconvection = h Tsurface T , [W/m2]
(b)
(2)
Picture 1. (a) Details of the VBL and the TBL, (b)
laminar and turbulent velocity profiles with corresponding
thicknesses, l and t, respectively and (c) different
temperature profiles at the TBL and corresponding TBL
thickness, T
where h is heat transfer coefficient, brackets define a
difference of the surface temperature, Tsurface, and a
stagnation temperature of the flow, T. The convective
heat transfer coefficient, h, depends on the air properties,
flow characteristics and quality of the model surface,
especially depends on its roughness.
Therefore, the sum temperature of the model surface is a
combination of different heat transfer types. The influence
of the heat transfer by radiation, from the other bodies
near the model, is not significant in considering flow and
temperature behavior.
Furthermore, the phenomenon of the momentum and heat
transfer, described by the Reynolds analogy is valid for
the air, while the Prandtl number is unit [1618].
kf
h = 1 c f V
, c flam 0.664 , c fturb 0.027
2
Re
Re
(c)
3. IR THERMOGRAPHY
(3)
IRT is based on the measurement of the thermal energy
radiated from the model surface. The measurement of
temperature distribution over the modelwetted surface by
IRT, differs from the other measuring technique. The
kf is a thermal conductivity of the fluid, v is a kinematic
viscosity, Re is Reynolds number.
Boundary layer (BL) flow conditions affect the heat flux
36
AHIGHSPEEDTRAINMODELTESTINGINT32WINDTUNNELBYINFRAREDTHERMOGRAPHYANDSTANDARDMETHODS
OTEH2016
radiation of a near bodies and heat sources has to be
involved as an influence on the measurements.
intensities from all of the three described influences
produce total radiation intensity defined as (4)[13]
All the bodies, by the temperature above absolute zero,
thermally radiates continually in time. The heat transfer by
radiation is specific in comparison to other two types,
conduction and convection. It is caused by an
electromagnetic emission and absorption, and is occurring
either in the vacuum. Furthermore, the velocity of thermal
radiation equals the speed of light with energy proportional
to a fourth power of the surface absolute temperature [13].
The total radiated energy, obtained from the Planks Law,
is described by the StefanBoltzmanns Law
Etotal = Eobj + (1 ) Eenvir + (1 ) atm
E T = T 4 , W2 ,
m
(5)
The Picture 2. represents the scheme of the IRT
measurement. The model, surrounded with the
environment, emits the radiation in the all directions. The
intensity of the radiation that is measured by the IRT
camera is described by the eq. (5). For the valuable results
four main parameters matters: thermal sensitivity of the
IRT camera (signal overstepping noise), scanning rate
(frame rate) for following the flow dynamics, resolution
of imaging (for high quality data recording), and intensity
of resolution [16].
(4)
where = 5,6697108 [Wm2K4] denotes Stefan
Boltzmann constant and T [K] is a surface absolute
temperature. Total radiated energy is lower for the real
body comparing to the black body [17].
Emitted radiation is dependent on the body temperature.
Radiation laws are defined for the thermally black body
[13], ideal body, but real bodies i.e. their surfaces, have
different characteristics. By that, radiation of the real
surfaces are affected by a surface material, surface
finishing quality, radiation wave length, angle of emission
or absorption, spectral distribution of the incoming
radiation and transparency. In the contrary to the black
body, real bodies behave according to their emittance and
the characteristics of IR radiation, like gray (=const.) and
nongray  spectral bodies =(). The spectral emissivity
coefficient, = E ( ) Eblack ( ) , presents the body ability
of radiation and is defined as a ratio of the emissive
powers of the real, E(), and the black body, Eblack(). As
the emissive power of the real body is less than the black
body, the emittance for the real body is in the range from
0 to 1, while for the black body is unit.
Picture 2. Illustration of the method of temperature
measurement by the IR thermography
The IR detector has a significant role in defining the
inputs, thus it is the source of the signals, for which is
required to have as low as possible noise. Beside the IR
detector, on the signal quality influences also the
difference of the phenomenon of flow transition. The
frame rate of the camera gives the possibility to follow the
flow dynamics. Unfortunately, after flow dynamics
influence on the thermal response, also the thermal
characteristic of the model, heat capacity and
conductivity, are influencing to overall phenomenon
dynamics. Knowing all these facts, the special care have
to be taken on the measurements of the temperature by
IRT, model manufacturing purposed for thermal imaging
and the environmental conditions adjustments [15,16].
IR thermography is a temperature measuring method that
involves use of the IRT cameras (Picture 3.) and
commonly is called thermal imaging. The IRT camera is
collecting the radiation from the three different sources:
(a) from the body surface of interest (dependence of the
body temperature); (b) from an environment in which the
body is placed for imaging (this radiation was actually
reflected from the body of interest but is sourced from the
nearbodies) and (c) from the atmosphere. The
atmosphere, by means of IRT, presents the air
environment in witch the body is immersed and the
atmosphere in witch the IRT camera is fixed (as both are
the sources of the radiation and the absorbers). Radiation
from the body surface is defined as Ebody , where: is
4. DESCRIPTION OF THE EXPERIMENT
Wind tunnel facility. The thermal imaging tests were
done at the closed circuit, semiopened lowspeed wind
tunnel (WT) T32, at the VTI, Belgrade. A facility was
intentionally built for testing aeronautical models, but
after test section adaptations, it was prepared for this nonaeronautical testing. Picture 3 shows the facility the from
the side view. The flow is continuous, forced by the DC
motor propeller manually controlled, in a range up to 70
m/s. A test section, TS, is opened, elliptical in a crosssection (1.8 m 1.2 m ) with length of 2 m. The internal
TS surfaces are coated with matte color. The estimated
turbulence factor at the T32 WT is TF=1.14.
emittance, and is transmissivity of the atmosphere, Ebody
is radiation intensity at the objectmodel temperature, Tobj,
corresponding to the black body. The radiation intensity
emitted from the near bodies is described as (1 ) Eenvir ,
where: (1) is surface emissivity, and Eenvir is the black
body environmental radiation intensity, at temperature of
the atmosphere. Further on, (1)atm is the radiation
intensity sourced by the atmosphere, where (1) is
atmosphere emissivity. Summarizing, the radiation
The testing set up in the cross view is presented in
Picture 4. The model is positioned at the center of the test
section, TS by the struts (front and rear). The ground was
made from a wood, coated with simple glossy lacquer. In
longitudinal direction, metal profiles were jointed to
37
AHIGHSPEEDTRAINMODELTESTINGINT32WINDTUNNELBYINFRAREDTHERMOGRAPHYANDSTANDARDMETHODS
OTEH2016
appropriate for these tests. The recipe of the emulsion
contained 30ml of the paraffin oil, 5ml of the Olean acid
and 10g of TiO2. The upper and side model surfaces were
covered with the emulsion spots.
prevent ground deformations under the flow influence
The experimental procedure was divided into the
temperature measurements and the visualization
separately. For each of the test types, the task was to set
the flow conditions through the shortest time, but
respecting the abilities of the power unit of the WT. The
significant note for testing procedure is connected with
the test run time and test brake duration. For visualization
test, standard test run were performed, restarting just after
model test preparation, but the test for thermal imaging
needed test brakes for model cooling down to the
environmental temperature. An each thermal imaging test
run started before the WT run by capturing the null image
for checking of the actual model thermal condition and
the results analyze.
Picture 3. Lowspeed semiopened wind tunnel
SHST model. The SHST model was designed for the
force measurements and flow visualization, made from a
piece of wood, 1m long, in geometry presented in
Picture 5. The SHST finishing coating consisted of one
base coating over the wooden surface and two layers of
black matte lacquer. The SHST surfaces were polished in
finishing process to uniform roughness.
5. RESULTS AND DISCUSSION
The test results were obtained for these cases: (A) the WT
stepping acceleration, the model was at the beginning on
the room temperature (thermal images were recorded after
the flow conditions were stabilized on each of the selected
velocity steps; (B) the WT continual deceleration (the
frame velocity data were marked while measuring) and
(C) WT continual acceleration (the framevelocity data at
selected velocities were marked while measuring).
The diagrams on the Picture 6. correspond to the
measuring lines L1L4 from the Picture 5 (L4 was placed,
from the perspective, along the SHST roof). A first along
the SHST side, while another 7 were in an array along
lines to the trailing edge. The Picture 6. shows
temperatures, along the measuring line L1 (spots C1C9)
during stepping acceleration, when SHST was at room
temperature at the beginning. Warming of the model is
linked to warming of the SHST surface, produced by
flowsurface thermalmomentum interaction (friction,
convection, and conduction) inside the boundary layer.
Furthermore, uniformity of temperatures is pointout to
the turbulent BL existence over the SHST sidewalls.
Picture 4. The SHST model and the IRT equipment set
for tests (L=2m)
Picture 5. The SHST model geometry
IRT camera. For the temperature measurements the
FLIR E40 SC camera was used. The main IRT camera
characteristics are: measuring range 20C to 650C (with
two selectable options from 20C to 650C or 20C to
120C) a resolution 160 x 120, thermal sensitivity <
0.07C, standard optics of 25 x 19, the measuring error
of +/ 2% or 2C, bolometer wit no cooling, frame rate
30Hz, manual focusing, zoom selection 12 times.
Thermal images, thermograms, were read by specialized
software FLIR Researcher and FLIR Tools.
Picture 6. Temperatures measured along L1, at spots
C1C9, V=055 m/s (from the thermograms).
The oil emulsion visualization was applicate as a most
Dependences of the temperature against velocity, again
38
AHIGHSPEEDTRAINMODELTESTINGINT32WINDTUNNELBYINFRAREDTHERMOGRAPHYANDSTANDARDMETHODS
over L1 and C1C9, is showing intensive temperature
increase of the SHST for V>40m/s (Picture 7).
OTEH2016
deceleration in range V=550 m/s and acceleration in
range V=055 m/s (from video capturing)
(a)
Picture 7. Temperatures along L1, C1C9, stepping
acceleration ,V=055 m/s (from imaging)
The tendencies of the surface temperature behavior,
measured along L1, during WT continual deceleration and
acceleration are following dynamic of the velocities
changes (Picture 8). Continual deceleration was manually
controlled by gradual decreasing of velocity, the surface
temperature followed it but with notation that the
temperature at the end of this process was higher than the
room temperature. After the time brake of 10 min
between runs, the WT was started again, now with
continual accelerating in dynamics to intentionally control
the velocity at exact values. After velocity was set, a
thermogram was recorded, and acceleration continued.
(b)
WT continual deceleration, SHSTjust after the test run,
from video sequences.
The Pictures 9ac represent the comparative view of the
flow visualization and thermograms for three velocities
40 m/s, 50 m/s and 55 m/s. Flow visualization justified
the turbulent BL existence while thermograms justified
the heating of the SHST surfaces during the acceleration.
(c)
Picture 9. Comparative view of the flow visualization
and thermograms at (a) 40 m/s, (b) 50 m/s and (c) 55 m/s
6. CONCLUSIONS
This paper presents the results of tests for experimentally
flow visualization and temperature measurements on the
model surface, using IRT, applied to the highspeed train
model in lowspeed wind tunnel.
Obtained IRT results are in good correlation with the oil
emulsion flow visualization. Measurements, made during
this experiment, could not be used for precise
determination of the boundary layer transition, because
the model is small in size, smoothsurface finished and
tested in the low speed flow. However, the IRT results
show the temperature changes during the wind tunnel
runs. The estimated temperature difference between
laminar and turbulent flow is less than 1C, what is less
than temperature arise due to an airsurface friction.
Picture 8. Temperatures along L1, C1C9, continual
39
AHIGHSPEEDTRAINMODELTESTINGINT32WINDTUNNELBYINFRAREDTHERMOGRAPHYANDSTANDARDMETHODS
OTEH2016
2D Flow Fields for the Bionic HighSpeed Train
Concept Designs Inspired with Aquatic and Flying
Animals, Proceedings of the 6th International
Scientific Conference on Defensive Technologies,
OTEH 2014 (2014) Military Technical Institute,
Belgrade, Serbia, 4449,
[10] Puhari,M., Luanin,V., Lini,S., Mati,D.: Research
Some Aerodynamic Phenomenon of High Speed
Trains in Low Speed Wind Tunnel, Proceedings of
the 3rd International Scientific and Professional
Conference CORRIDOR 10  A sustainable way of
integrations, October 25th (2012) Belgrade, Serbia,
Risti,S.,
Stefanovi,Z.,
Kozi,M.,
[11] Lini,S.,
Ocokolji,G.: Experimental and Numerical Study of
SuperCritical Flow Around the Rough Sphere,
Scientific Technical Review, 65 (2) (2015) 1119.
[12] Kozi,M., Risti,S., Katavi,B., Lini,S., Risti,M.:
Determination of the Temperature Distribution on
the Walls of Ventilation Mill by Numerical
Simulations of Multiphase Flow and Thermography,
Proceedings of the 5th International Congress of
Serbian Society of Mechanics, June 1517. (2015)
Arandjelovac, Serbia,
[13] Risti,S., PoliRadovanovi,S.: Termografija u
zatiti kulturne batine, Institut GOA, Jan 15, 2013.
[14] Carlomagno,G.M., Cardone,G.: Infrared thermography for convective heat transfer measurements,
Exp Fluids, 49 (2010) 11871218.
[15] Crawford,B.K.,
Duncan,Jr.G.T.,
West,D.E.,
Saric,W.S.: LaminarTurbulent Boundary Layer
Transition Imaging Using IR Thermography, Optics
and Photonics Journal, 3 (2013), 233239,
dx.doi.org/10.4236/opj.2013.33038
[16] Simon,B., Filius,A., Tropea,C., Grundmann,C.: IRThermography for Dynamic Detection of LaminarTurbulent Transition, 18th International Symposium
on the Application of Laser and Imaging Techniques
to Fluid Mechanics, July 47, (2016) Lisbon,
Portugal
[17] Lienhard,IV,J., Lienhard,V,J.: A Heat Transfer Textbook, Phlogiston Press, Cambridge Massachusetts,
2008.
[18] Schlichting,H.: BoundaryLayer Theory, 7th edn.,
McGrawHill, New York, 1979. ISBN 0070553343.
[19] Anderson,J.D.Jr.: Fundamentals of Aerodynamics,
McGrawHill Series in Aeronautical and Aerospace
Engineering, 2nd ed., 1991,
Further development of model design and testing
methods, are recommended. The application of IRT in
wind tunnel experiments and use the high sensitive type
of IRT camera can record thermograms that are more
precise, over the higher resolution and thermal sensitivity,
thus it is recommended for the BL research at the low
speed wind tunnels.
ACKNOWLEDGMENTS
The authors are grateful to the Ministry of Education,
Science and Technological Development, Republic of
Serbia, for supporting this work as a part of the researches
through the financed Projects TR35045 and TR 34028.
REFERENCES
[1] Rauo,B.: Bionics in Design, Faculty of Mechanical
Engineering University of Belgrade, (Belgrade)
eBook on CD (in Serbian), 2014.
[2] Pope,A., Harper,J.: LowSpeed Wind Tunnel Testing,
Wiley, Jan 1, 1966.
[3] Mrkalj,N., umonja,S.: Ispitivanje modela sa
prostrujavanjem u aerotunelu T32, Scientific
Technical Review, XLVI (45) (1996) 5159.
[4] Ocokolji,G., Damljanovi,D., Rauo,B., Isakovi,J.:
Testing of a Standard Model in the VTIs Largesubsonic Windtunnel Facility to Establish Users
Confidence, FME Transactions 42 (2014) 212218.
[5] Raghunathana,R.S.,
Kimb,H.D.,
Setoguchi,T.:
Aerodynamics of highspeed railway train, Progress
in Aerospace Sciences, 38 (2002) 469514.
[6] Baker,C.: The Flow Around HighSpeed Trains,
BBAA VI International Colloquium on: Bluff Bodies
Aerodynamics & Applications, July 2024 (2008),
Milano, Italy,
[7] Baker,C.J., Brockie,N.J.: Wind Tunnel Tests to
Obtain Train Aerodynamic Drag Coefficients:
Reynolds Number and Ground Simulation Effects,
Journal of Wind Engineering and Industrial
Aerodynamics, 38 (1991) 2328.
[8] Cheli,F., Rocchi,D., Schito,P., Tomasini,G.: Steady
and moving high speed train crosswind simulations.
Comparison with windtunnel tests, Proceedings of
the 9th World Congress in Railway Research, May
2226 (2011) WCRR Lille
[9] Linic,S.,
Rasuo,B.,
Kozic,M.,
Lucanin,V.,
Puhari,M.: Comparison of Numerically Obtained
40
AERODYNAMICS OF THE HIGH SPEED TRAIN BIOINSPIRED BY A
KINGFISHER
SUZANA LINI
Institute Gosa, Belgrade, sumonja@yahoo.com
BOKO RAUO
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, brasuo@mas.bg.ac.rs
MIRKO KOZI
Military Technical Institute, Belgrade, Serbia, mkozic@open.telekom.rs
VOJKAN LUANIN
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, vlucanin@mas.bg.ac.rs
ALEKSANDAR BENGIN
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, abengin@mas.bg.ac.rs
Abstract: With the intention to contribute to aerodynamic optimization of transport vehicles, the high speed train was
biomimic with a kingfisher. Presented bionic concept design was defined after a series of hydrodynamic and numeric
experiments of bionic model of the kingfisher. The aerodynamic characteristics of the high speed train were observed in
three configurations, under various railing conditions, as follows: in a free flight (up to critical local conditions), in a
presence of ground and passing through the tunnel. Obtained results presents the main aerodynamic characteristics of
the bionic highspeed train, and in further, analyzed from the view of a design source. Adequate results were compared
with concept designs bioinspired with other animals showing the best performances after biomimic with the kingfisher
up to velocities of 400km/h.
Keywords: aerodynamics, bionic, CFD, optimization.
model geometry, in 2D space, that is hypothetic, but
represents the characteristics of the initial selfsimilar
profile in longitudinal direction under advanced
conditions. All the others selfsimilar nose/tail cones
crosssections, on the future 3D BHST, were assumed to
be much sharper and of less area than the longitudinal
one. Presented cases were wallbounded and introducing
the flows more complex than in the reality. In the first
place, the idea was to apply experiences and knowledge
from aircraft aerodynamics and design in various ranges
of flight that were not represented up to know, to a design
process of the highspeed trains. Combining selected
geometries in two orthogonal and longitudinal directions,
created upon their flow responses is expected to give in
further research a final bionic design ready for wind
tunnel testing and numerical justifications [18,19], with
intention to predict the flow using a new method until real
scale rail run [20].
1. INTRODUCTION
Although very frequently pointed biomimicry with the
kingfisher as a model of aerodynamic modeling of the
highspeed train nose [1,2,3,4,5], at the contemporary
literature in a field of biomimicry and bionics [6] no
further details were reported. With the motivation to
apply biomimimcry on the industrial design, in this work
on the bionic highspeed train, BHST [7,8], we found
necessary to develop knowledge [9] and collect data [10]
about the phenomenon of the kingfisher in plungediving
[11], as a design base and to justify the artificial case by
the hydrodynamic tests [12,13,14]. The same parametric
model used for manufacturing the bionic kingfisher in a
freefall waterentry, was used for design of the numerical
BHST in real scale, after adjustment of its contour.
Furthermore, the BHST was forced to run in different
hypothetic conditions. To understand the base
aerodynamic characteristics of the BHST, with drag and
pressure distribution in focus [15], the BHST was
released to a free flight from subsonic to transonic
velocity ranges [16]. Afterwards, the BHST was grounded
to run in open rail, and at last to pass the infinite tunnel.
The flow conditions were set to the ranges of interest
needed for further design adjustments. One may note that
this work in based on parameters of the flow velocity and
In this work, from the BHST we have required answers
about geometry quality from upon the flow reaction [7,8].
The available sources commonly are based on the
transient flow characteristics over the 3D models, for
observing the passing train through the tunnel, just a
small number are covering 2D or axissymmetric cases,
like is the observation of the tube train [21].
41
AERODYNAMICSOFTHEHIGHSPEEDTRAINBIOINSPIREDBYAKINGFISHER
OTEH2016
selection was made for reaching good manufacturing
quality and quality nonmacro video captures of the
phenomenon during the hydrodynamic tests.
2. NUMERICAL MODEL OF THE BHST
The numerical model of 2D BHST was created by reengineering of a kingfisher longitudinal profile,
contouring layered images from the Nature. The software
PRO/Engineer WF4 was used for creation of 2D body
contour. The contour, adopted as a longitudinal crosssection of maximal area, was translated into the formatted
text file, imported to the CFD space for mesh generation,
by the ANSYS Fluent Package. In this process, one data
source was a selected profile photo of the kingfisher,
Natural History Museum exhibit (Picture 1), while other
two, were used for positioning of the model relative to the
water, reconstructing the body pose the kingfisher during
water entry (Picture 2a)[7,8,9,11]. In Picture 2b, the
hydrodynamic model was shown with signed support
items. The wooden model simulated the kingfisher body
in the shape and mass [22].
The test model simulated the shape and dynamic
characteristics of the kingfishers rigid body. The center
of gravity was estimated by the PRO/Engineer tools and
those data were applied to the test model. Picture 2b is
shown the scheme of the model with its equipment: 1the
bodyof the bionic model made from the layers of balsa, 2fixation directioning strings, 3releasing string, 4lead
weight inside the model, at the estimated center of
gravity, setting the rigid body dynamic similarity, 5fixation of the string.
3. HYDRODYNAMIC EXPERIMENT
The hydrodynamic tests were based on methods used for
testing a marine structures [12,13,14]. The test model was
set to accomplish the freefall water entry with zero
deadrise angle, vertically, by fixed sliding strings. From
the height of 150mm over the water surface (measured
from the beak tip), the test model was released to slide
along strings and to enter water in a tank filled with clear
and calm water with average temperature of 260C. During
the water entry the videos were captured by a Samsung
ES95 (30fps), for about 10 repeats. The representative
sequence of the bionic kingfisher water entry is shown in
Picture 3.
Picture 1. The female kingfisher (museum exhibit from
the Natural History Museum, prepared in year 1939.)
Picture 3. Water entry of the bionic kingfisher model
(a)
(b)
The average estimated impact velocity, from the video
sequences, was 0.6m/s and a result shown good
correlation with the images from the Nature [11]. This
experiment justified, the statements from the literature
[2,3,4], the norippling water surface behavior during
impact. Furthermore, the biological design of the
kingfisher beak was shown norippling water surface over
the whole beak length at velocity about 0.6m/s, and by
that means classified the geometry acceptable for
application on the bionic highspeed train, BHST.
Picture 2. Models of the kingfisher (a) a reengineered 3D body and (b) a 2D bionic model for the hydrodynamic
tests (1wooden bionic model, 2directioning strings, 3releasing string, 4lead weight, 5 fixation of the strings)
The layered images were combined to form a plungediving pose of the kingfisher. A selection of body size,
dependent on genetic and environmental conditions, was
the maximal distance between the beak tip and tail end.
Reference line followed the overlapping line of the beak
parts. Adopted model length was 190mm and weight
180g. One may note that model was overdimensioned for
domestic environment (about 120mm), but such a
The longitudinal contour of the BHST contained two
faced and symmetrical power cars, in gross length of 50m
and height of 3m, with nose and tail geometries scaled
42
OTEH2016
AERODYNAMICSOFTHEHIGHSPEEDTRAINBIOINSPIREDBYAKINGFISHER
from the bionic numerical kingfisher. The nose and tail
contours, from the free ends were crossed with the BHST
train body shape by tangent radiuses, both, over the upper
and lower sides, resulted in the nose elongation of 4.471.
4. CFD EXPERIMENT
The CFD experiment was made by use of the ANSYS
Workbench 12 [23]. Geometry of the computational space
was made by the ANSYS Geometry software with
imported BHST contour in a form of organized text file
data.
Picture 5. The mesh around the BHST in the presence of
the ground and a detail with inflation layer cut
The CFD observations of the BHST were made for three
configurations: in free flight, on the open rail and passing
an infinitive tunnel, and observed under conditions at
three nominal Mach numbers M=0.2; 0.3 and 0.4.
The BHST bounded with tunnel requested much denser
mesh to answer the mesh quality requirements. Created
inside the domain with the same longitudinal dimensions
as others, height of 8m, the mesh consisted of 1537287
elements finally after the adaption (Picture 6). Beside the
fixed dimension of the tunnel length it will be treated, in
an analyze, as an infinite because no changes of geometry
conditions would occur under the set of steady flow
conditions. Mesh quality is described by maximal squish
of 0.57, skewness 0.78 and aspect ratio 18.
4.1. Geometries and meshes
The elliptical domain surrounded the BHST model in free
flight. As shown in Picture 4, smaller axes is a symmetry
axis, placed horizontally. The inlet domain boundary is
150m from the BHST nose tip, while its outlet boundary
is 300m behind the BHST nose. The vertical halfaxis
height is 290m. For better mesh quality, the domain was
parceled into smaller surfaces following the BHST
contour. The resulted unstructured mesh consisted of
221284 cells, skewed up to 0.64 with aspect ratio up to
50.
Picture 6. The mesh inside the infinite tunnel
4.2. Numerical methods
For the purpose of understanding the phenomenon and
compare the results the first CFD calculations were made
with the same set up, as suggested [23,24] that close
results are workable even for differently sourced solution
methods nowadays (pressure or densitybased).
Nevertheless, in some cases the results have had not
follow the same behavior, why they were substituted.
Therefore, in those cases with assumed neglectable
compressibility effects, the pressure based solution
method was selected, and vice verse, were the
compressibility was assumed  the densitybased solution
methods. In accordance to the solution methods, the
pressurevelocity coupling was selected to SIMPLE,
spatial discretization defined for gradients as the GreenGauss Node Based, standard pressure and second ordered
other parameters from the pressure to the energy.
Picture 4. Geometry of the domain around the BHST in
the free flight
The ground presence was treated through the rectangular
domain, keeping the same dimensions along the ground
(inlet distance from the nose was 150m, outlet of 300m),
height was 75m with the similar parceling as previous.
The BHST was lifted from the ground for 0.4m. Picture 5
represents the details of the mesh. The mesh for nearground configurations consisted of 383438 cells without
adaption, and with introduced adaption 1030505 cells (in
the compressible cases). Firstly made nonadapted mesh
was arranged to be denser over the BHST body and over a
ground by support of the two, elliptical and flat, bodies of
influence. Later adaption, through a solver tools, was
made by a rectangular field. All the mesh constructions
involved 10th layered inflation layer over the BHST
wetted surfaces.
Numerical models in use were the k Realizable, RKE,
and SpalartAllmaras, SA. First one was used for a
treatment of the wallbounded flows over the BHST
passing the tunnel and incompressible flows over the
BHST on the open rails, reasoning to ensure the quality
wall treatment, involving Non Equilibrium Wall
Function. The use of RKE was not motivated with the
calculation time economy as the problem are far less
complex than BHST passing the tunnel, for example,
under velocity of M=0.4 it was required to run over 3100
iterations to rich the convergent solution. The SA was
used in the cases of the free flight and compressible flow
on the open rails, motivated with quality/time
43
AERODYNAMICSOFTHEHIGHSPEEDTRAINBIOINSPIREDBYAKINGFISHER
OTEH2016
optimization of the calculations thus the priority was
BHST in the tunnel. Furthermore, either one RANS
model was in use, if the compressible flow assumed the
energy equation was involved, the air was set as an ideal
gas, of viscosity described by the Sutherland third order
function. Boundary conditions were involved as
following: inlet  pressure far field, PFF; outlet pressure
outlet, ground  moving wall and top symmetry wall
while the tunnel walls were both moving walls. All the
moving walls have had equal velocities as the actual
BHST, treated as noslip.
In cases of density based solver, the implicit formulation
with RoeFDS flux type was selected, as well as the
Green Gauss Node Based discretization together with the
second ordered flow and modified turbulent viscosity
functions. The calculations were however supported with
the solution steering, set to pick the Courant numbers in a
range of 0.01 to 200, and default explicit underrelaxation
factor of 0.75.
Picture 8. Cp vs. M, BHST in free flight, lower side
The ground effect on the BHST was presented in Picture
9. The presence of the shock wave underneath is
noticeable on the first sight while the pressure distribution
over the upper side is similar over the range of M. The
phenomenon was occurred at M=0.4, but its initiation
seems to be started at M=0.3, according to actual
observations, from the compressible waves. One may note
the local similarity under the tail to the flow behavior
through the nozzle. Nevertheless, here the long channel
was a kind of the entrance tank, continually filled with a
flow from the entrance cone, formed by the nose curve,
with a gradual decrease of height.
5. RESULTS AND DISCUSSION
The functions of the pressure coefficient, Cp [25], plotted
against Mach number in a range of subsonic and transonic
velocities, from M=0.2 to M=1, for the BHST in the free
flight, were shown in Pictures 7 and 8 (by order for upper
and lower BHST side). The investigation, imagined to
overview flow character by applying, if it may be called,
Mach lens. An increase of the M was expected to left
traces, from initial cp disturbances to the shock waves,
favorable out of the nose or tail zones.
Picture 9. Cp vs. M, BHST on the open rail, upper and
lower side
Picture 7. Cp vs. M, BHST in free flight, upper side
The good agreement in cp behavior with similar airfoil
was present [16], while the flow disorders were spotted in
zones where the BHST nose is narrowing the BHST
body. The increase of M, up to 0.7, caused the presence of
compressible waves over the upper BHST surface.
Further M increase gradually transform the waves to
shock waves, in range 0.7<M<0.8, making them stronger
and moving along the stream from the entrance of the tail
cone to its tip. On the other side, increase of the M just
over 0.80.9 started to produce the shock waves of similar
strength as previous, near the BHST trailing edge (Picture
9) as a response on the sharper lower tail side, smaller
semiangle of the tail cone.
The comparison of cp against M, over the wetted BHST
body passing the tunnel was shown in Picture 11. The
presence of shock waves, at M=0.4, underneath the BHST
is noticeable and compared to the case of BHST on the
open rail, the stronger one was occurred in the tunnel than
just the ground. If compared to flows over the BHST in
the infinite tunnel similarity was found between flows
under the train, in difference that wallbounded flow
decreased the cp in further and moved it to the tail zone.
From the upper side, however, compressible zone was
formed over the central upper surface.
From the other side, cp distribution over the tunnel ground
44
OTEH2016
AERODYNAMICSOFTHEHIGHSPEEDTRAINBIOINSPIREDBYAKINGFISHER
(Picture 12) was shown the similarity in flow behavior in
comparison with the tubetrain [21].
compressibility and at last formation of the shock waves
underneath the BHST. Further investigations involving
increase of M were not reasonable without either drastic
design changes of the BHST underneath contour, in the
manner of sailfishlike or barracudalike designs, or
better, fully closing the channel for simulation purposes.
Comparing cd for kingfisherlike designs, one near ground
is lower that one in the tunnel, up to M<0.3, after which
due to wallbounding from both sides cd become smaller
during passing tunnel that on the open rail.
Picture 10. Cp vs. M, BHST passing the tunnel, lower
side
Picture 12. Cd vs. M for various designs in the presenc of
grouns and the kingfisherlike for three configurations
free flight, open rail and passing tunnel
6. CONCLUSIONS
In this work the BHST was set to imagined conditions for
a series of designs. The kingfisherlike bionic design was
in the focus. The kingfisher beak design justification was
done by the hydrodynamic tests and afterwards scaled and
adjusted to the real scaled BHST. The numerical
observations have had the aim to find the answers linked
to the potential design problems. The present results of
the hydrodynamic tests were encouraging thus a
development of test procedure and analyze will be
continued in the future work. For presented kingfisherlike BHST design, was found that the method pointed to
zones that has to be redesigned, mostly on the lower side
of the BHST. However, the method using Mach lens is
applicable in the early stages of BHST designing, during
the profile selection phase, and is effective and timesaving. The aerodynamic characteristics of the bionic
kingfisherlike design shown the good potentials and
technological applicability, next to it is the barracudalike
design. Other observed designs have had either not
desirable aerodynamic characteristics or they are not
suitable for conventional rail vehicle construction.
However, kingfisherlike design presented the stabile
characteristics over the M domain, but further
improvements are necessary. These improvements are
referred to both the design and designing method by
means they should lead to avoidance of the shock waves
underneath the BHST in operating Mrange. A further
grading of the velocity, reasonable solution will be
observed in complete grounding of the BHST, so as it
would become the complex parametric wedge; the
Picture 11. Cp vs M, infinite tunnel walls: tunnel top and
tunnel ground
The main aerodynamic characteristic for the highspeed
train is a drag, thus it is directly related to the power
assumption, here represented by the nondimensional
coefficient, cd [25], referred to the referent crosssection
area of 3m2. In Picture 12, cd was plotted against free
stream M. The character of the plots for the BHST
running freely and on the open rail are similar, but
displaced by means of both parameters, cd and M, giving
the advantage to the free flight, as it was expected.
Comparing to the airfoils imposed to the free subsonic
and transonic flows, described by the gasdynamic [17],
results of this work showed a good agreement. The flow
at the infinite tunnel, have had the largest drag
coefficients in the range of applied M. As one may note,
the range of M in the case of the BHST passing tunnel is
narrow after selection, M=0.20.4, compared to other two
kingfisherlike configurations. The ground effects caused
the flow complexity, presence of high rates of
45
AERODYNAMICSOFTHEHIGHSPEEDTRAINBIOINSPIREDBYAKINGFISHER
numerical calculations would become simpler and less
advanced in the selection phase. After all, designflow
connection might be improved, and the research field
spreader, by applying experiences in a field of highspeed
aerodynamics, due to similarity of phenomenon never the
less to the end purposes.
OTEH2016
at Belgrade, 2016.
[10] *Nature History Museum, Museum exhibit of the
kingfisher prepared for the research use, 1939.
[11] Sawer,P.: Wildlife Photography/Solent, via Daily
Mail, goo.gl/B14OKP, with a written permission
[last accessed on 22/02/2016]
[12] Hereman,W.: Shallow water waves and solitary
waves, Mathematics of Complexity and Dynamical
Systems, Springer, ed Meyers R A, New York, 2012
152032.
[13] GhazizadeAhsaee,H., Nikseresht,A.H.: Numerical
Simulation of Two Dimensional Dynamic Motion of
the Symmetric Water Impact of a Wedge, IJMT, 1(1)
(2013) 1122.
[14] Faltinsen,O.: Hydrodynamics of Highspeed Marine
Vehicles Cambridge, Cambridge University Press,
New York, 2005.
[15] Raghunathana,R.S.,
Kimb,H.D.,
Setoguchi,T.:
Aerodynamics of highspeed railway train, Progress
in Aerospace Sciences, 38 (2002) 469514
ACKNOWLEDGMENTS
The authors are grateful to the Ministry of Education,
Science and Technological Development, Republic of
Serbia, for supporting this work as a part of the researches
through the financed Projects TR35045 and TR 34028
(2011.2016). Authors are grateful to Mr. Marko Rakovic,
biologist, curator ornithologist, and the Natural History
Museum at Belgrade for kind help, access to exibits and
knowledge, crucial for understanding the natural behavior
of the kingfisher and its numerical modeling. The authors
are grateful to Mr. Paul Sawer, wildlife photographer, and
the Solent, publisher, on a kindness to approve the use of
the artworks in this research which helped creation of the
detailed kingfisher shape.
[16] Liepmann,H.W.,
Roshko,A.:
Elements
of
Gasodynamics, Galcit Aeronautical Series, John
Willey and sons, London,1957.
[17] Dobrovolskaya,Z.N.: On some problems of
similarity flow of fluids with a free surface, Journal
of Fluid Mechanics, 36 (1969) 80529.
[18] Puhari,M., Luanin,V., Lini,S., Mati,D.: Research
Some Aerodynamic Phenomenon of High Speed
Trains in Low Speed Wind Tunnel, Proceedings of
the 3rd International Scientific and Professional
Conference CORRIDOR 10  A sustainable way of
integrations, October 25th, 2012., Belgrade, Serbia,
[19] Lini,S.,
Risti,S.,
Stefanovi,Z.,
Kozi,M.,
Ocokolji,G.: Experimental and Numerical Study of
SuperCritical Flow Around the Rough Sphere,
Scientific Technical Review, 65 (2) (2015) 1119.
[20] Lucanin,V., PuharicM., Milkovic,D., Golubovic,S.,
Linic,S.: Determining the influence of an air wave
caused by a passing train on the passengers standing
at the platform, International Journal of Heavy
Vehicle Systems, 19 (3) (2012) 299313
[21] TaeKyung,Kim, KyuHong,Kim, HyeokBin, Kwon:
Aerodynamic characteristics of a tube train, J. Wind
Eng. Ind. Aerodyn. 99 (1) (2011) 1871196.
[22] Dumont,E.R.: Bone density and the lightweight
skeletons of birds, Proc. R. Soc. B. 277 (2010) 2193
2198
References
[1] Bhushan,B.: Biomimetics: lessons from nature an
overview, Phil. Trans. R. Soc. A 367 (2009) 1445
1486.
[2] Kobayashi,K.: 2005 JFS Biomimicry Interview
Series: No.6., Shinkansen Technology Learned from
an Owl?,  The story of Eiji Nakatsu JFS Newsletter
31, goo.gl/HOZrjL [last accessed on 20/03/2015]
[3] *Ask Nature Shinkansen Train, Highspeed train
silently slices through air, goo.gl/aSfpQ5 [last
accessed on 20/03/2015]
[4] McKeag,T.: Auspicious Forms: Designing the Sanyo
Shinkansen 500Series Bullet Train, Zygote
Quarterly (2012) 1433.
[5] Anders,J.B.: Biomimetic Flow Control, AIAA20002543, NASA Langley Research Center. Fluid (2000),
1922 June 2000, Denver, CO.
[6] Rauo,B.: Bionics in Design, University of Belgrade,
Belgrade, eBook on CD., (in Serbian), 2014.
[7] Linic,S.,
Rasuo,B.,
Kozic,M.,
Lucanin,V.,
Puhari,M.: Comparison of numerically obtained 2D
flow fields for the bionic high speed train concept
designs inspired with aquatic and flying animals,
Proceedings of the 6th International Scientific
Conference on Defensive Technologies  OTEH
2014, The Military Technical Institute, (2014)
Belgrade, 4449.
[8] Linic,S., Rasuo,B., Kozic,M., Lucanin,V., Bengin,A.:
DragCoefficient Behavior of the BioInspired High
Speed Train Design, Proceedings of the 5th
International Congress of Serbian Society of
Mechanics, Arandjelovac, June 1517 (2015) Serbia.
[9] Personal communication with Mr. Marko Rakovic,
research assistant from the Natural History Museum
[23] *Theory and UsersGuide of the ANSYS Fluent 12
Documentation, ANSYS Inc., 2009.
[24] Veersteg,H.K., Malalasekera,W.: An Introduction to
Computational Fluid Dynamics The Finite Volume
Method, Longman Scientific & Technical, New
York, 1995.
[25] Anderson,J.D.: Fundamentals of Aerodynamics, II
ed., McGrawHill, 1991.
46
OBSERVATIONS ON SOME TRANSONIC WIND TUNNEL TEST RESULTS
OF A STANDARD MODEL WITH A TTAIL
DIJANA DAMLJANOVI
Miltary Technical Institute (VTI), Belgrade, didamlj@gmail.com
ORE VUKOVI
Miltary Technical Institute (VTI), Belgrade, vdjole@sbb.rs
ALEKSANDAR VITI
Miltary Technical Institute (VTI) (retired), Belgrade, sasavitic@gmail.com
JOVAN ISAKOVI
College of Applied Engineering Studies, BelgradeZemun, jisakovic@tehnikum.edu.rs
GORAN OCOKOLJI
Miltary Technical Institute (VTI), Belgrade, ocokoljic.goran@gmail.com
Abstract: As a part of a periodic health monitoring of the wind tunnel structure, instrumentation and flow quality, a
series of tests of an AGARDC calibration model was performed in the 1.5 m T38 trisonic wind tunnel of the Military
Technical Institute (VTI) in Belgrade. The tests comprised measurements of forces and moments in the transonic Mach
number range with the purpose of comparing the models obtained aerodynamic characteristics with those from other
wind tunnel laboratories, in accordance with an adopted procedure for standard models testing. Interfacility
correlations were based on test results of physically the same model in the 5ft trisonic wind tunnel of the National
Research Council (later operated as National Aeronautical Establishment) of Canada, in the 1.2 m trisonic wind tunnel
of the Romanian National Institute for Scientific and Technical Creation and in the T38 wind tunnel during the
commissioning period. Analysis of correlated test results confirmed a good flow quality in the T38 test section, good
condition of wind tunnel structure and instrumentation, and the correctness of the data reduction algorithm. Small
differences were observed in the pitching moment coefficient data obtained in the normal and inverted model
configurations, and it has preliminary been concluded that the effect may have been caused by a slight asymmetry of
flow in the rear part of the wind tunnel test section, the AGARDC model being known for the high sensitivity of the
pitching moment to local conditions.
Keywords: wind tunnel, transonic flow, standard model, aerodynamic characteristics.
easier to detect (from anomalies in the wind tunnel test
results) if the shock waves reflected from the walls of the
wind tunnel test section are passing too close to the rear
end of the model. The existence of the tail also makes this
model more sensitive than AGARDB to flow curvature
in the wind tunnel test section.
1. INTRODUCTION
The Military Technical Institute (VTI) in Belgrade has
established a procedure for windtunnel data quality
assurance, primarily based on periodic testing of the
AGARDB standard model [1]. The procedure comprises
windtunnel testing, maintenance of a database of
standard test results and a statistical control on the test
data, [24]. Considerations and directives recommended
in the procedure have now been applied and implemented
in testing of the AGARD standard model C with a Ttail.
The intention of the research was to start the statistical
control on the database windtunnel results for this model.
The obtained and analyzed results will serve to ascertain
the stability of the measurement process and help in
future tests of similar configurations.
The configuration was originally designed for calibration of
transonic wind tunnels [57] and primarily used in them.
Unfortunately the database of published test results is
somewhat smaller than the one for the AGARDB model.
Beside the review data available in [7] comprising results
from AEDC, Boeing, ONERA and NLR, which,
unfortunately, is of very poor legibility, available test results
include [8] those from the Canadian 5ft trisonic blowdown
wind tunnel of the National Research Council (NRC), later
operated as National Aeronautical Establishment (NAE) [9],
those from the Romanian 1.2 m trisonic blowdown wind
tunnel [10,11] of the National Institute for Scientific and
Technical Creation (INCREST, now INCAS  National
Institute for Aerospace Research) and tests made during the
AGARDC model configuration differs from the betterknown AGARDB configuration by the addition of a rearbody segment with a Ttail. The longer body of the
AGARD model C and the existence of the Ttail make it
47
OBSERVATIONSONSOMETRANSONICWINDTUNNELTESTRESULTSOFASTANDARDMODELWITHATTAIL
commissioning of the T38 wind tunnel of VTI [12]. The
intention of the authors is to extend the amount of data and
make it available to the community.
OTEH2016
The support sting for the AGARDC model is identical to
the sting for the AGARDB model, having a length of 3D
aft of model base and a diameter of 0.5D. In order to
reduce cost and produce more versatile wind tunnel
models, actual designs of AGARDB and AGARDC are
sometimes realized as an AGARDB configuration to
which a body segment with the Ttail can be attached at
the rear end to form the AGARDC configuration.
2. STANDARD MODEL WITH A TTAIL
AGARDC standard model, a derivative of the wellknown standard AGARDB model, is an ogivecylinder
with a delta wing and a horizontal and a vertical tail in the
Ttail configuration, Picture 1.
The AGARDC wind tunnel calibration model, used in T38, was supplied by Boeing. Model size (of 115.8 mm
dia.) was chosen with respect to the tunnels test section
size. Model had been used in previous T38 wind tunnel
calibrations and tests in other wind tunnels, and there is a
database with which to compare the obtained results [12].
3. TEST FACILITY
The T38 test facility at the Military Technical Institute
(VTI) in Belgrade is a blowdowntype pressurized wind
tunnel [13] with a 1.5 m 1.5 m square test section,
Picture 3. For subsonic and supersonic tests, the test
section is with solid walls, while for transonic tests, a
section with porous walls is inserted in the configuration.
The porosity of walls can be varied between 1.5% and
8%, depending on Mach number, so as to achieve the best
flow quality.
Picture 1. AGARDC model with a body diameter of
115.8 mm
At the AGARD Wind Tunnel and Model Testing Panel
meeting in Paris, France, in 1954, it was agreed [7] to add
a third model configuration to the family of AGARD
standard calibration models, by extending the body of the
AGARDB by 1.5 diameters and by adding a Ttail.
Mach number in the range 0.2 to 4.0 can be achieved in
the test section, with Reynolds numbers up to 110 million
per meter. In the subsonic configuration, Mach number is
set by sidewall flaps in the tunnel diffuser. In the
supersonic configuration, Mach number is set by the
flexible nozzle contour, while in transonic configuration,
Mach number is both set by sidewall flaps and the flexible
nozzle, and actively regulated by blowoff system. Mach
number can be set and regulated to within 0.5% of the
nominal value.
The horizontal tail has an area equal to 1/6 of the wing
area. Sections of the vertical and horizontal tail are
circular arc profiles defined identically to the profile of
the wing. Forward of the 1.5D body extension, the
geometry of the AGARDC model is identical to that of
the AGARDB: an 8.5D long solid body of revolution
consisting of the 5.5D long cylindrical segment and a
nose with the length of 3D.
Also, the position of the moments reduction point (the
aerodynamic centre) is the same as on AGARDB. All its
dimensions are given in terms of the body diameter D so
that the model can be produced in any scale, as
appropriate for a particular wind tunnel, Picture 2.
Picture 3. The T38 test facility in VTI, Belgrade
Stagnation pressure in the test section can be maintained
between 1.1 bar and 15 bar, depending on Mach number,
and regulated to 0.3% of nominal value. Run times are in
the range 6s to 60s, depending on Mach number and
stagnation pressure.
Model is supported in the test section by a tail sting
mounted on a pitchandroll mechanism by which desired
aerodynamic angles can be achieved. The facility supports
both stepbystep model movement and continuous movement of model (sweep) during measurements. Positioning
accuracy is 0.05 in both pitch and roll.
Picture 2. A drawing defining the geometry of the
AGARDC standard model and its sting fixture
48
OBSERVATIONSONSOMETRANSONICWINDTUNNELTESTRESULTSOFASTANDARDMODELWITHATTAIL
OTEH2016
Canadian NRC/NAE 5ft trisonic wind tunnel and results
[11] from the Romanian 1.2 m 1.2 m INCREST trisonic
wind tunnel facility. The early set of the T38 windtunnel
test data are also correlated [12]. All results are from the
tests with the 115.8 mm dia. model.
A good agreement of the correlated test results confirms a
high quality of air flow in the T38 test section, good
condition of the wind tunnel instrumentation and the
correctness of the data reduction algorithm, Pictures 58.
Picture 5. Interfacility correlation in the drag force
measurement
Picture 4. Pitchandroll model support mechanism
4. RESULTS AND DISCUSSION
VTI is among the laboratories that have adopted the
practice [1] of periodic testing of a standard model every
couple of years in order to provide a continued confidence
in the reliability of measurements in their wind tunnels.
The adopted procedure for determining the overall T38
windtunnel data quality and verification in the standard
testing [1] includes a number of steps, but here,
consideration is made only from the two points: (1) interfacility and (2) the test section symmetry correlations.
Comparable tests results are presented in the wind axes
system in the form of graphs, showing drag force
coefficient, lift force coefficient, pitching moment
coefficient and base pressure coefficient as function of the
angle of attack in the wind axes system. Test results are
given for the model aerodynamic center, Picture 2. Model
reference length is mean aerodynamic chord.
Picture 6. Interfacility correlation in the lift force
measurement
4.1. Sets of VTI test results
A small set of early AGARD model C test data obtained
during the commissioning of the T38 wind tunnel existed
in the VTI database, as the model B is being more often
used for wind tunnel calibration and verification, [1,14].
Tests had been performed at the transonic Mach numbers
0.7 to 1.05 at angles of attack from 2 to +13 traversed
in a continuousmovement mode.
Picture 7. Interfacility correlation in the pitching
moment measurement
An additional test campaign with this model in the T38
wind tunnel was executed later, comprising measurements
at Mach numbers 0.7 to 1.15. The angle of attack range
was 4 to +10 traversed in a continuousmovement
mode.
4.2. Interfacility correlation
The test data were analyzed based on correlations with
data from other experimental aerodynamics laboratories.
The reference sets of data were results [8] from the
Picture 8. Interfacility correlation in the base pressure
measurement
49
OTEH2016
OBSERVATIONSONSOMETRANSONICWINDTUNNELTESTRESULTSOFASTANDARDMODELWITHATTAIL
Analysis of the measured aerodynamic coefficients from
the point of test section symmetry was done for two Mach
0.7 runs at the two opposite roll angles: 0 (modelupright), Picture 9, and 180 (modelinverted), Picture 10.
The large deviation of the pitching moment coefficient
obtained in the INCREST wind tunnel is obviously a
result of the reduction of the moments to a reference point
different from that used in other tests. Unfortunately, the
location of the reference point used in INCREST tests is
not known, so the data could not be recomputed for the
proper reference point.
Mach 0.7 data in wind axes system at the aerodynamically
same angles of attack from modelupright and modelinverted runs were compared. CAD renderings of the
115.8 mm dia. AGARDC model in the both upright and
inverted configurations in the T38 test section, given in
Pictures 9 and 10, present aerodynamically the same +10
angle of attack in the wind axes system.
Relatively large scatter of the base pressure coefficients in
the correlations (Picture 8) can probably be explained by
the fact that absolute pressure transducers of high range
and, therefore, lower absolute accuracy, were used for the
base pressure measurement in the NRC/NAE and in early
VTI tests, while the newer VTI tests were performed
using a differential pressure transducer of a lower range
and higher accuracy. The method of basepressure
measurement in the INCREST wind tunnel is not known.
Comparison of test data obtained in these two runs
showed that there was no noticeable effect on the lift and
the drag aerodynamic coefficients. It was seen that the
magnitude of differences in the force (drag and lift,
Pictures 11 and 12) and pressure measurements (Picture
13) that can be attributed to asymmetry of the test section
are comparable to the total measurement uncertainties [1]
confirming a good symmetry of the test section. It should
be noted that the test section was calibrated and that the
determined flow angularities in vertical and horizontal
planes were uptodate.
Forebody drag force coefficient
It should also be noted that, because of the smaller size of
the INCREST wind tunnel relative to NCR and VTI wind
tunnels (1.2 m vs. 1.5 m) the blockage of the AGARDC
model in tests [11] was higher than in VTI and NAE tests,
which may have affected the results, including the
measurement of the base pressure.
4.3. Correlation from the point of symmetry
The test data were analyzed from the point of test section
symmetry, more exactly test results were checked from
runs with model both in upright and inverted position.
0.14
0.12
AGARDC model, Mach 0.7
Porous wall test section
T38 (model upright)
T38 (model inverted)
0.10
0.08
0.06
0.04
0.02
0.00
6
4
2
10
12
Angle of attack
Picture 11. Correlations from the point of test section
symmetry in the drag force measurement
Lift force coefficient
0.8
Picture 9. CAD rendering of the 115.8 mm dia. AGARDC model, upright configuration, T38 test section, +10
angle of attack
AGARDC model, Mach 0.7
Porous wall test section
T38 (model upright)
T38 (model inverted)
0.6
0.4
0.2
0.0
0.2
0.4
6
4
2
10
12
Angle of attack
Base pressure coefficient
Picture 12. Correlations from the point of test section
symmetry in the lift force measurement
0.12
0.13
0.14
0.15
0.16
0.17
AGARDC model, Mach 0.7
Porous wall test section
T38 (model upright)
T38 (model inverted)
0.18
0.19
0.20
6
4
2
10
Angle of attack
Picture 10. CAD rendering of the 115.8 mm dia.
AGARDC model, inverted configuration, T38 test
section, +10 angle of attack
Picture 13. Correlation from the point of test section
symmetry in the base pressure measurement
50
12
OBSERVATIONSONSOMETRANSONICWINDTUNNELTESTRESULTSOFASTANDARDMODELWITHATTAIL
Contrary to the drag coefficient data and lift coefficient
data, there is some difference between the pitching
moment coefficients obtained in the normal and inverted
positions (Picture 14). It is known [7] that the existence of
the Ttail makes the AGARDC model very sensitive
(more than the model B) to flow curvature in the test
section. Indeed, comparable differences were not
observed in the tests of the shorter AGARDB model in
VTI. Therefore, the differences in the pitching moment
can be attributed to a small amount of flow curvature in
the rear part of the transonic test section, possibly caused
by the asymmetry of the model support mechanism
(Picture 9 and Picture 10). Further investigation of the
observed effect is indicated and the feasibility is being
considered of eliminating this asymmetry by differential
porosity setting of the upper and lower wall of the
downstream part of the transonic test section that is
located in the model cart of the T38 wind tunnel.
It will also be of interest to correlate the obtained data
with those from the large subsonic wind tunnel of VTI,
where the standard model C has also recently been tested
following the adopted procedure. Therefore, the future
tests of the AGARDC model in the T38 wind tunnel
should encompass a wider span of Mach numbers, not
only in the transonic range, but in the subsonic range as
well. It is expected that, after the introduction of a new T38 wind tunnel control system and fine tuning of all the
wind tunnel systems and subsystems, a new set of data
will be referential for future correlations and verifications.
References
[1] Damljanovic, D., Isakovic, J., Rauo B., "T38 WindTunnel Data Quality Assurance Based on Testing of
a Standard Model", Journal of Aircraft, 50 (4)
(2013) 11411149.
[2] Recommended Practice: Calibration of Subsonic and
Transonic Wind Tunnels, AIAAR0932003, AIAA.
[3] Reed T.D., Pope T.C., Cooksey J.M., Calibration of
Transonic and Supersonic Wind Tunnels, NASA
Contractor Report 2920, Vought Corporation, 1977.
[4] Hemsch M., Grubb J., Krieger W., Cler D., "Langley
Wind Tunnel Data Quality Assurance: Check
Standard Results", AIAA 20002201, Proceedings of
the 21st AIAA Advanced Measurement Technology
and Ground Testing Conference, 2000.
[5] Specification for AGARD Wind Tunnel Calibration
Models, AGARD memorandum, AGARD, 1955.
[6] Wind Tunnel Calibration Models, AGARD Specification 2, AGARD, 1958.
[7] Hills R. (ed.), A Review of Measurements on AGARD
Calibration Models, AGARDograph 64, Aircraft
Research Association Bedford, England, 1961.
[8] Report on Tests Conducted on NACA 0012, and
AGARDB and C Models in the NAE 5 ft Blowdown
Wind Tunnel During Training of VTI Personnel:
NovDec 1981, DSMA Rept. No. 4001/R84, 1983.
[9] NRCCNRC Information; Aerodynamics: 1.5m
1.5m Trisonic Blowdown Wind Tunnel, IARAL03e
National Research Council Canada, 2005.
[10] Munteanu, F., "INCAS Trisonic Wind Tunnel",
INCASBulletin, No.1/2009, INCAS National Institute for Aerospace Research Romania, 2009.
[11] The calibration of the transonic and supersonic test
section using the AGARD model B and C, Report:
RLST14, National Institute for Scientific and
Technical Creation, Bucharest, Romania, 1979.
[12] Isakovic, J., Zrnic, N., Janjikopanji, G., "Testing of
the AGARD B/C, ONERA and SDM Calibration
Models in the T38 1.5m 1.5m Trisonic Wind
Tunnel", Proceedings of the 19th ICAS congress,
1994, pp. 19.
[13] Elfstrom, G.M, Medved, B., "The Yugoslav 1.5m
Trisonic Blowdown Wind Tunnel", Paper 860746CP, AIAA, 1986.
[14] Damljanovic, D., Vitic, A, Vukovic, ., Isakovic J.,
"Testing of AGARDB calibration model in the T38
Trisonic Wind Tunnel", Scientific Technical Review,
56 (2) (2006) 5262.
Pitching moment coefficient
0.08
AGARDC model, Mach 0.7
Porous wall test section
T38 (model upright)
T38 (model inverted)
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
6
4
2
10
12
Angle of attack
Picture 14. Correlations from the point of test section
symmetry in the pitching moment measurement
Pitching moment coefficient
For comparison, the available data on the differences in
pitching moment coefficient for the normal and inverted
model position in the NAE/NRC 5ft wind tunnel are
given in Picture 15.
0.11
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
AGARDC model, Mach 0.7
Porous wall test section
NRC/NAE (model upright)
NRC/NAE (model inverted)
4
2
10
12
OTEH2016
14
Angle of attack
Picture 15. Correlation from the point of test section
symmetry in the pitching moment measurement in the
Canadian facility
5. CONCLUSION
The intention of the research was to start the statistical
control on standard AGARDC model test data in
accordance with the procedures adopted in VTI and to
expand the relatively meagre published reference data.
There is a need for further tests of this model in order to
investigate the preliminary conclusions related to the
pitching moment coefficient. Simulations performed by
computational fluid dynamics software tools can be
helpful in understanding the observed phenomena.
51
NUMERICAL AND EXPERIMENTAL ASSESSMENT OF TRANSONIC
TURBULENT FLOW AROUND ONERA M4 MODEL
JELENA SVORCAN
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, jsvorcan@mas.bg.ac.rs
DIJANA DAMLJANOVI
Military Technical Institute, Belgrade, didamlj@gmail.com
DRAGAN KOMAROV
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, dkomarov@mas.bg.ac.rs
SLOBODAN STUPAR
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, sstupar@mas.bg.ac.rs
NEBOJA PETROVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, npetrovic@mas.bg.ac.rs
Abstract: Experimental investigation of transonic flow at Mach number of 0.84 around a subsonic/transonic transport
calibration model ONERA M4 at four different anglesofattack has been conducted in T38 trisonic blowdown wind
tunnel of the Serbian Military Technical Institute. Experimental results include relative pressure and pressure
coefficient distributions along three wing sections. Obtained aerodynamic performance data present a good basis for
CFD studies. Both experimental and numerical setups and processes are briefly described. Numerical simulations are
performed in ANSYS FLUENT 16.2 with several different turbulence models employed. Two different sets of results are
presented and compared. Pressure distribution along the wing surface, sonic bubble on the wing suction side and wingtip vortex have been investigated in more detail and presented in form of contours and isosurfaces.
Keywords: transonic flow, experiment, CFD, turbulence models, wingtip vortex.
1. INTRODUCTION
Acquiring accurate numerical data (on boundary layer
transition characteristics at transonic speeds) to
supplement experimental data can be extremely useful for
possible improvement of performances at cruise regimes.
Numerical experiments may present a powerful and
inexpensive tool that can be and is used for development
of improved wingbody designs with increased
aerodynamic performances or reduced fuel burn, noise
and emissions [6, 8].
At transonic flow conditions many flow phenomena such
as pressure fluctuations, shockboundary layer interaction,
etc. appear which make measuring and simulating these
types of flows difficult [1, 2]. At high subsonic speeds,
sonic bubbles (supersonic regions) terminating in shock
waves appear around the airfoil. At sufficiently large
pressure rise across the shock, shockinduced separation
of the boundary layer occurs. As stated in [2], for a
turbulent boundary layer, this starts when the local Mach
number just upstream of the shock ranges in 1.251.3.
Furthermore, in most cases, experimental data cannot
provide a complete insight into the complex transonic
flow field. Visualizing different flow regions around the
body (sonic zones, transition or separation zones, wingtip
vortex, etc.) can help in better understanding of the
complex flow physics.
Additional difficulties in performing these experiments
arise from the existence of boundaries in wind tunnels.
For that reason, ONERA has built a number of standard
models for use in evaluating Reynolds number and
blockage effects in various wind tunnels [3]. An extensive
collection of aerodynamic performance data has been
compiled over the years [35].
The analysis presented in the paper focuses on transonic
flow field around the wing of the standard transport
calibration model ONERA M4. Experimentally obtained
pressure distributions along wing cross sections are
compared and complemented by numerically obtained
pressure distributions, Mach number distributions,
vorticity fields, etc. at different anglesofattack (AoA).
Although measurements of flows around standard models are
primarily done for the purpose of validation of the testing
facility and empirical correction procedures, they can also
serve for the establishment of guidelines for the development
of a sufficiently accurate numerical setup and the assessment
of numerical models (e.g. turbulence models) [6,7].
52
NUMERICALANDEXPERIMENTALASSESSMENTOFTRANSONICTURBULENTFLOWAROUNDONERAM4MODEL
OTEH2016
2. EXPERIMENT
Experimental part of the presented research has been
performed in T38 trisonic blowdown wind tunnel of the
Serbian Military Technical Institute. Dimensions of the
test section are 1.5m 1.5m. In this wind tunnel, Mach
numbers in the range 0.2 to 4 and Reynolds numbers up
to 110 million per meter are achievable. For transonic
tests, a section with porous walls is inserted in the wind
tunnel configuration.
The values of flow quantities measured in the wind tunnel
and used in numerical simulations are given in Table 1.
Table 1. Transonic flow conditions, M = 0.84
p0 [bar]
p [bar]
q [bar]
T0 [K]
MRe
V [m/s]
1.524
0.960
0.474
282.6
2.52
265.1
Figure 2. Model in the wind tunnel
The stagnation pressure in the test section was measured
by a Mensor quartz bourdon tube absolute pressure
transducer pneumatically connected to a pitot probe in the
settling chamber of the wind tunnel. Range of this
transducer used was 7bar. The static pressure in the test
section was measured by a Mensor quartz bourdon tube
absolute pressure transducer pneumatically connected to
an orifice on the test section sidewall. The range of this
transducer was 1.75bar. The nonlinearity and hysteresis of
the transducers used were typically 0.02% F.S.
Subsonic/transonic transport calibration model ONERA
M4 is a wellknown, standard configuration used to
investigate wind tunnel transonic performance [5]. It
consists of a fuselage with wing and tail surfaces.
Fuselage is shaped like an ogive in the fore part (x/L <
0.2935), followed by a cylinder (0.2935 x/L < 0.6654)
and ending in an elliptic boat tail. Wing and tail cross
sections are the same: "peaky" type symmetric with a
maximal relative thickness of 10.5% occurring at the
37.5% chord location (fig. 1). Wings are swept, tapered,
untwisted with a setting angle of 4 and a dihedral angle
of 3.
The stagnation temperature was measured by a RTD
probe in the settling chamber. The accuracy of this
transducer was approximately 0.5K.
Pressure distribution was measured using an
electromechanical scanning device with a solenoiddriven
valve. A piezoresistive differential pressure transducer
was used to measure local pressures.
The data acquisition system consisted of a Teledyne 64
channel front end controlled by a computer. The frontend channels for flow parameters transducers were set
with 10Hz, fourthorder low pass Butterworth filters and
appropriate amplification. The frontend channel for
pressure transducer was set with 100Hz lowpass filter of
the same type.
The scanning device operates by rotating a valve that
connects modepressure measuring ports sequentially to a
single differential pressure transducer. The main benefit
of this concept is that many local pressures could be
measured with a single transducer, which greatly
simplified the test installation, put less load on the data
acquisition system, etc. Main drawbacks lie in the fact
that the pressures were not measured simultaneously and
in the fact that after each portswitching of the rotating
valve, certain time was required for the pressure to settle
in the tubing. The pressuresettling time effectively
limited the scanning speed to about 10ports/second,
although there was mechanically capability of scanning at
about 40ports/second.
Figure 1. Model geometry
Pressure measurements have been conducted along three
different wing sections denoted by S1, S2 and S3 in fig. 1.
Sections S1 and S3 are parallel to the fuselage axis, while
section S2 is normal to the wing line of maximum
thickness (located at the 37.5% of the chord). Measuring
equipment was built into the model. Numbers of ports
distributed along the sections are as follows: 39 ports per
S1 (27 at the suction and 12 at the pressure side), 41 (29 +
12) per S2 and 39 (27 + 12) per S3.
In the test section the model is supported by a tail sting
mounted on a pitchandroll mechanism by which
different anglesofattack can be achieved, fig. 2. Here,
four different anglesofattack were considered: 3.3,
1.57, 0.2 and 2.06.
3. NUMERICAL SETUP
Since the quality of the computational grid is extremely
important, the meshes were created according to general
53
NUMERICALANDEXPERIMENTALASSESSMENTOFTRANSONICTURBULENTFLOWAROUNDONERAM4MODEL
recommendations of the AIAA High Lift and Drag
Prediction Workshops [9, 10]. Generated meshes are built
around a halfmodel. The support system is not taken into
account. It probably should be done in later stages of the
investigation for better results. The surrounding domain is
in the form of a halfsphere, with the farfield boundary
located approximately 100 cref around the model
(R = 11m).
OTEH2016
symmetry plane.
4. RESULTS AND DISCUSSION
Prior to the discussion, it should be repeated that the wing
angleofattack differs from the global AoA by the value
of the wing setting angle of 4. For that reason, wing AoA
are actually 0.7, 2.43, 4.2 and 6.06. Given that the last
value is quite high for the investigated flight regime, it is
not surprising that the numerical results diverge from
experimental data at higher AoA. Other authors also
confirmed that while at lower anglesofattack, even with
URANS models (SA turbulence model), at meduimsized
meshes (around 2 million cells), good prediction can be
achieved, results at higher AoA tend to deteriorate [10].
As a part of the grid convergence study, a family of
parametrically similar meshes was generated (same
topology, same stretching factors). They are all hybrid
unstructured and threedimensional. Coarse grid (C)
numbers around 860.000 cells, medium (M)  around
1.450.000 cells, and fine (F)  around 2.600.000 cells. Thirtyforty layers of prismatic cells exist around the model walls
resulting in dimensionless wall distance around 3 and 1, y+
3 and y+ < 1, for the coarse and fine grid respectively. Grid
spacing normal to symmetry plane is larger.
4.1. Pressure distribution and aerodynamic
coefficients
The mesh along the wing surfaces is mappedfaced with a
biased chordwise distribution of element size, fig. 3.
Cells around the leading and trailing edge are
approximately 0.5%c for medium mesh. Wing trailing
edge is modeled as blunt.
Figs. 46 illustrate relative pressure distributions along the
three crosssections for different turbulence models and
anglesofattack.
Figure 3. Computational Fgrid on the wing
Numerical simulations were performed in ANSYS
FLUENT 16.2 where governing flow equations for
compressible, viscous fluid were solved by finitevolume
method. Reynoldsaveraged NavierStokes (RANS)
equations were closed by a oneequation SpalartAllmaras
(SA) and twoequation k SST (kwSST) turbulence
models. SA model incorporates modified turbulent
viscosity equation, while kwSST model presents a
combination of standard k model near the walls and k
model in the outer layer. Fluid, air, was considered as
ideal gas whose dynamic viscosity changes according to
the Sutherland law.
Figure 4. Relative pressure distribution at crosssection
S1 at AoA = 1.57;  SA on Mmesh,  kwSST on Cmesh
Densitybased implicit solver was used. Spatial
discretizations were of the second order. Gradients were
obtained by the least squares cellbased method. CourantFriedrichsLewy (CFL) number was set to 5. Numerical
simulations were performed until fluctuations of
aerodynamic coefficient became less than 0.1%.
Noslip boundary conditions were assigned to all model
surfaces. Dirichlet boundary conditions concerning
velocity and pressure were imposed on farfield boundary,
while zero Neumann boundary conditions (zero normal
velocity and normal gradients) were defined on the
Figure 5. Relative pressure distribution at crosssection
S2 at AoA = 0.2;  SA on Fmesh,  SA on Mmesh
Although refining the mesh increases the accuracy of
numerical results, the end of sonic zone (terminating in a
54
NUMERICALANDEXPERIMENTALASSESSMENTOFTRANSONICTURBULENTFLOWAROUNDONERAM4MODEL
OTEH2016
shock wave) is hard to simulate on all meshes by the
RANS equations. This is particularly evident on higher
AoA at crosssection S1 near the wing tip.
Crosssection S3, nearest to the fuselage, seems to be
simulated most accurately.
Figure 8. Relative pressure contours along the model at
AoA = 1.57 at Mmesh, SA model
Figure 6. Relative pressure distribution at crosssection
S3 at AoA = 3.3;  SA on Mmesh,  kwSST on Cmesh
Although discrepancies between the two sets of data exist,
global numerical data (in the form of normal force
coefficient presented in fig. 7) agree well with
experimental data obtained on ONERA model M5 at the
same Mach and Reynolds numbers at different wind
tunnels [3]. Both models possess the same geometrical
features, with the M5 model being somewhat bigger with
the fuselage length of 1.058m. Numerical results
presented in fig. 7 were obtained on Mmesh with SA
turbulence model.
Figure 9. Pressure coefficient contours along the wing at
AoA = 3.3 and 0.2 at Mmesh, SA model
4.2. Velocity field
Flow field around the model can also be presented by
Mach number contours, figs. 10 and 11. Variations in the
shape of the sonic bubble appearing primarily at the
suction side of the wing nicely illustrate the change of
aerodynamic loading with the increase of AoA, fig. 12.
Figure 7. Normal force coefficient Cn for the whole
model at different AoA, experimental data taken from [3]
Relative pressure contours over the whole model and
pressure coefficient contours over the wing are presented
in figs. 8 and 9 respectively. At both AoA = 1.57 and
AoA = 0.2 the superposition/composition of shock
waves (both weak and strong) appearing at the suction
side of the wing is well illustrated.
Figure 10. Mach number contours at three crossplanes at
AoA = 3.3 and 1.57 at Mmesh, SA model
55
NUMERICALANDEXPERIMENTALASSESSMENTOFTRANSONICTURBULENTFLOWAROUNDONERAM4MODEL
OTEH2016
Figure 11. Mach number contours at three crossplanes at
AoA = 0.2 and 2.06 at Mmesh, SA model
4.3. Wing tip vortex
As a result of pressure differences between wing pressure
and suction sides a vortex trails downstream from the
wing tip. The study of this effect is important since these
vortices can be the source of additional drag, noise and
hazard [7]. Due to significant velocity and pressure
gradients, it is still difficult to accurately predict wing tip
vortices numerically.
Fig. 12 illustrates the vortices appearing one, two and
threechords behind the wing at different AoA in the form
of xcomponent of vorticity, x, obtained on Mmesh by
SA model. Although employed turbulence model greatly
affects the presented results since turbulent kinetic energy
distribution directly defines vortex shape and size (its
maximum should be at the core center), this effect can
sufficiently accurately be captured by isotropic turbulence
models [7].
Angleofattack significantly influences the generation of
vortices. At higher anglesofattack, the vortices occur not
only because of the wing tip but also due to the complex
composition of sonic bubbles terminating in shock waves
appearing at the suction side of the wing (i.e. mixed
shock/angleofattack induced separation occurring near
the wing tip).
Figure 12. Sonic bubble at the suction side of the wing
and xvorticity x [s1] in three planes behind the wing at
AoA = 3.3, 1.57, 0.2 and 2.06 respectively
Vortices circulate counterclockwise as viewed from the
front. At higher AoA vortices decay more slowly and
significantly broaden. The existence of wing tip vortices
is captured 3 chords downstream.
56
NUMERICALANDEXPERIMENTALASSESSMENTOFTRANSONICTURBULENTFLOWAROUNDONERAM4MODEL
OTEH2016
Sciences, 37 (2001) 147196.
[3] Binion, T. W. Jr., "Tests of the ONERA calibration
models in three transonic wind tunnels", AEDCTR76133, Arnold Air Force Station, Tennessee, 1976.
[4] Nakakita, K., Kurita, M., Mitsuo, K., "Development
of the pressuresensitive paint measurement for large
wind tunnels at Japan Aerospace Exploration
Agency",
24rd
International
Congress
of
Aeronautical Sciences, Yokohama, Japan, 2004.
[5] Yoshida, K., Ueda, Y., Noguchi, M., "Experimental
and numerical analysis on lift and transition
characteristics of ONERAM5 configuration model",
25th International Congress of Aeronautical
Sciences, Hamburg, Germany, 2006.
[6] Kroll, N., AbuZurayk, M., Dimitrov, D., Franz, T.,
Fuhrer, T., Gerhold, T., et al., "DLR project DigitalX: towards virtual aircraft design and flight testing
based on highfidelity methods", CEAS Aeronaut J, 7
(2016) 327.
[7] antrak, . S., Kushner, L. K., Heineck, J. T., "Timeresolved stereo PIV investigation of the NASA
Common Research Model in the NASA Ames Fluid
Mechanics Laboratory 32 by 48in indraft wind
tunnel", Center for Turbulence Research Annual
Research Briefs, (2014) 179191.
[8] Deere, K. A., Luckring, J. M., McMillin, S. N.,
Flamm, J. D., "CFD Predictions for Transonic
Performance of the ERA Hybrid WingBody
Configuration", AIAA Science and Technology
Forum and Exposition, San Diego, CA, 2016.
[9] Vassberg, J. C., "A Unified Baseline Grid about the
Common Research Model WingBody for the Fifth
AIAA CFD Drag Prediction Workshop", 29th AIAA
Applied Aerodynamics Conference, Honolulu, HI,
2011.
[10] Ochi, A., Shima, E., "A Hybrid Unstructured Grid
System for Viscous and Inviscid Aerodynamic
Analysis", 23rd International Congress of
Aeronautical Sciences, Toronto, Canada, 2002.
5. CONCLUSION
A comparative experimental and numerical investigation
of transonic flow around ONERA M4 standard model has
been performed. Satisfactory coincidence between locally
measured and computed pressure distributions has been
achieved.
Numerical simulation is an important tool in aircraft
design since it requires less time and resources than other
investigations producing similar scope of results. It
complements wind tunnel testing and provides valuable
additional results and fluid flow visualization. However,
there are still many limitations, especially in flight
regimes such as transonic, and additional investigations
must be performed. Good practice to validate numerical
setup and make sure that the obtained computed results
are in expected range of accuracy is through comparison
with the experimental data for calibration models.
Transonic regimes are extremely important for
contemporary airliners. Sonic zones terminating in shock
waves and wing tip vortices can significantly increase
drag and deteriorate aerodynamic performances. Any
attempt made in their successful simulation can lead to
improved aircraft design. Therefore, despite the fact that
the steady RANS models used in the study are not able to
completely capture identified complex flow phenomena,
presented numerical study could be a good basis for more
detailed and precise further investigations.
ACKNOWLEDGEMENT
The paper is a contribution to the research TR 35035
funded by the Ministry of Education, Science and
Technological Development of the Republic of Serbia.
References
[1] Takakura, Y., Ogawa, s., Wada, Y., "Transonic windtunnel flows about a fully configured model of
aircraft", AIAA Journal, 33(3) (1995) 557559.
[2] Lee, B. H. K., "Selfsustained shock oscillations on
airfoils at transonic speeds", Progress in Aerospace
57
COMPUTATIONAL ANALYSIS OF HELICOPTER MAIN ROTOR BLADES
IN GROUND EFFECT
ZORANA TRIVKOVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, zposteljnik@mas.bg.ac.rs
JELENA SVORCAN
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, jsvorcan@mas.bg.ac.rs
MARIJA BALTI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, mbaltic@mas.bg.ac.rs
DRAGAN KOMAROV
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, dkomarov@mas.bg.ac.rs
VASKO FOTEV
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, vfotev@mas.bg.ac.rs
Abstract: Numerical investigation of an isolated representative helicopter main rotor has been performed in ANSYS
FLUENT 16.2. In general, flow field around the rotor is unsteady, threedimensional, complex and vortical. Such a
simulation requires substantial computational resources. Ground effect, which improves the aerodynamic performances
of the rotor, represents an additional challenge to numerical modeling. In this study, flow field is computed by Unsteady
Reynolds Averaged NavierStokes (URANS) equations. Both Frame of reference and Sliding mesh approaches were
employed to model the rotor rotation. Obtained results are compared to results obtained by simpler, sufficiently reliable
models such as Momentum Theory (MT) and Blade Element Momentum Theory (BEMT). Presented results include fluid
flow visualizations in the form of pressure, velocity and vorticity contours and the values of aerodynamic coefficients.
Keywords: helicopter, rotor, RANS, ground effect, power coefficient.
field is quite irregular over the rotor disc, and variations
of flow quantities per angular coordinate also exist.
1. INTRODUCTION
Flow field around helicopter blades is highly complex [16].
Although many flow phenomena are present (e.g.
unsteadiness, tip vortex formation, 3D dynamic stall, bladevortex interaction, shock/boundarylayer interaction etc.) [5],
because of the unique characteristics of helicopters,
extensive experimental and numerical research is constantly
being conducted for the purpose of improvement of their
aerodynamic performances. Several worldwide projects,
e.g. HELISHAPE, HART II, GOAHEAD, have been
performed in the last two decades [26].
Another interesting phenomenon is that the thrust of a
helicopter rotor, operating at constant power, increases as
it approaches the ground since the development of the
rotor wake is constrained [1, 7]. Such an effect is
extremely
important
when
determining
rotor
performances and has been studied both experimentally
and numerically [79] but is still not fully understood [1].
Because of the great complexity of the problem in question,
several simplifications of the study had to be adopted.
Although aspect ratio is high for helicopter blades, they were
considered rigid. Coning angle of the blades (as a
consequence of mutual aerodynamic and inertial loads) was
neglected. Collective pitch, necessary for achieving different
values of thrust, was estimated by BEMT. The effect of
helicopter fuselage was implicitly included in forward flight
computation through necessary values of angleofattack.
Unfortunately, these limitations make comparison to
experimental data more difficult and less accurate.
Since hover is the basic, simplest and most important
flight regime of a helicopter, hover performances are an
extremely important issue in the helicopter design process
("a dimensioning condition" as stated in [3]). Obtaining a
usable, sufficiently accurate numerical solution is a
difficult task [3]. Several models different in complexity
exist, that can be used for simulation of axissymmetric
flow field in hover and vertical flight. They include
momentum models, combined bladeelementmomentum
models and full NavierStokes equations capable of
capturing the changes of flow quantities along the blade.
The paper is structured as follows. A short description of
the representative rotor used in computations is given in
the next section. This is followed by a short description of
the used analytical and numerical models, and adopted
numerical setup. Results for both hover and forward
Compared to hover, forward flight regime is even more
complicated. Rather than being axissymmetric, the flow
58
OTEH2016
COMPUTATIONALANALYSISOFHELICOPTERROTORINGROUNDEFFECT
flight condition are presented in the section "Results and
discussion". In the end, short concluding remarks are
given.
3.1. Momentum Theory (MT)
Detailed equations can be found in [1]. Conservation laws
are applied in a quasionedimension integral formulation
to a control volume surrounding the rotor and its wake.
This simple approach enables a first level analysis of the
rotor thrust and power without having to consider blade
characteristics. In modified MT, actual power required to
hover presents the sum of induced and profile power:
2. MODEL DESCRIPTION
The representative model of a main helicopter rotor was
taken from [2] where a description of a conducted
experimental investigation on performances of two
different rotors, baseline and BERPtype, in the Langley
Transonic Dynamics Tunnel (TDT) can be found.
Experimental data, obtained in hover and forward flight
over a nominal range of advance ratios from 0.15 to
0.425, were used for validation of numerical models.
CP = CPi + CP0 =
CT3/ 2 Cd0
2
(1)
Power can be obtained from required power coefficient as
P = CP2R5, and thrust as T = CT2R4 where is air
density and rotor angular velocity.
Representative main rotor model consists of 4 rectangular,
1.428m long blades. Blades use two U.S. airfoils: 10%
thick RC(4)10 (r/R 0.84) and scaled 8% thick RC(3)08 (r/R 0.866), fig. 1. A smooth transition is made
between these two different airfoil shapes. Solidity of the
rotor is = 0.101. Twist distribution is triple linear, fig. 2.
In forward flight, due to forward speed and existence of
fuselage, additional members appear:
CP = CT +
Cd 0
8
(1 + 4.65 2 ) + 12 Af 3 .
(2)
Although this model assumes uniform distribution of flow
quantities across disc rotor, by "correctly" estimating the
values of induced power correction factor , section
profile drag coefficient Cd0 and fuselage equivalent
wetted area f, it is possible to obtain sufficiently accurate
estimations of required power coefficient. Here, these
values were adopted in accordance with available
experimental data.
By replacing the rotor by a simple source with an image
source to simulate the ground effect [7], for constant
power, ratio of thrust in (IGE) and out of ground effect
(OGE) in hover can be presented by the corresponding
ratio of induced velocities. Assuming uniform
distributions over the rotor area, the equation becomes:
Figure 1. Main rotor model
TIGE
1
=
TOGE P = const . 1 ( R 4 z ) 2
(3)
where R is rotor radius and z is rotor height off the
ground.
Figure 2. Twist distribution
Nominal test conditions are defined by advance ratio ,
tip Mach number MT (constantly kept at 0.628), rotorshaft angleofattack and blade collective pitch angle .
Since the values of the last two angles were not known,
they were estimated by BEMT. In hover ( = 0), data
were obtained at z/d = 0.83 where z is the distance from
windtunnel floor to rotor hub. Numerical simulations
were also performed for another, smaller relative distance
z/d = 0.25 and the two sets of data were compared.
3.2. Blade Element Momentum Theory (BEMT)
This hybrid approach combines the basic principles from
both the blade element and momentum theory. As a
result, if the blade twist distribution is known, it is
possible to solve the induced velocity distribution.
Afterwards, with the known airfoil aerodynamic
characteristics, it is possible to estimate thrust and power
increments along the blade. However, these integrals
become quite complicated for forward flight and here this
approach was used only for hover.
3. NUMERICAL APPROACH
As previously stated, since no set of data is complete and
can provide detailed information in a short amount of
time, several different analytical and numerical
approaches, ranging from fast approximate to highfidelity, were applied.
Necessary airfoil aerodynamic characteristics were taken
from [10, 11].
To account for the ground effect in hover, the BEMT
results were corrected according to [1]:
59
OTEH2016
COMPUTATIONALANALYSISOFHELICOPTERROTORINGROUNDEFFECT
CP = kG CPi + CP0 , kG =
while the forward flight condition is better represented by
moving meshes since periodic unsteadiness during one
rotation can be more accurately captured. Here, both
approaches were used.
1
. (4)
0.9926 + 0.0379(2 R z ) 2
3.3. Unsteady Reynolds Averaged NavierStokes
equations (URANS)
Dirichlet boundary conditions concerning velocity and
pressure were imposed on inlet and outlet boundaries. Noslip boundary conditions were defined on blade and floor
surfaces. Angular velocity of 149rad/s, resulting in fixed
MT, was assigned to rotor zone.
Since two different distances from ground were
considered (z/d = 0.25 and z/d = 0.83) at several different
thrust coefficients (different pitch distributions), several
different computational grids had to be created. All
generated meshes are unstructured, threedimensional and
prismatic extending from 0.83 (0.25) to 1.5 rotor
diameters along the zaxis and 5 blade lengths in both xand ydirections, fig. 3. They contain two fluid zones,
rotor and stator, and a total number of cells of
approximately 2 million. This number was adopted after a
grid convergence study. Meshes are additionally refined
around the blades, fig. 4. Dimensionless wall distance
around the blades is below 5, y+ < 5. Trailing edge of the
blades is modeled as blunt.
Numerical simulations were performed in ANSYS
FLUENT 16.2 where governing flow equations for
compressible, viscous fluid were solved by finitevolume
method. Unsteady Reynoldsaveraged NavierStokes
(URANS) equations were closed by a twoequation k
SST turbulence model. Fluid, air, was considered as ideal
gas whose dynamic viscosity changes according to the
Sutherland law.
Pressurebased coupled solver was used. Gradients were
obtained by the least squares cellbased method. Spatial
discretizations were of the second order. Where needed
(unsteady simulations of isolated rotor in forward flight),
temporal discretization was of the first order. CourantFriedrichsLewy (CFL) number was in the range 15.
4. RESULTS AND DISCUSSION
Obtained results are grouped according to the flight
condition. Rotor hover performance was computed at 5
different collective angles. Tip Mach number was kept
constant in all performed simulations.
4.1. Hover
Obtained relations between thrust coefficient CT and
required power coefficient CP at z/d = 0.83 and z/d = 0.25
are presented in figs. 5 and 6 respectively.
Figure 3. Example of a generated mesh; blue  pressure
inlet, red  pressure outlet, black  ground, yellow interface between rotor and stator
Figure 4. Mesh along the blade surfaces
In order to decrease the number of cells and better resolve
the fluid flow it is customary to generate only a part of the
mesh around a single blade and define periodic boundary
conditions at the sides. This approach is particularly
applicable in axissymmetric hover and vertical flight
conditions. However, since in this study forward flight
condition was also considered, complete meshes were
generated and used for both flight cases.
Figure 5. CP = f(CT) at z/d = 0.83
Although experimental data is available only for z/d =
0.83, all numerical models clearly capture the increase in
aerodynamic performances in ground vicinity, i.e. for the
same thrust less power is required. BEMT results very
well correspond to experimental data. FLUENT results
somewhat underestimate the rotor performance although
the character of the relation CPCT is accurately captured.
Apart from the computational grid appearance, flight
condition also dictates the numerical approach. Isolated
rotor in hover and vertical flight can successfully be
simulated by steady flow in a rotating frame of reference
60
OTEH2016
COMPUTATIONALANALYSISOFHELICOPTERROTORINGROUNDEFFECT
This small discrepancy can, at least partially, be explained
by the existence of wind tunnel walls (not present in
numerical model). Another explanation may be the
numerical setup, i.e. mesh density, turbulence model, the
use of the steady frame of reference approach, etc.
comparison, computed plots are consistent with results
obtained by other authors [5, 12]. Greater pressure
difference resulting in increased thrust is evident for
z/d = 0.25 at inner parts of the blade. Near the blade tip,
for smaller z the velocity increases and pressure
decreases.
Behavior of the wake from the hovering rotor in ground
effect can be presented by streamlines, fig. 9. Slipstream
expansion near the surface is obvious.
Figure 6. CP = f(CT) at z/d = 0.25
Numerical results obtained in FLUENT can also be
validated against a known relation marked as eq. 3, fig. 7.
A plot of the thrust ratio in hover versus relative height
from the ground has been drawn for different geometries
(i.e. different collective pitch angles ). Although small
deviations exist (in particular for higher collective pitch
angles), it can be concluded that the trend of the change
has been successfully captured. The ground effect is
particularly important for z/R < 1, i.e. z/d < 0.5.
Figure 9. Streamlines from the hovering rotor for one
collective angle at z/d = 0.83 and z/d = 0.25
Flow field is affected by the vicinity of the ground, both
slipstream and induced velocities change, resulting in
altered power and thrust coefficients. These changes can
also be illustrated by vorticity fields, fig. 10.
Figure 7. TIGE/TOGE = f(z/R) for different geometries
An illustrative comparison of fluid flows at two different
distances from the ground can be made by comparing
pressure coefficients Cp along two spanwise locations on
the blade, r/R = 0.775 and r/R = 0.945, for a single
geometry (collective angle), fig. 8.
Figure 10. Vorticity contours in [s1] in midplane for one
collective angle at z/d = 0.83 and z/d = 0.25
4.2. Forward flight
The ground effect on rotor performance in forward flight
is also important. However, the flow field around the
rotor gets even more complicated [1]. Flow characteristics
also greatly depend on the forward speed, i.e. advance
ratio , and for the same geometry (collective angle)
thrust coefficient increases in forward flight when
compared to hover. For that reason, forward flight
simulations were performed on 3 different model
geometries, for two distances from the ground z/d = 0.83
and z/d = 0.25, and a single advance ratio = 0.1.
Figure 8. Chordwise Cp distributions at 2 cross sections
for one collective angle at z/d = 0.83 () and z/d = 0.25 ()
In order to accurately simulate forward flight condition, it
was necessary to use sliding mesh approach. Pressure farfield boundary conditions defining values of pressure,
Full line denotes z/d = 0.83 while dashed line refers to
z/d = 0.25. Although no experimental data is available for
61
COMPUTATIONALANALYSISOFHELICOPTERROTORINGROUNDEFFECT
velocity and turbulence quantities were assigned to outer
domain surfaces. Time step corresponds to an angular
increment of 5. A great number of rotations (around 10)
was necessary for attaining quasiconvergence of thrust
and power coefficients.
OTEH2016
Blade sectional pressure coefficients at the radial
positions x/R = 0.775 and x/R = 0.945 during one
revolution with the angular increment of = 60 are
presented in figs. 13 and 14 respectively. Full line denotes
z/d = 0.83 and dashed line refers to z/d = 0.25. Again,
computed results are comparable the other published
results [6, 13]. Since the advance ratio is low, overall flow
variations are smaller at the inner part of the blade.
Results computed in ANSYS FLUENT, marked by
square symbols in figs. 11 and 12, were compared to
experimental data (where available) and MT results.
Again, power coefficients computed by URANS are
somewhat higher than those by MT, and both sets of
numerical results seem higher than experimental values.
However, the character of the relations seems to be well
captured although discrepancies increase with the increase
of advance ratio (i.e. for = 0.1). Ideally, for thorough
analysis a complete map of aerodynamic performances
should be generated (for various and ). However, since
the simulations require large amounts of time, at this stage
of the study, only individual results are presented (as
discrete points on the graphs). One color refers to a single
value of thrust coefficient.
Figure 13. Chordwise Cp distributions at x/R = 0.775
Figure 11. CP = f() at z/d = 0.83
Figure 12. CP = f() at z/d = 0.25
Although thrust coefficient are different, smaller amounts
of required power at the lower distance from the ground
(z/d = 0.25) for low advance ratios are evident (i.e. for the
same CP much higher CT can be achieved). Unfortunately,
no experimental data is available for the employed model
of the helicopter rotor at low advance ratios.
Figure 14. Chordwise Cp distributions at x/R = 0.945
Main differences in the two ground distances can
primarily be seen at the advancing side of the rotor. At the
retreating side, pressure distributions seem quite similar.
62
OTEH2016
COMPUTATIONALANALYSISOFHELICOPTERROTORINGROUNDEFFECT
Flow structures in slow forward flight can be illustrated
by streamlines, fig. 15. Region of flow recirculation
formed upstream of the rotor at z/d = 0.25 is noticeable.
Employed CFD solver seems able to reproduce the
transient flow and loading of the blade.
References
[1] Leishman,J.G.:
Principles
of
Helicopter
Aerodynamics, 2nd ed., Cambridge University Press,
New York, 2006.
[2] Yeager,W.T. Jr.: Noonan, K. W., Singleton, J. D.,
Wilbur, M. L., Mirick, P. H., Performance and
Vibratory Loads Data from a WindTunnel Test of a
Model Helicopter MainRotor Blade with a PaddleType Tip, NASA TM 4754, Hampton, Virginia,
1997.
[3] Pomin,H., Altmikus,A., Buchtala,B., Wagner,S.:
Rotary Wing Aerodynamics and Aeroelasticity, in
High performance Computing in Science and
Engineering
2000,
SpringerVerlag
Berlin
Heidelberg, 2001.
[4] Beaumier,P., Bousquet,J.M..: Applied CFD for
analyzing aerodynamic flows around helicopters,
24th International Congress of the Aeronautical
Sciences, Yokohama, Japan, 2004.
[5] Barakos,G., Steijl,R., Badcock,K., Brocklehurst,A.,
Development of CFD capability for full helicopter
analysis, 31st European Rotorcraft Forum, Florence,
Italy, 2005.
[6] Antoniadis,A.F., Drikakis,D., Zhong,B., Barakos,G.,
Steijl,R, Biava,M. et al.: Assessment of CFD methods
against experimental flow measurements for
helicopter
flows,
Aerospace
Science
and
Technology, 19 (2012) 86100.
[7] Cheeseman,I.C., Bennett,W.E.: The Effect of the
Ground on a Helicopter Rotor in Forward Flight, R.
& M. No. 3021, London, 1957.
[8] Ganesh,B.: Unsteady Aerodynamics of Rotorcraft at
low Advance Ratios in Ground Effect, Ph.D. thesis,
Georgia Institute of Technology, 2006.
[9] Pulla,D.P.: A study of helicopter aerodynamics in
ground effect, Ph.D. thesis, The Ohio State
University, 2006.
[10] Bingham,G.J., Noonan,K.W.: TwoDimensional
Aerodynamic Characteristics of Three Rotorcraft
Airfoils at Mach Numbers from 0.35 to 0.90, NASA
TP 2000, Hampton, Virginia, 1982.
[11] Noonan,K.W.: Aerodynamic Characteristics of Two
Rotorcraft Airfoils Designed for Application to the
Inboard Region of a Main Rotor Blade, NASA TP
3009, Hampton, Virginia, 1990.
[12] Jarkowski,M.,
Woodgate,M.A.,
Barakos,G.N.,
Rokicki,J.: Towards consistent hybrid overset mesh
methods for rotorcraft CFD, Int. J. Numer. Meth.
Fluids, 74 (2014) 543576.
[13] Biava,M., Khier,W., Vigevano,L.: CFD prediction of
air flow past a full helicopter configuration,
Aerospace Science and Technology, 19 (2012) 318.
Figure 15. Streamlines from the rotor in forward flight
for one collective angle at z/d = 0.83 and z/d = 0.25
5. CONCLUSION
Aerodynamic performances of an isolated model
helicopter main rotor in ground effect (at two distances
from the ground) obtained by several different analytical
and numerical approaches, ranging from fast approximate
to highfidelity detailed solutions, were compared. For
most cases, satisfactory outcome (agreement with
experimental data) was accomplished given the fact that
relatively limited success has been achieved in correctly
predicting rotor performance in ground effect when
compared to experimental results [1]. This is due to the
problem complexity and strongly viscous nature of the
rotor IGE problem.
None of the employed numerical models can fully capture
the complexity of the flow. However, through their
combination and comparison with available experimental
data many useful pieces of information can be extracted.
Presented results have contemporary importance and
enable the development of a more efficient rotor design.
They also provide insight into complex flow fields around
a representative rotor in hover and forward flight (two
quite different, but equally important flight regimes).
Although additional work is necessary, it is possible to
use presented numerical setups to assess possible
increase of aerodynamic performances in ground effect.
ACKNOWLEDGEMENT
The paper is a contribution to the research TR 35035
funded by the Ministry of Education, Science and
Technological Development of the Republic of Serbia.
63
SIMULATION OF ROLL AUTOPILOT OF A MISSILE WITH
INTERCEPTORS
MILAN IGNJATOVI
Military Technical Institute, Belgrade, milan.ignjatovic@hotmail.rs
MILO PAVI
Military Technical Institute, Belgrade, cnn@beotel.rs
SLOBODAN MANDI
Military Technical Institute, Belgrade, msmanda@open.telekom.rs
BOJAN PAVKOVI
Military Technical Institute, Belgrade, bjnpav@gmail.com
NATAA VLAHOVI
PhD studies: School of Electrical Engineering, University of Belgrade,
Work place: Military Technical Institute, Belgrade, natasha.kljajic@yahoo.com
Abstract: In this paper a nonlinear simulation of roll autopilot of a missile using interceptors as actuators is presented.
Transfer function of a missile roll velocity used in this paper is a first order transfer function. Autopilot design was done
without nonlinear block. Simulation is developed using computer software. Results of a simulation, with several different
values of changing parameters are given.
Keywords: simulation, missile, roll, autopilot, interceptors.
1. INTRODUCTION
GA =
Although most guided missiles use fins for stabilizing and
controlling missile flight, some use interceptors.
Interceptors are surfaces intercepting the air flow. Their
actuators are electromagnets causing their movement.
K
T s + 1
(1)
where K (K FI ) = 14 , and T (TFI ) = 0.32 .
Transfer function of interceptors is given with steady state
value equals unity:
In this paper we have developed simulation for roll autopilot
of an air to land guided missile which has interceptors.
GAKT =
Missile has four fixed wings. There are six interceptors
mounted on the rear side of the wings. They are all paired,
so there are three pairs. One pair of interceptors is used for
roll stabilization and is mounted on one pair of wings. Other
two pairs are mounted on the other pair of wings and serve
for yaw and pitch control. Surfaces of roll pair are moved in
opposite direction while surfaces of yaw and pitch pairs are
moved in the same direction.
1
Ta s + 1
(2)
Where Ta = 0.002 .
NL_Func block is where the command signal going to
actuators (eta) is calculated from command signal
(zeta). This calculation is given in Section 3.
dis (zeta_dis) is disturbance signal equivalent to surface
Simulation is developed using Matlab and Simulink
software package.
deflection due to fault in construction.
K and GR are parameters of roll autopilot. They are
determined in design procedure given in [1], which assumes
ailerons as control surfaces. Values of determined
parameters are equal to 5.7 and 0.5 respectively
2. SIMULATION MODEL
Simulation model is given in Picture 1. Roll velocity
transfer function is given as in [1] as a first order:
64
OTEH2016
SIMULATIONOFROLLAUTOPILOTOFAMISSILEWITHINTERCEPTORS
Picture 1. Simulation model
3. COMMAND SIGNAL CALCULATION
t c = f ( , Ts , t)
Interceptors can have only two steady state values, 1 and 1.
It means that interceptor is either all the way on one side of
the wing (position 1) or all the way on the other side
(position 1).
(4)
Command time is limited, so that it cannot be larger than
command sample time, and with that command signal
effect is practically limited from 0 to 1:
So, in order to control the missile with interceptors, we must
change the time during which interceptors are in position 1
and in position 1.
t + Ts , 0 < < 1
tc = t + Ts , 1
t
, 0
Here, the internal structure of NL_Func block from Picture
1. is explained. NL_Func block has as its inputs command
signal , simulation time t and command sample time Ts .
Its output is command signal . Time_past input and output
is the same simulation global variable which is used to
remember the time of last command calculation. This form
of global variable manipulation is done because of software
specifics.
(5)
Basically what we are doing is modulating command
sample period with percentage of period determined from
command signal thus linearly transporting signal to
command signal going to actuators , from 0 to 1 (with
saturation at 0 and 1) to 1 and 1.
Command signal is calculated periodically. This period
should not be too small, because we are restricted with
actuator's time constant, which is as we said 2 [ms], and
with time needed for forming aerodynamic forces due to
interceptors position change.
This period is called command sample time ( Ts ) . In Section 4.
it is shown how different values of Ts affect the output value.
On Picture 2. one command sampling period is shown.
Command_time input and output is the same global variable
which is used to remember the last calculated command
time tc value.
We can see that before time reaches command time tc ,
command signal has value 1, and after this moment value
1. Here, it is given mathematically:
1, t < t c
1, t > t c
Picture 2. tc dependence
(3)
4. SIMULATION RESULTS
In this Section we presented results for three different values
of command sample time.
Command time ( tc ) is calculated as a function of command
signal ( ), command sample period ( Ts ) and time ( t ):
On Picture 3, we can see step responses to /4 step input for
65
OTEH2016
SIMULATIONOFROLLAUTOPILOTOFAMISSILEWITHINTERCEPTORS
command sample time values of Ts = 5 [ms] , Ts = 15 [ms]
and Ts = 30 [ms] .
Picture 5. Ts = 15 [ms]
On Picture 5. we see the change of and a for
Ts = 15 [ms] .
Picture 3. Step responses
Step responses change with changing command sample
time.
We can see that value of is beginning to fluctuate when it
reaches steady state value because G A output is starting to
react to change in a . It is not expected for real system to
behave like this.
With smaller values of command sample time transient
process is faster, response has less overshoot and larger
static error.
With larger values of command, sample time delay is added
to response, overshoot is larger and static error is smaller.
Picture 6. Ts = 30 [ms]
On Picture 5. we see the change of and a for
Ts = 30 [ms] .
Picture 4. Ts = 5 [ms]
In Picture 4. we can see the change in command signal
It can be seen that for this high value of Ts , interceptors are
able to achieve the given command. and fluctuation is
increased due to added delay to the system.
and interceptor movement a for Ts = 5 [ms] . We have
said that command signal is practically limited from 0 to
1. Here, it is represented as it is calculated, without
limitations.
Shorter command sample time has better transient process
and larger static error, while larger command sample time
has worse transient process and smaller static error.
Up to approximately 0.31 [s] a is not modulated
because time constant of interceptors is too small and
interceptor cannot achieve the demanded change.
5. CONCLUSION
When response reaches settle time we can see that command
signal has value of 0.5, which as we said, corresponds to
zero command . This means that for half of the command
sample time interceptors will be in position 1 and for the
other half in opposite position.
In this paper, one way of simulating roll autopilot of a
missile using interceptors is presented. We have shown
simulation results for three different values of command
sample time. Changing command sample time has influence
on step response of the roll autopilot.
66
OTEH2016
SIMULATIONOFROLLAUTOPILOTOFAMISSILEWITHINTERCEPTORS
[2] Garnell,P.: Guided Weapon Control System, Second
Edition, Pergamon Press, New York, 1980.
[3] Palm,W.III: System Dynamics, McGrawHill, New
York, 2014.
References
[1] Program APD Theoretical Manual, Military
Technical Institute, Belgrade, 2003.
67
DESIGN OF THE MAIN PIVOT ON THE FORCED OSCILLATION
APPARATUS FOR THE WIND TUNNEL MEASUREMENTS
MARIJA SAMARDI
Military Technical Institute, Belgrade, majasam@ptt.rs
DRAGAN MARINKOVSKI
Military Technical Institute, Belgrade, marinkovskid@ikomline.net
DUAN URI
Military Technical Institute, Belgrade, dusan.curcic@vti.vs.rs
ZORAN RAJI
Military Technical Institute, Belgrade, zoran.rajic@vti.vs.rs
ABDELWAHID BOUTEMEDJET
Military Academy, Belgrade, abdelwahed1954@gmail.com
Abstract: The main pivot on the forced oscillation apparatus for dynamic measurements in the T38 wind tunnel is
described in this paper. Design of such element is complicated by restricted space inside the wind tunnel models. The
pivot of the T38 forced oscillation apparatus is formed from a pair of symmetrical crossflexures. Two types of the
crossflexures are presented: crossflexures with uniform crosssection of the strips and crossflexures with variable
crosssection of the strips. Stress analysis of the crossflexures showed that strips with variable crosssection much
better matched strict requirements of the dynamic wind tunnel measurements.
Keywords: flexure pivot, oscillations, wind tunnel, dynamic stability derivatives.
A special kind of flexure pivot is crossflexure pivot,
Picture 2. It has a bisymmetrical geometry and contains
two leaf springs of equal dimensions crossing at their
midpoints [14]. These pivots permit a high rotational
accuracy to obtain via compact, reliable and maintenancefree design with limited production costs. The crossflexure pivots are superior to conventional joint in
controlling an oscillatory motion. They are characterized
by high compliance with respect to the inplane rotational
degree of freedom and high stiffness in other, secondary,
degrees of freedom. These characteristics make them very
useful in dynamic wind tunnel experiments with forced
oscillation motion [5,6].
1. INTRODUCTION
The idea of supporting the moving parts of sensitive
apparatuses on thin strips of metal or other elastic
material rather than on other types of pivots and bearings
is not new. The simplest and most common type of
flexure pivot consists of a tin metal strip which is free to
bend. Frequently two of these, one at right angles to the
other, are machined out of a rod, Picture 1. This provides
ball and socket action in roads subjected to tension or
compression and having only a negligible amount of
moment.
Picture 1. Flexure pivot
Picture 2. Crossflexure pivot
68
OTEH2016
DESIGNOFTHEMAINPIVOTONTHEFORCEDOSCILLATIONAPPARATUSFORTHEWINDTUNNELMEASUREMENTS
strips were selected in design of this elastic element.
2. DYNAMIC EXPERIMENT IN THE T38
WIND TUNNEL
Performance parameters of the apparatus for the pitching
experiments, according to the expected aerodynamic
loads in the T38 wind tunnel dynamic experiment are:
 model oscillation amplitude: 0.25  1.5
 model oscillation frequency: 115 Hz
 maximum normal force Rm =1800 N
 maximum axial force: Rt = 5600 N
 maximum side force: Rs = 3000 N.
The main task of the dynamic experiment in wind tunnels
is to obtain modelscale dynamic stability information of
the aircraft at realistic Reynolds and Mach numbers. The
forced oscillation techniques are the most often used for
the dynamic experiments [7]. The apparatus for the
dynamic measurements in the T38 wind tunnel is a fullmodel forced oscillation apparatus with the primary
angular oscillation around the wind tunnel model
transversal axis. The wind tunnel model is forced to
oscillate at constant amplitude. The apparatus is
distinguished by the capability to measure aerodynamic
reaction in the primary and secondary degrees of freedom.
The front part of the apparatus is shown in Picture 3.
3. THE CROSSFLEXURE PIVOT
In the first design concept of the pivot the crossflexures
were considered as rectangular strips of equal width and
thickness along the strips. The first design of the pivot
and load on the crossflexure pivot in the horizontal plane
is shown in Picture 4, where Rs is aerodynamic side force,
Ms is bending moment, c.p. is center of pressure, e is
distance of the internal flexure strip axis from longitudinal
axis of the apparatus, l is length of the each strip, f is
distance of the outside flexure strip axis from longitudinal
axis of the apparatus, c is distance between two blocks of
the crossflexures xp is distance from the centre of
pressure to the centre of the crosssection.
The bending moment at the centre of the inner surface of
the mowing block is:
Picture 3. Apparatus for the T38 wind tunnel dynamic
measurement [8]
c
M S = Rs ( x p )
2
The structural rigidity is a very important requirement in
any forced oscillation apparatus. The elastic element has
to be designed in such way that permits the primary
motion of a model, but its high stiffness in other degrees
of freedom is necessary to withstand the aerodynamic
loads with negligible deflection. This relation between
relatively high compliance in the primary degrees of
freedom and high stiffness in the secondary degrees of
freedom is especially important in the measurement of the
cross and crosscoupling derivatives [9]. Because of all
these requirements the crossflexure pivot is chosen for
the main pivot on the T38 wind tunnel forced oscillation
apparatus. The special shape of the individual flexure
(1)
The bending moments at the end of strips, MA and MB,
are:
MA = MB =
Rs l
8
(2)
The total direct forces in the one pair of strips in
horizontal plane YA and YB, are:
YA = YB =
Picture 4. Aerodynamic load in horizontal plane
69
Rs l
Ms
+
4 (e + f )
2 (e + f )
(3)
DESIGNOFTHEMAINPIVOTONTHEFORCEDOSCILLATIONAPPARATUSFORTHEWINDTUNNELMEASUREMENTS
OTEH2016
Maximum bending stress in strips in horizontal plane, A,
is:
A =
(4)
YA M A
+
Ah wh
where Ah is crosssection area of strips and wh is section
modulus of strips.
Table 1 lists values of the total forces; bending moments
and bending stress for one pair of strips in horizontal
plane obtained according Equation (4).
Table 1. Results of design analysis in horizontal plane
One pair
strips in
horizontal
plane
Ms
[Nm]
Y A= Y B
[N]
MA=MB
[Nm]
[N/mm2]
85.5
1943.3
21.75
789.8
Picture 6. Maximum stress in the strips in horizontal
plane
A load applied in the stress analysis done by NASTRAN
NX V9 is shown in Picture 5. Yawing moment at the
centre of the crossflexure pivot, Ms = RS xp, is generated
by side force RS. Side force and yawing moment at the
centre of the crossflexure pivot were simulated by side
force which acted at the front surface of the mowing
block of the pivot.
The stress analysis showed that the aerodynamic loads
capability may be improved by designing with variable
crosssection of the strips, Picture 7. The largest thickness
in the centre and the largest width at the end of the strips
were chosen for achieving the greatest loadscarrying
capability for a different allowable stress and restricted
space inside the wind tunnel models. Both the elastic
elements, with uniform and variable crosssection of the
strips, are equal in general dimensions: distance between
moving and immobile block, lent of the flexures,
distances of the internal and outside flexure strips axis
from longitudinal axis of the apparatus.
Picture 7. Crossflexure pivot with the variable crosssection strips
In the stress analysis of the strips with the variable crosssection load was applied in the same way as in the stress
analysis of the strips with uniform crosssection. For the
simulated loads in the horizontal plane, maximum normal
stress, 1, is obtained closely at the centre of strips and
stress value is more than two times lower than in strips
with uniform crosssection, Picture 8.
Picture 5. Load applied to the crossflexure in horizontal
plane
Increasing stiffness of the elastic element in the horizontal
plane has not led to noticeable increase in stiffness in the
primary degree of freedom [10].
For the simulated load maximum bending stress is
obtained at the end of strips, Picture 6.
70
DESIGNOFTHEMAINPIVOTONTHEFORCEDOSCILLATIONAPPARATUSFORTHEWINDTUNNELMEASUREMENTS
OTEH2016
sensors or can be connected in the one sensor. Basic
characteristics of the primary oscillatory motion sensor
are shown in Table 2, where FS is sensor full scale.
The primary oscillatory motion sensor enables good
signal of the model primary motion. This is very
important in the data reduction procedure. In most cases
signals from the sensors of the secondary oscillations as
well as signals from the excitation moment sensors are
seriously contaminated by the noise generated mainly by
flow unsteadiness. As a nose level can be several times
higher than that of a desired signal, it is generally very
hard task to extract them adequately. Knowledge that the
desired signals from these sensors for excitation moment
and secondary oscillations are coherent with the primary
motion permits the use of crosscorrelation technique.
Crosscorrelation technique is especially suited to
applications where a clean reference signal coherent with
the one that needs to be extracted from the noise is
available. The signal from the primary oscillatory motion
sensor is noisefree. The amplitude of this signal is
determined by applying autocorrelation functions [11].
The primary oscillatory motion signal from the sensor
realized on the crossflexure pivot is used as reference
signal in order to determine the secondary oscillations and
amplitude of the excitation moment using the crosscorrelation functions.
Picture 8. Maximum stress in the strips with variable
crosssection in horizontal plane
3.1. The primary oscillatory motion sensor
In the T38 forced oscillation experiments the amplitude
of the excitation moment, amplitude of the model angular
oscillatory motion and phase shift between these
quantities have to be measured. The amplitudes of the
excitation moments are measured by the strain gauge
balance.
5. CONCLUSION
The amplitudes of the wind tunnel model angular
oscillatory motion are measured with the primary
oscillatory motion sensor located on the crossflexure
pivot.
One of the main tasks in design of the crossflexure pivot
on the T38 forced oscillation apparatus is to achieve the
greatest possible stiffness in horizontal plane and side
force capability for a given overall dimension of the
elastic element. These dimensions are specified by
available space within the wind tunnel models for
dynamic experiments, i.e. by dimensions of the dynamic
balance and actuator arm of the apparatus. At the same
time, the pivot should provide large capability of the
aerodynamic normal force which is dominant component
of the aerodynamic load in experiments. Since the crossflexure pivot enables primary oscillatory motion of a
model this large capability of the normal force should not
diminish the high compliance in the primary (pitch)
degree of freedom. Two different shapes of the flexure
strips are considered: the first analysis was done for
uniform crosssection along strips and the second analysis
was done for strips with larges thickness in the centre and
the largest width at the ends.
Picture 9. Primary oscillatory motion sensor
Table 3. The results of the analyses
Maximum bending stresses in strips [N/mm2]
NASTRAN NX V9 results
Vertical
Horizontal
Strips with uniform
plane
plane
crosssection
580
790
Vertical
Horizontal
Strip with variable crossplane
plane
section
660
350
Table 2. The primary oscillatory motion sensor
characteristics
Sensor measuring
Maximum error
range
[% FS]
[]
1.5
0.25
Hysteresis
[% FS]
0.15
There are two primary oscillatory motion sensors on each
side of the crossflexure pivot, Picture 9. Measuring
bridges are formed from 350 foiltype strain gauges
(Vishay Micro Measurements TK06S075P350/DP).
These measuring bridges can be used as individual
Results of the crossflexure stress analyses are shown in
Table 3 [10]. For the same given normal force, the values
of maximum bending stress in vertical plane for the strips
71
DESIGNOFTHEMAINPIVOTONTHEFORCEDOSCILLATIONAPPARATUSFORTHEWINDTUNNELMEASUREMENTS
with uniform and variable geometry in the pitch plane are
approximately equal. But, for the same given side force,
maximum bending stress in the strips with variable crosssection is more than two times lower. The variable crosssection of strips provided significant increase in stiffness
of the crossflexure pivot in horizontal plane. It is very
important that this increase in stiffness has not led to
noticeable increase in stiffness in primary degree of
freedom. Such design of the elastic element provides
required amplitudes of a model oscillatory motion in the
T38 wind tunnel experiments.
OTEH2016
[5] AEDC, Von Karman Gas Dynamics Facility (VKF),
http://wwwnimr.org/systems/images/vkf.html.
[6] National Research Council, Canada, http://www.nrccnrc.gc.ca/eng/solutions/facilities/wind_tunnel_index
.html
[7] OrlickRckeman,K.J.: Review of techniques for the
determination of dynamic stability parameters in
wind tunnels, AGARDLS114, The Advisory Group
for the Aerospace Research and Development,
NATO Research and Technology Organization,
Brussels, Belgium, 1981.
[8] Samardi,M., Anastasijevi,Z., Marinkovski,D.:
Comparison of the T38 wind tunnel data obtained
by static and dynamic tests, in Proceedings of 6th
International Scientific Conference on Defensive
Technologies OTEH 2014, Belgrade, Serbia, (2014)
2125.
[9] Samardi,M., Anastasijevi,Z., Marinkovski,D.,
Isakovi,J.: Measurement of the crosscoupling
derivatives due to pitching in the high Reynolds
number blowdown wind tunnel, in Proceedings of
29th Congress of the International Council of the
Aeronautical Sciences, ST. Petersburg, Russia
(2014) 18.
[10] Samardi,M., Marinkovski,D., Anastasijevi,Z.,
uri,Z., Raji,Z.: An elastic element of the forced
oscillation apparatus for dynamic wind tunel
measurements, Aerospace Science and Technology,
50 (2016) 272280.
[11] Samardi,M.,
Isakovi,J.,
Milo,M.,
Anastasijevi,Z., Nauparac,B.D.: Measurement of
the direct damping derivative in roll of the two
calibration missile model, FME Transaction, 41
(2013) 189194.
ACKNOWLEDGMENT
This study was supported by the Military Technical
Institute (VTI) and Ministry of Education, Science and
Technological Development of Serbia (project number
TR 36050).
References
[1] Young,W.E.: An investigation of the crossspring
pivot, J. Appl. Mech., 11 (1944) A113A120.
[2] Zelenka,S., Bona,DeF.: Analytical and experimental
characterization of highprecision flexural pivots
subjected to lateral loads, Precs. Eng., 26 (2002)
381388.
[3] Hongzhe,Z., Shusheng,B.: Accuracy characteristics
of the generalization crossspring pivot, Mech.
Mach. Theory, 45 (2010) 14341448.
[4] Hongzhe,Z., Shusheng,Y., Jingjun,Y., Guanghua,Z.:
The accurate modeling and performance analysis of
crossspring pivot as a flexure module, in
Proceedings of ASME IDETC/CIE , Brooklyn, New
York, USA (2008) 17, ASME Paper No.
DETC200849694.
72
PRELIMINARY AERODYNAMIC COMPUTATION OF LONG
ENDURANCE UAV WING
ABDELWAHID BOUTEMEDJET
Military Academy, Belgrade, abdelwahed1954@gmail.com
MARIJA SAMARDI
Military Technical Institute, Belgrade, majasam@ptt.rs
ZORAN RAJI
Military Technical Institute, Belgrade, zoran.rajic@vti.vs.rs
Abstract: Wing preliminary aerodynaic computation of a low speed, long endurance unmanned aerial vehicles (UAV ), is
formulated as a single objective aerodynamic optimization.During this process,the wing planform parameters are optimized
with maximization of endurance .Four design variables from aerodynamics discipline namely taper ratio,aspect ratio,wing
loading and wing twist are taken in concediration.3D wing aerodynamic analysis is performed by the XFLR5 panels
method code.In the optimization process ,genetic algorithm is used to perform the optimal solution under defined
requirements of preliminary computation.
Keywords: UAV,Wing endurance,Optimization,Genetic algorithm,Panel method.
1. INTRODUCTION
2. PROBLEM FORMULATION
Aircraft design is a discipline of aeronautical engineering
different from the analytical disciplines such as
aerodynamics, structures, controls and propulsion. The
design process for an UAV is divided into three major
phases: the conceptual design, preliminary design and the
detailed design. The conceptual design in which the basic
question of configuration, size, weight and performance
based on the mission specifications and requirements are
presented in [13]. The preliminary design then advances
these concepts by individually designing and sizing of the
major components of an aircraft. The UAV aerodynamic
design starts with preliminary computation to satisfy the
performance requirements.
The main goal of the UAV design is to ensure long
endurance flying under specified requirements. The wing is
the main part that defines UAV performances. Only wing
design is considered for optimization study. In this
optimization process, the design aims at maximizing the
endurance, which is an aerodynamic aspect. Basing on this
aspect, objective function, design variables and constraints
are involved to formulate the optimization problem. The
airfoil SD7062 is chosen in this paper to get the optimized
shape.
2.1. Objective function
The choice of the objective function in any optimization
problem is dectated by the design requirements of the
aircraft. One may find objectives such as life cycle cost or
profit. Since these functions are usually very hard to connect
to the design variables via objective functions, lower level
related objectives may be used. For UAV, achieving
maximum endurance is required for the most aircrafts, the
fact that allows it to accomplish missions specified to
UAVs, also endurance is an ideal lower level optimization
objective for the aerodynamic designer. An initial estimate
of which can be obtained from the Breguet endurance
equation:
Many approaches have been used to realize the preliminary
computation. In this paper a numerical method for
optimizing aerodynamic has been explored. These
numerical methods are classified into one of three general
categories: inverse methods, gradientbased methods and
genetic algorithms (GA). The inverse method in
aerodynamic design seeks to determine the aerodynamic
shape for a specified surface pressure distribution. The
general idea associated with gradient methods consists of
the determination of the optimization's objective, the
parameterization of the geometry, and the computation of
the direction in the design space [4].
C 3/ 2
E= L
2 S 1 1
Wf
c CD
Wi
The GA is used to perform the optimization. Such method is
adequate for non smooth design spaces which may contain
many local optimums. It is used also for multipointpoint
design computations. One of the disadvantages of the GA is
expense where the number of function evaluations required
for such method is exceeded [5, 6].
(1)
where:
is propulsive efficiency, c is specific fuel consumption, CL
73
OTEH2016
PRELIMINARYAERODYNAMICCOMPUTATIONOFLONGENDURANCEUAVWING
is the lift coefficient, CD is the drag coefficient, Wi and Wf
are aircraft weights at the beginning and the end of the
cruise segment respectively.
h2 (t ) = VS ( t ) VS *
(4)
h3 (t ) = VM ( t ) VM *
(5)
The optimization of CL3/2/CD will contribute to optimizing
endurance, so the objective function is the maximization of
the rate CL3/2/CD [7].
where t = [ AR WS TR ]
2.2. Design variables
3. OPTIMIZATION PROCESS
The parameters related to the wing planform are considered
as the design variables. The aspect ratio (AR), wing Loading
(WS), taper Ratio (TR), twist angle () are chosen to reflect
the effect of aerodynamic discipline during the 3D wing
optimization process. The following figure shows the wing
shape:
3.1. Aerodynamic analysis
Vortex Lattice Method (VLM) is used for the wing analysis.
The wing is defined as a set of panels, that determined by
the following parameters: length, root, tip chord and
dihedral angle. The principle of a VLM is to model the
perturbation generated by the wing by a sum of vortices
distributed over the wings planform. The strength of each
vortex is calculated to meet the appropriate boundary
conditions, i.e. non penetration conditions on the surface of
the panels. The viscous drag is estimated by interpolation of
XFoil pregenerated polars, by the CL value resulting from
the linear VLM analysis. These code is well validated with
tunnel test experiments and other CFD codes [8].
Cr is the root chord, Ct is the tip chord, b is the wing span, S
( Cr + Ct ) b
2 , and W is the UAV
is the wing area, S = 2
weight.
The performance constraints are evaluated using a
programmed function as a simple computation for a
preliminary computation. The aerodynamic characteristics
obtained are used as a part of the inputs of this function to
evaluate the performance constraints.
3.2. Genetic algorithm
Picture 1. Design variables
Genetic algorithm is a search algorithm based on natural
selection and genetics, was first used by John Holland. The
Genetic algorithm presented in this paper utlizes three
operators; passthrough, crossover, and mutation [9]. In this
algorithm; gens, chromosomes and fitness represent
respectively: design variables, design condidates and
objective function. After the design space is defined, the
next step is to form an initial population of 50
chromosomes. The values of genes corresponding to each
chromosome are chosen randomly between fixed limits. For
each formed chromosome, fitness function is computed
using endurance function evaluation. Ranking process is
used after fitness computation, the most fit individual takes
the first ranks until the last rank corresponding to population
size. The passthrough operator is used to enable 10 % of
the fittest chromosomes to pass to the nexte generation
without any change.A simple random crosse over operator is
utilized. A determined number of chromosomes (20% of
population) undergo a modification using this operator;
where all selected chromosome genes are combined together
to forme new individuals. The other common operator is
mutation, in which a subset of genes are chosen randomly,
where their values are changed to produce 70 % of the new
individuals.
2.3. Constraints
There is a set of constraints usually associated with a wing
design. In this paper only aerodynamic constraints are
considered to formulate optimization problem. This
constraints are imposed on the performance parameters of
the UAV. The aerodynamic constraints are imposed on:
Rate of climb (ROC)
Stall speed (VS)
And maximum speed (VM)
This constraints arise from the tactical requirements, where
the UAV has to be hand launch at speed of 10 m/s,
accomplishing their missions at the height of 300 m and to
have maximum speed about 120 km/h because of the
propulstion system power limitations. This maximum speed
is delivered by an electrical motor.
2.4. Mathematical formulation
The 3D wing design process explored in this paper can be
formulated as a classical optimization system as following:
C 3/ 2
f (t ) = max L
CD
(2)
3.3. Fitness function evaluation
Constraints are:
h1 (t ) = ROC ( t ) ROC
Rapid computation is evaluated using panels method, The
XFLR5 flow solver is applied to compute aerodynamic
coefficients. The fitness function is evaluated using the lift
(3)
74
OTEH2016
PRELIMINARYAERODYNAMICCOMPUTATIONOFLONGENDURANCEUAVWING
and the drug coefficients. To make the process of
optimization more simple and rapid, the flow solver is
decoupled from the process, and the programmed genetic
algorithm is implemented with a trained artificial neural
networks data that gives an approximation of aerodynamic
coefficient. The data base used to train artificial neural
networks is obtained from a computation set, evaluated
using XFLR5.
Picture 2. Artificial Neural Network
3.4. Optimization
The optimization is performed using the encoded algorithm
given by the diagram shown in Picture 3. First the airfoil
SD7062 with thickness of 14 % is chosen to generate the
planform of the optimized wing. This airfoil design work
was conducted by David Wood and his research group. The
large relative thickness and the highlift characteristics of
the SD7062 were what first triggered the selection of this
low Reynolds number airfoil for small UAV [10]. The wing
parameters are determined using the defined objective
function. Optimization model is formed essentially off ANN
block, performances evaluation block, and the GA
optimization block. The ANN is applied to determine lift
and drug aerodynamic coefficients, and stall speed. The
performances block is used to compute rate of climb and
maximum speed. The GA block is used to evaluate the
optimization process and determination criteria satisfied.
Picture 3 Optimization process
4. RESULTS AND DISCUSSION
The optimized shape of the desired wing is illustrated in the
following figure. The wing is given by the distribution of
pressure coefficient and the stream line. The flow structure
takes the shape of parallel stream lines behind the wing
planform, which is suitable for good fly performances where
the drug value is small because of the uniform flow. At the
two ends of the wing it is easy to observe the formation of
3D structures which are wingtip vortices formed because of
the pressure gradient between the two surfaces of the wing.
Picture 4. Wing planform
75
OTEH2016
PRELIMINARYAERODYNAMICCOMPUTATIONOFLONGENDURANCEUAVWING
The aerodynamic design variables of the obtained solution
using GA are shown in Table 1:
Table 1. Wing design variables
wing Loading (WS)
9.25
aspect ratio (AR)
10.5
Taper Ratio (TR)
1.57
twist angle ()
2.01
The weight of the UAV is fixed to be about 7 Kg, basing on
the obtained parameters; the geometrical characteristics of
the wing are resumed in Table 2:
Table 2. Wing geometry variables
wing area (S)
0.756 m2
Wing span (b)
2.818 m
Root chord (Cr)
0.351 m
Tip chord (Ct)
0.292 m
Picture 6. ROC variations with speed
The performances of the evaluated wing after the process of
GA optimization are given in Table 3.
Table 3. Wing performances
Endurance factor ( CL3/2/CD)
Rate of climb (ROC)
Maximum speed (VM)
Stall speed (VS)
5. CONCLUSION
23.453
7.48
41 m/s
10 m/s
Results indicate that the genetic algorithm is easy to
implement and extremely reliable, being relatively
insensitive to design space. The presented paper resumes the
preliminary aerodynamic computation for a small UAV
under defined requirements using panels method to resolve
flow around the wing and GA to evaluate the optimized
wing shape.
The Picture 6 represents the variation of the endurance
coefficient with angle attack (alpha). The maximum value of
this coefficient takes place for an angle of attack about 5o. In
real cases, cruising angle is less than 5o, according to this
diagram, till an angle of attack of 2.5o, it is illustrated that
the endurance coefficient stills have relatively high values.
References
[1] Daniel P. Raymer, Aircraft Design: A Conceptual
Approach American, Institute of Aeronautics and
Astronautics, USA, 1989.
[2] Egbert Torenbeek, Synthesis of subsonic air plane
design, Delft University Press, The Netherlans,1976.
[3] Jan Roskam, Airplane design,The University of Kanas,
Lawrence, 2000.
[4] Terry L. Host, Thomas H.Pullin, "Aerodynamic Shape
OptimizationUsing A RealNumberEncoded Genetic
Algorithm", 19th applied aerodynamics conference,
2001.
[5] Obayashi.S,
Tsukahara.T,
"Comparison
of
Optimization Algorithms for Aerodynamic Shape
Design", AIAAJ, 35(1997) 14131415.
[6] Bock, KW, "Aerodynamic Design by Optimization,",
AGARD CP,1990.
[7] A.Sbester, A.I.J.Forrester, Aircraft aerodynamic
design, John Wiley & Sons, Ltd, 2015
[8] XFLR5 v6.02 Guidelines
http://www.xflr5.com/xflr5.htm
[9] Gendreau, Michel, Potvin, JeanYves, Handbook of
metaheuristics, springer, 2003.
[10] Christopher A.Lyon, Andy P.Broeren, Philippe
Gigure, Ashok Gopalarathnam, and Michael S.Selig
Summary of LowSpeed Airfoil Data, SoarTech
Publications Virginia Beach, Virginia, 1995.
Picture 5. Endurance variations with attack angle
The maximum rate of climb and maximum speed are
obtained from the diagram shown in Picture 6.
76
SECTION II
Aircraft
CHAIRMAN
Professor Dragoljub Vuji, PhD
Mirko Kozi, PhD
DEVELOPMENTS IN HEADUP DISPLAY TECHNOLOGY FOR BASIC
AND ADVANCED MILITARY TRAINING AIRCRAFT
ROBERT WILSEY FRAES
Director Total Reaction Ltd, Representing Esterline CMC Electronics Inc. Knighton, UK, treaction@btinternet.com
Abstract: Since the 1960s HeadUp Display development has revolutionized military flight training. A HUD allows the
pilot to fly headout by reference to angle of attack and velocity vector as well as the primary flight display information.
Digital navigation information and weapon aiming solutions can also be also displayed. Todays trainer HUDs can
mimic the displays seen on frontline fighters permitting costeffective Leadin Fighter training.
Keywords: Fighter, Trainer, HUD, Avionics.
no magnification, projecting two collimated concentric
aiming rings. This allowed the pilot to move his head in
combat and still be able to see focused aiming rings. With
his eye position at five inches from the sight, the Field of
View (FOV) was 20 degrees. Collimation is the
projection of parallel light rays so that the focus is on
infinity. This allows the aiming pipper or ring to be
focused on infinity, matching the outside world and thus
reducing aiming errors due to parallax between the sight
and the target. Thus the pilot will see a focused aiming
solution when his eyes are focused on a distant aircraft. A
major advance in technology was the adoption by the
Germans of the Oigee reflector gun sight manufactured by
Optical Antal Oigee of Berlin, introduced during the last
months of the First World War and fitted to the Fokker
Dr.1 triplane and Albatros D.V biplane fighters. This
allowed the pilot greater freedom to move his head
without the visual obstruction of a tube, whilst the
illuminated pipper was projected onto a glass screen.
During the interwar years the reflector sight was further
refined. After the outbreak of the Second World War the
gyroscopic or gyro gun sight (GGS) was developed in
194143 at RAE Farnborough, which fed information on
the rate of turn and skid into the sight to adjust the
position of an illuminated graticule of six diamonds which
could be set to correspond with the wingspan of the target
by rotating the throttle grip, known as stadiametric
ranging. This took much of the guesswork out of
deflection shooting. The Germans in particular produced
some advanced examples of gyro sights. The RAF
experimented with a radar projector system for their De
Havilland Mosquito night fighters fitted with the AI Mk
IX radar in 1944 which projected the radar image of the
target together with the gunsight graticule and an artificial
horizon onto the pilots windshield using a cathode ray
tube and lenses. This was probably one of the first true
aircraft HUDs as it displayed flight attitude information
together with target and aiming solutions but was never
adopted. The Blackburn NA.39 strike fighter, later named
the Buccaneer, was the first British military aircraft to
enter service with a HUD in 1958, manufactured by
Cintel, developed for very lowlevel antishipping strikes.
1. INTRODUCTION
th
th
Over the last 12 years the introduction of 4 and 5
Generation frontline fighters has resulted in a revolution
in basic and advanced military flying training. The cost of
operating frontline fighters has increased dramatically
and has led to the necessity of downloading flying
training to less expensive platforms such as LeadInFighter Trainers (LIFT) and Advanced jet Trainers (AJT).
In turn this has led to a knockon effect of downloading as
much as possible of advanced flying training to less
expensive Basic or Intermediate trainers. A key element
in this process has been the development and use of the
allimportant HeadUp Display (HUD).
Picture 1. Example of typical F/A18 HUD symbology
during a dogfight.PD.
2. DEVELOPMENT OF THE HUD
Probably the first use of an optical device by pilots was
the collimated gun sights used on First World War
biplanes. The Aldis sight was adopted by RFC pilots in
19161918 on their S.E.5a, Sopwith Camel and Spad
S.XIII biplane fighters. The French developed a similar
Chretian collimated sight. The Aldis sight consisted of a
hermetically sealed tube, resembling a telescope but with
79
DEVELOPMENTSINHEADUPDISPLAYTECHNOLOGYFORBASICANDADVANCEDMILITARYTRAININGAIRCRAFT
OTEH2016
Target identification improved.
Safer instrument approaches with head up
approaching decision or minimum descent height.
Easier scan whilst formation flying and airtoair
refueling.
Safer low speed and hover VSTOL operations when
outside visual cues are critical.
Reduces the need for a Weapon System Officer
(WSO or backseater) in ground attack and air
defence aircraft.
Typically, the HUD is controlled by an Up Front Control
Panel (UFCP) attached to the front lower HUD body
which, with a combination of soft buttons and LED or
LCD window displays, allows the moding of the HUD to
be changed together with the insertion of waypoints and
communication frequencies. The UFCP also controls the
brilliance of the HUD and UFCP displays with test
functions and day night lighting switches.
Picture 2. Aldis gunsight on WW1 Royal Aircraft
Factory S.E.5a fighter. The Vintage Aviator Ltd, New
Zealand
A quicker and more intuitive means of changing the HUD
display known as Master Moding (first developed on
the F16) enables the pilot to switch from navigation
based display to airtoground to airtoair modes by a
simple press of a Hands on Throttle and Stick (HOTAS)
button on the throttle grip. The displays in the HUD and
on the MultiFunction Displays (MFDs) can be preprogrammed by the pilot for each mode as he desires.
The power supply controls and software for the HUD
reside either in the HUD or run from a dedicated card
inside an open architecture Mission Computer. This
computer, if dedicated to driving the HUD only, is
sometimes referred to as the HUD Symbol Generator
(HSG).
The HUD with associated UFCP and HUD camera are
mounted on the dashboard of the panel on a specially
manufactured tray attached to primary aircraft structure.
This is carefully designed to align with the pilots design
eye position (DEP). It is imperative that the HUD cannot
move in its mounting (except for adjustment), that the
combiners clear the canopy bow for HUD fitting and
removal, that the combiners are clear of the windshield
and canopy in the event of the canopy flexing due to bird
strike, and that the front face of the UFCP is clear of the
ejection seat line. The resultant is a very accurate
placement of symbology, often 5 to 2 milliradians,
depending on its position in the FOV.
Picture 3. Ferranti MkIIC Gyro Gunsight in a Spitfire
MkIX towards the end of the Second World War. PD
3. THE HUD IN FRONTLINE AIRCRAFT
The LTV A7 Corsair was the first operational US aircraft
to be fitted with a HUD, the HUD Weapon Aiming
System (HUDWAS), produced by MarconiElliott. The
HUD, which has equipped most frontline aircraft since
the 1970s, allows pilots to tactically fly their aircraft,
whether it be whilst conducting airtoground close air
support missions or during airtoair combat missions by
reference outside the aircraft with primary flight
information displayed in their field of view together with
computer generated weapon aiming solutions. However,
there are additional advantages that the HUD brings to the
pilot which include:
Increased tactical awareness with increased lookout.
Improved flight safety, especially at low level and in
poor visibility.
Simplified navigation with waypoints and targets
boxed in the pilots field of view and with steer
points and distance to go displayed.
Flight by reference to Angle of Attack.
Use of Velocity Vector (or energy) information.
Picture 4. Internal arrangement of a typical Refractive
HUD. Esterline CMC Electronics.
80
DEVELOPMENTSINHEADUPDISPLAYTECHNOLOGYFORBASICANDADVANCEDMILITARYTRAININGAIRCRAFT
OTEH2016
TFOV. The symbology is usually designed so that
primary flight information lies within this IFOV, but other
conformal symbols may move beyond the DEP IFOV.
This necessitates the pilot moving his head to see symbols
that have moved toward the edge of the TFOV. This
behavior is often described as like looking through a
hole in a fence. The hole is not large enough to see the
entire TFOV from the DEP. However, moving closer to
the HUD permits more to be seen, and moving laterally or
vertically permits a shift to see other parts of the display
as required. The projected imagery is traditionally green
phosphor P53 which has the attributes of lasting longer
than other greens and is also tuned to the peak wavelength
of human vision. The refractive HUD represents a lower
cost and a lighter solution compared to a holographic
HUD and is particularly suited to trainers, groundattack
aircraft and 4th Generation fighters.
4. ADVANTAGES OF THE HUD FOR
MILITARY TRAINING AIRCRAFT
The software that draws the HUD symbology on the HUD
combiner can today be written so that the HUD will
mimic the symbology and functionality of various frontline aircraft. This, for instance will allow BAE Systems
HUD on the RAAF Hawk Mk127 to display similar
symbology to the RAAF F/A18 Hornet with the same
button presses. The pilots HUD view is also recorded by
means of a digital mission recorder. A miniature HUD
camera, now usually placed forward of the HUD
combiners, records the pilots view of the outside world
whilst the HUD symbol generator or a mission computer
superimposes the HUD symbology onto the video
recording. Green phosphor symbology is difficult to film
and this computer generated superimposition results in a
clearer, more focused, recording and improved exposure.
This feature allows for timely and accurate post training
sortie debriefs from the instructor. The flying instructor in
the rear seat must also have access to what the student is
seeing through the HUD and this is achieved in one of
three ways:
A second rear seat HUD.
A rear seat conformal HUD Repeater.
A rear seat HUD repeater MFD menu page.
The format of the rear seat HUD display depends on the
design of the rear cockpit and to what degree it is angled
above the front ejection seat headbox. Both the
McDonnell Douglas TAV8B Harrier and Eurofighter
Typhoon T1/T3 have separate rear seat HUDs with a
FOV largely above the front ejection seat headbox. LIFT
and advanced trainers are more usually equipped with a
dedicated HUD repeater monitor which is mounted
conformally above the rear dash enabling the instructor to
land the aircraft using the repeater and peripheral vision if
required. The third option is to display the HUD view in
an MFD menu page. This is a low cost solution but
requires the instructor to refer to a headdown display.
Picture 5. Esterline CMC Electronics SparrowHawk 25
degree Refractive HUD Esterline CMC Electronics
7. REFLECTIVE HUD TECHNOLOGY
The reflective HUD (also known as a diffractive HUD or
pupil relaying system) has a single, large diameter, oblong
combiner which gives a larger TFOV than a refractive
HUD. From the Design Eye Position, the IFOV of the
reflective HUD usually covers almost the entire TFOV.
Head movement is possible whilst retaining view of all
symbology within a defined eye motionbox or design
eye box, but when the head is moved outside this
imaginary box during high energy maneuvering all
symbology may disappear. This is not as disadvantageous
as it sounds and a pilot soon learns where to place his
head in order to regain the symbology. The relay lenses
required to adjust the light to be displayed on the single
combiner are complex and heavy. The combiner itself
acts as a collimating lens and an added advantage is that
the combiner can be placed further away from the pilot
than with a refractive HUD. The large TFOV provided by
the single combiner provides a more practical solution for
displaying raster images such as Forward Looking InfraRed (FLIR), Enhanced Vision System (EVS) or Synthetic
Vision on the HUD. The reflective HUD is relatively
5. HUD FIELD OF VIEW
For all types of HUD, the term Instantaneous Field of
View (IFOV) describes the overall size and shape of the
symbol space that can be seen from any specific head
position. The IFOV typically changes as the pilots head
position changes. The term Total Field of View
(TFOV) describes the fixed overall limits within which
the HUD can display symbols. The TFOV does not
change with head position.
6. REFRACTIVE HUD TECHNOLOGY
The traditional refractive HUD uses a Cathode Ray Tube
(CRT) and lenses to project collimated symbology onto
either one or two flat combiners. The use of a second
combiner increases the vertical dimension of the IFOV.
However, due to space constraints in the cockpit, the size
of the optics is limited and the IFOV from the cockpit
design eye position (DEP) is typically smaller than the
81
DEVELOPMENTSINHEADUPDISPLAYTECHNOLOGYFORBASICANDADVANCEDMILITARYTRAININGAIRCRAFT
heavy and expensive and more suited to 4.5 and 5th
Generation frontline fighters where a large IFOV is
required.
OTEH2016
but essential sorties. Neither the F35 Joint Strike Fighter
(JSF) nor the F22 Raptor have a twoseater training
variant and the current estimate of the cost per hour for
the F35A is approximately US$42,000 (compared with
US$20,000 per hour for the F16C) [8]. It is estimated
that 210 flying hours per year will be required to keep an
F35 pilot fully combat ready [10]. There is therefore an
urgent requirement for Leadin Fighter Trainers (LIFT) to
mimic 5th Generation Fast Jets. Much of operational
conversion training and Squadron continuation training
can then be downloaded to a suitably equipped LIFT. The
LIFT must be able to mimic the HUD, HMD symbology
and the wide touchscreen display technology of the 5th
generation frontline aircraft with fidelity.
8. THE HUD AS PART OF A DIGITAL
COCKPIT SOLUTION
The HUD and UFCP are closely integrated with the
HOTAS, weapon delivery, MFD selection and displays,
GPS navigation and sensors. The HUD is the key to the
integrated fighter and trainer cockpit which has two
additional levels of redundancy. Primary Flight
information is displayed in the HUD, on the PFD page of
the MFDs and typically on a standalone electronic
standby instrument system which runs off its own
independent power supply in the event of a total systems
failure. Flybywire and carefree handling has made the
pure flying skills required to fly 5th Generation aircraft
less critical whilst systems management, information
processing and decision making under high Gloads have
become increasingly essential. Most of these skills can be
learned on a less capable and lower cost platform using
training software to mimic the symbology and systems
management of the frontline platform.
Virtual training, mission debriefing tools and simulations
can accurately replicate expensive advanced sensors and
mission systems. For example the requirement for an
expensive and highly capable radar can be replaced for
training purposes by computer generated virtual radar
together with virtual defensive systems, including RWR,
chaff and flares. Related virtual HUD radar symbology
items such as a Target Designator box, locator line and
text boxes can be displayed using Virtual Training System
(VTS) software. The 5th Generation fighter such as the F22 is designed to engage multiple adversaries with beyond
visual range (BVR) AIM120 AMRAAM missiles. Thus
it is wasteful in terms of both training time and money to
practice unrepresentative 1v2 airtoair combat. It is also
prohibitively expensive to get 14 highperformance
aircraft airborne to act as targets. VTS will enable the
LIFT student to engage, for instance, 12 virtual
adversaries during a training mission at the cost of one
hours flight time in a LIFT. The combination of these
features means that valuable flying training hours can be
downloaded from a US$42,000 per hour frontline fighter
to a US$3,000  9,000 per hour LIFT/advanced jet trainer.
In response to these requirements Esterline CMC
Electronics has recently integrated a Digital HUD with a
20 x 8 inch Large Area Display (LAD) touch screen,
replicating the 5th Generation cockpit. The large area
display can show multiple windows and includes
synthetic vision. This, and other future LIFT glass
cockpits, will not only help solve the 5th Generation
training problem, but will allow frontline squadron pilots
to remain in current operational readiness without using
very expensive frontline aircraft.
Picture 6. BAE Systems 35 x 25 degree reflective HUD
on Eurofighter Typhoon. Eurofighter GmbH
9. THE FUTURE OF THE HUD
The most important recent change to the conventional
HUD has been the replacement of the CRT with a digital
light engine (DLE). CMCs digital SparrowHawk HUD
was unveiled at the Farnborough Airshow 2012 as part of
CMCs next generation cockpit (Cockpit4000 NexGen).
It utilizes a Digital HUD, replacing a conventional CRT
with a digital light engine, eliminating the requirement for
a high voltage power supply to the CRT. The introduction
of the DLE will mean that the lifelimited CRT
component will no longer have to be replaced, typically
every 2,000 hours, saving expensive downtime. In
addition to improved MTBF the symbology will also burn
brighter and not suffer from fade as did the CRT based
symbology. This in turn will improve the raster display
performance of the HUD.
10. THE HUD IN FLIGHT SIMULATORS
With todays emphasis on realistic simulation, a modern
full motion flight simulator can replicate the real aircraft
with fidelity. This includes the HUD when fitted to the
real aircraft, but there is a technical adjustment that has to
be made to the HUD optics since the outside world is
usually a computergenerated scene projected onto a
curved wraparound screen. For simulator applications the
HUD symbology focus therefore has to be specially
tuned. Such focus tuning is effective with domes of 10
meters radius or greater. With the increasing importance
on providing compact visual systems in modern
simulators, it becomes difficult to refocus the HUD
without introducing undesirable display artificialities that
With the introduction of the 5th Generation aircraft, frontline jets will prove to be too expensive to fly on anything
82
DEVELOPMENTSINHEADUPDISPLAYTECHNOLOGYFORBASICANDADVANCEDMILITARYTRAININGAIRCRAFT
increase as the distance to the dome decreases. One solution
is to project the HUD symbology with the visual scene and
forgo use of real HUD symbol generation. This works but
there are lingering concerns as to whether or not training on a
HUD without the relationship between IFOV and TFOV
being present might introduce negative training elements that
must be overcome in the real aircraft.
OTEH2016
HMD which is useful for the accurate delivery of
boresight weapons and for instrument approaches. The
HUD/HMD combination also offers an effective backup
in the event of HMD failure, especially in the case of
night or poor weather recovery to a carrier. The cost and
complexity of HMDs will be beyond the budget of many
smaller nations. There is thus a continuing requirement
for HUD and it is unlikely that the HMD will replace a
HUD in training aircraft in the near future.
11. IMPACT OF HELMET MOUNTED
DISPLAYS
12. CONCLUSION
The introduction of Helmet Mounted Displays (HMD)
was first used operationally by the South African Air
Force on the Dassault Mirage F1AZ and later the Atlas
Cheetah. Russia and Israel were quick to develop the
concept further and Elbits DASH was the first HMD to
see service in the West. Development of HMDs for 4th
and 5th Generation fighters is progressing although there
have been considerable technical challenges with issues
including:
Weight and ejection safety
Lag in displaying information
Jitter
Alignment of the pilots eye and HMD
Incorporating night vision into the HMD
In summary HeadUp Displays are now a core part of
military training aircraft avionics. Training the 5th
Generation fighter pilot on frontline aircraft will be
unaffordable and thus a new generation of Leadin Fighter
Trainers will be required. They must be able to mimic the
handling, feel and mission systems of the frontline. As
much of the advanced training syllabus as possible should
flow down to basic training in order to make the best
economic use of valuable and dwindling assets. Integrated
glass cockpit technology, together with the use of flight
simulators, replaces expensive flying hours and increases
the effectiveness of training the modern combat pilot.
Despite the development of the HMD the HUD is likely
to continue to play a vital part in training for many years.
Advanced fighters such as the F/A18E Super Hornet,
Eurofighter Typhoon, Dassault Rafale and the Lockheed
Martin F22 Raptor rely on the HUD for primary flight
information whilst the HMD gives additional tactical
situational awareness and offaxis weapon cueing. The
Lockheed Martin F35 Lightning II JSF will be the first
aircraft designed to rely solely on an HMD, the ESA
Visions Systems Helmet Mounted Display System
(HMDS), without a HUD. All other current HMD
equipped frontline fighters such as the Typhoon, with the
BAE Systems Head Equipment Assembly (HEA) Striker
HMD, continue to use a HUD. Currently a HUD will
generally produce more accurate symbology than an
Picture 8. Example of a modern LIFT Trainer Cockpit.
CMC Electronics new Cockpit4000 NexGen, showing
28 degree digital HUD (top) and an early demonstration
20 x 7 inch Large Area Display touchscreen. Esterline
CMC Electronics.
Picture 7. Elbit Systems of America/VSI International
HMDS for F35 JSF. PD
83
DEVELOPMENTSINHEADUPDISPLAYTECHNOLOGYFORBASICANDADVANCEDMILITARYTRAININGAIRCRAFT
OTEH2016
[7] Clarke, R Wallace: British Aircraft Armament:
Volume 2 RAF Guns and Gunsights from 1914 to the
present day Patrick Stephens 1994.
[8] Drew, James: F35A cost and readiness data
improves in 2015 as fleet grows, Flightglobal 02
February 2016.
[9] Gunner, Jerry: Netherlands & the F35, Air Forces
Monthly July 2014.
[10] HUDWAC Flight International, 14 November 1974.
[11] Newman, Richard L.: HeadUp Displays, Avebury
Aviation 1995.
References
[1] http://simhq.com/forum/ubbthreads.php/topics/32726
19/Re:_Aldis_gunsight
[2] Kopp, C.: The Modern Fighter Cockpit, Australian
Aviation & Defence Review March 1981.
[3] Wilsey, R.: Developments in Glass, Cockpit
Technology Front Line Defence Vol 10 no 1.
[4] Penney, S.: Honing the Hawk, Flight International
Magazine 25 Feb 2003.
[5] Wood, RB & Howells P.J.: HeadUp Displays, The
Avionics Handbook CRC Press 2001.
[6] Croft, John: Helmet Mounted Displays: Adding Night
Vision, Aviation Today 1 Sep 2006.
84
FLIGHT PERFORMANCE DETERMINATON OF THE PISTON ENGINE
AIRCRAFT SOVA, COMPUTER PROGRAM "SOVAPERF"
NEMANJA VELIMIROVI
YUGOIMPORTSDPR, Belgrade, Serbia, velimirovicnemanja@yahoo.com
KOSTA VELIMIROVI
Military Technical Institute, Belgrade, Serbia, kolevelimirovic@yahoo.com
Abstract: The method for the estimation of the piston, armed or non, aircraft performances is presented in this paper.
"SOVAPERF" new computer program is used for the calculation of basic and special performances of the piston engine
aircraft. Nonlinear model total energy which served as the computer program is basis for performance calculation. The
computer program "SOVAPERF" consists of several modules: Module for determining the mass aircraft, Airport Module based on current meteorological conditions determines the data of the airport. This module requests data: frontal wind speed,
slop and quality of the runway, Atmosphere Module  data calculations are made for the altitude. It works for all conditions
from polar to tropical. Aerodynamics Module  loaded aerodynamic characteristics of clean, takeoff and landing
configurations. This module also presents data on the aerodynamic characteristics of the launcher under the wing and
weapons, Engine Module  provides information on the performance and fuel consumption in function of the regime and the
height of flight. Performance Module  computes the minimum and maximum speed, climb and ceiling. Turn Module turnaround data calculation. Takeoff and Landing Modules  presents all data related to the takeoff and landing. The data
refers to the characteristic length, velocity, time and fuel consumption. Cruise Module  optimal parameters of cruising. Start,
Taxi and Combat Modules  compute fuel consumption and other data. Range Module  the most complex module of the
computer program "SOVAPERF". This module calls all the listed modules and determines the maximum length of the flight
profile. The method and its results are illustrated by the numerical example. SOVA armed, aircraft performances are
presented in numerical example. Used programming language is MATHCAD 14.
Keywords:. Aircraft, piston engine, performances, computer program.
1. INTRODUCTION
The production of the domestic aircraft SOVA (SDPR,
UTVAPanevo) has started. SOVA (Picture 1) (Table 1) is
allmetal, "side by side" arrangement of four sets, lowwing,
fixed landing gear, singleengine, intended for initial training
and selection, sport flying  tourism and light groundattack
aircraft. Power plant is Lycoming IO390A1 B6 210HP
maximum constant power (Picture 2). Aircraft SOVA can be
armed with guns, bombs and rockets. SOVA has excellent
flying characteristics. SOVA is perfect trainer, due to
exceptional features at low speed flying, and a strike plane of
good maneuverability, needed in combat conditionals [3, 4].
SOVA can be used for reconnaissance, patrolling and antiterrorism operations. Aircraft SOVA new electronic
equipment ("glass cockpit") (Picture 3) is equipped. A "glass
cockpit" is an aircraft cockpit that features electronic (digital)
instrument displays, typically LCD screens, rather than the
traditional style of analog dials and gauges. The "glass
cockpit" has become standard equipment in modern aircraft.
During the process of airplane modernization, design and
exploitation the great attention is focused on the
performance [11]. For aircraft performance calculations it
is necessary to have a "tool". "SOVAPERF" is new
"tool"computer program for the calculation of piston
engine aircraft flight characteristics. This paper presents
aircraft SOVA performance calculation.
Picture 1. Aircraft SOVA (SDPR, UTVAPanevo,
Serbia), exhibition Partner 2015
85
FLIGHTPERFORMANCEDETERMINATONOFTHEPISTONENGINEAIRCRAFTSOVA,
Table 1. Characteristics of the aircraft SOVA
Basic trainer: max. takeoff
1250kg
weight
(4 pilots, max. internal fuel)
Weight of empty aircraft,
750kg
equipped
Max. internal fuel
124 kg
Wing surface area
14.63m2
Propeler Hartzell,two1.93m
blade,allmetal, constant
speed,diameter
Power plant, max power, Air
Lycoming IO390A1
conditioning provisions,
B6, 210HP
electronic ignition
OTEH2016
4. Aerodynamics Module  loaded aerodynamic
characteristics of clean, takeoff and landing
configurations [5]. This module also presents data on
the aerodynamic characteristics of the launcher,
missiles, bombs and gun.
5. Engine Module  piston engine: provides information
on the performance and fuel consumption in function
of the regime and the height of flight [7].
6. Performance Module  computes the minimum and
maximum speed, climb and ceiling.
7. Turn Module  turnaround data calculation.
8. Takeoff and Landing Modules  presents all data
related to the takeoff or landing. The data refers to the
characteristic length, velocity, time and fuel
consumption.
9. Cruise Module  optimal parameters of cruising.
10. Start, Taxi and Combat Modules  computes fuel
consumption and other data.
11. Planning module.
Range Module  the most complex module of the
computer program"SOVAPERF". This module calls all
the listed modules and determines the length and time of
the flight profile [3], [4].
Picture 2. Lycoming IO390A1 B6
Picture 3: Aircraft SOVA  glass cockpit, exhibition
Partner 2015
2. COMPUTER PROGRAM "SOVAPERF"
"SOVAPERF" computer program (Picture 4) is used for
the calculation of basic and special performances aircraft
equipped with piston power plant. It is new computer
program. Nonlinear model total energy which served as
the computer program was basis for performance
calculations [1], [2], [6], [8]. Plane is considered as a
material point[9]. Used programming language is
MATHCAD 14.
The program consists of several modules [12], [13]:
1. odule for determining the mass aircraft,
2. Airport Module  based on current meteorological
conditions determine the data of the airport. This
module requests data: frontal wind speed, slop and
quality of the runway.
3. Atmosphere Module  data calculations are made for
the altitude. It works for all conditions from polar to
tropical.
Picture 4. "SOVAPERF" computer program  flowchart
3. NUMERICAL EXAMPLE
Some aerodynamic characteristics, flight capability and
other performances of the aircraft SOVA (Picture 5)
(Weight 1279 kg, configuration 1pilot+2xBomb FAB100
86
OTEH2016
FLIGHTPERFORMANCEDETERMINATONOFTHEPISTONENGINEAIRCRAFTSOVA,
M80, Picture 6, Table 2) are presented in numerical
example. Diagram, Picture 7, shows the coeficients
resistance and aerodynamic lift of configuration 2xBomb
FAB100 M80. Picture 8 presents sea level and altitude
performance. Picture 9 is diagram ofminimum fuel flow
versus power. Picture 10 presents propeller efficiency
versus advance ratio for different power coefficient..
Speed altitude envelope (maximum speed, minimum
speed, minimum speed with flaps) is presented in Picture
11. Thrust available and thrust required, for different
heights and velocity at maximum engine rating is
presented (Picture 12). Specific excess power for diferant
heights and velocity at maximum engine rating are
presented in Picture 13. Distances of takeoff are
presented in Picture 14. Combat flight profile (aircraft
configuration: 1pilot+2xBomb FAB100 M80, reserve
fuel 10%) is shown in Picture 15 [10]. Geometric
characteristics of the flight profile are shown in the Table
3. Weather conditions and "altitude" airport (Standard
Atmosphere, STA) are presented in Table 4. Some
calculated data of optimal flight profile are presented in
Table 5 [9, 14].
Picture 8. Lycoming IO390A1 B6, Sea level and altitude
performance
120
100
fuelflowt
80
60
40
20
50
100
150
200
250
Brakhpt
Picture 5. Aircraft SOVA+2xBomb FAB100 M80
Picture 9. Lycoming IO390A1 B6, min. fuel flow(lb/hr)
versus power (hp)  Power (80210hp)
( Jtt , Cpt 1)
( Jtt , Cpt 2)
Picture 6. Bomb FAB100 M80
( Jtt , Cpt 3)
Table 2. Bomb FAB100 M80 specifications
Bomb FAB100 M80
Diameter
230 mm
Length
1490 mm
Mass bomb
117 kg
Mass warhead
39 kg (TNT)
( Jtt , Cpt 4)
( Jtt , Cpt 5)0.6
( Jtt , Cpt 6)
( Jtt , Cpt 7)
1.5
Cz
0.8
( Jtt , Cpt 8)0.4
1
0.5
0
0.2
0.1
0.2
0.5
1.5
Jtt
Cxavn( Cz , Ikonf , Itert , borbter )
Picture 10. Variable pitch installed propeller (Hartzell
D=1.93 m), Efficiency () versus advance ratio Jtt() for
different power coefficient Cp
Picture 7. Aircraft SOVA Coeficients resistance Cxavn
and aerodynamic lift Cz, configuration 2x Bomb FAB100 M80, velocity V=50 m/s, height h=1000 m
87
OTEH2016
FLIGHTPERFORMANCEDETERMINATONOFTHEPISTONENGINEAIRCRAFTSOVA,
410
hwmaxbih 3103
hwmaxbih
hwmaxbih
210
110
0
100
150
200
250
SZ=583.m
Vmaxstacih , Vminstacih , Vminsletanjestacih
Picture 14. Takeoff, configuration 1pilot+2xBomb FAB100 M80
Picture 11. Aircraft SOVA, True speed altitude envelope,
configuration 1pilot+2xBomb FAB100 M80, height
h(0.04000m), velocity V (0.0250.0km/h), maximum
engine rating
310
2.510
210
Spl=745.m
Rxav1( Vleta)
Rxav2( Vleta)
Rxav3( Vleta)
Rxav4( Vleta)
Telise1( Vleta)
Picture 15. Range, length of the flight profile
Table 3. Data flight profile
Height cruising 1 (5)
Height holding (6)
Height penetration 1 (8)
Combat (9)
Height penetration 2 (10)
Height cruising 2 (12)
Telise2( Vleta)
Telise3( Vleta)
Telise4( Vleta) 1.510
110
100
150
200
Table 4. Meteo data airport
250
Temperature, airport (C)
Air pressure, airport (Pa)
Wind head, airport (m/s)
"Altitude" airport STA (m)
Vt( Vleta)
Picture 12. Aircraft SOVA, Thrust available Telise (N)
and thrust required Rxav (N), for different heights
(0.,500.,1000.,1500. m) and velocity Vt (0.250.0km/h),
maximum engine rating
w1 ( Vletasvs1)3
w3 ( Vletasvs3)2
w5 ( Vletasvs5)
1
40
50
15.0
1.013*105
0.0
5.096
Table 5. Calculated characteristics of the optimal flight
profile
Distanc.
Height (m) Time (min)
(km)
1
Start
5.
5.
2
Taxi
5.
5.
3
Takeoff
5.
0.464
0.745
4
Climb 1
5.1500.
10.537
25.529
5
Cruising 1
1500.
75.059
195.45
6
Holding
1500.
15.0
7
Planning 1
1500.300. 4.155
10.4
8
Penetr. 1
300.
14.417
50.0
9
Combat
300.
3.0
10
Penetr. 2
450.
11.736
40.0
11
Climb 2
450.1300.
2.99
6.898
12
Cruising 2
1300.
87.495
222.7
13
Planning 2
1300.5.
5.25
11.538
14
Landing
5.
0.426
0.63
1500m
1500 m
300 m
300 m
450 m
1300 m
60
Vletasvs1 , Vletasvs3 , Vletasvs5
Picture 13. Aircraft SOVA, Specific excess power
w(m/s), configuration 1pilot+2x Bomb FAB100 M80 for
heights (0.,1000.,2000. m) and velocity Vletasvs (3560m/s), maximum engine rating
88
FLIGHTPERFORMANCEDETERMINATONOFTHEPISTONENGINEAIRCRAFTSOVA,
1
2
3
Start
Taxi
Takeoff
Climb 1
5 Cruising 1
6 Holding
7 Planning 1
8 Penetr. 1
9 Combat
10 Penetr. 2
11
Climb 2
12 Cruising 2
13 Planning 2
14 Landing
OTEH2016
maksimalnog taktikog radijusa naoruanog klipnoelisnog aviona, SYMOPIS2010, Tara, Serbia
2010.
[5] Velimirovi,K.,
Velimirovi,N.:
Odreivanje
maksimalnog taktikog radijusa klipnoelisne
bespilotne letelice, program BELI ORAO, SYMOPIS2011, Tara, Serbia 2011.
[6] Bajovi,M.,
Velimirovi,K.,
Molovi,V.,
Velimirovi,N.: Analiza aerodinamkih koeficijenata
na osnovu aerotunelskih i letnih ispitivanja, OTEH
2009, Military Technical Institute, Belgrade , Serbia,
2009.
[7] Bajovi,M., Velimirovi,K., Molovi,V.: Estimacija
performansi klipnoelisnog aviona, OTEH 2007
Military Technical Institute, Belgrade, Serbia, 2007.
[8] Velimirovi,K., Velimirovi,N.: Tactical UAV
PEGASUS with in flight adjustable propeller,
COMPUTER PROGRAM WHITE EAGLE 2, OTEH
2010 Military Technical Institute, Belgrade, Serbia,
2010.
[9] Velimirovi,K., Velimirovi,N.: Flight performance
determination of the turboprop aircraft, Fourth
Serbian Congress on Theoretical and Applied
Mechanics, Vrnjaka Banja, 2013.
[10] Velimirovi,K.,
Velimirovi,N.:
Odreivanje
maksimalnog taktikog radijusa naoruanog turboelisnog aviona, program KOBACPERF SYMOPIS2014, Divibare, Serbia 2014.
[11] Velimirovi,K., Velimirovi,N.: Tactical UAV
PEGASUS as a platform to carry missiles, OTEH
2011 Military Technical Institute, Belgrade, Serbia,
2011.
[12] Ili,D., Devi,V., Velimirovi,K., Antoni,V.,
MilenkoviBabi,M., Sari,Z.: Analysis of Lasta
aircraft improvement by integration of a turboprop
power plant, OTEH 2014 Military Technical
Institute, Belgrade, Serbia, 2014.
[13] Velimirovi,K.: Avion sa turboelisnom pogonskom
grupom: proraun performansi leta, monografija,
Military Technical Institute, Belgrade, Serbia, 2015.
[14] Velimirovi,K.: Taktika bespilotna letelica sa
klipnoelisnom pogonskom grupom: proraun
performansi leta, monografija, Military Technical
Institute, Belgrade, Serbia, 2013.
[15] Velimirovi,K., Velimirovi,N.: Flight performance
determination of the turbjet aircraft, Serbian
Congress on Theoretical and Applied Mechanics,
Aranelovac, Serbia, 2015.
Absolute
Instrm. Aircraft
Rotat.
manifold
Speed
per min.
mass
pressure
(km/h)
(min1)
(kg)
(in Hg)
1279
21
2300
1276
20
2000
144.2
1275
28.4
2670
141.21275 28.323.5 2700
139.4
158.4
1266
21.5
2450
147.6
1232
20.4
2350
145.5
1225
15
1100
205.0
1225
27.3
2600
1213
26
2650
200.0
977
23.6
2650
135.7969
26.824.1 2700
131.7
162.0
967
18.85
2450
129.1
933
15
1100
129.2
932
16.015.0 1200
Reserve fuel (10%)= 12.4 kg,
Range = 281.2 km
Time=240.5 min
4. CONCLUSION
Program "SOVAPERF" based on total energy method is
used for piston engine aircraft flight characteristics
calculation. Program is very useful for: aircraft design,
writing pilot instructions and planning missions. Program
is fast enough and reliable. Accuracy of results depends
on the loaded engine and aerodynamic characteristics of
the aircraft, launcher, missiles, bombs and gun. Method
and its results are illustrated by numerical example for
piston engine SOVA aircraft in software MATHCAD 14.
Results of calculation performances SOVA aircraft are
presented in the paper. Calculated flight performances and
capabilities shows a quality of piston engine SOVA
aircraft.
References
[1] Renduli,Z.: Mehanika leta, Vojnoizdavaki i
Novinski Centar, Belgrade , Serbia. 1987.
[2] Smetana,F.: Flight vehicle performance and
aerodynamic, AIAA WrightPatterson Air Force
Base, Ohio. 2003.
[3] Upravljanje avionom V53, VTUP, Beograd 1991
[4] Velimirovi,K.,
Velimirovi,N.:
Odreivanje
89
POSSIBLE APPROACHES TO EVALUATION OF TRAINING AIRCRAFTS
USED IN FLIGHT SCREENING
SLAVIA VLAI
Military Academy, Belgrade, slavisavlacic@yahoo.com
FRANC HUDOMAL
International Test Pilot School, Canada, franchudomal@yahoo.com
ALEKSANDAR KNEEVI
Military Academy, Belgrade, aleksandarknezevic75@gmail.com
Abstract: Flight screening or a selection of future pilots is a very sensitive and complex phase in the process of flight
training and pilots education. Most of the known flight screening processes are based on the same requirements but the
aircrafts which are used for selection are different. This paperwork describe some experiences regarding to the aircraft
types used in the flight screening and explains the possible approaches to evaluation of an aircraft which can be
implemented in this flight training phase.
Keywords: Flight screening, training aircraft, evaluation.
become military pilots [1]. The second stage is basic
flight training. It practices fundamental skills of handling
the plane and guiding the plane through the airspace. It
consists basing handling, aerobatic flying, navigation
flying, and sometimes formation flying. The total flight
time in this phase of training is between 80130 flight
hours, depending on type of plane and syllabus.
Generally, piston and turboprop powered plane are used,
and the jet plane in this phase becomes rarity. Jet training
planes is typical for the advanced flight training, as a third
stage. In every case, the main tool in the process of
making a pilot, throughout the flight training, is an
adequate training plane or simply trainer. The flight
screening or abinitio training is the very first stage of
flight training. It is a flyingbased assessment of potential
candidates. This is intended to establish whether the
student has the necessary aptitude to become a military
pilot in a reasonably short time. It will eliminate, for
example, those who get airsick or lack coordination or
judgment. Experienced instructors make qualitative
assessments of applicants. This goal is almost the same in
all air forces throughout the world.
1. INTRODUCTION
Flight screening or a selection of future pilots is a very
sensitive and complex phase in the process of flight
training and pilots education. Most of the known flight
screening processes are based on the same requirements
but the aircrafts which are used for selection phase are
different. The difference emerges as a consequence of
budget constraints, attitudes regarding to flight training,
specific demands on pilot training, syllabus, political and
industry reasons and some other, less important
circumstances. Diversity of criteria has a great impact on
decision on which type of aircraft can or even must be
used in flight training system. But also, there are some
aircraft characteristics which cannot be ignored and some
categories of training aircraft which can be avoided in
flight screening. This paperwork describes some
experiences regarding to the aircraft types used in the
flight screening and explains the possible approaches to
evaluation of an aircraft which can be implemented in this
flight training phase. This comes in the moment when
domestic air force has to change its existing fleet of light
piston engined trainers intended for pilot selection.
Table 1: Piston engine trainer characteristics
2. FLIGHT SCREENING
Aircraft
Flight training is a course of study used when learning to
pilot an aircraft. Given the expense of military pilot
training, air forces typically conduct training in phases to
washout unsuitable candidates. The cost to those air
forces that do not follow a gradated training is not just
monetary but also in lives. Flight training is generally
performed in three stages: primary, basic and advanced.
Grob 115
Da 20
Zlin 242
G120
Lasta
139
93
149
190
224
Performance
Powerplant, kW
Length, m
7.54
7.16
6.94
8.11
7.97
Wingspan, m
10.00
10.87
9.34
10.18
9.70
12.20
11.61
13.86
13.3
12.9
Wing area, m
The first stage, primary or abinitio training, is used to
filter those students who lack the aptitude to quickly
90
POSSIBLEAPPROACHESTOEVALUATIONOFTRAININGAIRCRAFTSUSEDINFLIGHTSCREENING
Empty weight,
kg
685
529
745
1100
850
Useful load, kg
320
271
250
360
200
185
256
236
319
310
Max.speed,
km/h
Rate of climb,
m/sec
Service ceiling,
m
Take off
distance, m
Landing
distance, m
5.3
5.08
5.5
6.5
8.5
3050
4000
5500
5486
6000
461
550
565
654
500
457
450
495
562
600
Range, km
1150
1013
1056
1176
Price, USD
400.000
250.000
250.000
1.300.000
800.000
OTEH2016
142/242, T67 Slingsby Firefly, Valmet L70, Saab
Supporter MFI 17, Grob G120A and Utva75. For many
decades after the World War II there was considerable
agreement that the primary training phase demanded an
aircraft of around 150 kW, with fixed gear and side by
side seating. The other two subcategories are more
familiar with military training and can cover up to the 100
flight hours of syllabus. First considerable trainers in
these subcategories have emerged during the 1970s and
80s. It was Italian SF.260 which has been followed by
TB.30 Epsilon, T35 Pillan and Lasta 1. They are also
characterized by more powerful engines of around 250
kW. For training aircraft, side by side seating has the
advantage that enables pilot and instructor seeing each
others actions, allowing the pilot to learn from the
instructor and the instructor to correct the student pilot.
For this task it is necessary to implement a training
airplane with appropriate characteristics. Trainer can not
be expensive for operation, complexity has to be on low
level, and flight characteristics must be benign for the
beginners. It must adequately support and validate the
pilot selection system by confirming which students are
likely to succeed in the later, more costly phases of pilot
training.
There are generally smaller numbers of primary trainers in
service than basic or advanced trainers. Nevertheless there
are significant fleets of certain types in operation; Grob 115
(205), Grob G120 (90 + 48 on order), T35 (91 + 9 on
order). In addition, the latest generation of Diamond DA20/40/42 and Cirrus SR.20/22 aircraft are being procured in
increasing numbers; DA20 (57 + 4 on order), DA40 (37),
DA42 (13), SR.20 (37), SR.22 (39) [3].
Historically, change has come extremely slowly to this
category. Many of todays airframes were designed 50
years ago and the biggest change that has occurred since
then was moving the tail wheel from the back of the lane
to the front. The only other major visible changes have
been to the navigation receivers in the cockpit. Also, in
the year period beginning in 2003, the general aviation
industry converted to from shipping no glass cockpits at
all to equipping approximately 90% of all new small
airplanes with glass cockpits [4]. However, assessment of
potential candidates for military pilot, does not necessary
need glass cockpit, especially in time limited abinitio
training. Digital cockpit inevitably demands more time for
initial technical preparations. That is the reason why some
of users insist on analogue cockpit [5].
3. TRAINING AIRCRAFTS
The main tool in the process of making a pilot, throughout
the flight training, is an adequate training plane or simply
trainer. There are many different types of training aircraft,
especially in the primary and basic flight training phase.
Also, there are many different divisions of the training
planes, but the most important and common used is the
division according to the type of powerplant, which
determines its main characteristics and performances [2].
The training aircraft mainly applied in ab initio training is
driven by piston engine. They are also called primary
trainers. Some of the piston engined trainer specifications
are shown in Table 1. This paperwork doesnt consider
ultra light aircrafts (ULA), because of their characteristics
and their irrelevance for serious system of military pilot
training.1
4. EVALUATION PREREQUISITES
Different air forces have a variety of specific prerequisites
and needs regarding to the trainer which is going to be
used for flight screening.
Airplane piston engines of today are, generally speaking,
simple, aircooled, horizontally opposed, fourstroke
internalcombustion devices with low operating speeds
and low specific output. Trainers driven by the piston
engine belongs to the sort of the most economical trainers.
Prerequisites can be divided in several groups as military,
political, economical and educational. There are some
postulates about the trainers which can be used in ab initio
training. According to the specific situation the trainer
choice can be hardly influenced by many reasons which
are not connected by trainer capabilities. Military users
usually have their so called tactical and technical
requirements, but industry holdstakers and political
influences can be decisive factor on a trainer selection,
also as budget constraints. For this reason, inappropriate
trainer can be implemented into the flight screening
process. In such cases, there can be pressure to use cheap
ULA plane, or a new trainer acquired for the next training
phase. Anyway, military requirements have to be priority.
Piston engined trainers are often divided based on the
aerodynamical and cockpit configuration by following:
fixed landing gear, side by side seats,
retracting landing gear, side by side seats,
retracting landing gear, tandem seats.
The first subcategory is intended for the very basics, and
it is represented by light aircraft not too dissimilar from
civilian training aircraft. Representatives are ZLIN1
Analyzing materials from the Military Flight Training
Conference held in March 2016th in London it can be seen that
relevant air forces with developed training system dont use, not
even consider implementation of ULA.
Except the tactical and technical requirements many other
factors have to be considered such as available money,
91
OTEH2016
POSSIBLEAPPROACHESTOEVALUATIONOFTRAININGAIRCRAFTSUSEDINFLIGHTSCREENING
number of candidates, education time,2 flight training
organization, meteorological conditions and other
resources intended for screening process. It is very
important to have in sight also the type and equipment of
the next trainer in the process. It is unacceptable to use a
trainer with digital cockpit in the flight screening and
aftermath transferring to analog cockpit.
The question therefore arises, as how an air force should
select a primary trainer that will give reasonably low
operating cost and acceptable steps in difficulty between
the different stages of training.
The Serbian Air Force expects such a process in the near
future. The airplane, which is used in the flight screening
of Serbian Air Force future pilots, is UTVA75,
domestically produced pistonengine trainer. This plane
belongs to the older generation of pistonengine trainers
and new trainer has to replace it before the end of 2017th
[6]. One of the main stream options is to use Lasta as a
primary trainer. Avoiding the option that Lasta trainer is
going to be used for primary training, because of its
complexity and consequently prolonged time of selection,
there are some other limitations which have to be
considered.
Picture 1: Utva75
According to existing experience and methodology of
aircraft evaluation there are some requirements and
prerequisites that decision makers have to have on their
minds.
Flight screening with a syllabus of only 12 flight hours
including VFR basics consider moderate performances
and complexity with fixed landing gear, spacious side by
side cockpit with military type dual controls3 and good
field of view in all maneuvers. Take off and landing
distance, as a climb performance is not of great influence.
But, the cruising speed of more than 130 kts is not
desirable. Control harmony throughout the envelope is
very important with good harmony of roll and pitch
forces. The priority list also includes good static and
dynamic stability and lateral/directional control without
PIO tendency. The plane must be a good platform for
teaching students for a critical task of trimming for
airspeed. Control forces have to allow an easy transition
to larger aircraft.
For over 30 years, UTVA75 covering the ab initio phase
training which consisted of 10 flight hours. For the last 15
years, UTVA75 also covered some parts of basic flight
training. This was imposed by loosing all Galebs G2. The
cockpit of UTVA75 was spacious, handling characteristics
were satisfying, stall behavior was benign but the main
restriction was impossibility of performing basic aerobatic
maneuvers including spin. Assessment of young pilots was
difficult especially to cadets who have to be streamed to the
fast jet trainer Supergaleb G4. Any kind of night or IFR
flying was impossible, so the first introduction night and IFR
sorties are on the advanced jet trainer.
Stall and spin performances represents the category for
itself. High angle of attack flight regime requires clear
stall warning. Easy recovery of stalling is also must be
priority. Conventional and predictable entry to the spin is
highly desirable as a mild spin rotation speed and
oscillatory in the first two spin turns. Mild spin rotation
speed per turn for such plane is 46 sec.
According to the existing Military aviation study
program, the maximum period of time intended for the
flight screening is two months including ground school
and technical introduction lectures about plane, its
construction, systems, manuals etc. It means that training
scope is maximum 12 flight hours (VFR only) and during
that time the plane with high performances and workload
can not be implemented. The plane which has retracting
landing gear, tandem cockpit with full digital equipment
and systems is not desirable and doesnt provide
possibilities for objective assessment. Such configuration
is usually linked to a heavier trainer. The beginners
workload is also high and candidate capabilities are
overstressed.
Primary trainer has to provide good glide performance
which allows forced landing procedures training.
Syllabus must be adjusted to provide the student time to
learn to fly safe approaches, so the plane has to provide it
in an acceptable time scope and manner. Up to 80kts final
approach speed is desirable. no power approach with no
flap is highly desirable.
In this particular case, the distance and altitude of training
airspace is not of big influence.
Analyzing UTVA75s (Picture 1) possible successor the
step forward has to be made with aerobatic and spin
performing availability. The successor has to be
considered in a light of a Lasta plane as a next step in the
flight training syllabus.
5. EVALUATION OF TRAINING AIRCRAFTS
USED IN FLIGHT SCREENING
Evaluation is a systematic determination of a subjects
merit, worth, significance, using criteria governed by a set
of standards.
Flight screening is usually a step before entrance in the
military pilot school or Military academy. In the case of Serbian
Air Force it couldnt last more than a two month.
Pilot sticks rather than yokes and separate engine control for
each pilot.
92
POSSIBLEAPPROACHESTOEVALUATIONOFTRAININGAIRCRAFTSUSEDINFLIGHTSCREENING
Evaluating a complex technological product such as
training aircraft needs the systematic approach.
OTEH2016
screening phase must be adequately supported by a
suitable trainer. It has to validate the pilot selection
system by confirming which students are likely to succeed
in the later, more costly phases of pilot training.
There are different approaches to the process of training
airplanes evaluation. Usually, it is done by national test
centers, specialized commercial testing organization or
adhoc commissions. Sometimes the evaluation is done
only by analyzing promo materials or under political
pressures which is unacceptable and very detrimental on
the training and education system.
Different models of flight training system consequently
mean a trainer with different characteristic. In our case,
the possible successor of UTVA75 has considered, with
all necessary prerequisites in this specific case. Having in
mind the syllabus scope, goal and available time for
selection combined with next step trainer characteristics
(Lasta) the profile of desirable primary trainer was
created. According to this, ULA and Lasta as primary
trainers were not considered because their characteristics
are not fit for this role. However, it doesnt mean that in
necessity such plane cant be used.
According to experts experience and reviews, the only
legal and proper way is a testing on a ground and on the
air. The scope of evaluation depends on different
circumstances especially on evaluation object itself. In the
case of primary trainer and its level of technical
complexity minimum three test sorties are required. The
composition of flight testing program considers totaling
two and a half flight hours, at least.
By highlighting the syllabus characteristics, equipment
profile, aircraft configuration, desirable performances and
handling characteristics, the basis for the thorough
evaluation is made. It shouldnt be allowed to evaluate a
plane with ad hoc commissions or only by analyzing
commercial proposals. The only proper way is evaluation
done by experienced experts (test pilots and engineers)
and experienced test organizations, included those
commercial based. The test program must include, at
least: design features, cockpit, performance, handling
qualities and systems. It requires minimum three test
sorties.
The test team must be experienced, with broad experience
on such airplane category. This is one of the most
important conditions for objective evaluation. Having test
facility doesnt necessary mean broad experience on such
plane category. Test pilot schools have different training
modules and courses providing training for test pilots,
flight test engineers, flight engineers and technicians
involved in flight testing. Schools also provide light
aircraft test pilot courses.
Usually, testing is divided on several groups of evaluation
elements:
 The role of the aircraft
 Design features
 Cockpit
 Performance
 Handling qualities
 Systems.
This methodology is not new and it doesnt bring any
novelty, but it must be followed with no exceptions.
Everything else could lead to the wrong choice and
misdecision.
References
[1] Braybrook,R., Valpolini,P.: Trainer order in
prospect, Armada International, 1/2011.
[2] Vlai,S.: Development perspectives of piston and
turboprop trainers and their COIN derivatives,
International Scientific Conference on Defense
Technologies OTEH 2012, Belgrade, 2012.
[3] Market Report 2016, Part I, Military Flight Training,
Defence IQ, London, 2016.
[4] Trescott,M.: Max Trescotts G1000 Glass Cockpit
Handbook, Glass Cockpit Publishing, Mountain
View, CA 94040, 2006
[5] https://www.flightglobal.com/news/articles/infocusgrobaircraftbullishovermoreg120tpsales373317/ accessed 30th July 2016th
[6] Vlai,S., Roenkov,S., Kneevi,A., Vlai,I.: Use
of the commercial software tools in the preparation
phase of the military pilot education and training, V
International Conference of Information Technology
and Development of Education (ITRO 2014),
Zrenjanin, 2014.
[7] Braybrook,R.: Trainers at a Cusp, Armada
International, 5/2009.
[8] The Market for Military FixedWing Trainer Aircraft
20112020, Forecast International, Newtown, USA,
2011.
Main evaluation elements inside groups are as follows:
Cockpit: cockpit access, cockpit size, reachibility of
controls, Student and instructor arrangement, field of
view, Command control dimensions and geometry.
Performance: Touch and go distance, climb
performance, cruise performance, glide performance
(propeller effects).
Handling qualities: Ground procedures, taxiing, takeoff,
maneuver flight, stall warning, stall characteristics,
aerobatics, spin entries and stabilized spins, spin
recoveries, approach, landing.
Systems: Landing gear, flaps, engine other existing
subsystems (for example INS/GPS, AHRS, EFIS, TCAS).
All those elements have to be closely connected with
tacticaltechnical
requirements
and
evolution
prerequisites.
6. CONCLUSION
The flight screening or abinitio training is the very first
stage of flight training. This is intended to establish
whether the student has the necessary aptitude to become
a military pilot in a reasonably short time. Flight
93
POSSIBLEAPPROACHESTOEVALUATIONOFTRAININGAIRCRAFTSUSEDINFLIGHTSCREENING
[9] Vlai S., Kneevi A., Peki N., Comparative
analysis of the analog and the digital cockpit of Lasta
training aircraft, International Scientific Conference
OTEH2016
on Defense Technologies OTEH 2014, Belgrade,
2014.
94
INTEGRATION OF TACTICAL  MEDIUM RANGE UAV
AND CATAPULT LAUNCH SYSTEM
ZORAN NOVAKOVI
Military Technical Institute, Belgrade, novakoviczoca@gmail.com
ZORAN VASI
Military Technical Institute, Belgrade, vti@vti.vs.rs
IVANA ILI
Military Technical Institute, Belgrade, ivilic76@yahoo.com
NIKOLA MEDAR
Military Technical Institute, Belgrade, nmedar9@gmail.com
DRAGAN STEVANOVI
Military Technical Institute, Belgrade, stevanovic.dragan.daca@gmail.com
Abstract: In this paper the selection of appropriate UAV catapult launch system from the supply on the world market for
existing tactical  medium range Unmanned Aerial Vehicle (UAV) in the Serbian Army is analyzed. A special emphasis was
placed on UAV structure strenght at longitudinal acceleration direction which are exposed onto the launch ramp of the
catapult. In addition, UAV accommodation to the catapult launch carriage is analyzed. The ultimate goal of this analysis is
to define the necessary changes to the structure of the existing UAV in order to successfully integrate it with the selected
catapult launch system.
Keywords: UAV, launching ,launching device, catapult, UAV integration redesign.
During their use, the advantages and disadvantages of each
type of individual LDs were recognised. This led to the
conclusion that an LD must be lightweight, must be able to
be operated with minimal personnel, and must have a small
storage volume. These factors need to be considered and
incorporated into the conceptual design of LDs. Also, the
UAV launching device must have the possibility to be set up
and to launch a UAV within fifteen minutes, [3]. The
important factor is the purchase price and the cost of LD
maintenance, which perhaps is crucial to the military
budget.
1. INTRODUCTION
Increasingly the development of unmanned aerial vehicle
(UAV or aircraft) introduces a new term: unmanned aerial
system (UAS) which is considered a hybride system. UAS
comprise UAVs (one or more), ground control station
(GCS), data processing system (DPS), launch and recovery
system. Each of these subsystems is being developed in
several directions so that the number of possible
combinations of hybrid UAS definition grows significantly.
The need to UAV catapult launch arose for the following
reasons. First, UAV catapult launching eliminates
requirements for a takeoff path. This is of crucial
importance in combat conditions when the availability of
the runway is very uncertain. Further, the propellers, being
of fixed pitch, would have to be designed for best
performance for takeoff. This would severely compromise
the performance of the aircraft in flight, [1]. In addition,
UAV takeoff from the runway requires additional fuel
capacity which increases the weight of the aircraft.
All the abovementioned concepts, regardless of their
relative advantages and disadvantages, are in operational
use in armed forces of a number NATO countries. They are
used in those situations when their advantages come to the
fore.
2. UAV LAUNCHING DEVICE SELECTION
The launching device main task is to hand over to UAV the
previously accumulated energy in its system, so that the
UAV at the time of leaving the catapult has a speed of at
least 15% greater than the stall speed for a given
configuration of the UAV. To have a successful takeoff, the
UAV should have sufficient lift force after the instant of
leaving the catapult when its own driving propeller achieves
stable flight take over, [4].
Several systems of launching devices (LDs) for unmanned
aerial vehicles have been developed so far. Existing LDs
could be grouped into six categories: (1) Pneumatic, (2)
Hydraulic, (3) Bungee cord, (4) Kinetic Energy, (5)
Electromagnetic, (6) Rocket Assisted Takeoff (RATO) and
other methods, [2], [3], [4].
95
INTEGRATIONOFTACTICALMEDIUMRANGEUAVANDCATAPULTLAUNCHSYSTEM
Table 1. List of LD essential requirements, [3]
Final Launch Velocity
Maximum Takeoff Mass
Operational Temperature
Range
LD Mass
Maximum Length of the
Launch Envelope
Acceleration at Launch
10G
OTEH2016
The launch parameters of PEGAZ (PEGASUS) UAV:
Max takeoff mass: ............................................250 kg,
Max takeoff speed: ...........................90 km/h (25 m/s),
Launching Angle Range
Launching Remote Control
Set Up Time < 15 minutes
require LD with appropriate launch performance and
solution should be sought in the selection of:
Pneumatic Catapult,
Hydraulic Catapult,
Hybrid HydraulicPneumatic Catapult,
Rocket Assisted Takeoff (RATO).
LD Disassembling for
storage
Number of the Set Up
Personnel
LD Safety, Reliability and
easiness to operate
From the current range of UAV LDs in the global market,
[2], [5], [6], which meet the criteria set out in the Table 1,
for PEGAZ the following UAV LDs could be taken into
consideration:
According to their launch performance pneumatic (or
hydraulic) LDs are perhaps the best because they have a
huge catapults launch power and because they can launch
multiple types of UAV (reconnaissance, aerial target)
designed for different missions. They are characterized by a
relatively uniform acceleration along the launch rail. The
disadvantage of this type of LDs is the complexity of their
construction, which reduces their reliability. The purchase
price of pneumatic (or hydraulic) LD is relatively high
compared to other LD concepts, as well as higher
maintenance costs.
1. MDS HERCULES pneumatic launcher (England)
RATO (Rocket Assisted Takeoff) launching device is
characterized by nearly zero length launch ramp and
relatively higher level of acceleration. This is the most
reliable drive for the UAV launch in extreme environmental
conditions, for higher mass of 500kg UAV and in conditions
of scarce space to launch (boat deck). However, rocket launch
has the disadvantage that the rocket bottle is rejected from
UAV (2 3) seconds after the start, which is unacceptable
from the ecological standpoint . Rocket bottles must be stored
in separate areas in the same conditions as explosive. This is
inconvenient as it imposes additional costs and liabilities. In
addition, the rocketpowered launch reveals a position that is
unfavourable in combat conditions, [3].
Picture 1. MDS HERCULES pneumatic launcher with
aerial target BANSHEE
Max UAV launch mass, [7], [2]...........................250 kg
Max UAV launch speed.......................................55 m/s
2. Aries RO01 or Aries ALPPUL LP02 pneumatic
launcher (Spain)
Bungee catapult systems employ the characteristics of
stored energy within high powered, highly elastic bungees
to launch UAV. Concept of LD with elastic cords is
characterized by the simplest design structure in relation to
other concepts of LDs. Because of its simple construction
they have the lowest purchase price and the lowest cost of
maintenance. Generally, bungee catapults are considered as
LDs of less launch power (max. takeoff weight 50 to 55kg),
[3], except the one that is in operational use in NATO and
that stands out for its launch performance (takeoff weight
140kg, takeoff speed 34m/s). Disadvantage of bungee
catapult is the initial jerk that can unfavourably affect the
sensitive payload of UAV, but it is successfully overcome
by additional fixing equipment inside the aircraft, [1].
Picture 2. Aries RO01 pneumatic launcher with UAV
SIVA
When choosing a catapult for Serbian army (regardless of
whether being purchased or produced) must be primarily
met all the criteria set out in Table 1, which must be in
accordance with the characteristics of UAV (maximum
launch velocity, maximum launch weight, maximum mean
acceleration in the axial direction that UAV withstands). We
should endeavour to catapult selection being based on the
multi catapult, with which it will be possible to launch the
current UAVs and UAVs that will eventually be the future
home development.
Picture 3. Aries ALPPUL LP02 pneumatic launcher
with UAV SIVA
96
INTEGRATIONOFTACTICALMEDIUMRANGEUAVANDCATAPULTLAUNCHSYSTEM
Max UAV launch mass, [8], [2]...........................360 kg
Max UAV launch speed.......................................34 m/s
OTEH2016
Launch mass........................................... (220275) kg
Launch speed............................................. (2138) m/s
3. ESCOZodiac HP3407 hydraulicpneumatic launcher
(USA)
Picture 6. Speed/Acceleration changes onto the catapult
ARCHER ramp
Picture 4. ESCOZodiac launcher HP3407
Max UAV launch mass, [9], [2]............................340 kg
Speed/Acceleration profile Pic. 7, is relating to UAV SIVA
launch with the launch parameters:
Launch mass........................................................ 300 kg
Launch speed.................................................... 33.6 m/s
Max UAV launch speed........................................33 m/s
4. ARCHER hydraulic launcher (Switzerland)
Picture 7. Speed/Acceleration changes onto the catapult
ALPPUL LP02 ramp
Picture 5. ARCHER hydraulic launcher
Max UAV launch mass, [10], [2].........................320 kg
Max UAV launch speed.......................................34 m/s
Aircraft RANGER mean acceleration on ARCHER launcher
rail is close to 5g while the mean acceleration of aircraft
SIVA on ALPPUL LP02 launcher is around 6g. It is
evident that the launch speed on both diagrams are
practically the same, about 33 m/s, wherein the launch mass
of aircraft SIVA is higher for approximatelly 50 kg relative
to the launch mass of RANGER aircraft. It can be concluded
that a lower level of mean acceleration from 1g on
ARCHER catapult is result in differences of aircraft launch
mass (~ 50 kg).
5. Launching of PEGAZ UAV by rocket assisted takeoff
catapult should be taken into consideration only if none of
the above launchers meet the requirements, (Table 1.).
3. ANALYSIS OF UAV STRUCTURE
STRENGTH ONTO THE LAUNCH RAMP
OF THE CATAPULT
Since the launching parameters of PEGAZ UAV are closer
to launch parameters of RANGER UAV, for this calculation
it could be adopted that the acceleration profile (Pic. 6) is
more credible in relation to the acceleration profile (Pic. 7).
No matter the launch speed of being ~ 33 m/s, Pic. 6, (much
higher than 25 m / s, which requires PEGAZ UAV), one can
expect a similar level of acceleration when the PEGAZ
UAV would be launched with a ARCHER catapult. This
conclusion should not reduce the quality of ALPPUL LP02
catapult, as the acceleration profile Pic. 7 refers to the
heavier aircraft in relation to the acceleration profile Pic. 6.
In selection of catapult for the aircraft PEGAZ UAV, the
first step is to check the strength of the UAV existing
structures during acceleration on the launch ramp of the
catapult. In addition, it is necessary to check the sensitivity
of UAV payload and its fastening inside UAV, but this
analysis is beyond the scope of this paper. The actual
analysis is performed according to the available acceleration
profiles of ARCHER ("RUAG"Switzerland), Pic. 6,
launcher and ALPPUL LP02 ("ARIES" Spain), Pic. 7.
Speed/Acceleration profile Pic. 6, is relating to UAV
RANGER launch with the launch parameters:
97
INTEGRATIONOFTACTICALMEDIUMRANGEUAVANDCATAPULTLAUNCHSYSTEM
OTEH2016
3.1. Necessary changes on PEGAZ structure for
catapult integration
PEGAZ UAV airframe is designed as monoplane, highwing with fuselage nacelle and two tail booms with one
horizontal tail and two vertical tails. Aerodynamic scheme
with pusher propeller enables safe maintenance and
operation before and after flights without dangerous contact
with rotating propeller. The aerodynamic scheme enables
payload and mounting of equipment into fuselage nose
section.
Picture 8. PEGAZ side layout (view without wing, tail
booms and tail units)
PEGAZ UAV airframe is made of modern composite
materials. Sandwich composite panels are used for airframe
parts with complex shapes in the outer UAV skins and for
simple internal structure as well. There are numerous local
stiffeners made of laminated fabrics and laminated wood
plates at the points of concentrated loads. Original project
design defined UAV airframe according to dominated flight
loads in flight envelope, and according to extreme loads that
can develop during landing on UAV landing gear and
parachute landing in the case of emergency. First phase of
UAV airframe conceptual design does not have tactical and
technical requirements for takeoff using catapult system or
landing using arrester hook or air bag. In other words,
design of parts, assemblies and the whole UAV airframe is
dimensioned according to aerodynamic and inertial loads
that are developed during flight.
Fuselage nacelle is consisted of four composite stringers
spreading along the whole fuselage, from nose frame
(Frame 1) to firewall frame (Frame 7). Stringers (longeron
type) with trapezoidal shapes in cross section are made of
foam core and two layers of glass fabrics. Positions of
stringers in fuselage cross section are chosen in a way to
fully encompass large fuselage openings, which reinforce
fuselage skin together with skin doublers. Fuselage frames
are cut in the points of stringers interconnections.
Fuselage cross section is hexagonshape having horizontal
upper and lower skins. Lateral skins of fuselage cross
sections do not have vertical sides. These complex shapes
seriously complicate integration of added structure
reinforcement intended to receive takeoff catapult driven
loads, avionics equipment units, and payload into UAV
airframe.
Fuselage nacelle design characteristics
Fuselage nacelle together with central part of wing makes
onepart technological and constructive assembly. Fuselage
nacelle is made of several sections, such as:
Nose front fuselage section with keelson assembled for
mounting of nose landing gear, avionics and part of
hydraulic brake installations.
Front central fuselage section is dedicated for embedding
of UAV flight navigation equipment, and for mounting
of reconnaissance equipment.
Central fuselage section is used for embedding of front
fuel compartment and part of hydraulic brake
installations.
Central fuselage with central part of wing section, part
and starboard, is used for embedding of integral fuel
tanks and for attachment of fittings for tail booms and
outer wing assemblies.
Rear central fuselage section is used for embedding of
rear fuel compartment and parachute compartment for
emergency landing, and for attachment fittings for main
landing gear composite bar spring.
Rear fuselage section is used for UAV engine and engine
equipment.
There is the opening in the lower fuselage skin allowing
nose landing gear movement during extension and
retraction. There is circular opening between frame 3 and
frame 4 for attachment of gimbaled optoelectronic payload
set. There are two hard points for attachment of main
landing gear composite spring bar, just behind the main
fuselage to wing bulkhead (frame 5). Main landing gear
composite spring bar is attached to modified lower fuselage
skin by attachment fittings and bolts. Attachment fittings
enable freely elastic deformations of composite spring bar
during landing, absorbing vertical UAV energy (Pic. 9). The
design of fuselage to landing gear attachments could not
provide robust receiving of longitudinal loads induced by
catapult system. Thats way the landing gear attachment
points are not taken into consideration for receiving catapult
takeoff loads.
Fuselage sections and UAV avionics attachment that meet
aircraft weight and balance define internal UAV
arrangement. Internal UAV structure is made of composite
sandwich panels (laminates with foam core) and appropriate
reinforcements at the points where the concentrated loads
are distributed. There are seven frames and two wing ribs at
each wing side in fuselage nacelle (Pic. 8).
Picture 9. Lower part of UAV fuselage with landing
gears and optoelectronic payload
98
INTEGRATIONOFTACTICALMEDIUMRANGEUAVANDCATAPULTLAUNCHSYSTEM
Other UAV airframe assemblies (outer trapezoidal wings,
tail booms, horizontal and vertical tails) are made of modern
sandwichtype laminated composite materials with
appropriate reinforcements at the points where the
concentrated loads act the airframe, [11], [12], [13].
OTEH2016
and the same length as its internal counterpart support made
of laminated wood reinforced with glass fabrics. Mechanical
connection of outer support and inner counterpart support is
made by three bolts.
Chosen solution for connnection of fuselage nacelle and
catapult launch carriage system at the frame four and six is
optimal in the sense of meeting different contradictory
requirements related to takeoff by catapult system, [14], [15]:
Design solution enables production of lighter, shorter and
compactmade catapult launch carriage assembly,
Center of gravity (CG) of the whole UAV is located
almost at the midpoint between the two supports,
Fuselage nacelle and catapult launch carriage
interconnection does not make any interference with
main landing gear spring bar and its bolt connections to
the frame structure. Main landing gear spring bar to
fuselage structure attachment is designed for vertical
loads generated during landing and not for loads
longitudinally distributed along the frame nacelle
generated by catapult system.
Analysis of possible locations of airframe
reinforcement and integration with catapult system
The optimal location for UAV structure and catapult launch
carriage interface unit must be chosen and designed carefully.
Fuselage nacelle is relatively long in comparison to the whole
UAV length, so the first solution for interface attachment
fittings positioned close to the first and last fuselage frame
was excluded from further consideration because the catapult
launch carriage should be in that case too long and
cumbersome. Two central fuselage frames, fourth and sixth,
were chosen as an acceptable solution for fuselage to catapult
launch carriage attachment and interface unit.
Picture 11. Redesigned structure elements at front
catapult launch carriage support
Main (front) catapult launch carriage support at the frame
four of the fuselage nacelle is designed for receiving of
horizontal and vertical loads (Pic. 11). Conceptually, the
main support is designed as one transversal steel tube which
is actually made of two tubes, port and starboard, connected
by one connecting muff at the fuselage symmetry plane
(center line). Diameter of the tubes is 25 mm and tube wall
thickness is 3 mm. At the fuselage center line, the axial
movement and rotation of the tubes are constrained by two
bolts and connecting muff. Connecting muff is attached to
newadded longitudinal fuselage rib at the fuselage center
line by bolt connections. The main vertical and horizontal
loads are transferred from the two tubes to the fuselage
structure through the two main supports with bushing, port
and starboard, made of laminated wood plate reinforced by
glass fabrics. Two main supports with bushing are shaped to
align to the inner surfaces of the fuselage skins and to the
lower trapezoidalshaped fuselage stringers. Main supports
have the hole close to the frame four. Metal bushing is
embedded into the main support hole. Inside diameter of the
metal bushing is with high tolerance to the outside diameter
of the steel tube. Metal bushing (on the port and starboard
side) has flange that acts as an axial limiter protecting from
hard impact of catapult launch carriage steel arms to fragile
UAV composite skin and its consecutive damage.
Picture 10. Fuselage internal structure layouts with
catapult launch carriage supports
Frame four of fuselage nacelle is used as the rear support of
the gimbaled optoelectronic payload platform, in the lower
part of the frame, and as the vertical support of two front
emergency parachute brackets, in the upper part of the
frame. According to original design, frame is made of 5 mm
light foam core and four glass fabric plies, two from each
side, and with local reinforcements made of laminated wood
plates at the areas where the concentrated load acts. Actual
modification of the frame meant additional cutting of the
frame in the lower area where the stringers and newadded
main wood support with bushing go through the frame (Pic.
11). Main wood support with bushing is the main structural
part that receives the main longitudinal load generated by
catapult launch system.
Frame six of fuselage nacelle is made of light foam core and
glass fabrics. Frame is used as the rear support of two rear
parachute brackets, in the upper part of the frame. Rear
support is designed to receive vertical loads (along zaxis) at
the frame plane, and not along the other axis. Frame
modification meant embedding of newadded counterpart
support made of laminated wood plate at the place where
the outer catapult launch carriage support is located.
Exterior catapult launch carriage support is made of
tetrafluoroethylene transversal bar with L cross section
Main landing gear conception and its position to the catapult
launch carriage supports require additional design efforts
related to resolving of kinematics of the catapult launch
99
INTEGRATIONOFTACTICALMEDIUMRANGEUAVANDCATAPULTLAUNCHSYSTEM
carriage arms that will avoid collision of landing gear,
fuselage structure, optoelectronic payload, and engine
propeller.
Strength calculation of attachment of catapult launch system
to UAV is done by using classical engineering methods.
Reserve factor, which represents ratio of ultimate stress for
material that was used and stress in the element of structure,
is greater than 1 on all elements for conection of the
fuselage nacelle and the catapult launch system, at the frame
four and six.
3.2. PEGAZ Strength Analysis
In case of takeoff using catapult launching device the
strength of structure of the aircraft PEGAZ UAV must be
taken into consideration. All joints of main parts of the
UAV structure and all elements that have attached
concentrated mass have to be designed for a load case when
inercial force is applied, in direction of x axis of the UAV.
3.3. Launch Carriage Interface Requirements
The main task of launch carriage interface is to support
UAV during the launching phase. Also launch carriage
mechanical interface is essential in the System Preparation
and Preflight State until launch. The UAV is held in place
during the preflight operations, including checks and engine
ground run. At the launch phase, the UAV is pushed to
flight speed and at the end of the launch phase, the UAV
leaves the launcher in the flight direction.
The value of inertial force is calculate by following:
Fin = nx g m
(1)
where nx=8 represents the factor of inertial load in direction
of x axis, and m represents mass of the vehicle.
Mechanical Interface Requirements, [10]
Besides that, the ultimate factor must be taken into
consideration, j=1.5. For all vital joints the ultimate factor
has to be multiply by the joint factor, jv=1.2.
Mechanical interface operating is sequentially distinguished
in three phases:
Loading the UAV onto the launcher.
The UAV is held in static position with only gravitational
forces and also combined with the engine thrust. This
requires longitudinal fixation of the UAV in both
directions.
The UAV is accelerated by the launcher and slides out of
the guidance at the end of the launch. Longitudinal
fixation of the UAV only in backward direction is
required.
To calculate strength of the joint between the UAV and
catapult launching device (for maximum mass of the UAV
m=250 kg) the value of the force that is introduced to the
joint is
Fx = j jv Fin = 35316 N
(2)
This force is applied in center of gravity, aligned with x axis
of UAV, in postive direction of the axis.
Physical Requirements
Force, Fx, is reduced to the main catapult launch carriage
support, at the frame four of the fuselage nacelle, which
causes bending moment My with intesity:
M y = Fx 0.214 = 7558 Nm
Dimensions
The dimensions of the interfacing elements of the PEGAZ
UAV (nominal weight of 250 kg) are given in pictures 10
and 11.
(3)
Forces
Front support is receiving all horizontal load, Fx, while
moment My is defined as a couple of vertical forces (Pic. 12)
that are accepted by front and rear support points:
Fz =
My
0.9
7558
= 8398 N
0.9
OTEH2016
Force for acceleration of PEGAZ UAV: max 22073 N,
corresponding to 9 g with 250 kg PEGAZ UAV mass,
acting in direction of the rail. This force to be
transferred by the forward hooks. The rearward support
serves for guidance of the UAV.
Weight force of the PEGAZ UAV: max 2500 N with
respect to ground vertical.
Center of PEGAZ Gravity Range: max 350 mm in
back of the forward hook.
(4)
4. CONCLUSION
The proposed redesign of PEGAZ UAV enables it to be
launched from any catapult whose max acceleration on the
launch rail is up to 8g. The launch parameters of PEGAZ
UAV (max launch speed 25 m/s, max launch mass 250 kg)
are less demanding than the launch parameters of RANGER
UAV or SIVA UAV. This confirms that the launch catapult
performance of ARCHER or ALPPUL LP02 makes
possible to launch PEGAZ succesfully. It is realistic to
expect that PEGAZ UAV will be exposed to maximum of
5g average acceleration on the launch rail of any of these
Picture 12. Forces acting on the catapult launch carriage
supports
100
INTEGRATIONOFTACTICALMEDIUMRANGEUAVANDCATAPULTLAUNCHSYSTEM
OTEH2016
[8] Catalogue of "Aries Ingenieria y Systemas" company,
ALPPUL
LP02
pneumatic
launcher,
Also, the proposed redesign of PEGAZ UAV and set
http://www.ariestesting.com/solutionsbyrequirements for launch carriage interface enable launch
applications/unmannedaerialsystems/uavlaunchers/
carriage adaptation on launchers ARCHER or ALPPUL LP[9] Catalogue of "Zodiac Aerospace" company,
02 according to modular dimensions of PEGAZ UAV.
http://s76586.gridserver.com/downloads/documents/hp
3003.pdf
References
[10] Catalogue of "RUAG Aerospace" company,
[1] Austin, R.: Unmanned aircraft systems:UAVS design,
http://ofp.gamepark.cz/_hosted/swissteam/Pictures_of_
development and deployment, John Wiley & Sons Ltd,
Swissmod/archer_flyer.pdf
Chichester, West Sussex, United Kingdom, 2010.
[11] Niu,M.C.Y.: (1992). Composite Airframe Structures,
[2] Novakovi,Z., Medar,N.: Lansirni sistemi bespilotnih
Conmilit Press Ltd., Hong Kong
letelica, Podaci o naoruanju, faktografska sveska,
[12] Aronsson,A.: (2005). Design Modeling and Drafting of
ISSN 18203426, VTI, Beograd, 2015.
Composite Structures, Master thesis, Lulea Univerisity
[3] FRANCIS,J.: Launch System for Unmanned Aerial
of Technology, 2005:060 CIV, ISSN: 1402 1617.
Vehicles for use on RAN Patrol Boats, Final Thesis
[13]
Willinger,M.:
(1989), Industrial Development of
Report 2010, SEIT, UNSW@ADFA
Composite Materials: Towards a Functional
[4] Novakovi,Z., Medar,N.: Design of UAV Elastic Cord
Appraisal, Composites Science and Technology 34,
Catapult, OTEH 2014 Symposium, MTI, Belgrade,
pp. 5371.
2014.
[14] Niu,M.CY.: Airframe Structural Design, Practical
[5] Jane's.: Unmanned Aerial Vehicles and Targets, Issue
Design Information and Data on Aircraft Structures,
29, 2007, Launch and Recovery, Chapter.
Conmilit Press Ltd., 1988.
[6] Unmanned Vehicle Handbook, Issue 21, SHEPHARD [15] Stinton,D.: The Design of the Aeroplane, BSP
2013.
str.:
164,165.
Professional Books, Oxford, 1983.
http://shephardftp.co.uk/digital_editions/sample%20pa
[16]
Bruhn,E.F.:
Analysis and Design of Flight Vehicle
ges/UVH_21_sample/index.html#/16/
Structures, TriState Offset Company, 1973.
[7] Catalogue of "Meggitt Defence Systems" firm,
http://www.meggittdefenceuk.com/PDF/Hercules%202 [17] STANAG4671
014_App%20Mod%201.pdf
two launchers.
101
CONTRIBUTION TO THE MAINTENANCE OF MI8 HELICOPTER IN
THE SERBIAN AIR FORCE
ZORAN ILI
Technical Test Center, Belgrade, zoranilic_65@yahoo.com
BOKO RAUO
Faculty of Mechanical Engineering, Belgrade, brasuo@mas.bg.ac.rs
MIROSLAV JOVANOVI
Technical Test Center, Belgrade, mjovano@sbb.rs
LJUBIA TOMI
Military Technical Institute, Belgrade, ljubisa.tomic@gmail.com
STEVAN JOVII
Technical Test Center, Belgrade, stevanjovicic@gmail.com
RADOMIR JANJI
Technical Test Center, Belgrade, lari32@mts.rs
NENKO BRKLJA
Technical Test Center, Belgrade, brkljacnenko@gmail.com
Abstract: Modern helicopters are equipped with Health and Usage Monitoring Systems (HUMS). Serbian Air Force
helicopters are not equipped with HUMS. One of the most effective ways to control state of helicopter is vibration
measurement, i.e. defining frequency spectrum using appropriate equipment. Based on the analysis of vibrations, measured
on helicopter, it is possible to determine the various types of failures and improper function of component parts or
aggregates. The basic problem is to determine vibration amplitudes that are characteristics of proper work, and which ones
are characteristics for incorrect operation of helicopter. In the case of frequency spectrum disorder due to reduction of
function of certain working elements, amplitudes of vibration will be increased on speeds of these elements. In such
circumstances, the only way to assess condition of helicopter is a comparative analysis of results previously measured
vibrations and present measurements. All measurements are should be done at the same places and equipment with the same
characteristics. This paper presents results of the analysis and vibration measurements, which were performed at
helicopters Mi8. It was stated a few examples of observed irregularities of individual parts and clear guidelines for taking
action to maintain are given.
Keywords: helicopter vibrations, blade, fault, vibration measurement, maintenance.
1. INTRODUCTION
Rotation of rotor helicopter in a horizontal plane generates
vertical lift force, and depending on its size, helicopter is
stationary in the air or moving up or down [1].
During rotor rotation, on helicopter are generated vibrations
caused by effect of the aerodynamic and mechanical forces.
Asymmetry of vertical lift force on the rotor, during its
horizontal rotation, causes vertical vibration, and asymmetry of
aerodynamic drag force on blades causes horizontal vibration.
In addition, separation of airflow at the ends of blades causes
both vertical and horizontal vibrations, upon high induced
velocities. The vibrations of helicopter may occur as a
consequence of the lack of dynamic and static balancing
blades. Main sources of vibration on helicopter, such as rotors,
are driveline, transmission and connections [2].
102
During life cycle of helicopter, such as the correct exploitation,
it is necessary, create maintenance that ensures high reliability
and integrity of the components and helicopter as a unique
complex systems. The current helicopter maintenance practice
requires a large number of parts to be monitored and replaced
at fixed intervals. This constitutes an expensive procedure
which contribute significantly to maintenance costs [3].
Maintenance of helicopters in the Serbian Air Force meets
those requirements. Unfortunately, maintenance significantly
lags behind the methods of maintenance helicopters in modern
armies of the world. The US army, for example, has developed
program that allows to definine safety system in real time for
different types of helicopters. This program is called Vibration
Management Enhancement Program (VMEP)  a program of
extended vibration management [4, 5]. Based on identified
spectrum of vibration, VMEP determines condition and
CONTRIBUTIONTOTHEMAINTENANCEOFMi8HELICOPTERINTHESERBIANAIRFORCE
OTEH2016
correctness of powertrain helicopters and elements of
transmission power and torque [5].
RPM shaft tail rotor at 2589 rpm (frequency f =43 Hz);
Serbian Air Force helicopters are not equipped by devices
that can continuously monitors vibrations in real time during
flight. When vibration phenomena, that are not common,
appear on helicopter, it is necessary to install a helicopter
measurement equipment, to measure vibrations and make a
comparative analysis of the vibration spectrum in which
there are disorders with normal (regular) vibration spectrum
[6].
Vibrations on helicopter can be divided into vertical (in
direction of vertical Z  axis) and lateral (in longitudinal
direction X axis, in direction of transverse Y  axis).
Characteristics of gears in main gearbox of helicopter.
In normal (regular) frequency spectrum dominant are
vibrations at BPF main rotor and its higher harmonics of
oscillation (Picture 1). Also, in frequency spectrum exist
vibrations at all working frequencies of rotating elements,
but these vibrations should have smaller amplitudes than
vibrations at BPF main rotor.
Technical Test Center engineers have made the experiment
and performed vibration measurements at frequencies from
0 Hz to 2 kHz on two Mi8 helicopters in elected positions.
Procedures and methods of vibration measurement have
been described in [7] and [8]. The aim of the experiment
was to demonstrate a part of extensive possibilities which
vibration analysis provide for rapid assessment of the
condition of helicopters. Vibrations were measured in
helicopters, which were attached to the ground during the
work, for several main blade rotor angles (USV), without
autopilot. Frequency spectrum is analyzed for the range
from 0 Hz to 2 kHz, but in this paper is shown spectrum
from 0Hz to 80Hz in which there are certain anomalies in
the frequency spectrum. The case of highfrequency
vibration spectrum which has certain anomalies in the
frequency spectrum is shown. The analysis of vibrations that
measured on two helicopters has determined concrete faults
and has given precise instructions for their elimination.
During regular vibration measurements, should monitor
changes in vibration spectrum on helikopter. In the case of
vibration spectrum disorders due to failure of an working
element, vibration amplitude would be increased at the
frequency of rotation of the element. Therefore, it is
necessary to know frequency of rotation of all working
elements of helicopter.
2. SOME ASPECTS VIBRATION ON
HELICOPTER
During operation of helicopters are always present
vibrations small amplitudes. Helicopter vibrations are
unsteady accelerations of any given location inside the
fuselage, e.g. on pilot seat, copilot seat or at a given crew
or passenger seat measured along three mutually orthogonal
axes (as a fraction of acceleration due to gravity, g). The
dominant source of helicopter vibration is the main rotor.
Frequency of vibration caused by the main rotor is at integer
multiples of the rotor RPM  1 per revolution (1/rev) is the
rotor RPM, then 2/rev, 3/rev and so on. In addition to the
main rotor, other sources of vibration are  the engine/fan
system, the main rotor transmission/driveshaft/gear system,
the tail rotor and its transmission system and loose
components that are a regular or external part of the aircraft.
Examples are four out of balance rotor blades, loose tail
fins, loose engine shaft mounts, unsecured canopy, landing
gear system or external weapons or cargo systems [9].
Picture 1. Harmonic oscilation of main blades
2nd harmonic b) 3rd harmonic
To determine reasons of increased occurrence of vibrations
on a helicopter, it is recommended to carry out a
comparative analysis of vibrations with disorders in
spectrum, compared to normal (regular) spectrum of
vibration.
Quick comparative analysis of vibrations on Mi8 helicopter
is done by comparing amplitudes of two characteristic
frequencies: the frequency at 3.2 Hz (frequency of main
rotor  RPM MR ) and the frequency at 16 Hz (Blade Pass
Frequency of main rotor  BPF). Values of amplitude of
vibration at frequency of rotation of main rotor should be
several times fewer than amplitude at BPF, or ARMS (RPM
MR)/ARMS (BPF) <<1.
Vibrations that are usually analyzed on Mi8 helicopter
occur at:
Revolutions per Minute main rotor (RPM MR) at 192193
rpm (frequency of main rotor f = 3.2 Hz);
Blade Pass Frequency (BPF) which suits to the product of
N * f (the number of blades N and the frequency of the
rotor f). The helicopter Mi8 main rotor is equipped with
5 blades so BPF is 16 Hz;
3. EXPERIMENTAL SETUP
The aim of the experiment, which was carried out by
engineers of Technical Test Center (TTC), was to measure
vibrations at Mi8 helicopter, in elected positions, at
RPM tail rotor at 1124 rpm (frequency f=18,733 Hz). Tail
rotor helicopter Mi8 is equipped with 3 blades so BPF is
56.2 Hz;
103
CONTRIBUTIONTOTHEMAINTENANCEOFMi8HELICOPTERINTHESERBIANAIRFORCE
OTEH2016
dBFASuite software is installed on this system, that is also
used to record signals from installed piezoelectric
accelerometers, as well as processing of the recorded data.
frequencies from 0 Hz to 2 kHz, and to show part of
extensive capabilities of vibration analysis for rapid
assessment of helicopter condition. Vibrations were
measured in helicopters, which were attached to the ground
during the work, for several main blade rotor angles (USV),
without autopilot.
Accelerometers No. 1, 2 and 3 are mounted at position 1
(down below the outside of fuselage), Picture 3.
Accelerometer No. 4 is mounted on the floor next to the
pilot seats. Accelerometer No. 5 is placed on the floor of
passenger compartment within the zone center of gravity of
the helicopter. Accelerometer No. 6 is positioned on the rear
side of the fuselage, at the beginning of the tail cone.
Table 1. Characteristics of NetdB12 system  01 Metravib
Input channels
No. channels
12 BNC
Resolution
24 bits
Voltage
AC/DC/AC ICP
20db: 14,1V (10V RMS)
0db: 1,41V (1V RMS)
Range
+20db: 141 mV (100mV RMS)
CHP
>105dB RMS full scale
Characteristics of piezouniaxial accelerometers that were
used are given in Table 2.
Table 2. Characteristics of accelerometers
Dynamic Sensitivity Position
o/n
Type
Axes
range (g) (mV/g) (place)
1 CTC 1351A X
10
500
1
2 CTC 1021A Y
50
100
3 CTC 1351A Z
10
500
4 CTC 1021A Z
50
100
2
5 CTC 1021A Z
50
100
3
6 CTC 1021A Z
50
100
4
Vibration spectrum was analyzed from 0 Hz to 2 kHz, but in
this paper is shown area in which there are certain
anomalies in the frequency spectrum.
4. RESULTS
Vibration spectrum at all helicopters was recorded in the
area from 0 Hz to 2 kHz, but in this paper is shown
spectrum from 0Hz to 80Hz, where are some anomalies.
Higher vibration spectrum is seen for concretely occurrence
where are certain anomalies in the vibration spectrum.
Vibrations on helicopter No.1 were measured twice.
Frequency spectrums, which were measured in the first
measurement on the helicopter, position 1, for several main
blade rotor angles (USV) 5, 7, 9 and 11, are shown in
Picture 4.
Picture 2. System NetdB12  01 Metravib in a helicopter
Testmeasuring equipment type NetdB12  01 Metravib,
was used for measuring vibrations on Mi8 helicopter on the
ground. Technical characteristics of the data acquisition
system are given in Table 1, and Picure 2 that shows the
system embedded in a helicopter.
Picture 4. Frequency spectrums for main blade rotor
angles (USV): a) 5 (1st measurement)
During measurement of vibrations in helicopter No.1,
according to subjective feeling of crew, the helicopter
vibrations were large, so measurement was interrupted.
Picture 3. Three piezoaccelerometers down below the
outside of fuselage  position 1
104
CONTRIBUTIONTOTHEMAINTENANCEOFMi8HELICOPTERINTHESERBIANAIRFORCE
OTEH2016
Picture 4. Frequency spectrums for main blade rotor
angles (USV): b) 7 c) 9 d) 11 (1st measurement)
This crew feeling is confirmed by diagram in Picture 4 d)
where are perceived increased levels of vibration in
spectrum up to 7 Hz.
Checking of the main rotor blades was carried out, after
vibration measurement, and has been found damage in
sealant between segments on Engineers from TTC have
repeated vibration measurement on this helicopter. a blade.
The blade was repaired and main rotor blades were adjusted
in track condition.
Picture 5 shows the frequency spectrum of vibrations,
measured on the same helicopter, after repairs the blade, in
position 1, for USV: 5, 7, 9 and 11.
Picture 5. Frequency spectrum for USV: a) 5 b) 7 c) 9
d) 11 ( 2nd measurement)
Picture 6 shows the frequency spectrum on helicopter No.2,
position 1, for USV 5, 7, 9 and 11.
105
CONTRIBUTIONTOTHEMAINTENANCEOFMi8HELICOPTERINTHESERBIANAIRFORCE
OTEH2016
5. DISCUSSION
Vibration measurement on helicopter No.1 was performed
twice. After the first measurement the following were
found:
Vibration at frequency 3.2 Hz, RPM MR, were dominant
throughout the whole main blade rotor angles (USV).
There are vertical and lateral vibrations;
Amplitudes at frequency 16 Hz, BPF, have usual values.
Vibrations at frequency of 16 Hz in the vertical direction
are dominant in relation to the lateral vibrations;
Vibrations at frequencies 56.2 Hz and 65 Hz are changed
with variation USV, Pictue ratio of amplitudes ARMS
(RPM MR)/ARMS (BPF) at frequencies 3.2 Hz and 16 Hz,
at all USV is unfavorable, which indicates a pronounced
imbalance of the main rotor.
Mentioned relationships of observed amplitudes, indicates
an imbalance of the main rotor and that one blade generates
more lift force than other blades.
After of corrective actions, vibration measurement in the
helicopter were performed once again, and have been found
the following:
Vibration at frequency 3.2 Hz, RPM MR, are significantly
reduced;
Vibrations at frequency 16 Hz, BPF, have become
dominant, especially in direction of Y  axis, for all USV,
followed by vibrations along Z  axis;
Vibrations at frequency 56.2 Hz and 65 Hz are reduced,
which shows that unbalance on the main rotor has an
impact on elements of tail rotor transmission.
The ratio of amplitudes ARMS (RPM MR)/ARMS (BPF) at
frequencies 3.2 Hz and 16 Hz is favorable for all USV, ie.
vibration level at 3.2 Hz is several times less than the level
of vibration on the16 Hz, in direction of all axes.
Based on presented analysis of vibration spectrum in
helicopter, which were measured before and after of
corrective actions, it can be concluded that the main rotor is
balanced, and that the level of vibrations reduced to
minimum, at all frequencies, except of BPF, which
remained the same.
On helicopter No.2, in the frequency range from 0 Hz to 80
Hz, have been found the following:
at frequency 3.2 Hz, RPM MR, the dominant are vertical
vibrations (Z  axis) for US) via 9. Also exist and lateral
vibrations (both X and Y axes), that are dominant for
USV 9;
at frequency 16 Hz, BPF, the dominant vibrations are in
the X  axis. This level of vibrations is occured for USV
5 and remains constant up to maximum USV angles;
at frequency 56.2 Hz, for all USV, tail blades pass
frequency is constant.
The ratio of amplitudes ARMS (RPM MR)/ARMS (BPF) at
frequencies 3.2 Hz and 16 Hz, for USV 5 is favorable and
is approximately 0.2 to 0.4 on all axes of the helicopter.
Further increase USV leads to disorder of the regular
vibration spectrum and increases the ratio of amplitudes.
Picture 6. Frequency spectrum on helicopter No.2, for
USV: a) 5 b) 7 c) 9 d) 11
106
CONTRIBUTIONTOTHEMAINTENANCEOFMi8HELICOPTERINTHESERBIANAIRFORCE
In direction of Y  axis the ratio ARMS (RPM MR)/ARMS
(BPF) has been already disrupted at USV angle 7, when its
value becomes greater than 1. The ratio ARMS (RPM
MR)/ARMS (BPF), in the direction of Z  axis, begins to be
distorted at angle of 9 USV, and at USV angle 11 this
relationship value is 1.
OTEH2016
helicopters, have developed programs for defining safety of
systems in real time, based on detection of changes in
vibration spectrums. Maintenance of helicopters in the
Serbian Air Force is done in traditional manner, without
continuous monitoring of vibration during operation.
6. CONCLUSION
The most favorable ratio of ARMS (RPM MR)/ARMS (BPF) is
held along X  axis and its value is less than 0.5 for all USV.
Based on the analysis of vibrations, it is possible to
determine the various types of failures and improper
function of component parts or aggregates. In the case of the
frequency spectrum disorder due to reduction of function of
certain working elements, the amplitudes of vibration will
be increased rotation speeds of these elements. Maintenance
of helicopters in the Serbian Air Force is done in traditional
manner, without continuous monitoring of vibration during
operation. In these circumstances, without Health and Usage
Monitoring Systems (HUMS) on helicopters, the only way
to assess condition of helicopters is a comparative analysis
of results previously measured vibrations and present
measurements.
The said ratio of amplitudes ARMS (RPM MR)/ARMS (BPF)
indicates that there is an imbalance of the main rotor, and
that one blade generates a greater lift force than the other
ones.
On this helicopter, occurrence of high levels of vibration , at
the oscillation frequency of 337.75 Hz has been spotted.
This frequency refers to the work (rotation) gear No. 10 in
the gear unit, ie. the bell of the main rotor shaft. In Picture 7
is given frequency spectrum from 0 Hz to 800 Hz measured
by an accelerometer No.5 placed on position No. 3 (the
floor of passenger compartment).
The analysis of vibrations measured on two helicopters, in
frequency range from 0 Hz to 80 Hz, it was found that
imbalance on main rotor of helicopter No.1 exists and that
one of blades generates more lift power than other blades.
Corrective activities were carried out on the main rotor,
because of this. Vibration measured after the corrective
actions, confirms that the disadvantages are eliminated, and
that the level of vibration in helicopter reduced to a
minimum at all frequencies, except at a frequency of 16 Hz,
BPF, which remained the same. In the analysis was carried
out a comparison of vibration amplitude ARMS(RPM MR)
and ARMS(BPF), at the oscillation frequency of 3.2 Hz and
16 Hz.
The similar analysis showed that on helicopter No.2, it is
necessary to perform validation geometry blades, balancing
and adjustment of main rotor blades in track condition, in
order to reduce the vibration of the main rotor. Because of
increased amplitude of vibration at a frequency of 337.75
Hz oscillation is necessary to review gear No. 10 in the gear
unit, ie. the bell of the main rotor shaft. During the
experiment were not carried out corrective actions on the
second helicopter.
Picture 7. The vibrations at rotation frequency of gear
No. 10 of the main rotor shaft helicopter No.2
In order to reduce vibration of the main rotor in direction of
Yaxis, it is necessary to perform balancing and adjustment
of main rotor blades in track condition.
The increase of vertical vibrations in the main rotor requires
validation geometry blades and adjustment of main rotor
blades in track condition.
Changes in vibration spectrums of helicopters in operational
use can not be fully defined through periodic vibration
measurements, so that need continuous monitoring
vibrations on helicopters during flights and throughout the
whole exploitation.
Because of the increased amplitude of vibration at
frequency of 337.75 Hz it is necessary to inspect gear No.
10 in the gear unit, ie. the bell of the main rotor shaft.
Removals of identified failures were not carried out during
the experiment.
This paper analyzes specified vibrations measured on two
helicopters. Vibrations were measured during the
experiment. Highlighted are possibilities which provide
monitoring of vibration to define the current status and
prediction of the future condition of the helicopter
maintenance.
6. References
[1] Nenadovi,M. Osnovi projektovanja i konstruisanja
helikoptera, Beograd, 1982. (in Serbian)
[2] Padfield,G.D.: Helicopter Flight Dynamics, Blackwell
Science Ltd. Cambridge, 1996.
[3] [Stupar,S., Simonovi,A., Jovanovi,M.: Measurement
and Analysis of Vibrations on the Helicopter Structure
in Order to Detect Defects of Operating Elements,
Scientific Technical Review, Vol.62, No.1, 2012,
pp.5863.
Vibrations are measured during operation,in exceptional
cases, on helicopters in the Serbian Air Force. For example,
that was the case for identification the cause of a
cancellation that has already occurred, so that could not be
solved in a classic way. Modern armies for different types of
107
CONTRIBUTIONTOTHEMAINTENANCEOFMi8HELICOPTERINTHESERBIANAIRFORCE
[4] Giurgiutiu,V., Grant,L., Grabill,P., Wroblewski,D.:
Helicopter health monitoring and failure prevention
through Vibration Management Enhancement
Program, presented at 54th Meeting of the Society for
Machinery Failure Prevention Technology, Virginia
Beach, VA, May 14, 2000.
[5] Grabill,P., Brotherton,T., Berry,J., Grant,L.: The US
Army and National Guard Vibration Management
Enhancement Program (VMEP): Data Analysis and
Statistical, presented at the American Helicopter
Society 58th Annual Forum, Montreal, Canada, June
1113, 2002.
[6] Ili,Z., Jovanovi,M., Jovii,S., Janji,R., Tomi,Lj.:
New opportunities for maintenance of helicopters in
the serbian air force, 19th International Conference
Dependability and Quality Management ICDQM
OTEH2016
2016, Prijevor, Serbia, 2930 June 206, pp 216223. (in
Serbian)
[7] Rakonjac,V., Filipovi,Z.: Merenje vibracija i
relevantnih parametara leta transportnog helikoptera
Mi8 sa revitalizovanim lopaticama noseeg rotora,
Vojnotehniki glasnik 6, 2004, pp.611621. (in
Serbian)
[8] Ili,Z., Rauo,B., Jovanovi,M.: Istraivanje uticaja
nivoa vibracija na pilotskom seditu u funkciji
ispravnosti motora ispitivanjem u letu aviona Lasta,
Tehnika 67 (6), 2012, pp 951963. (in Serbian)
[9] Datta,A.: Fundamental Understanding, Prediction and
Validation of Rotor Vibratory Loads In SteadyLevel
Flight, doctoral dissertation, Faculty of the Graduate
School of the University of Maryland, 2004.
108
ON THE EFFECTIVE SHEAR MODULUS OF COMPOSITE HONEYCOMB
SANDWICH PANELS
LAMINE REBHI
University of Defence in Belgrade, Military Academy, Belgrade, Serbia, rebhi.lamine@gmail.com
MIRKO DINULOVI
University of Belgrade, Faculty of Mechanical Engineering, Belgrade, Serbia, mdinulovic@mas.bg.ac.rs
PREDRAG ANDRI
cole Polytechnique Fdrale de Lausanne, Institute of Mechanical Engineering, Lausanne, Switzerland,
predrag.andric@epfl.ch
MARJAN DODI
University of Defence in Belgrade, Military Academy, Belgrade, Serbia, marjan.dodic@va.mod.gov.rs
BRANIMIR KRSTI
University of Defence in Belgrade, Military Academy, Belgrade, Serbia, branimir.krstic@va.mod.gov.rs
Abstract: In the present paper the effective shear modulus of composite plates with honeycomb core is determined. This
elastic coefficient represents one very important property especially in constructions subjected to torsion and combined
bending  torsion. The structures investigated in this research consisted of top and bottom plates (of different types of carbon
fibers, T300, AS4 in epoxy matrix) and honeycomb core (HexWeb engineered core). Starting from classical lamination
theory the effective shear modulus of top and bottom plates was determined for each ply in the stack up sequence. These
plies were lumped into a single composite layer for different fiber orientations and plies thicknesses. Elastic coefficients
for the HexWeb engineered core were obtained using Master and Evans relations for the equivalent properties of
honeycomb cores.
To verify this approach finite element method was used to determine the displacement, stress and strain field on composite
plates with honeycomb core. Two types of models were compared: the initial model where all the material components,
plates and the core were modeled with their intrinsic properties and lumped model with calculated effective elastic
coefficients.
It was found that the method of effective shear modulus calculation can be successfully used in situations where top and
bottom plates are symmetric or quasi isotropic in general.
Keywords: Shear modulus, Equivalent material properties, Composite material, Honeycomb sandwich panel, Finite element
analysis.
1. INTRODUCTION
Sandwich composites are widely used in aerospace
structural design, mainly for their ability to substantially
decrease weight while maintaining mechanical performance.
It has long been known, that separating two stiff materials
(facings or stress skins) with a lightweight material (core)
increases the structures stiffness and strength.
The faces carry the tensile and compressive stresses in the
sandwich. Faces can be manufactured from metallic material
(such as Al 3003, Al 5052, Al 5056), however polymer matrix
composites are being used more and more as face materials.
Composites can be tailored to fulfill a range of demands like
anisotropic mechanical properties, freedom of design, excellent
surface finish, etc. Faces also carry local pressure. When the
local pressure is high, the faces should be dimensioned for the
shear forces connected to it.
108
The cores function is to support the thin skins so they do not
buckle (deform) inwardly or outwardly, to keep them in
relative position to each other and to carry shear stresses. To
accomplish this, the core must have several important
characteristics. It has to be stiff enough to keep the distance
between the faces constant. It must also be so rigid in shear that
the faces do not slide over each other. The shear rigidity forces
the faces to cooperate with each other. If the core is weak in
shear, the faces do not cooperate and the sandwich will lose its
stiffness It is the sandwich structure as a whole that gives the
positive effects. However, the core has to fulfill the most
complex demands. Strength in different directions and low
density are not the only properties the core must have. Often
there are special demands for buckling, insulation, absorption
of moisture, aging resistance, etc. The core can be made of a
variety of materials, such as wood, aluminum (Al 3003, Al
5052, Al 5056), variety of foams (Corecell M Foam) and
polymer matrix composite materials (Figure 1).
OTEH2016
ONTHEEFFECTIVESHEARMODULUSOFCOMPOSITESANDWICHPANEL
To keep the faces and the core cooperating with each other,
the adhesive between the faces and the core must be able to
transfer the shear forces between them. The adhesive must
be able to carry shear and tensile stresses. It is hard to
specify the demands on the joints.
particles) and the floc (short fibres). These two components
are mixed in a waterbased slurry and machined to a
continuous sheet. Subsequent hightemperature calendering
leads to a dense and mechanically strong paper material
(Figure 2).
During this manufacturing process, the longer floc fibres
may align themselves in direction of the paper coming off
the machine, which leads to orthotropic mechanical
structure, or the fibers may randomly distribute themselves
in the paper resulting in a 2D random composite structure (it
is assumed that the paper is manufactured to be relatively
thin, hence 2D structure results). The latter case will be
considered in this paper.
Further processing of this paper material to form hexagonal
cells is commonly carried out using the adhesive bonding
method, meaning that the bonded portion of two adjacent
paper sheets is held together by adhesive. This method
inevitably leads to double the wall thickness of the bonded
cell walls if compared to the unbonded free cell walls
(Figure 3).
Figure 1. Composite panel with honeycomb core
construction
It is of great importance to predict during the design phase
the properties of above mentioned construction. This usually
means to accurately estimate all elastic coefficients (Eij, Gij,
ij, i,j=x,y, Youngs modulii, Shear modulii and Poissons
ratios) based on sandwich panel geometry (core cell
geometry, cell wall thickness, facings thicknesses, etc.) and
panel constituents material properties.
In the present work the shear modulii (inplane Gxy and outofplane, Gxz and Gzy) of the panel with composite
hexagonal core and composite faces are investigated and the
model for these properties calculation is proposed.
2. NOMEX PRODUCTION
First, the manufacturing process of core will be analyzed.
The emphasis will be given to composite material cores.
Nomex is a trademarked of nonmetallic paper (honeycomb
core basic building block) material, which is well known for
its excellent mechanical and other properties relevant for
aerospace applications.
Figure 3. Double wall thickness
After the paper sheets are bonded and shaped to hexagonal
cells, the resulting honeycomb block is dipped in liquid
resin (usually phenolic based) and subsequently oven cured.
This dippingcuring process is repeated until the desired
density of the core is achieved. The resulting resin coating
of the paper leads to a layered material with an orthotropic
or 2D anisotropic ductile center layer (core paper) and two
isotropic very brittle outer layers (resin layers).
3. ANALYTICAL MODEL OF SHEAR
MODULUS FOR THE HONEYCOMB CORE
In order to relate stress and strain in the structure the
constituent material properties have to be known and elastic
coefficient in the stiffness matrix have to be calculated. For
the structures investigated in this paper the determination of
the necessary elastic coefficients is very complex due to
complex geometry of the panel a furthermore complex
structure of the constituent materials i.e. core and facings.
The analyst relies on the manufacturer data or has to
perform tests if the all required values of the elastic
Figure 2. Honeycomb (hexagonal) core production
process scheme
In general, the paper is a composite that consists of two
forms of polymers, the fibrids (small fibrous binder
109
OTEH2016
ONTHEEFFECTIVESHEARMODULUSOFCOMPOSITESANDWICHPANEL
coefficients are not supplied. Secondly numerical studies
can be performed, requiring large and cumbersome finite
element models if the equivalent model approach is not
taken. This method consists of lumping several composite
layers into a single (lumped) layer with same characteristics
rendering the same stress  strain (although averaged) and
displacement fields for the same boundary conditions. It is
true that this approach still requires numerical modeling of
the structure, however, the models based on equivalent
material properties are much smaller in size (less DOFs),
hence requiring less confrontational resources, computing
time and yielding stressstrain and displacement fields much
faster compared to detailed (micro mechanics models). This
is the reason why the validation and determination of the
equivalent material properties (especially when the
composite panels with honeycomb core are in question).
The literature survey on the subject matter has revealed that
the development of equivalent composite models in the
focus of many researchers. For example, Master and Evans
model for equivalent Youngs modulii in fiber and cross
fiber directions, Ashbys model for equivalent inplane
shear modulus are only few of several existing models.
However, all this models assume that the starting material is
isotropic. For example, in Master and Evans model one of
the required input variables is Ef which represents the
Youngs modulus of the paper. This is directly applicable
for honeycomb cores where the basic building material is
isotropic (for example: hexagonal aluminum cores), in cases
where the cores are manufactured from composite materials
(section 2), the equivalent properties of the building core
material have to be determined first before using one of the
already established and proven material equivalent material
properties.
(2)
As described in previous section, the paper core is dipped in
resin (several times until the desired density of the paper
used for honeycomb structure is achieved).
Figure 4. Core paper with added isotropic resin layers
This process adds rigidity to the paper and equations (1) and
(2) have to be modified to account for this effect. Resin
properties (Modulus of elasticity, Shear Modulus) are
known and classical lamination theory (CLT) can be applied
in this case. Also resin itself is considered to be isotropic.
Theretofore, it follows using CLT, stresses in each core
paper layer can be expressed in the following form:
If the paper material is manufactured in a such a way that
the resulting material is anisotropic. Consists of polymer
matrix and randomly distributed fibers (in a 2D plane) the
effective Youngs modulus of such a structure can be
determined based on Christensen and Waals model [1]
x Q1 Q2 0 x
y = Q2 Q1 0 y
xy 0 0 Q6 xy
Christensen and Waals examined the behavior of a composite
system with a threedimensional random fiber orientation. Both
fiber orientation and fibermatrix interaction effects were
considered. For low fiber volume fractions, the modulus of the
3D composite was estimated to be:
Vf
E f + [1 + (1 + m ) V f ] Em
6
E2 D
2 (1 + 2 f
Inplane shear modulus for the paper of the composite
honeycomb structure can be calculated.
3.1 Honeycomb core with 2D random fibers
E2 D =
2D =
(3)
Where,
Q1 =
Ei
E
, Q2 = i 2i , Q6 = Gi
1 i2
1 i
(4).
In equations (4), index i denotes the elastic coefficient for
the corresponding layer, for example if the central midlayer
is in question the value of Ei is calculated according to
equation (1) and all other layers (resin) assume their
intrinsic values of E. Same applies to shear modulus and
Poisson ratio in equations (4).
(1)
Where Vf is the fiber volume fraction in the composite, m is
the matrix Poissons ratio Ef an Em are the Youngs modulii
of the fiber and matrix phase respectively. Since it can be
considered that the core composite (paper) is isotropic
(fibers are evenly distributed) the basic relation between
Youngs modulus (E), Shear modulus (G) and Poissons
ratio holds. Manera [2] proposed approximate equations to
predict the elastic properties of randomly oriented short
fiber composites. Using this approach, it can be shown that
for this type of composites (thin plates, with randomly
distributed short fibers) Poissons ratio is =1/3. Using
relation:
Expressing relation (3) in terms of inplane resultant forces
one can obtain:
N x t / 2 x
Ny =
y dz
N xy t / 2 xy
Where t denotes thickness of the layer (Figure 4)
110
(5)
OTEH2016
ONTHEEFFECTIVESHEARMODULUSOFCOMPOSITESANDWICHPANEL
Ei1
Ei 2 i 21
, Q = Q21 =
,
1 i12 i 21 12
1 i12 i 21
Ei 2
Q22 =
, Q = Gixy
1 i12 i 21 66
For the whole layup it follows:
Nx
Ny =
N xy
Q11 =
x
y dz
t / 2
xy
t /2
i =1
(6)
Coefficients Qij for the midlayer (core paper) can be
calculated using standard rule of mixture theory (ROM).
Using rule of mixtures theory shear modulus Gxy is obtained
using known relation:
In the previous relation n denotes the total number of layers
(core paper and resin layers), therefore it can be written:
Nx
Ny =
N xy
zk
[Q] { } dz
(7)
G12 =
i =1 zk 1
(8)
Matrix A is stiffness matrix and all coefficients Aij (i=1,2,6
j=1,2,6) can be easily calculated once all the properties of
all the layers are known.
t /2
Aij =
t /2
Qij( k ) dz = Qij( k ) ( zk zk 1 )
G12 =
(9)
k =1
Since the layup in this case can be considered to be
symmetric, it can be concluded that the effective shear
modulus for this structure can be expressed as:
eff
Gxy
=
A66
h
(13)
(V f + 12 Vm ) G f Gm
( GmV f + 22 VmG f )
(14)
Correction coefficients 12 and 22 depend on the fiber type
used in the core paper and are given for typical fibers used
in honeycomb structure production.
22
12
(10)
In equation (10) h denotes total thickness of the laminate
(core paper and resin layers).
Carbon
0.500
0.400
Glass
0.516
0.316
Aramid
0.516
0.400
Once elastic coefficients (equation 14) are computed
effective shear modulus for this type of composite (core
paper with resin layers) is calculated using equation (10).
3.2. Honeycomb core with orthotropic material
During certain manufacturing process, the floc fibers can be
aligned in the direction of the paper coming off the machine.
This leads to orthotropic mechanical properties of the paper.
3.3. Effective inplane and outofplane shear
moduli of hexagonal honeycomb cores
Many authors have developed theoretical approaches for
obtaining the equivalent material properties of honeycomb
cores [3]. The nine core material properties are: two inplane Youngs moduli Ex, Ey, the outofplane Youngs
modulus Ez, inplane shear modulus Gxy, outofplane shear
moduli Gxz, Gyz, and three Poisson ratios nxy, nxz, nyz. Using a
sensitivity analysis, Schwingshackl et al. [3] reported the
major influence of the outofplane shear moduli Gxz, Gyz on
displacement, stress and strain fields of the honeycomb core
type thin plates. One of the analytical approaches mentioned
in this work [3] was developed by Gibson and Ashby [4].
They described the honeycomb core material as a cellular
solid made up of an interconnected network of solid
structures which form the edges and faces of the cells. These
formulae were later slightly modified by Zhang and Ashby
[5] to include double thickness walls for the outofplane
values, Ez, Gxz, and Gyz.
Figure 5. Core paper (orthotropic) with added isotropic
resin layers
For this type of composite equation (3) assumes the
following form:
x Q11 Q12 0 x
y = Q21 Q22 0 y
xy 0
0 Q66 xy
G f Gm
V f Gm + Vm G f
Where Vf and Vm are fiber and matrix volume fractions, and
Gf and Gm are shear modulii of the fiber and matrix phases.
This relation, as experiments have shown, tends to
underestimate the value of G12, and semiexperimental
HalpinTsai theory is used instead. According to this theory
the G12 modulus can be obtained using the following
relation:
Or rewriting the previous relation,
{ N } = [ A] { }
(12)
(11)
The honeycomb cell geometry is presented in the following
figure (Figure 6) where l and h indicate the length of the
hexagon face; b indicates the height; t indicates the
Where Qij are given in the following form:
111
ONTHEEFFECTIVESHEARMODULUSOFCOMPOSITESANDWICHPANEL
thickness of the face; indicates the semiangle between
two faces.
OTEH2016
used. Two FEA models were constructed. First mezoscale
model, where complete structure (composite panel with
honeycomb core) was modeled using plate finite elements
(based on Kirchoffs thin plate theory). In this model each
honeycomb cell (Figure 9) was modeled using plate
elements according to geometry which corresponds to
geometry of HexWeb engineered core, with HRH 10
Nomex, with cell size h=13 [mm]. Dimensions of the
analyzed model were 250 mm x 100 mm x 25 mm. The
plate was clamped at one end, where all translations and
rotations were prevented. Displacement was measured at the
top plate, free end semispan. The software used was
ANSYS.
Figure 6. Hexagonal honeycomb core and cell geometry
The equations (15)  (18) represent the equivalent material
outofplane shear moduli according to [5], and are depicted
in the following figure (Figure 7).
Figure 7. Hexagonal honeycomb core outofplane shear
moduli (Gxz and Gyz)
()
cos t G eff
xy
h / l + sin l
(15)
eff
t Gxy
()
(16)
2
eff
Gxzupper = 1 h / l + 2 sin t Gxy
2 ( h / l + sin ) cos l
()
(17)
eff
Gxz = Gxzupper + 0.787 ( Gxzupper Gxzlower ) Gxy
b/l
(18)
G yz =
Gxzlower =
h / l + sin
(1 + 2h / l ) cos l
Figure 9. Hexagonal honeycomb core Finite element
model
Facings (stress skins, Figure 10) where thin plates, 1 mm
thick made of carbon T300. Composite layup for the
facings (lower and upper) was [0o/45o/45o/90o].
The inplane shear modulus is presented in the following
figure (Figure 8).
Figure 10. Hexagonal honeycomb core with facings and
refined core mesh finite element model
Figure 8. Hexagonal honeycomb core inplane shear
modulus (Gxy)
In the second verification finite element model (model 2),
the core was modeled using solid elements, where material
properties (elastic coefficients in stiffness matrix) were
determined using the equivalent material approach, as
described in section 3.
The equivalent inplane shear modulus for the honeycomb
composite core may be expressed by the following equation
(19).
Gxy = 0.2910
Ef t
a
(19)
Displacement field was calculated for both models and
results were compared. Boundary conditions under which
displacements of the plate tip were determined were the
same in both cases. The plate was clamped at one end (all
degrees of freedom of end nodes for both translation and
rotation are set to zero), whereas the loading moment was
applied at the plates free end. These boundary conditions
4. NUMERICAL MODEL
In order to verify the validity of the proposed model for
shear material properties of composite honeycomb panel
numerical approach, using finite elements approach was
112
OTEH2016
ONTHEEFFECTIVESHEARMODULUSOFCOMPOSITESANDWICHPANEL
applied, enforced honeycomb core and facings twist, where
shear moduli of the complete structure have an important
role and great influence on the complete structure
displacement.
and time is necessary. Many researchers have focused on
finding methodology to find equivalent material models for
different types of plates with honeycomb cores with
different cell geometries, sizes and constituent materials.
The results are shown in the following figure (Figure 11)
In this paper Shear moduli (inplane and outofplane) for
the hexagon honeycomb core made of different types of
composite materials (with aligned and randomly distributed
fibers) were investigated. The equivalent material model
was presented. To verify equivalent material model validity,
the numerical approach was deployed, by creating two
different types of models. The true scale model and
lumped model where the core material of the honeycomb
was defined using equivalent material properties.
Displacement field was calculated for both models and
results compared. It was found that in the linear region both
models yield same results. Since the computing time for the
lumped model is significantly lower its application is
recommended. In nonlinear region the equivalent material
model yields lower results when compared to the truescale
model which is probably due to very complex core crushing
mechanism, and if the results are required in this region the
truescale model has to be developed.
Figure 11. Displacement of plate tip as a function of
applied load
References
5. CONCLUSION
[1] Christensen,R.M., Waals,F.M.: Effective Stiffness of
Randomly Oriented Fiber Composites, Journal of
Composite Materials, Vol. 6, 1972, 518532.
[2] Manera,M.: Elastic properties of randomly oriented
short fiberglass composites, Journal of Composite
Materials, Vol. 11, 1977, p. 235247
[3] Schwingshackl,C.W., Aglietti,G.S., Cunningham,P.R.:
Determination of honeycomb material properties,
Journal of Aerospace Engineering 19 (2006) 177183
[4] Gibson,J., Ashby,F.: Cellular Solids: Structure and
Properties. Cambridge University Press, 2001
[5] Zhang,J., Ashby,M.: The outofplane properties of
honeycombs, International Journal of Mechanical
Sciences 34 (1992) 475489.
In the present paper Shear properties of the honeycomb
plate were investigated. Approach based on the material
equivalent model was proposed for the honeycomb cores
made of different types of composites. In general, due to
complex geometry of hexagonal composite cores, prediction
of the elastic coefficients for these type of structures is
relatively complicated. One of the approaches is to use
numerical approach starting from constituent material
properties and creating honeycomb true scale model in order
to determine stressstrain and displacement fields. As
experiments have confirmed, this approach renders correct
results, however these (true scale) models require quite a
long development time (each honeycomb cell is modeled
independently with large number of elements and hence
large number of degrees of freedom), furthermore, in order
to solve and obtain desired fields large computing power
113
EFFICIENT COMPUTATION METHOD FOR FATIGUE LIFE
ESTIMATION OF AIRCRAFT STRUCTURAL COMPONENTS
STEVAN MAKSIMOVI
Military Technical Institute, Belgrade, Serbia, s.maksimovic@mts.rs
MIRJANA URI
Military Technical Institute, Belgrade, Serbia, mina.djuric.12@gmail.com
ZORAN VASI
Military Technical Institute, Belgrade, aig135@mts.rs
OGNJEN OGNJANOVI
Military Technical Institute, Belgrade, ognjenognjanovic@yahoo.com
Abstract: This work considers efficient computation method for total fatigue life estimation of aircraft structural
components under general load spectra. Total fatigue life of aircraft structure or structural components can be divided
in two parts. The first part represents initial fatigue life up to crack initiation in which low cyclic fatigue properties are
used and second part represents fatigue life during crack growth in which dynamic material properties are used. To
obtain efficient computation method in this work the same low cyclic material properties, for determination of initial
fatigue life and for residual life during crack growth, are used. Based on the strain energy density (SED) theory, a
fatigue crack growth model is developed to predict the lifetime of fatigue crack growth for single or mixed mode cracks.
Presented computation results are shown that crack growth method based on strain energy density approach is in a
good agreement with conventional Forman`s approach. This computation method for total fatigue life prediction can be
used in practical design of metal structural parts of helicopter tail rotor blades as well as in service life extension of the
G4 Super Galeb aircraft structures.
Keywords: aircraft structure, total fatigue life estimation, SED, low cyclic fatigue properties.
1. INTRODUCTION
Many failures of structural components occur due to
cracks initiating from the local stress concentrations.
Attachment lugs are commonly used for aircraft structural
applications as a connection between components of the
structure. In a lugtype joint the lug is connected to a fork
by a single bolt or pin. Generally the structures which
have the difficulty in applying the failsafe design need
the damage tolerance design1,2. Methods for design
against fatigue failure are under constant improvement. In
order to optimize constructions the designer is often
forced to use the properties of the materials as efficiently
as possible. One way to improve the fatigue life
predictions may be to use relations between crack growth
rate and the stress intensity factor range. To determine
residual life of damaged structural components here are
used two crack growth methods: (1) conventional
Forman`s crack growth method and (2) crack growth
model based on the strain energy density method. The last
method uses the low cycle fatigue properties in the crack
growth model1,2. In Figure 1 and 2 are shown
representative structure of helicopter and aircraft in which
presented computation procedures can be used. Figure 1
shows the composite helicopter tail rotor blades with
metal fittings.
Figure 1. The helicopter tail rotor
In Figure 2 is shown part of fuselage structure aircraft
Super Galeb G4.
114
OTEH2016
EFFICIENTCOMPUTATIONMETHODFORFATIGUELIFEESTIMATIONOFAIRCRAFTSTRUCTURALCOMPONENTS
estimation. Fatigue life estimation is defined using
number of cycles, Nf, when starts initial crack in critical
zone of structural element.
The complete results of fatigue life estimation up to crack
initiation of structural element is given in Table 2.
2.2. CRACK GROWTH PHASE
2.2.1 Determination of Fracture Mechanics Parameters
of Lug Structural Components
Here are considered cracked aircraft attachment lugs,
Fig.3. Once a finite element solution has been obtained,
Fig.4, the values of the stress intensity factor can be
extracted from it. For determination Stress Intensity
Factors (SIF) of cracked aircraft attachment lugs here two
approaches are used: (1) Method based on Jintegral
approach and (2) Method based on extrapolation of
displacements around tip of crack.
Figure 2. Structure of Super Galeb
2. FATIGUE LIFE ESTIMATION
Total fatigue life of aircraft structural components can be
divided in two phases:
Crack initiation phase
83.3
Crack growth phase
t = 15 mm
2.1. Crack initiation phase
R20
Relation Morrowa
44.4
(i)
160
For determination initial fatigue life of structural
components can be used various relations. Here are used
two relations: (i) Morow and (ii) SmithWatsonToper
relations. Both these relations are based on using low
cyclic material properties
R41.65
Relation of Morrow curve for low cyclic fatigue has the
next form [4]
F = 6371.6 daN
Figure 3: Geometry of cracked lug
= f m N b + N c
f
f
f
2
E
(1)
The difference this curve and basic curve for low cyclic
fatigue is that it takes into account the effect of mean
stress m, to modify only of the elastic component of the
total amplitude of deformation.
(ii)
SmithWatsonTopper relation
In relation SmithWatsonToppers (SWT) curve of low
cyclic fatigue [5]
SWT = max E =
2
=
( ) ( N f )
'
f
2b
+ E
'
f
'
f
(N f )
b+c
Figure. 4. Finite Element Model of cracked lug with
Von Misses stress distribution
(2)
The pathindependent Jline integral which was proposed
by Rice [11] is defined as
The effect of the mean stresses is included using the next
relation
max = m +
2
J =
(3)
W d x
Ti
ui
d s
x1
(4)
where W is the elastic strain energy density, is any
contour about the crack tip, Ti and ui are the traction and
displacement components along the contour and s is archlength along the contour, x1 and x2 are the local
coordinates such that x1 is along the crack.
Mark SWT in (2) refers to SmithWatsonToppers
parameter. SWT parameter is used to include the effects
of the mean stresses in fatigue life estimations. It means
that SWT parameter includes the effects of the mean
stress m on fatigue life estimation in fatigue life
115
OTEH2016
EFFICIENTCOMPUTATIONMETHODFORFATIGUELIFEESTIMATIONOFAIRCRAFTSTRUCTURALCOMPONENTS
purpose conventional Forman crack growth model and
crack growth model based on strain energy density
method are used. Material of lugs is Aluminum alloy
7075 T7351 with the next material properties: m=432
N/mm2 Tensile strength of material, 02=334 N/mm2,
KIC=2225 [N/mm3/2], Dynamic material properties
(Forman`s constants): C=3* 107, n=2.39; Low cycle
fatigue material properties of material are: f =613 P,
2.2.2. Crack growth models
Conventional Forman`s crack growth model is defined in
the form3
C ( K )
da =
dN (1 R ) K C K
n
(5)
f =0.35, n =0.121.
where: a is crack length, N is number of cycle, KC is the
fracture toughness C, n are experimentally derived
material parameters. The strain energy density method
(SED) can be written as [610]
da = (1 n ) ( K K )2
I
th
dN 4 EI n f f
2.2.3.2. Determination of stress intensity factors
Here are determined stress intensity factors of cracked
lug, defined in Figures 3 and 4 using different
computation methods. The stress intensity factors (SIF`s)
of cracked lugs are determined for nominal stress levels:
g = max=98.1 N/mm2 and min=9.81 N/mm2. The
corresponding forces of lugs are defined as, Fmax= g (w2R) t = 63716 N and Fmin= 6371.16 N, that are loaded of
lugs. For stress analysis of cracked lug in this paper
contact pin/lug finite element model is used. Results of
SIF`s are given in Table 1. Good agreement between FE
and analytic results is evident. Two crack growth models
of cracked lug using relations (5) and (6) are shown in
Fig.5.
(6)
where: f is cyclic yield strength and f  fatigue
ductility coefficient, KI is the range of stress intensity
factor,  constant depending on the strain hardening
exponent n , I n  the nondimensional parameter
depending on n . Kth is the range of threshold stress
intensity factor and is function of stress ratio i.e.
K th = K th 0 (1 R )
(7)
Kth0 is the range of threshold stress intensity factor for
the stress ratio R = 0 and is coefficient (usually,
= 0.71).
2.2.3. Numerical validations
Here are considered two crack growth models: (i) cracked
aircraft lug and (ii) Cracked plate with circular hole and
initial crack.
2.2.3.1. Crack growth model of aircraft lug
Figure 5. Comparisons crack growth behavior using SED
and Forman`s methods
Subject of this analyses are cracked aircraft lugs under
cyclic load of constant amplitude and spectra. For that
Table 1. SIFs of cracked lugs using Jintegral approach and Displacement Extrapolations
Path of integral on
locations w
Crack Length
Jintegral approach
Method of Displacements Extrapolation
= 5mm
Jintegral
KI
KI
KII
KIII
2D
0.581
65.582
69.592
5.7803
0.000
Path 1 w=0.0 mm
3D
0.609
67.136
60.439
10.003
5.967
Path 2 w=4.75 mm
3D
0.637
68.669
66.090
8.221
2.099
Path 3 w=7.50 mm
3D
0.640
68.809
66.180
8.234
0.000
Analytic solution6: I = 65.621
Results of FEM6: I = 68.784
116
EFFICIENTCOMPUTATIONMETHODFORFATIGUELIFEESTIMATIONOFAIRCRAFTSTRUCTURALCOMPONENTS
OTEH2016
Table 2. Fatigue life estimation of structural components using inhouse software up to crack initiation (plate with
central hole under load spectrum)
2.2.3.3 Crack growth model of plate with hole and initial
crack under load spectrum
complete results of residual life estimation of cracked
plate with circular hole and one crack under load
spectrum are given in Table 3.
The second example of crack growth model is plate with
circular hole and one crack subjected load spectrum. The
Table 3. Fatigue life estimation of structural components using inhouse software during crack growth (plate with
central hole and initial crack under load spectrum)
117
EFFICIENTCOMPUTATIONMETHODFORFATIGUELIFEESTIMATIONOFAIRCRAFTSTRUCTURALCOMPONENTS
OTEH2016
[2] Sehitoglu,H., Gall,K., Garcia,A.M.: (1996), Recent
Advances in Fatigue Crack Growth Modelling, Int. J.
Fract., 80, pp.165192.
[3] Forman,R.G., Kearney,V.E., Engle,R.M.: Numerical
analysis of crack propagation in cyclic loaded
structures, J. Bas. Engng. Trans. ASME 89, 459,
1967.
[4] Morrow,J.D.: Fatigue properties of metals. In:
Fatigue design handbook Sec. 3.2. SAE advances in
engineering, Warrendale, PA; 1968. p.219.
[5] Smith,KN, Watson,P, Topper,TH.: A stressstrain
functions for the fatigue of metals. J Mater
1970;5:76778.
[6] Maksimovic,K.: Strength and Residual Life
Estimation of Structural Components Under General
Load Spectrum, Doctoral Thesis, Faculty of
Mechanical Engineering, 2009 (in Serbian).
[7] Maksimovi,S.,
Vasovi,I.,
Maksimovi,M.,
uri,M.: Residual life estimation of damaged
structural components using lowcycle fatigue
properties, Third Serbian Congress Theoretical and
Applied Mechanics, Vlasina Lake, 58 July 2011, pp.
605617, Organized: Serbian Society of Mechanics,
ISBN
9788690997336,
COBISS:SRID
187662860, 531/534(082).
[8] Maksimovi,S.,
Kozi,M.,
SteticKozi,S.,
Maksimovic,K.,
Vasovic,I.,
Maksimovic,M.:
Determination of Load Distributions on Main
Helicopter Rotor Blades and Strength Analysis of the
Structural Components, Journal of Aerospace
Engineering, Vol.27, Number 6, November /
December 2014.
[9] Boljanovi,S., Maksimovi,S.: Fatigue crack growth
modeling of attachment lugs, International Journal of
Fatigue, International Journal of Fatigue 58 (2014),
6674.
[10] Maksimovic,S., Doi,R., Vasi,Z.: Application NDI
Methods in Process of Extension of Operation Life of
Aircraft Structures, OTEH 2014.
[11] Rice,J.R.: Elastic fracture mechanics concepts for
interfacial
cracks.
Journal
of
Applied
Mechanics1988; 55:98103.
The complete computation results of plate with central
hole made from dural 2024 T4 are given in tables 2 and 3.
Computation results are in good agreement with
experiments. Total number of blocks up to crack initiation
is 38.4 (Table 2) and experimental 45. Total number of
blocks during crack growth is 5 (Table 3) and
experimental is 5. Here is evident good agreement
computation with experimental results.
3. CONCLUSIONS
This investigation is focused on developing efficient and
reliable computation methods for total fatigue life
estimation. In this investigation are presented two
computation methods for the fatigue life estimation: (i)
Fatigue life estimation up to crack initiation using low
cyclic material properties and (ii) Fatigue life estimation
of structural elements during crack growth using low
cyclic material properties too. Both computation methods
are based on using the low cyclic material properties.
Computation results are compared with own experimental
results. Good agreement between computation with
experiments is obtained in domains crack initiation and
crack growth phases.
A special attention has been focused on determination of
fracture mechanics parameters of structural components
such as stress intensity factors of aircraft cracked lugs.
Computation prediction investigations for fatigue life
estimation of an attachment lug under load spectrum were
performed. In this investigation the SED crack growth
model based on low cycle fatigue material properties is
included. Comparisons of the predicted crack growth rate
using strain energy density (SED) method with
conventional Forman`s model points out the fact that this
model could be effective used for residual life
estimations.
References
[1] Maksimovic,S.,
Posavljak,S.,
Maksimovic,K.,
NikolicStanojevic,V., Djurkovic,V.: Total Fatigue
Life Estimation of Notched Structural Components
Using LowCycle Fatigue Properties, J. Strain
(2011),
47
(suppl.2),
pp.341349,
DOI:
10.1111/j.14751305.2010.00775.x
118
ANALYSIS OF AIRCRAFT STRUCTURES CROSS SECTION
BOGDAN S. BOGDANOVI
UTVA AI, Panevo, bogdanovic00@gmail.com
DARIO A. SINOBAD
UTVA AI, Panevo, 318dsinobad@gmail.com
TONKO A. MIHOVILOVI
UTVA AI, Panevo, tonkojetonko@gmail.com
Abstract: Definition of normal and shear stresses produced by bending and twist in closed semi monocoque
structures, as structures mostly used in aircraft structures is considered in this paper. Calculation is realized using well
known theoretical background such as fundamental beam theory, shear flow definition in statically indeterminate
structure (single redundancy). Buckling behavior of thin sheet is also taken into account. As results effective cross
section is defined and normal stresses in longitudinal parts, stringers, and shear stresses in skin, thin sheets, between
stringers, are obtained. Complete procedure for the whole calculation is formed, and some results are shown.
Keywords: effective width, normal stresses, shear stresses, shear center.
1. INTRODUCTION
2. STRESSES DEFINITION FOR STRINGER
SKIN CONSTRUCTION
General load acting on aircraft as a whole comprises air
loads, inertia loads, power plant loads, landing gear loads
etc. All kinds of load are distributed on parts of aircraft
structure as:
forces (axial, transverse), and
The cross section of some main part of structure such as
wing, fuselage or tail units will be considered, for
obtaining the whole procedure. These cross sections are
composed of:
moments (bending, torsion).
stringers (longerons), and
skin.
Every part of structures is exposed to some component
part of general loads acting on aircraft, and they all
together should keep the whole aircraft in equilibrium.
Mostly used shape of parts in primary, and also in
secondary, structures of modern aircrafts is closed semi
monocoque type, which could be described as reinforced
shell, or tube. Basic elements of this type are:
The procedure, itself, is separated in two parts; it is
definition of:
normal stresses, and
shear stresses.
stringers or longerons, beams
2.1. Necessary data
frames or bulkheads, ribs, and
Before the procedure starts some set of necessary data
must be known. This set comprises:
skin (sheet).
defined components of loads acting on element cross
section
Stringer (longeron)  skin construction makes outer shape
of structural part, but it must be supported by frames (in
fuselage), or ribs (in wing). All of these elements must be
well designed to take and transfer part of load that acts on
them; it means that stresses in elements, produced by
loads, should be below the allowable defined for
elements' shape and material. A stresses definition
procedure for such constructions and obtained results will
be shown in this paper. The procedure will cover only the
linear range, but it is very convenient to be arranged for
nonlinear range of material behavior.
defined all geometrically data
Components of loads acting on structural element cross
section are (see Picture 1):
bending moments about any pair of mutually
orthogonal axes of airplane coordinate system, lying in
the cross section plane My, Mz,
twisting moment about third axis of airplane coordinate
system normal to the cross section Mx, and
119
OTEH2016
ANALYSISOFAIRCRAFTSTRUCTURESCROSSSECTION
transverse forces along direction of the same mutually
orthogonal axes, lying in the cross section plane Fy, Fz
In this discontinuous structure all cross section area is
concentrated on stringer positions, without any material
between them. Every approximation consists of several
steps:
effective skin, sheet, width and area definition,
calculation of total area on stringer position,
definition of new C.G. position and corresponding
stringer coordinate referring to new C.G. position
inertia calculation, corresponding to new C.G. position,
normal stress calculation,
final effective width calculation
a. Effective width and area definition
Effective skin area is defined as parts of skin situated fore
and aft of every stringer, and it is obtained by multiplying
effective width with corresponding skin, sheet, thickness.
It is supposed that normal stress value in whole effective
skin area and the stringer is the same. First of all
effective width will be defined. There are two ways how
effective width can be defined, depending on stringer
normal stress type:
Picture 1. Cross section simplified loads and geometry
normal stress is tension, and
Geometrical data are:
normal stress is compression.
coordinate of stringer positions usually referring to the
airplane coordinate system axes situated in cross section
plane (Y,Z),
a1. Normal stress is tension
stringer cross section area
In this case whole skin, length between two successive
stringers is effective. The effective width which belongs
to any of them is one half of total length; it means:
skin length between two stringers (b),
skin depth thickness, (t), and
stringer efficiency coefficient (k)
weffi = weffi +1 =
2.2. Normal stresses calculation procedure
bi ,i +1
2
(1)
It should be noted that the same expression is used in
defining effective width on the beginning of the first
approximation for every stringer.
Procedure for normal stress definition is iterative; it
means that there are several approximations of same type.
Minimum number of approximations is two. Each of them
starts, begins, using corresponding geometrical and load
data, and as final results:
a2. Normal stress is compression
In this case effective width length is calculated using
expression:
new C.G. position
inertia values about mutually orthogonal axes
skin, sheet, effective width, and
weff ,i = C t
stringer normal stresses
are obtained. Then, these results are compared with the
same kind results of previous approximation. If the
difference is bigger than some given value the next step is
necessary, if not the procedure is stopped. Results that are
compared are:
st ,i
(2)
This is effective width on fore or aft side of stringer. Here
are:
C.G. position, and
C[ ]
coefficient, usually its value is 0.85
t [cm]
skin thickness
E [bar] Youngs modulus, and
maximum stringer normal stress value
[bar] normal stress value in stringer
2.2.1. Approximation details
b. Calculation of total area on stringer position
First of all it should be noted that cross section structure
made of stringer and continuous skin area is, for
calculation purpose, substituted by a discontinuous one.
Total area on stringer position is composed of two parts:
As [cm2] stringer area, and
120
OTEH2016
ANALYSISOFAIRCRAFTSTRUCTURESCROSSSECTION
Aw [cm2] effective width area
e. Normal stress calculation
It means:
At = AS + AW
The fundamental beam theory is used as a basis for
normal stress calculation. In application this theory is
little bit modified using so called K method as very
convenient for practical use. This method enables normal
stress calculation using data referred to any pair of
mutually orthogonal axes; data already defined. Used
expression is:
(3.1)
In this calculation the stringer coefficient of efficiency
(K) should be taken into account. This coefficient defines
part of total stringer area which is included in bending
moments transferring. Near cutouts in structure this
coefficient value is zero, on certain distance far from
opening, is one. Finally, expression for calculation of total
area on stringer position is:
At = K ( AS + AW )
b = ( K3 M z K1 M y ) yi
( K 2 M y K1 M z ) zi
(3.2)
Coefficient values K1, K2, K3 are obtained by expressions:
K1 =
c. New C.G. position
c1 C.G. position
K2 =
Coordinates of cross section C.G. position, on the
beginning, referring to the airplane coordinate system, are
defined using well known expressions:
YC .G. =
Ai yi
, Z C .G. =
Ai zi
K3 =
I yz
2
I y I z I yz
Iz
,
2
I y I z I yz
(8)
Iy
2
I y I z I yz
(4)
f. Final effective width calculation
At the end of each approximation final effective width is
calculated, using expression (2). Obtained final effective
width values are compared with the corresponding width
values from the previous approximation, and the smaller
one is used in further calculation as input data for next
approximation; that is:
Here are:
yi [cm], zi [cm] stringer coordinate along airplane
coordinate system axes
Calculation of new C.G. position, using (4) in every
iteration gives the difference between C.G. coordinates of
two successive approximations as a results.
weff , i +1 < weff , i weff , i +1
c2. Corresponding stringer coordinate referring to new
C.G. position
weff , i +1 > weff , i weff , i
Having defined new C.G. position, coordinate system is
just translated from previous C.G. position to new one.
Stringer coordinate in new coordinate system are obtained
as:
yi = yi 1 YC .G. , zi = zi 1 Z C .G.
(7)
g. Normal stress final values
These values are obtained from the last approximation,
when the calculation is stopped.
(5)
2.3. Shear stresses calculation procedure
d. Inertia calculation
There are two steps in shear stresses calculation
procedure:
Inertia values are calculated for pair of mutually
orthogonal axes passing through cross section centroid
using well known expressions set:
I y [cm ] =
( z Z
I z [cm 4 ] =
( y Y
I yz [cm 4 ] =
( y Y
i
C .G .
shear flow values definition, and
dividing shear flow by corresponding thickness and
obtaining shear stress.
) Ai
2
C .G .
C .G .
)2 Ai ,
Total shear flow acting on cross section skin is generally
produced by two load components:
(6)
part of flow produced by transverse force acting along
axes of airplane coordinate situated in cross section
plane (Y,Z axes), and
) ( zi Z C .G. ) Ai
part of flow produced by twisting moment (pure
torsion) acting about some cross section point, usually
origin of coordinate system for the first approximation
121
OTEH2016
ANALYSISOFAIRCRAFTSTRUCTURESCROSSSECTION
2.3.1. Shear flow produced by cross section transverse
forces
y A
F K F ) z A
qx = ( K 3 Fy K1 Fz )
( K2
(11)
Considered cross section is closed contour shape. It
means that shear flow calculation, is statically
indeterminate problem to the first degree. The problem
will be solved assuming that all transverse force pass
through cross section shear center. It means that:
Fy [daN], Fz [daN] force values along airplane
coordinate system axes
there is not any cross section angular deformation
(twist), and
b. Static indeterminate shear flow definition
Here are:
sum of moments, produced by all cross section shear
flows, for any cross section point must be zero
Static indeterminate shear flow has constant value around
whole cross section circumference. In this procedure, the
calculation of static indeterminate shear flow value is
realized using condition (9). In expanded form angular
deformation (twist) could be written:
The expression which describes first condition is:
=0
(9)
Here is:
1
2 AG
q ds
t
(12.1)
1
2 AG
q t b
(12.2)
rad angular twist of cross section
cm
Second condition is described using expression:
=0
(10.1)
Here are:
q [daN/cm]
shear flow value
A [cm2]
cross section area
G [bar]
material elastic shear modulus
ds [cm]
infinitesimal part of cross section
circumference
Mq [daNcm] total sum of moments produced by
cross section shear flows
b [cm]
Mqi [daNcm] sum of moments produced by cross
section indeterminate shear flows
part of skin, sheet, width between
two successive stringers
t [cm]
corresponding thickness of skin part
or, in expanded form:
M qi
M qd
=0
(10.2)
Here are:
Using (9) expression (12.2) became:
Mqd [daNcm] sum of moments produced by cross
section determinate shear flows
1
2 AG
Usually, force components dont pass through cross
section shear center, so in that case the condition (10.1)
will be not satisfied and some twisting moment could
exist. If it happens, shear flow produced by this moment
should be added to shear flows produced by force
components, for obtaining final shear flow values.
Problem solution consists of several parts (steps):
q t b = 0
Sum on the left side could be spliced in two:
1
2 AG
q t b + 2 A1 G q t b = 0
d
(13)
Here are:
static determinate shear flow definition,
qd [daN/cm] values of static determinate shear flow
static indeterminate shear flow definition,
total shear flow calculation,
qi [daN/cm] values of static indeterminate shear flow
calculation of moment produced by total shear force
components flow, and
By rearranging (13) it will be obtained:
shear center position definition
a. Static determinate shear flow definition
qd b
+ qi
t
bt = 0
* note: qi is constant value, so it can be put out of
summation sign.
For calculation purpose closed cross section contour is
cut on chosen place and transferred into determinate
structure. Shear flow in such structure will be calculated
using again K method. Following equation will be
used:
Finally, expression for indeterminate shear flow value
calculation is:
122
OTEH2016
ANALYSISOFAIRCRAFTSTRUCTURESCROSSSECTION
q b
t
=
b
t
(14)
+ Fy zsc = 0 ,
zsc =
Total shear flow values on any skin part are obtained
simply by adding corresponding values qi and qd:
d
q = q +q
O
qFz
+ Fz ysc = 0
Shear center coordinates are:
c. Total shear flow calculation
O
qFy
M
Fy
O
qFz
Fz
(19)
Expression for shear flow value calculation produced by
cross section torsion is:
d. Calculation of moment produced by total shear flow
from force components
qMt =
Moment produced by total shear flow force components
for chosen point (O) is defined by expression:
O
q
, ysc =
2.3.2. Shear flow produced by cross section torsion
(15)
M = q A
O
qFy
(20)
2 A
Here is:
M [daNcm] = M + M
(16)
total cross section
twisting moment  torsion
Here are:
A [cm2] total cross section area
A [cm ] part of cross section area, corresponding to
the shear flow value between two successive
stringers
3. NUMERICAL EXAMLE
As an example, normal and shear stresses in airplane
fuselage cross section are calculated. Obtained results are
shown in Tables 1, 2 and 3. Procedure is realized using
EXCEL software.
If:
O
q
(17)
Table 1. Stringer normal stress values
obtained value should be added to the already known
torsion moment value (input data Mx).
Str.
e. Shear center position
[bar]
CL
Total shear flow values produced by transverse force
components are obtained under assumption that these
force components pass through shear center. Usually, it
isnt a case and force components are situated to some
other point (O) and because of this result (17) is obtained.
Mostly, this point (O) is origin of airplane coordinate
system axes in the cross section plane.
86.61
76.50
25.97
4.09
67.51
87.73
Value:
87.73
67.51
4.09
25.97
10
76.50
86.61
O
q
could be canceled only by transferring force components
from point (O) to shear center. In this case should be
written:
M = M
O
q
O
qFy
O
qFz
CL
(18)
Table 2. Shear flow sums and shear stress values in skin
sections between stringers
Str
q
[daNcm]
[bar]
Here are:
MqFyO, MqFzO [daN*cm] total shear flow values
corresponding to the force components Fy, Fz
Both sums on the right side of (18) must be canceled
separately, and it gives:
CL0
01
12
123
2.7E15
8.912871
16.0749
3.4E14
178.2574
321.4979
OTEH2016
ANALYSISOFAIRCRAFTSTRUCTURESCROSSSECTION
23
34
45
56
67
78
89
910
100
0CL
Control
18.02914
18.66437
11.94374
2.7E15
11.9437
18.6644
18.0291
16.0749
8.91287
2.7E15
1.3E12
Table 3. Shear center position
360.5828
373.2874
238.8749
5.5E14
238.875
373.287
360.583
321.498
178.257
3.4E14
Ye.c.
[cm]
1.15576E15
Z e.c.
[cm]
0
4. CONCLUSION
The procedure presented in this paper is a very
convenient, easy and quick way for obtaining a lot of
reliable results. These results are very useful as in global,
so in detail, stress analysis of airplane structure parts.
References
[1] Bruhn,E.F.: Analysis and design of flight vehicle
structures, Tri state offset company, Ohio 1973
124
STRESS CALCULATION OF NOSE GEAR SUPPORT WITH ASPECT OF
WELDING OF AEROSPACE STEEL 15CRMOV6
ALEKSANDAR PETROVI
Utva avio industrija d.o.o. pealeks@gmail.com
BOGDAN S. BOGDANOVI
Utva avio industrija D.O.O. bogdanovic00@gmail.com
ALEKSANDAR STANAEV
Utva avio industrija D.O.O. aleksandarstanacev@outlook.com
Abstract: Nose gear support stress calculation of airplane SOVA and welding specification definition of welded
connections are considered in this paper. During design of new plane SOVA aroused need for modification of Utva75 nose gear support. Lack of information regarding welding 15CrMoV6 steel made us create welding procedure
specification and to check them experimentally.
Calculation of stress is realized with FEM analysis. Results are used for welded connections calculation and for
creating proper welding procedure specification.
Welding analysis is done experimentally and numerically. Experiment is done on standard specimens. Specimens are
made from aerospace steel 15CrMoV6 (1.7734). Result are compared, and shown in paper.
Keywords: Nose gear support, Welding, TIG, 15CrMoV6, FEM.
Filler materials for the 1.7734 steel are 8CrMo12
(8CD12) or steel 15CDV6 (1.7734.2). Procedure when
filler material has degrading characteristics than base
material is called undermatching. Additional information
for welding these types of steels can be found in [1].
1. INTRODUCTION
The design and certification of aircraft SOVA, derived
from the basic aircraft U  75 have made it necessary to
define the stress state in the nose gear support (NGS).
While doing so, we take into account the data related to
the exploitation of aircraft U  75, such as the occurrence
of cracks on the NGS, and implemented, modification.
Table 1. Chemical components of 1.7734 steel
Element
Min %
Max %
C
0,12
0,18
Si
0,2
Mn
0,8
1,1
S
0,015
P
0,02
Cr
1,25
1,5
Mo
0,8
1
V
0,2
0,3
Production of this assembly is done by welding and due
to the specific production it is necessary to detail the
technological process and determine the appropriate
welding parameters. For this purpose it was carried out:
Stress analysis of NGS;
Experiment, identifying the best possible welding
parameters and materials.
Stress analysis is done using finite element analysis
(FEM).
In order to optimize production, testing of welded
specimen of steel 1.7734 in condition 1.7734.4 and
1.7734.5 were carried out. Mechanical characteristics of
steel in specified conditions were given in Table 2, further
information can be found in [4].
During the experiment data that was varied are:
State of material
Post weld heat treatment (PWHT) of welded joints
Table 2. Mechanical properties of steel 1.7734.4 and
1.7734.5
Properties
1.7734.4
1.7734.5
0.2% Proof Stress [N/mm2]
550
790
Tensile Strength [N/mm2]
700
9801180
Elongation [%]
12
10
Hardness
[205 HV] 29 [HRC]
The steel used for NGS production is aviation steel
15CrMoV6 (1.7734). This steel is low alloy, low carbon
steel with excellent tensile strength, hardness and good
weldability. The basic purpose of this steel is
manufacturing of pressure vessels, rocket motor, and the
missile with solid fuel, auto industry, railway industry and
aviation. Chemical characteristics of this steel are given in
Table 1.
125
CALCULATIONOFNOSEGEARSUPPORTWITHASPECTOFWELDINGAEROSPACESTEEL15CrMoV6
OTEH2016
In pictures 4 to 6, stress values on welded joints between
planed reinforcement and existing (basic) NGS structure
are shown.
2. STRESS ANALYSIS
Calculation of NGS is done in accordance with the
requirements of [5] and the corresponding input values
are shown in the Table 3.
Table 3. The limit force components
For aft loads
For forward
For side loads
loads
Z+ comp.
5051 N
5051 N
5051 N
X+ comp.
4041 N
X comp.
2021 N
Y comp.
3536 N
Stress Analysis was carried out by FEM. To eliminate the
error induced by density of mesh, several mesh sizes were
employed, until the convergence of stresses and
displacements on characteristic points was achieved [2].
Picture3. Stress state in zone of previous crack
occurrence of cracks for side loads
In Pictures 1 to 3 obtained results for defined load in the
zone of previous crack occurrence in the UTVA75
plane for several loads condition are shown.
Picture4. Stress state in zone of welded joints for aft
loads
Picture1. Stress state in zone of previous crack
occurrence of cracks for aft loads
Picture2. Stress state in zone of previous crack
occurrence of cracks for forward loads
Picture5. Stress state in zone of welded joints for forward
loads
It was observed that the maximum VonMisses stress
value, which generally does not exceed the ultimate
values for a given material, occur in zones where cracking
occurred on unmodified NGS. It can be concluded that the
additional reinforcement should completely eliminate the
occurrence of zones with high stresses
Results show that maximum stress is induced at welded
joint for side loads. Normal stress values in that zone will
be used when calculating welded joints.
126
OTEH2016
CALCULATIONOFNOSEGEARSUPPORTWITHASPECTOFWELDINGAEROSPACESTEEL15CrMoV6
Welding process 141 (TIG) defines procedure of joining
metals by melting basic and filler metal using heat. Heat
is achieved with electric arc between Tungsten electrode
and base metal. Filler metal in wire form is immersed in a
liquid bath and due to the high heat melts and mixes with
the base material. Inert gas used for welding procedure is
Argon. Welded plates after PWHT are shown in Picture 8.
Prescribed parameters from
specification are given in Table 5.
welding
Table 5. Welding parameters
Filler material
Filler diameter [mm]
Current [A]
Voltage [V]
AC/DC current
Welding speed [cm/min]
Gas flow [l/min]
Picture6. Stress state in zone of welded joints for side
loads
3. EXPERIMENT DETAILS
procedure
1.7734.2
1,6
5060
12
DC3
9,5
The experiment takes into account the effect of PWHT on
the state of the weld joints and the base material. The butt
weld joint using 1.2 mm thick sheet, which is most often
used in this assembly, is examined. Standard specimens
were used, in accordance with [3]. Specimens are
obtained from the welded plates. Groove weld is designed
with a gap of 1mm and welding sequence was done with
one pass.
Following tests were carried out:
Tensile tests
Hardness tests
Overview of specimen used for obtaining mechanical
properties of welded joints is given in Table 4.
Table 4. Specimens and configuration for testing
MATERIAL
PWHT
1.7734.4
1.7734.4
1.7734.5
1.7734.5
600C/1.5h
580C/4h
600C/1.5h
580C/4h
Picture8. Welded plates
No. of
No of
specimens for specimens for
tensile tests hardness tests
3
2
3
2
5
2
5
2
Specimens are made of welded plates, cut and milled to
dimensions in accordance with the [3]. Picture 9 shows a
form of specimens made for tensile test. On picture 10
dimensions of specimens are shown.
Plates that are used for specimen production are shown in
Picture 7.
Picture9. Specimens for tensile testing
Picture7. Groove before welding
127
CALCULATIONOFNOSEGEARSUPPORTWITHASPECTOFWELDINGAEROSPACESTEEL15CrMoV6
OTEH2016
Picture11. Hardness for specimens of steel 1.7734.4
Changing the parameters of PWHT from 580C/4h to
600C/1h results in lower measured hardness values:
4,3% for weld metal
Picture10. Specimen dimension
0,6% for HAZ
4. RESULTS
Data measured for steel 1.7734.5 steel specimens yields
2.8% hardness drop in the weld seam and 10% hardness
drop in the HAZ for the same PWHT parameter change.
Hardness testing is done by Rockwell method.
Measurements are performed in zones of base metal, weld
seam and heat affected zone (HAZ). Data of hardness
testing for butt joint are given in Table 6.
Specimens
number
Base metal
Weld seam
HAZ
1.7734.5
580C/4h
1.7734.5
600C/1,5h
1.7734.4
580C/4h
1.7734.4
600C/1,5h
Specimen
configuration
Table 6. Values of hardness testing
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
28
28
28
28
29
28
28
29
30
30
30
29
29
28
28
32
33
32
32
33
41
41
41
40
36
36
44
43
41
44
35
35,5
35,5
37
36
35
38
38,5
35
38
34
37
37
35
33
40
34
34
36
33
28
28
29
28
29
29
32
32
31,5
33
Picture12. Hardness for specimens of steel 1.7734.5
Tensile testing of 1.7734.4 steel specimens was done on
the machine SHIMADZU at the Military Technical
Institute and in UtvaAI for steel 1.7734.5 specimens.
Test is shown in picture 13, and specimens after testing
are shown in picture 14 and 15.
Data from Table 5 is presented in picture 11 for WH
1.7734.4 steel and in picture 12 for WH 1.7734.5 steel.
From Picture 11 it can be concluded that the hardness of
the steel 1.7734.4 is highest in the zone of weld seam, as
expected.
Picture13. Tensile testing machine SHIMADZU
128
CALCULATIONOFNOSEGEARSUPPORTWITHASPECTOFWELDINGAEROSPACESTEEL15CrMoV6
OTEH2016
Picture14. Specimens of steel 17734.4
Picture15. Specimens of steel 17734.5
The results of the tensile tests are given in table 7 for steel
1.7734.4 and in table 8 for steel 1.7734.5. The test was
performed on three specimens for steel 1.7734.4, and on
five specimens for steel 1.7734.5. Tensile strength is
defined by the expression:
Data from table 7 in the form of graphs are presented in
picture 16 and picture 17. Data from table 8 are presented
in picture 18. From graphs a comparative overview of
observed values for different PWHT can be seen.
Changes in PWHT from 580C/4h to 600C/1h of
specimens made of steel 1.7734.4 shows increase of
tensile strength by 1.25% on the place of brake and
increase of tensile strength by 12.5 % on the weld seam.
F
RM = M
a b
Where:
RM [N] maximal tensile force
a [mm] specimen thickness
b [mm] specimen width
Table 7. Values of tensile testing for steel 1.7734.4
Material /
PWHT
1.7734.4
600C/1,5h
1.7734.4
580C/4h
Specimens
number
Maximum force
[KN)
1
2
3
1
2
3
13,2551
13,6046
13,2761
12,336
12,6201
12,3069
Place of brake
Tensile
Yield
strength
strength
[MPa]
[MPa]
804
864
826
890
823
888
804
781
826
804
823
846
Weld seam
Tensile
Yield
strength
strength
[MPa]
[MPa]
515
479
644
598
641
595
516
476
547
502
511
489
Table 8. Values of tensile testing for steel 1.7734.5
Material /
PWHT
1.7734.5
600C/1,5h
1.7734.5
580C/4h
Specimens
number
Maximum force
[KN]
1
2
3
4
5
1
2
3
4
5
15
14,6
15,3
14,8
15,2
15,6
15,3
15,4
14,9
14,9
Place of brake
Tensile
Yield
strength
strength
[MPa]
[MPa]
979
972
997
978
992
984
974
971
964
942
129
Weld seam
Tensile
Yield
strength
strength
[MPa]
[MPa]
564
547
671
544
652
614
591
655
514
568
CALCULATIONOFNOSEGEARSUPPORTWITHASPECTOFWELDINGAEROSPACESTEEL15CrMoV6
OTEH2016
5. CONCLUSION
The paper presents the calculation of stress in the
structure and welded joints of NGS, with load condition
defined in [5]. Stress calculation was made by FEM.
Results of the analysis of NGS show the need for further
modification in order to reduce stress in critical areas of
the assembly. It was concluded that additional stiffening
can prevent the occurrence of a local high stress
concentration.
Another aspect of paper relates to the testing of welded
joints on standard specimens in order to determine the
best welding procedure in terms of strength of materials
and production optimization.
.
Picture16. Tensile strength for specimens of steel
17734.4
Standard specimens of steel 1.7734 were tested. During
testing changing of parameters of PWHT were conducted
in order to verify mechanical characteristics of welded
joints and to check possibility for production
optimization.
Parameter change of PWHT for specimens made of steel
1.7734.5 gives increase of tensile strength by 1,7% on the
place of brake and increase of tensile strength by 1.2 % on
the weld .
The measured yield strength of specimens made of steel
1.7734.4 presented in picture 17 shows that the change in
the PWHT from 580C/4h to 600C/1h increases yield
strength by 0.3% on place of break, and by 4.3% on weld
seam.
It is shown that changing the parameters of PHWT from
580C/4h to 600C/1h should not weaken welded joints
on standard specimens. From point of production
optimization it is shown that NGS could be manufactured
of steel 1.7734.5 with parameters of PWHT defined as
600C/1,5h, which would cheapen the production and
shorten the time of manufacture.
Presented results are consistent but because of small
number of specimen the results should be considered as
preliminary.
Further test should be done using a sufficient number of
specimens with possibility to vary several other
parameters in order to properly define welding procedures
of welded joints necessary for production of NGS.
References
[1] GOA Institute, Literature for IWE curs
Moduo2, 2014
[2] Singriesu S. R., "The Finite Element Method In
Engineering Fourth Edition", Elsevier Butterworth
Heinemann, 0750678283.
[3] Standard ASTM E8.
[4] WerkstoffLeistungsblatt, Schweibraer, ChromMolybdanVanadiunVergutungsstahl,
Blache,
Platten und Bander1.7734, March 1982
[5] European Aviation Safety Agency, Certification
Specifications for Normal, Utility, Aerobatic, and
Commuter Category Aeroplanes CS23
Picture17. Yield strength for specimens of steel 17734.4
Picture18. Tensile strength for specimens of steel
17734.5
130
INFLUENCE OF PILOTS AVERAGE BODY MASS INCREASING ON
BALANCE OF LIGHT PISTON TRAINING AIRCRAFT
ZORICA SARI
Military Technical Institute, Belgrade, vti@vti.vs.rs
ZORAN VASI
Military Technical Institute, Belgrade, vti@vti.vs.rs
VOJISLAV DEVI
Military Technical Institute, Belgrade, vti@vti.vs.rs
BORIS GLAVA
Military Academy, Belgrade, risbo@vektor.net
Abstract: The procedure of calculation and experimental determination of aircraft mass and balance, design limitation
defined by international civil and military airworthiness regulative and requirements, statistical analysis of average pilot
body gain weight (mass) during the last sixty years, and their influence on balance of light piston trainer aircraft were
shown in the paper. The practice used for aircraft mass and balance calculation during the phase of aircraft conceptual
design, and experimental determination of aircraft mass and balance after production of prototype and series aircraft were
explained in first part of the paper. The short overview of related definitions and design limitations valid in international
civil and military requirements was shown. The statistical analysis of pilots body gain weight during the last several
decades was shown in second part of the paper. The importance of recording and updating of exact average pilots body
mass data assigned for piloting the specific aircraft and its conformance to current airworthiness and requirements were
explained. The real pilots mass data as well as mass and balance data of produced aircraft were used.
Keywords: aircraft, mass, balance, design, pilot, body dimensions.
years the airworthiness requirements in the world have been
changing, amending, and developing, all of these in order to
improve the flight safety.
1. INTRODUCTION
There are many factors that lead to efficient and safe
operation of aircraft. Mass (hereinafter: weight) and balance
is one of the most important factors affecting safety of
flight. An overweight aircraft, or one whose center of
gravity (CG) is outside the allowable limits, is inefficient
and dangerous to fly. The responsibility for proper weight
and balance control begins with the engineers and designers
and extends to the pilot who operates and technicians who
maintains the aircraft [1]. There are several equally
important elements related to weight and balance: the proper
calculation during aircraft design, the weighing of the
produced aircraft ready to fly, the maintaining of the weight
and balance records, and the proper loading of the aircraft
before and during the flight.
Design, production, maintenance and usage during the
service of an aircraft are complex activities and require
completely dedication of all involved in these activities.
Any inaccuracy in any one of these activities nullfies the
purpose of the whole effort.
There have not been significant and frequent changes in
airworthiness requirements in weight and balance sections
concerning the values for standard pilot and flight crew
weight. In the meanwhile, the requirements related to pilot
set (helmet, parachute, suit, etc.) have been changing
continuously becoming overall heavier, more complete with
the aim of becoming more safe. Moreover, historically the
average (mean) body weight is becoming greater, so the
human body weight increased by about 33% and the average
height increased by 10.11 cm during last fifty years [9].
During this period, the differences in weight between values
of standard pilot weight in airworthiness requirements and
actual pilots weight have become significant. For that
reason the designers have to take into account exact data of
current pilots weights or use different design margin of
safety in order to provide safe and quality flight of aircraft.
Keeping in mind that aircraft service life is always more
than 25 years, the designer must take into consideration the
future pilot generations that will fly the aircraft.
This paper is a part of research and development activities
related to design of light piston trainer aircraft for basic and
primary training that have been occurring in Military
Technical Institute for last sixty years. During all of these
Recent reports from the USA Aviation Safety Reporting
System (ASRS) provides some insight on the safety
implications of continued reliance on outdated human
strength/control force data [18]. Recent reports show that
131
INFLUENCEOFPILOTSAVERAGEBODYMASSCHANGINGONBALANCEOFLIGHTPISTONTRAINERAIRCRAFT
using of obsolete data of pilots anthropometric measures
during design of new aircraft could provoque adverse
consequences to pilot and aircraft safety. Hence, nonavailability of anthropometric data that is representative of
the target population should be considered as a safety
hazard [18].
OTEH2016
for example, minimum fuel capacity for continuous
operation for certain time duration, required minimum crew,
full capacity of oil, etc. The minimum weight is also
important and it is defined in [4] as the minimum sum of
empty weight (aircraft, fixed ballast, unusable fuel, full
operating fluids, such as oil, hydraulic fluid, other fluids
required for normal operation of aircraft systems), the
weight of 77 kg for each occupant for normal and commuter
category aeroplanes, and 86 kg for utility and acrobatic
category aircraft, and the weight of fuel necessary for
certain time of operation defined by specific airworthiness
requirements. For military pilots there are additional flight
equipment such as standard parachute (8 kg), helmet (1.9
kg), and pilot suit and shoes (2 kg) with a total weight
greater than 12 kg.
Motivation of the authors of the paper has emerged from the
identification of mentioned phenomenon and professional
need to indicate it and to technically research the topic for
future generation of designers. Aim of the paper is to
emphasize the need for careful approach to this design
challenge and establish the basis for new design
methodology that will enable the right way for design of
safe aircraft with optimal weight and balance which will
meet stability and controllability requirements, taking into
account body weight gain and stature height of current and
future pilots, and actual airworthiness requirements.
APS weight of the aircraft is defined [4] as aircraft
prepared for service weight (a fully equipped operational
aircraft, empty, without crew, fuel or payload). Arm is the
horizontal distance from the Datum (reference) plane
(chosen by designer) to the CG of a chosen item. Center of
gravity (CG) is defined as the point at which an aircraft
would balance if suspended. Technically, at the CG the
sum of all weight of force moments about the CG is equal to
zero. It means that if the weight of any component is Wn and
its distance from a reference datum is xn, then
The overview of design limitation defined by international
civil and military airworthiness regulative and requirements
was shown in Section 2. The weight engineering and
technical basis of weight and balance calculations and
predictions during design of the new aircraft were shown in
Section 3. Section 4 describes the practical activities that are
necessary during experimental measuring of weight and
balance of fabricated aircraft. Statistical analysis of average
pilot body gain weight (mass) during the last fifty years was
explained in Section 5. Discussion and conclusion of the
topic were presented in Section 6 and Section 7.
wx w1 x1 w2 x2 ... wn xn = 0
(1)
Where w is the weight of the whole and x the distance of
the CG from the datum. CG limits (acceptable foreandaft
limits of the CG) is defined as extreme locations of the
center of gravity within which the aircraft must be operated
at a given weight [6]. Fig.1 shows the view of the weight
and balance terms and measures.
2. WEIGHT ENGINEERING
Weight prediction is the engineering task of accurately
predicting the weight of an aircraft, well in advance of the
time the actual weight can be determined by placing the real
aircraft on scales. Aircraft weight prediction is always a
mixture of rational analysis and statistical methods.
Statistical weight equations for many components are
usually written in exponential form which will be seen in
weight prediction [3].
Importance of weight and balance in designing of safe
aircraft is emphasized in [4] and these are defined to be
implied that each stated requirement must be met at each
appropriate combination of weight and center of gravity
(CG) within the range of loading conditions. Compliance of
all requirements must be met and proved by tests or by
calculations, and by systematic investigation of each
probable combination of weight and center of gravity. The
load distribution on aircraft during flight must not exceed
the selected limits, the limits at which the structure is
proven, and the limits defined by adequate flight
requirements.
Figure 1. Main aircraft weights (pilots, equipment,
engine, fuel, etc.) and its arm distances from Datum plane
The CG must be established in both the longitudinal and the
vertical direction. The aircraft designer must demonstrate
that in all likely loading conditions the actual CG will
remain inside the CG limits, without undue penalties in the
form of loading restrictions [3]. The location of the CG
may be expressed in terms of length from a Datum plane
specified by the aircraft designer (manufacturer), or as a
percentage of the M.A.C. (the chord of a rectangular wing,
with the same span, and having similar pitching moment
characteristics [6]). Usually the M.A.C. is a datum to which
the aerodynamic center of the wing is referred and is the
primary reference for longitudinal stability considerations.
2.1. Weight and balance terms
Concerning the weight limits, the maximum weight of the
aircraft is defined as the highest authorized weight of the
aircraft and all of its equipment as specified for an aircraft
and must not be more/less than the specified values. These
values are driven by different parameters, such as the lift
that wing can provide under operation conditions, the
structural strength, and other flight requirements related to,
132
INFLUENCEOFPILOTSAVERAGEBODYMASSCHANGINGONBALANCEOFLIGHTPISTONTRAINERAIRCRAFT
OTEH2016
2.2. Effect of Location of Center of Gravity on
Stability
3. CALCULATION OF WEIGHT AND
CENTER OF GRAVITY
The stability and control characteristics of an aircraft
determine its ability to fly smoothly and a constant
attitude and air speed, to recover from the effects of
atmospheric disturbances, and to respond adequately to the
control of the pilot [5]. Stability in an aircraft is always
desirable, but not necessarily in the maximum possible
degree. Stability has a direct effect on controllability. The
controllability of an aircraft determines the ease of operating
its controls and/or its responsiveness to the controls.
Mechanical stability of the aircraft has two components:
static stability and dynamic stability. One of the factors that
affect the static, longitudinal stability of an aircraft is the
location of the center of gravity (CG) relative to the neutral
point (the point about which the pitching moment of the
aircraft remains constant, even when the angle of attack of
the aircraft is changed) of the aircraft. If the CG of an
aircraft is located ahead of the neutral point (for aircraft
with classical aerodynamically scheme, such as example
aircraft light piston aircraft; Figure 1), the aircraft
possesses positive static longitudinal stability.
Aircraft design is a series of compromises. Every design
alternative has a weight effect (positive or negative). Weight
control can be described as the process by which the
lightest possible aircraft is derived within the constraint of
the governing design criteria [3]. During the design of the
aircraft, in the very beginning of design process, in
conceptual phase of design, it is impossible to calculate the
exact solutions for many aircraft main characteristics and
parameters at the first attempt. The most successfully
aircraft designers make first aircraft sketch as a vision of the
side view of the aircraft and fuselage that we are most
aware of on the ground, because it embodies most of the
practical and aesthetic lines which attract or repel as at the
first sight [6]. Another good reason for drawing the
fuselage first is that it has to fit the human body, the size
and proportions of which dominate cockpit and cabin shape
and seat arrangements, windscreen and window shapes,
sizes and angles (Fig.1).
There is a range or CG limits, generally ahead of the neutral
point and the center of pressure of the aircraft, in which the
CG must be located to preserve safety of flight. In this range
there is an optimum point at which the aircraft is the most
stable, most controllable, and most effective. Any location
of the CG either forward or aft of this optimum point
endangers safety by reducing the stability, controllability,
and effectiveness of the aircraft to an extent comparable to
its distance from the optimum point [3].
Excessive stability (CG far forward from neutral point)
reduces maneuverability and renders the aircraft stiff, or
heavy, to the controls. Too little stability (CG approaching
neutral point, that is to say, approaching neutral stability)
renders the aircraft too sensitive and too responsive to the
controls, reduces the control feel received by the pilot, and
increases the risk of overcontrol by the pilot. The designer
must achieve a satisfactory balance, or compromise,
between stability and controllability. Therefore, it is very
important to design the optimal position of the CG relative
to neutral point for all possible loads and configurations.
Figure 2. Center of gravity envelope diagram
The estimation of the weight of a conceptual aircraft design
is a critical part of the design process. The weights engineer
interfaces in this phase with all other engineering groups,
and serves as the referee during the design process [2].
There are many levels of weight analysis. In the beginning
designer must use crude statistical techniques for
estimating the aircraft weight. More sophisticated weights
methods that come later on in design process, estimate the
weight of the various components of the aircraft and then
sum for the total aircraft empty weight [2]. Commonly
used method uses detailed statistical equations for the
various aircraft components. Aircraft components are
usually grouped in structured group (wing, tail planes,
fuselage, landing gear, air induction system), equipment
group (flight controls, instruments, hydraulic, pneumatic,
electrical, avionics, etc.), propulsion group (engine,
accessory gearbox, exhaust system, cooling provision,
engine controls, fuel system, tanks, etc.), and payload group
(pilot, fuel, oil, cargo, ammunition, pylons, expendable
weapons, etc.) in order to make calculation more easier.
When an aircraft is rigid, stability can be measured in terms
of fixed geometrical factors, like configuration and location
of CG. Real aircraft, however, bend and twist. Their skins
deform, and there is a constant aero elastic interaction
between what the pilot and other disturbances do to the
apparent constants [10]. For example, fuel is consumed and
the CG moves. Aerodynamic pressures suck the skin out
here and push it in there, altering profiles of the wing, tail
surfaces and fuselage. In all of these circumstances CG
margins cease to have strict relevance. However, all
mentioned situations, and much more, such as effects of
drag, thrust and propwash, are out of the scope of this paper
and it will not be here explained.
If balance is not properly controlled, the CG may be in
such a position that it will impose loads upon the structure
of the aircraft that are substantially higher than those
computed for the normal CG location [3].
133
INFLUENCEOFPILOTSAVERAGEBODYMASSCHANGINGONBALANCEOFLIGHTPISTONTRAINERAIRCRAFT
In a group weight calculation, the distance to the weight
datum (arbitrary reference point) is included, and the
resulting moment is calculated. These are summed and
divided by the total weight to determine the actual CG
location. The CG varies during flight as fuel is burned off
and weapons expended [2]. To determine if the CG
remains within the limits established by an aircraft stability
and control analysis, a CGenvelope plot is drawn during
aircraft design (Fig.2). The CG must remain within the
specified limits during the whole flight.
OTEH2016
4.1. Weight determination
Aircraft basic empty weight (BEW) is the sum of three
recorded pickup weights on three supports:
w = w1 + w2 + w3 ,
(2)
Where is:
w1 recorded forward support weight
Fig.2 shows diagram with three curves (two pilots in the
aircraft; one pilot in front cockpit; one pilot in rear cockpit) that
describe successive positions of aircraft CG during a typical
flight, starting from takeoff, retracting of landing gear,
spreading of fuel with reducing the weight of the aircraft,
extension of landing gear, and landing with final weight
Wlanding. Diagram is given for standard pilot weight. In the case
of pilot body weight gain more than its standard weight, or
calculated weight, at least three possible critical points could be
identified, as it is shown on the diagram (Fig.2).
w2 , w3 recorded rear support weight
Aircraft weight with fuel must be checked on the same way.
All weight determination must be checked at least three
times.
4. EXPERIMENTAL DETERMINATION OF
WEIGHT AND BALANCE
Every produced aircraft must be experimentally weighed to
establish an accurate base for weight and balance control in
the flight stage. In establishing the basic weight and CG of
each aircraft a complete equipment inventory must be
conducted, [3]. All of the required equipment must be
properly installed, and there should be no equipment
installed that is not included in the equipment list. Aircraft
weight must be experimentally weighed inside a hangar
(where wind cannot blow over the surface and cause
fluctuating or false scale readings), [1]. Aircraft equipment
completeness must be checked and approved by official
person. Aircraft must be horizontally leveled (its level flight
attitude) standing on the jacks having weight scale pickups.
The recommended practice for aircraft weighing is
measuring on three pickup scales with extended landing
gear, all command surfaces positioned in neutral position
and closed canopy (Fig.4).
Figure 4. Forward support scales; aircraft ready to be
weighed
4.2. Determination of Empty Weight Center of
Gravity
Position of the aircraft empty weight CG is calculated using
the equation (3):
x=
There are two basic types of scales used to weigh aircraft:
scales on which the aircraft is rolled so that the weight is
taken at the wheels, and electronic load cells type scales
where a pressure sensitive cell are placed between the
aircraft jack and the jack pads on the aircraft [1] (Fig.5).
Electronic load cells are usually used when the aircraft is
weighed by raising it on jacks.
w1 D + ( w2 + w3 ) D3
,
w
(3)
Where is:
D  Distance of the forward weighing point(s) from the
datum (reference) plane,
D3  Distance of the rear weighing point(s) from the datum
(reference) plane.
D = A D1 ,
(4)
D3 = D + D2 ,
(5)
A  Distance of the M.A.C leading edge point from the
datum (reference) plane,
D1  Distance of the forward weighing point from M.A.C.
leading edge point,
D2 Distance between two weighing points, forward and
rear (Figure 3).
Figure 3. Center of gravity determination
134
INFLUENCEOFPILOTSAVERAGEBODYMASSCHANGINGONBALANCEOFLIGHTPISTONTRAINERAIRCRAFT
OTEH2016
the accommodation of approximately the 20th percentile
(measured in that time) female of about 1.5 m tall and
larger, [2]. This obviously affected and still affects the
detailed layout of cockpit controls and displays, and aircraft
balance, [11].
Anthropometric data are required and necessary as a basic
input in the human factors engineering of arranging of
military aircraft cockpits. However, according to [7], there
are several common errors to be avoided by designers
when they apply anthropometric data to design of cockpit
arrangement. These are: (1) designing to the midpoint
(50th percentile) or average (standard men) (The 50th
percentile or mean shall not be used as design criteria as it
accommodates only half of the pilots), (2) the misperception
of the typical sized person, (3) generalizing across human
characteristics, and (4) summing of measurement values for
like percentile points across adjacent body parts [7].
Figure 5. Weighing equipment; electronically scale
When the CG of an aircraft falls outside of the limits, it can
usually be brought back in by using ballast. Temporary
ballast in the form of lead bars must be secured to the
airframe so it cannot shift its location in the flight.
Temporary ballast can be also used in the case when the
weight of the pilots exceeds the calculated weight and when
it endangers the aircraft balance. In the case of a repair or
modification (alteration) when the aircraft CG falls outside
of its limits, permanent ballast (usually blocks of lead) can
be installed [1].
5.1. Historical data
Many countries in their air forces occasionally organize
extensive surveys of aviators and pilots anthropometric
dimensions in order to have accurate measures for designing
or flying on trainer, transport or fighter aircraft. These
measuring have been conducted to different purposes, such
as cockpit arrangement, pilot suit making, or any other
purpose. For example, according to [13], an anthropometric
survey finished in 1966 encompassed 200 Royal Air Force
and Royal Navy aircrew flying F4 Phantom. The 44 human
body measurements were taken on each subject (pilot), such
as age, weight, standing height (stature), chest girth, torso
hoop, ankle, knee pivot height, wrist height, axillary height,
knee height, sitting height, arm reach, shoulder breadth,
shoe size, and many other. Subjects were measured wearing
only their own underpants. Some subjects were measured
more than once as a check on repeatability of measurement
by the same operator and as a comparison of performance
between the two operators who shared the task of
measuring. A percentile tables for each of the separate
dimensions was presented in the survey. Tables include
mean, standard deviation, coefficient of variation and range.
5. PILOT ANTHROPOMETRIC DATA
The anthropometric design requirements for pilot stature in
airplanes and helicopters in USA were defined by Civil Air
Regulations [CAR 4b.353(c) and CAR 7.353(b)] in
1950s. CAR stated that airworthiness compliance shall be
demonstrated for individuals ranging from 157.8 cm to
182.4 cm in height [18]. These design heights for stature
remained in effect until 1975 when the regulation was
changed, just for transport category airplanes, to increase
the maximum crewmember heights to be considered from
182.4 cm to 190.0 cm for the design of cockpit controls.
The traditional aircraft design requirement for a 5th
percentile female to a 95th percentile male that have been
adopted by the aviation industry, was based on historical
military standards. The design of aircraft for the 5th to 95th
percentile is only valid if there is a specific definition of the
population basis reflecting the target pilot population, and if
the specific relevant body measurements and muscular
strengths is defined [18]. However, an increasing number
of aviation designs have utilized today a multivariate
approach of body dimension combinations versus univariate
percentiles [20].
A comparison of some of the anthropometric data obtained
from the 1966 survey with the data provided by the 1944
survey of British military aircrew is also given in [13].
Comparison table indicates that in the 22 year period
between two surveys significant changes have been taken
place. For example, pilots average weight has increased by
8.6 kg and stature (height) by 2.97 cm, etc. The maximum
frequency distribution for the pilots height is about 177.5
cm, and for the pilots weight is about 73 kg [13]. Height
range is 161.8 cm minimum and 195 cm maximum.
Weight range is 57.1 kg minimum and 108.4 kg
maximum.
Raymer in [2] argues that in aircraft and cockpit conceptual
design it is first necessary to decide what range of pilot
sizes to accommodate. For most military aircraft, the
design requirements include accommodation of the 5th to the
95th percentile of male pilots (i.e. a pilot height range of
1.66 m to 1.86 m in that time). Due to the expense of
designing aircraft that will accommodate smaller or larger
pilots, the services exclude such people from pilot training.
Today women pilots now enter the Serbian military flying
profession in a certain numbers. Raymer predicted in 80ties
that modern (in 21st century) military aircraft would require
In [14] results of the 19871988 anthropometric survey of
1744 men and women (of different races, such as,
Caucausian White, Afroamerican Black, Hispanic,
Asian/Pacific, and others) of US Army personnel were
presented in the form of the summary statistics, percentile
data and frequency distributions. Dimensions given in this
report include 132 standard human measurements (direct
and derived) made in the course of the survey.
135
INFLUENCEOFPILOTSAVERAGEBODYMASSCHANGINGONBALANCEOFLIGHTPISTONTRAINERAIRCRAFT
Bradtmiller in [16] made comparison between civilian and
army people that were measured in period between 19871988 and 20102012. It was concluded that stature (height)
was not changed, but the weight increased drastically. He
concluded that changes over time are similar between US
army people and civilians. The conclusion was also noticed
that for new design purposes it is more useful to use 3D
scans (94 human dimensions) of human body that was
developped under the project ANSUR II, [16].
OTEH2016
According to literature, the results indicate that significant
differences in body dimensions (height, weight, etc.) do
exist among men of different aeroratings and different ages
(cadets, young officers or aged officers).
5.2. Serbian anthropometric data
One of the most recent domestic historical research and
surveys were conducted by Glava in his doctoral thesis,
[9], where the anthropometric data were presented for last
119 years. According to [9], the first systematic data of
human heights (stature) and weights in Serbian military
sector (soldiers, cadets and officers) where recorded during
1897 and 1898. The 2022 yearsold recruits had average
height 168.7 cm and weight 63.7 kg. Since those years to
the present day the average height has increased for 12.5
cm. Moreover, the average body weight of Yugoslav
working age people was 77.2 kg in 1964 and Serbian 87.3
kg in 2014. The average weight of officers in 1980 and 1999
compared to the present day average weight is increased by
9 kg and 3.9 kg, respectively.
According to researches published in [9], the human average
height of current generation of cadets (21.1 yearsold) is
the same as the average height of aged people (42.7 yearsold), which is 180.96 cm. In addition, the human average
weight of current generation of cadets (measured wearing
only their own underpants, without uniforms) is 78.5 kg
compared to 88.1 kg of the average weight of aged people
(lower and senior rank officers). Researches also shown that
the average heights are in the range from 174.65 cm to
187.27 cm, and the average weights are in the range from
75.68 kg to 99.38 kg for the 68% of Serbian officers.
Comparing these anthropometric data to other recent
domestic unpublished data it can be seen that the
anthropometric measures of current military pilots are
slightly different to data presented in [9].
Figure 6. Officers' body gain weight; 19442014
For the purpose of this paper only two dimensions, body
weight and height (stature), were analyzed given that the
pilot body weight mainly affects the aircraft balance and
pilot body height mainly affects the cockpit arrangement
and indirectly aircraft balance, are shown on Figures 6 and
7. Fig.6 shows officers (army and air forces) body weight
evolution according to anthropometric surveys conducted
since 1944 to 2012 in UK and USA air forces. Fig.7 shows
officers height (stature) evolution according to the same
surveys, [1314, 1617, 1923]. Additionally, the data of
Serbian officers anthropometric measures (weight and
height) were presented on Figures 6 and 7, marked as 2014
data, using data published in [9].
Average body height of current Serbian cadets is
approximately the same as the current younger and senior
officers, and it is not statistically significant [9]. The
difference between the body weights is statistically
significant and it is expected that the current cadets and
young officers will increase their weight by 10 kg in next
three decades. Concluding, the newcoming generation will
be, if not taller, but with greater body weight, with
corresponding (insignificant) need for new cockpit
arrangement and (considerable) influence on weight and
balance of aged aircraft.
5.3. Digital Human Modelling
Since the 1980s, 3D scanning technology has emerged as a
tool to measure the size and shape of the human body as
well as the linear dimensions that traditional anthropometry
provides. The technology has improved Digital Human
Modeling (DHM) and CAD applications required for
equipment and workstation design. 3D scanning technology
gives users the ability to extract new measurement
information after the data has been gathered from the
subjects, as well as the advantage of using 3D computer
models for concept visualization or for rapid prototyping
[20].
Figure 7. Officers' height (stature) evolution; 19442014
For many years the standard man weighed 77 kg, but people
are qrowing bigger. The western country male is heavier
and bulkier than his preWorld War 2 counterpart. Racial
variations in human size and shape can make a considerable
difference to weight and balance calcualtions, to fuselage
proportions, and to cockpit arrangement. The average
weight in 80ties is nearer 82 kg clothed, with deviations of
14%. Geometrical measurements can also vary 5% [10].
136
INFLUENCEOFPILOTSAVERAGEBODYMASSCHANGINGONBALANCEOFLIGHTPISTONTRAINERAIRCRAFT
Digital human modeling tools are used to reduce the need for
physical tests and to facilitate proactive consideration of
ergonomics in virtual product and production development
processes [15]. DHM tools provide and facilitate rapid
simulations, visualizations and analyses in the design process
when seeking feasible solutions on how the design can meet set
ergonomics requirements. DHM software includes a digital
human model, also called a manikin, i.e. a changeable digital
version of a human. An important part of DHM systems is
anthropometry, the study of human measurements, and the
functionality of creating human models.
New scanning technologies along with more sophisticated
statistical methods for matching/forecasting valid data from an
existing database represent a significant improvement over
legacy 1D (i.e. tape measure, calipers) body measuring
techniques. Body scanning technology has proven to be less
expensive, faster, and a more reliable way to measure,
especially for large surveys, [18]. The obvious advantages of
3D scan images for undertaking the measures of target (pilot)
population include the design of betterfitting aircraft cockpits,
and faster response time between measurement and delivery to
the design offices. Threedimensional human analogues created
from scanned images have an almost infinite number of uses in
the design of workplaces such as aircraft cockpits where
accommodation, lines of vision, and ability to reach hand and
foot controls can be tested on a computer screen. Fully dressed
and equipped pilots with helmets can also be scanned for input
into CAD models for assessing workplace interactions [19].
As an anthropometric tool, 3D scanning complements
traditional methods (1D) in two ways. First, a pilotparticipant
can be scanned in a matter of seconds, and the scanned images
become permanent records. Users can return to them as many
times as needed to extract new dimensions or to employ them
in the creation of computer models. Second, the relationship
of one dimension to another, or to several other dimensions, is
clearly apparent. This aids in understanding body shape, as well
as body size [19].
Choi et al. in [20] in 2011 made an extensive survey on 640
male and 60 female aircrew measuring 60 traditional body
dimensions. The measured dataset was compared to 1967
USAF and the Joint Strike Fighter datasets and the results of
these comparisons show that the aircrew population is
growing heavier and exhibiting some increased
measurements related to increased mass. Some of the
measured dimensions are shown in Figures 6 and 7.
OTEH2016
Fig.8 shows a standard pilot (male) without parachute as it
is defined in [10] with dimensions of a standard pilot given
in [10] dated in 1955.
6. DISCUSSION
According to authors experience, there are several
recommendations in the area of collection of relevant data
of pilot population measurements, such as:
Execute regular measuring surveys to all pilots and
focused groups of pilots (helicopter pilots, fighter pilots,
transport aviation pilots, cadets, female, male, etc.),
Make measuring surveys using fast and reliable modern
optical digital 3D scanning technology which gives the
ability to extract new measurement information,
Use mathematical statistical methods and prediction
methods for analysis of focused pilots group in the case of
the lack of uptodate measured data,
Use modern computer methods for 3D digital human
modeling (DHM) and CAD applications for design,
Use new multivariate approach of body dimension
combinations insted of unvariate percentiles in the case
whe there is not a specific definition of the population
basis reflecting the target pilot population, and if the
specific relevant body measurements and muscular
strengths is not defined.
Focused group and measured human dimensions must be
in accordance to the functions that must be accomplished
and equipment to be managed.
Recommendations in the area of aircraft design are as
follows:
Make more iterations for weight and balance calculations
during aircraft conceptual design,
Use modern 3D computer methods for design of aircraft
structure, equipment, pilot cockpits, cockpit arrangement
and weight and balance engineering,
Measure all produced aircraft experimentally after any
changes, modifications or equipment changes.
During aircraft design always have in mind that the
aircraft would be in service for next 40 years with pilots
(and crew and passengers) which will be probably heavier
and taller than their counterparts in the moment of the
aircraft design. According to that, it is necessary to take it
into account during weight engineering and cockpit
arrangement.
7. CONCLUSION
This paper emphasizes the importance of recording and
updating of exact average pilots body mass data of pilots
population assigned for piloting the specific aircraft. The
paper explains the basic methods that are used during
aircraft design and regular maintenance during service in the
area of weight engineering and balance calculation. The
paper also stresses possible discrepancies or conformance to
current airworthiness and requirements in the same area.
Figure 8. Geometrical 3D construction of a standard man
(without parachute) [10]
Aircraft designers often utilize legacy anthropometric
databases that may not reflect the present target pilot
137
INFLUENCEOFPILOTSAVERAGEBODYMASSCHANGINGONBALANCEOFLIGHTPISTONTRAINERAIRCRAFT
population. Although most current aviation designs have
utilized legacy anthropometric data (dated thirty or more
years ago), there are a sufficient number of valid existing
anthropometric databases that include body dimension
measurements and weight (mass) measurements compiled
from data sets gathered through large surveys that are
representative of the target pilot population. Furthermore,
new scanning technologies, 3D modeling using Digital
human modeling (DHM) tools, along with more
sophisticated statistical techniques for matching/ forecasting
target populations from legacy anthropometric databases,
have enhanced the ability to effectively and efficiently
gather these data on a periodic basis, hence anthropometric
data that are not representative of the pilot population
should no longer be accepted as a limitation in aircraft
design, [18].
OTEH2016
[10] AvP970 2, Book 2, App 8, Design Requirements for
Aircraft for the Royal Air Force and Royal Navy,
Ministry of Aviation, 1955.
[11] Anonymous, Anthropometry of Flying Personnel,
Wright Air Development Center, TR 52321, 1954.
[12] DODHDBK743A, Military Handbook Anthropometry
of U.S. Military Personnel, Department of Defence,
USA, 1991.
[13] Simpson, R.E., Bolton, C.B., An Anthropometric
Survez of 200 R.A.F. and R.N. Aircrew and the
Application of the Data to Garment Siye Rols,
Engineering Physics Dept., R.A.E., Farnborough,
1970.
[14] Gordon, C.C., Churchill, T., Clauser, C.E., Mc
Conville, J.T., Tebbetts, I., Walker, R.A., 1988
Anthropometric Survey of US Army Personnel:
Methods and Summary Statistics, Technical Reports
NATICK/TR89/044, United States Army Natick
Research, Development and Engineering Center
Natick, 1989.
[15] Brolin, E., Hanson, L., Hogberg, D., Ortengren, R.,
Conditional Regression Model for Prediction of
Anthropometric Variables,
[16] Bradtmiller, B, Anthropometric Change in the US
Army: Using the 2012 Army Data for Civilian Design,
The 19th Annual Applied Ergonomics Conference,
2016.
[17] Gordon, C., US Army Anthropometric Survey
Database: Downsizing, Demographic Change, and
Validity of the 1998 data in 1996, Technical Report
NATICK/TR97/003, 1996.
[18] Joslin, R.E., Examination of Anthropometric Databases
for Aircraft Design, Proceedings of the Human Factors
and Ergonomics Society 58th Annual Meeting, 2014.
[19] Gordon, C.C., Blackwell, C.L., Bradtmiller, B.,
Parham, J.L., Barrientos, P., Paquette, S.P., Corner,
B.D., Carson, J.M., Venezia, J.C., Rockwell, B.M.,
Mucher, M., Kristensen, S., 2012 Anthropometric
Survey of U.S. Army Personnel: Methods and
Summary Statistics, Technical Report NATICK/TR15/007, 2014.
[20] Choi, H.J., Coate, A., Selby, M., Hudson, J.,
Whitehead, C., Aircrew Sizing Survey 2011, Technical
report, AFRLRHWPTR20140113, 2014.
[21] MILSTD1472D, Human Engineering Design Criteria
for Systems, Equipment and Facilities, 1989.
[22] Weight, Height, and Selected Body Dimensions of
Adults, USA 19601962, National Center for Health
Statistics, 1965.
[23] Hudson, J.A., Zehner, G.F., Robinette, K.M., JSF
Caesar: Construction of a 3D Anthropometric Sample
for Design and Sizing of Joint Strike Fighter Pilot
Clothing and Protective Equipment, AFRLHEWPTR20030142, United States Air Force Research
Laboratory, 2003.
Authors firmly advocate a multidisciplinary approach to the
process of aircraft design in the area of weight and balance
engineering and cockpit arrangement. It means that the
designers must use modern multivariable optimization of
many significant parameters that have to drive optimal
aircraft design. Instead, using of old and obsolete pilots
measures during design of new aircraft could lead to
jeopardizing of aircraft safety. The authors also emphasize
the need for establishing a systematical methodology for
continual recording of anthropometric data for focused
population (pilots) which will contribute the optimal aircraft
design process.
References
[1] FAAH80831, Aircraft Weight and Balance
Handbook, US Department of Transportation, Federal
Aviation Administration, Flight Standards Service,
2007.
[2] Raymer, D.P., Aircraft Design: A Conceptual
Approach, AIAA Education Series, Air Force Institute
of Technology, 1999.
[3] Niu, M. CY., Airframe Structural Design, Practical
Design Information and Data on Aircraft Structures,
Conmilit Press Ltd., 1988.
[4] CS 23, Certification Specifications for Normal, Utility,
Aerobatic, and Commuter Category Aeroplanes,
European Aviation Safety Agency, Annex to ED
Decision, 2012.
[5] Crabs, C. C., Aircraft Stability and Control, Aerospace
Engineer and Pilot, Aircraft Operations Branch.
[6] Stinton, D., The Design of the Aeroplane, BSP
Professional Books, Oxford, 1983.
[7] HFDS, Anthropometry and Biomechanics, (Amended
2009), Chapter 14, 2003,
[8] Jenkinson, L.R., Marchman III, J.F., Aircraft Design
Projects, Butterworth Heinemann, Oxford, 2003.
[9] Glava, B., Motor skills, morphological status and life
habits among members of Serbian Army Forces,
(doctoral thesis in Serbian), Faculty of Sport and
physical education, University of Belgrade, 2015.
138
PROTOTYPE SOVA DEVELOPMENT:
AIRCRAFT LYFE CYCLE EXTENSION
VANJA STEFANOVI
Military Technical Institute, Belgrade, vti@vti.vs.rs
MARIJA BLAI
Military Technical Institute, Belgrade, vti@vti.vs.rs
MARINA OSTOJI
DOO UTVA AI, Panevo, majce74@yahoo.com
TONKO MIHOVILOVI
DOO UTVA AI, Panevo, tonkojetonko@gmail.com
DRAGAN ILI
Jugoimport SDPR, Belgrade, dragan_ilic01@yahoo.com
Abstract: With the aim of a new aircraft development, single engine fourseater aircraft with fixed landing gear,
analysis of the existing structure has been conducted. Analysed subject was earlier produced UTVA 75 structure, which
has been used as the basis of the new aircraft prototype. Elements of the structure were tested with different NonDestructive Testing (NDT) methods. This paper describes condition of existing structure and its test results, as well as
necessary repairs and replacement of the damaged parts, in order to extend life cycle of the structure.
Keywords: life cycle extension, aircraft structure, structure repairs, NDT methods.
1. INTRODUCTION
Making a general purpose airplane, suitable for any
purchaser or any use, is impossibility. However, it is
frequently possible to arrange a design which would simplify
future changes without sacrificing either structural or
aerodynamic efficiency or taking a weight penalty [1]. The
design process must not only address interactions between
traditional aerospace disciplines (e.g. aerodynamics,
structures, controls, propulsion), but should also account for
life cycle disciplines (e.g. economics, reliability,
manufacturability, safety, supportability, etc.). These
disciplines can bring a variety of uncertainties of differing
natures to the design problem, especially as innovation
occurs within and amongst the disciplines [2]. In a process of
upgrading an existing one, some of the traditional aerospace
disciplines are less reviewed, due to the already defined
aerodynamic with the current structure. Nevertheless
controls and propulsion are often revised, to follow
requirements of the customers and to be concurrent on the
market. When it comes to life cycle disciplines in a
process of upgrading, they should be considered completely
from the start (e.g. manufacturability of the prototype and
possible serial production component supportability, etc.)
and then decision should be made whether the new product
is affordable or not.
With fruitful tradition of designing and manufacturing metal
structure aircrafts (Galeb G2, Supergaleb G4, Utva 75,
Lasta and now Sova), mostly for training military pilots,
139
Serbia has vast amount of experienced experts all
generations, who can deal with new challenges in this area.
Aircrafts were designed and produced for our needs and also
numbers are sold to foreign countries. Positive feedback
from customers all around the world, and fact that many
aircrafts are still in operational use confirms our production
capabilities and proficient design approach to different type
of aircrafts. In a case of Utva 75 aircraft (picture 1), 140 were
produced and plenty are still in use.
Picture 1. Utva 75 aircraft
All metal Utva 75 aircraft, certified according to FAR 23,
is sidebyside seating configuration with an advantage
that pilot and instructor can see each other's actions. With
the conversion in fourseater aircraft, it will be allowed to
future pilots, being in the back seats during the flight, to
learn from the instructor and pilot by watching. Special
OTEH2016
PROTOTYPESOVADEVELOPMENT:AIRCRAFTLYFECYCLEEXTENSION
advantage of this aircraft version (fourseater Sova),
which can be used for initial training of several trainees,
is considerably reducing the cost of the training. That is
the reason why nowadays popularity of the fourseater
trainer is in increase all around the world.
NDT methods were used in a process of examining
structure for lifecycle extension. All parts of the structure
that are planned to be kept and especially vital ones were
subject of inspection with different methods, performed
by the firm RD Diagnostics and approved by Military
Technical Institute. Following NDT methods were used in
a process of testing structure.
With the aim of Utva 75 aircraft conversion into fourseater aircraft Sova (Utva 75 A41M), process of creating
prototype and modification of the existing structure is
shown in this paper.
Visual inspection (VT) is probably the most widely
used of all the nondestructive tests. VT was
conducted on all parts, using tools: a magnifying
glass zoom 5x and endoscope. Magnifying glass was
used to estimate the depth of corrosion in places
where it was discovered.
Endoscope (ES) was used for testing all inaccessible
places, the inner surfaces of the wings and the fitting
interior, the inner surfaces of the fuel tank, fins,
flaps, elevator and rudder, in order to detect hidden
defects and corrosion.
Penetrant test (PT) has its fundamental purpose of
increasing the visibility contrast between a
discontinuity and its background. It was carried out
on some fuselage fittings and pedals in order to
detect surface defect and cracks.
Edy Curent Test (ECT) is well suited for the
detection of service induced cracks usually caused
either by fatigue or by stress corrosion. Inspection
can be used with minimum of part preparation and a
high degree of sensitivity. ECT was performed on all
connections that are not going to be replaced. That
includes connections of wing and fuselage structure
e.g. rivets, bolts, the connection of all fittings and
connections of vertical and horizontal stabilizers. All
magnetic parts are also included in order to detect
cracks and surface defects.
Ultrasonic test (UT) permits the detection of small
flaws with only single surface accessibility and is
capable of estimating location and size of the defect
and also can be used for thickness measurement,
where only one surface is accessible. UT was
performed on the chosen rivets, especially those
which connect fittings with the rest of the structure,
axles and joints in order to determine defect of
homogeneity.
2. STRUCTURE LIFECYCLE
EXTENSION
Nowadays conversion of the aircraft, with upgraded
avionics and other systems, are the most economical
approach and frequently applied solution. Whether its a
case of business jet conversion to military trainer [3], as it
was in The United States Navy, or it was a decision of an
aircraft conversion within an air force with more limited
budget, it is moneytime saving solution without loss of
efficiency.
With some past experience of Utva 75 conversion in fourseater aircraft, it was stepped into feasibility study and
conclusion was made. Utva 75 structure meets
performance goals of a new aircraft with some changes in
the cockpit area, upgraded avionics and with redesign of
propulsion and controls, prototype could be shortly
finalized. Decision was made that the prototype is going
to be conversion of the Utva 75 offtheshelf structure,
presented in picture 2.
Picture 2. Future prototype structure
In our case, UTVA Aircraft Industry has main
responsibility for designing and producing all necessary
parts and has to ensure the project finishes on time.
Military Technical Institute and RD Diagnostics d.o.o.
were engaged in a process of examining existing structure
and possible life cycle extension.
Every element that has been inspected was firstly
detached from the rest of the structure, if it was possible,
then cleaned and prepared for testing.
2.2. Structure testing
2.1. Non Destructive Testing Methods
The aircraft chosen to be a prototype was made of main
assemblies of already produced Utva 75, whole metal
structure. Wing, horizontal and vertical stabilizer were in
storage and never used but fuselage, elevator and rudder
were in use for short period of time. The testing was
conducted in the facilities of the UTVA AI with a help of
personnel from the factory. They performed first
measures by checking the geometry of the aircraft.
Structure testing was next step of the extending lifecycle
process. In picture 3 is shown fuselage of Utva 75 aircraft.
Non Destructive Testing methods (NDT) are widely used
technics to probe structures and materials either before
they enter use or as a part of maintenance program.
Whether the aim is to observe microstructures of the
welds, measure the thickness of material or detect induced
cracks in a material NDT methods are appropriate choice
for fast and efficient inspection. NDT is now compulsory
for many aerospace firms and is a vital part of the
production process.
140
OTEH2016
PROTOTYPESOVADEVELOPMENT:AIRCRAFTLYFECYCLEEXTENSION
Picture 5. Left and right wing damaged zone
The main defect was found in the bushing of the front
spar left wing fitting (picture 6  right). Irregularities were
found with UT and described as a lack of homogeneity.
Also some mechanical damages were found on the front
spar right wing fitting (picture 6  left).
Picture 3. Fuselage in the UTVA AI facilities
As mentioned earlier, Military Technical Institute and RD
Diagnostics d.o.o. started inspection according to already
formed plan. In this plan main structural elements, one by
one, are anticipated to be subject of the analysis. The
structure was tested with different NDT methods.
In picture 4 main structural assemblies of the examined
structure are presented (wing, fuselage, vertical and
horizontal stabilizer, ailerons, elevator, and rudder). Some
of the major defects of examined structure, classified by
main assemblies, are presented in the further text.
Picture 6. Left left wing fitting mechanical defects
Right right wing bushing defect of homogeneity
Vertical stabilizer was in storage and its inner
structure was tested with ES and was healthy. Brackets
were tested with UT and had no inner flaws but some
mechanical damages on the surface at the lower bracket
were found. Location of the bracket and irregularities
are shown in picture 7.
Picture 4. Main examined structural elements and
assemblies
Wing was never in operational use; nevertheless the
process of examining was detailed. NDT methods that
have been applied here are: VT and ES for outer and
inner surfaces and elements (skin, spars, ribs), then
ECT for bolts, joints and UT for fittings, rivets and
axels inspection.
Location of the above mentioned defects are presented in
the picture 5.
Picture 7. Left  vertical stabilizer  damaged zone
Right  lower bracket, mechanical defects
141
OTEH2016
PROTOTYPESOVADEVELOPMENT:AIRCRAFTLYFECYCLEEXTENSION
Rudder was in operational use and had corrosion on
the torsion tube. NDT methods confirmed that the
damages are only on the surface, and with adequate
treatment and protection it can be easily removed.
Location of the torsion tube and its damage is presented
in picture 8.
Elevator was in operational use and had some
mechanical deformation of the skin which was noticed
by VT. Deformation of the upper skin near the outer rib
was appeared as a consequence of the wrong
manipulation. Place of the deformed skin is presented in
a picture 11 and mechanical damages in picture 12.
Picture 8. Left rudder  damaged zone
Right torsion tube corrosion
Picture 11. Horizontal stabilizer  damaged zone
Horizontal stabilizer had traces of surface corrosion
on the fittings that connect stabilizer with frame no.11
in the fuselage. Location of the fitting is presented in
the picture 9.
Picture 12. Deformed skin of the horizontal stabilizer
Picture 9. Horizontal stabilizer  damaged zone
NDT methods have confirmed that inner material of the
fitting is without irregularities and corrosion is on a
surface only. Damages of the fitting are shown in the
picture 10.
Fuselage was in operational use and demanded more
dedication and work in order to examine whole
structure from the first to the last frame. The important
information is that the greatest numbers of
modifications are performed on the fuselage, and the
most of the structure from the cockpit was changed
during the process of examination. All NDT methods
were used in order to locate defects of the material.
Main tests were performed on the fittings that connect
fuselage with wing and horizontal and vertical stabilizer
and engine mount with the fuselage. They were placed
on the first and eleventh frame. UT, ECT, PT confirmed
that inner structure of the material (fittings, bushings,
brackets, bolts, axels) was flawless. There were some
traces of corrosion on the first frame fittings, but it was
easily removed. Mechanical damages that have been
detected are presented in the next pictures, location in
the picture 13 and mechanical irregularities on the left
fitting and right front skin in picture 14.
Picture 10. Corrosion of the horizontal stabilizer fitting
142
OTEH2016
PROTOTYPESOVADEVELOPMENT:AIRCRAFTLYFECYCLEEXTENSION
seats and enlarged cabin doors.
Main modification was performed in the cabin space
(seats, instrument panels, controls, etc.). There is a new
digital instrument panel, pilot sticks and engine controls
(picture 16); redesigned pilot seats area, with addition of
central and side console which are also used as a support
for pilot seats. Pilot seats can be set in six position
providing suitable accommodation for any pilot.
Picture 13. Fuselage  damaged zone
Picture 15. Unchanged structure of Sova aircraft
Picture 14. Left  mechanical damages of the fitting
Right right skin damages
After inspection, the report has been written. It includes
results of examined structure and requirements for
mandatory repairs. Experts have confirmed, after control
inspection of the structure that all necessary repairs are
conducted in compliance with report requests.
Consequently, lifecycle of the structure has been
extended for 10 years.
3. PROTOTYPE DEVELOPMENT
Picture 16. CAD model of redesigned cabin space
New fourseater airplane was required to perform all of
the training missions. It can be used for reconnaissance,
photoshooting and other operation depending on
equipment. One special advantage of this aircraft version
equipped with advanced digital displays is that it can be
used for initial training of several trainees, reducing
considerably the cost of the training
Second approach was upgrading of aircraft systems
making the aircraft concurrent on the market; redesigned
primary and secondary control system with electrical trim
and flaps actuators. Pedals in the cabin area are
redesigned and adjustable according to pilot needs.
Completely new fuel system, with new boost pump and
fuel selector with four positions (left, right, both tank and
fourth position for fuel off). There is a new hydraulic pipe
line with the same non retractable landing gear as on
Utva 75. Completely new ventilation system and
electrical air condition unit.
Whole metal structure Utva 75 with a sidebyside seating
configuration is a basis of the new fourseater airplane
Sova. At the same time with extension of a structure lifecycle, the conversion project has been conducted. In a
process of prototype development two main approaches
can be mentioned: modification of the structure and
revision and upgrade of the aircraft systems.
Changes including reinforced structure and mass increase
consequently led to revision in a propulsion. New engine
has been integrated, Lycoming IO390A1 B6 (max
power 210 hp) paired with appropriate two blades
Hartzell propeller. It improves aircraft maneuverability
comparing with former Utva 75. Engine mount was tested
Structure revision consists of minimum changes on the
wing and empennage (picture 15) but central area of the
fuselage is reinforced due to two additional passenger
143
OTEH2016
PROTOTYPESOVADEVELOPMENT:AIRCRAFTLYFECYCLEEXTENSION
then slightly modified according to new engine (picture
16).
Picture 18. Sova  Technical demonstrator
Time and costs that have been used in a process of
developing
feasibility
study,
then
designing
modifications, producing elements, mounting them then
integration of new aircraft systems are considerably less
than design and production of a completely new aircraft.
Not to mention confirmation of efficiency and
maintainability of 40 years long operational use of the
Utva 75, and the importance of that pieces of information
in a process of development the aircraft, make this
conversion affordable solution.
Picture 16. CAD model front Sova aircraft section
4. CONCLUSION
After inspection of the Utva 75 structure, lifecycle
extension has been confirmed. Conversion, as a choice of
moneytime saving solution, has been performed on the
existing structure. Project included redesign of the central
fuselage part with emphasis on the cabin area
modification. New control and propulsion systems are
integrated. Whole project is supported by modern design
software (CATIA V5) (picture 17).
REFERENCES
[1] Niu,C.Y.: Michael, Airframe structural design, Hong
Kong Conmilit Press LTD., Hong Kong, 1999.
[2] Dimitri,N.,
Mavris,D.,
DeLaurentis,A.:
A
probabilistic approach for examining aircraft
concept feasibility and viability, Aircraft Design,
Volume 3, Issue 2, June 2000, Pages 79101
[3] Lyle Jr,J.W.: Converting A Citation Business Jet to a
military trainer, Aircraft Design, Volume 1, Issue 1,
March 1998, Pages 5160
[4] Milutinovic,S.: Konstrukcija aviona, Beograd, 1970
[5] Allen,R.: Aircraft conversions: specialist freighter
conversions work expands, Aircraft Engineering and
Aerospace Technology, 2001, Vol. 73, Issue 6,
[6] Boeing to convert five MD11s to freighters, Aircraft
Engineering and Aerospace Technology, 2004, Vol.
76, Issue 2
[7] Threebladed prop conversion for the Piper
Comanche 260C, Aircraft Engineering and
Aerospace Technology, 2003, Vol. 75, Issue 5
[8] Vasic,Z., Blai,., Stefanovi,V.: Reconstruction of
aircraft structure with the aim of optimizing and
extending aircraft lifecycle, OTEH 2012, Belgrade
2012., ISBN 9788681123584.
Picture 17. CAD model of Sova aircraft
Sova (Utva 75 A41M), as a conversion of Utva 75, with
modification of cockpit area, controls, propulsion will
result new CS 23 certification (Certification Specification
for Normal, Utility, Acrobatic and Commuter Category
Airplanes).
144
SOME ASPECTS OF THE DIFFERENT TYPES WIRELESS SENSORS
IMPLEMENTATION WITHIN AIRBORNE FLIGHT TEST
CONFIGURATION
ZORAN FILIPOVI
Lola Institute, Belgrade,zoran.filipovic@li.rs
VLADIMIR KVRGI
Lola Institute, Belgrade,vladimir.kvrgi@li.rs
DRAGOLJUB VUJI
Military Technical Institute, Belgrade, dragoljub.vujic@vti.vs.rs
Abstract: Design and implementation of the Flight Test Instrumentation Systems (FTI) is very important for successful
evaluation the testing prototype of aircraft. The configuration of FTI System is based on the General Plan of Testing of
any new aircraft (or significant improvement to an existing aircraft). The system represents a fully integrated approach
to flight test systems which addresses the endtoend requirements from airborne data acquisition and real time flight
monitoring through aircraft performance and stability/control analysis. Implementtation of wireless sensor networks
(WSNs) communication within airborne FTI configuration or aircraft structural health monitoring system, introduces a
number of challenges such as guaranteeing reliable transfer of the sensor data and time synchronization of the remote
nodes. Special attention has given on the use of MicroElectroMechanical Systems (MEMS) and Surface Acoustic
Wave (SAW) sensors technology. This paper addresses some aspects of WSNs acquisition, the associated challenges
and discusses approaches and solutions to these problems.
Keywords: Flight Test Instrumentation, wireless sensor networks, surface acoustic wave sensors, microelectromechanical systems, sensor node.
1. INTRODUCTION
The main purpose of the Flight Test Instrumentation (FTI)
System is to acquire data about the operation of the test
vehicle and provide the data for processing during post
flight analyses. For development, verification and
certification phase, aircraft must be equipped with a
complete FTI System which is consisting of three
components:
 Airborne Data Acquisition System (DAS),
 Compatible Ground Telemetry Station (GTS),
 Post Test Data Processing System (DPS).
High environmental acceptability of FTI  equipment
(temperature, vibration, humidity .)
EMC compatibility [2].
The airborne DAS in the early days consisted of pilots
notes, photographs of cabin instrumentation, and basic
sensor data. Todays systems consist of thousands of data
acquisition modules, recording subsystems, and the
associated cabling that is required to support these
systems. And the need for additional sensors continues to
increase as aircraft systems become more sophisticated
and technologies evolve. As a result, FTI systems must be
designed to reduce overall test program time. Moreover,
size, weight, and power (SWaP) are major concerns for
aircraft manufacturers, as demands for more fuel efficient
aircraft rise. A major contributor to aircraft development
(and assembly) timeframe, cost, and overall weight is
cabling. Minimizing the amount of cabling in an aircraft
may considerably reduce its development time, cost, and
weight. Therefore, a wireless sensor network (WSNs)
based airborne FTI system must be small that can be
relatively easy installed in a variety of configurations and
locations on aircraft, stores, or other test articles without
significantly impacting the performance of the system
being tested [3], [4].
Design and implementation of the FTI System is very
important for successful evaluation of the testing
prototype of aircraft. The configuration of FTI System is
based on the General Plan of Testing of any new aircraft
(or significant improvement to an existing aircraft) and
missile [1], [2]. Numerous operational aspects must be
considered as part of the FTI system design like:
 Small in size and open architecture, reduced wiring,
 Easy to install (modularity adaptable, plug and play),
 Low power consumption,
 Low maintenance, easy to handle (selftesting)
 MTBF(Mean Time Between Failure) very high,
 Optimized data acquisition and data recording,
 Long lasting calibration or no calibration necessary,
 Total costs for installation very low,
This paper describes some approaches to addressing these
challenges and achieving a useful FTI system using
different tapes of WSNs. Many challenges need to be
145
SOMEASPECTSOFTHEDIFFERENTTYPESWIRELESSSENSORSIMPLEMENTATIONWITHINAIRBORNEFLIGHTTESTCONFIGURATION
overcome before WSNs can be effectively and
successfully deployed in FTI, including throughput,
latency, power management, electromagnetic interference
(EMI), and band utilization considerations.
OTEH 2016
applications. The first is to have a wireless link from a
Data Acquisition Unit DAU to sensors.
2. AIRBORNE WSNs FTI CONFIGURATION
The airborne part of FTI system support the measurement
different parameters and its transfer to the GTS using
specific airborne infrastructure consisting of cables,
connectors and signal conditioning electronics. The
introduction of a new hardware element into a nonWSNs
based test system requires any number of these
infrastructure pieces, thereby almost guaranteeing a
negative effect on the SWaP constraints while also
increasing the material cost of the system. Furthermore,
each new physical connection introduces a new potential
point of failure and a maintenance item that requires more
time when checking/auditing the system [1].
Picture 1. Wireless FTI Networks
For the second case, one or more remote DAUs can be
connected via wires to sensors, but connect to the rest of a
airborne DAS by a wireless link. This has, to an extent,
been a strategy that has been used in rotorcraft where a
DAU resides on a rotor hub and is connected using slip
plates to the rest of the system. This still requires a
physical link however and a wireless RF link would
remove the need for such an oft problematic electromechanical solution.
Considering these effects, the integration of WSN based
approaches into instrumentation test systems offers key
benefits. First, WSNs based replacement elements are
inherently designed to require fewer physical connections.
That results in positive effects on the SWaP constraints and
physical robustness of the test system. Furthermore, the lack
of physical connections requires WSNs to be adaptive and
flexible. As a result, introducing new elements into a WSNs
based instrumentation system will require less installation
time, maintenance, and documentation.
Ethernet technology offers numerous benefits for
networked FTI (Systems Integrated Network Enhanced
TelemetryiNET) such as increased data rates, flexibility,
scalability and most importantly interoperability owing to
the inherent interface, protocol and technological
standardization [1].
Also, WSNs provide new opportunities to extend the
sensing and monitoring capabilities throughout the
lifecycle of a system. The remote accessibility of WSNs
provides a means to calibrate and configure many sensors
and devices without direct physical access. These WSNs
will communicate its data to a central unit installed within
the aircraft (Picture 1). Picture 1 shows a diagram of the
Ethernet based FTI data acquisition systems including
WSNs and a number of the data acquisition units (DAUs).
There are two principal wireless use cases for FTI
The focus of this paper is in some aspects of integration
of WSNs wireless within FTI network. Picture 2 shows a
system diagram of a small onboard WSNs
instrumentation and telemetry configuration.
Picture 2. Onboard WSNs instrumentation and telemetry configuration
146
SOMEASPECTSOFTHEDIFFERENTTYPESWIRELESSSENSORSIMPLEMENTATIONWITHINAIRBORNEFLIGHTTESTCONFIGURATION
The system is comprised of a group autonomous WSNs
spatially distributed that monitor physical or
environmental conditions, such as temperature, sound,
vibration, pressure, motion and so on. These WSNs
transmit the acquired data to a wireless controller module
(WCM) placated in the main airborne DAU including and
the others necessary modules.
OTEH 2016
OLM  Operational Loads Monitoring,
RHUMS  Rotorcraft Health and Usage Monitoring,
SHM  Structural Health Monitoring.
Implementing such wireless communication introduces a
number of challenges such as guaranteeing reliable
transfer of the sensor data and time synchronization of the
remote nodes. The key design challenges for the airborne
instrumentation/telemetry are: remote sensor, battery
selection, frequency study including radio and antenna
selection [3], [4] [5].
The (WCM) provides a selfdiscovery mechanism for
identifying the sensors in its network, facilitating the
sensor calibration process, programming acquisition
variables in the sensor units, and providing two way
communication with all the sensors in its network. The
installation of the WCM as part of the data system may
require an external remote antenna for communication
with the wireless sensors. The main function of WMC is
to receive wireless data from the sensors, and format the
data in a format required by the host DAU. The WMC
implements command and control with the sensor units,
uniquely addresses each sensor in its network, controls
multiple sensors simultaneously, and provides access
capability to the health status of each sensor unit in its
network. On power up or as required, the controller
programs the sensitivity and sample rate of the sensor
channels. Periodically the sensor will append additional
data to the sensors. This data includes the state of the
discrete sensor inputs, temperature within the sensor
module enclosure as well as other required
housekeeping signals. As required, the controller can
send commands to each sensor under its control,
instructing it to change the logic state of its discrete
output lines [5].
3. WIRELESS COMMUNICATIONS
STANDARD FOR WSNS
There are numerous standards that are used in the
implementation of wireless sensor for aerospace
applications like:
All information about the measured electrical and nonelectrical quantities during the flight testing of the aircraft
is transmitted in real time via aircraft telemetry
transmitter (XTMRS band transmitters operate from 2.2 2.45 GHz) and the corresponding antennas (2 or 3) that
are positioned in special places on the external structure
of the aircraft.
IEEE 802.11 is a set of standards that operate
primarily in the 2.4 GHz and 5 GHz bands. This
standard is commonly known as WiFi, and widely
used in the office and at home. There are a number of
protocols defined using this standard, currently the
most widely used being 802.11g and 802.11n.
Bluetooth (IEEE802.15.1) operate in Bluetooth
ranges from 2.4  2.483 GHz with 79, 1 MHz wide
RF channels (generally, 2.408  2.480 GHz is the
operating range), transmitting up to 1Mbps
(Bluetooth 2.0) or up to 2 to 3 Mbps for Bluetooth
2.0 + EDR (Enhanced Data Rate).
Zigbee (IEEE802.15.4) is a networking standard
generally used for low data rate, home automation
and industrial control applications such as lighting
control. It is based on the IEEE 802.15.4 radio
standard and supports applications that require
periodic short data transfers up to 250 Kbit/s over
distances to 75 m.
These technologies were evaluated in terms of:
standardization, maturity, bandwidth, range, security, and
network join time.
As available space in aircraft weapon bays, small
weapons, and unmanned vehicles becomes extremely
limited, the miniaturization of remote sensors and
telemetry units becomes critical. The basic requirements
that must satisfy airborne part of FTI are:
 Flexible and modular instrumentation and telemetry
system that can be installed on an aircraft or other
test article without the need for permanent
modifications.
 The ability to remotely activate/deactivate and
reconfigure remote sensors system such as (changing
resolution and/or sample rate) and to interface with a
larger airborne FTI network.
 The FTI equipment must be easily removed after test
completion with minimal impact on the operational
configuration [1], [2].
In terms of power consumption and device size, Zigbee
and the latest version of Bluetooth (2.0) are comparable.
The power requirement of Bluetooth 2.0 devices is half
that of previous versions. Zigbee devices are specified to
operate over distances of up to 100 meters as are
Bluetooth class 1 devices. Bluetooth class 2 devices are
specified for distances up to 10 meters. Both wireless
communication systems have their own security
protocols. Zigbee security is defined by a 128 AES plus
application layer security; Bluetooth security uses 64 and
128 bit encryption. The different security systems must be
evaluated in the aircraft environment. One advantage of
Zigbee devices is their connection time. Zigbee can join
existing networks in less than 30 ms, whereas, Bluetooth
2.0 devices require approximately 35 seconds. Zigbee
has the unique ability to receive information from a
device outside the network, even if this receiver is not the
principal controller, and then forward the information
(without reading the data it is transmitting). The device
simply looks at the destination and sends the data on to
WSNs are used not only during the process of testing
prototypes of the aircraft under the complex FTI system
but also in cases of equipping aircraft during their
exploitation with additional acquisition systems in order
to extend its lifetime. For this purposes, there are several
Usage Monitoring programs such as:
147
SOMEASPECTSOFTHEDIFFERENTTYPESWIRELESSSENSORSIMPLEMENTATIONWITHINAIRBORNEFLIGHTTESTCONFIGURATION
the controller. This capability may be useful in situations
where a direct line of sight between an outlying sensor
and controller is not possible. It is one of the reason why
Zigbee has been chosen as a more perspective for further
investigation in FTI WSNs.
OTEH 2016
Wireless sensor unit with integrated transducer and
battery
Wireless sensor unit with external transducer and
uses external power supply (28 V)
In both cases, sensors digitizes and encodes the
measurement data for wireless transmission to a WCM
including and the others necessary modules within FTI.
The sensor node is fully programmable via the wireless
connection from the WCM. The sensor programmable
variables include channel sample rate, gain, offset, and
antialias filter cutoff frequency characteristics. All
calibration and signal conditioning will be preset within
the unit. It includes an onboard nonvolatile memory for
storage of critical information such as module identifier,
channel setup, transducer calibration data, etc. (In
accordance with IEEE 1451.2, "Standard for a Smart
Transducer Interface for Sensors and ActuatorsTransducer to Microprocessor Communication Protocols
and Transducer Electronic Data Sheet (TEDS) Formats").
4. WIRELESS SENSORS TECHNOLOGY
Wireless sensors are generally divided in 4 categories:
 Active sensors are powered by a battery. They
cannot be chosen for measurements in harsh
environments such as under high humidity and
temperature due to the battery and electronic
components sensitivity.
 Semi active sensors, also called radiofrequency
identification (RFID) are powered by strong ambient
energy such as inductive coupling. They can only
operate at low temperatures.
 Optical sensors can be used to measure temperatures
up to 2000C but they present problems of sensitivity
and dimensions.
 Passive sensors such as acoustic transponders or
wireless electromagnetic resonators do not require a
source of power. The signal backscattering is
performed in a frequency or time domain RADAR
interrogation. They present significant advantages
compared to previous sensing methods and can be
used up to 1300C [6], [7].
Industrial, military, automotive and aerospace
applications require accurate, small size, maintenance free
and low cost sensing systems. MicroElectroMechanical
Systems (MEMS) is the integration of mechanical
elements, sensors, actuators, and electronics on a common
silicon substrate through micro fabrication technology.
MEMS brings together siliconbased microelectronics and
micromachining technology, making possible a smaller,
more highly integrated and lower power transducer than
would have otherwise been possible. Sensor network
nodes are devices that incorporate communications,
processing, sensors and power sources within a small
package. Use MEMS technology enables production of
lowcost, lowpower multifunctional sensors having very
small size and light weight. MEMS technology whereby
micro actuators, microelectronics and other technologies,
can be integrated onto a single microchip. The critical
physical dimensions of MEMS devices can vary from
well below one micron on the lower end of the
dimensional spectrum, to several millimeters.
Picture 3. Triaxial accelerometer with builtin transducer
and battery
The sensor unit (Picture 3 includes a digital signal
controller capable of the following actions:
 Generating a data sample rate clock,
 Initiating data sampling,
 Controlling the A/D converter and axis channel
multiplexer,
 Collecting axis channel data from the A/D converter,
 Implementing an IIR filter algorithm to provide the
required data output sample rate and frequency
response,
 Interfacing with the onboard wireless radio [5].
The sensor unit includes provisions to collect auxiliary
data such as sensor temperature and/or the state of two
digital input and two digital output lines. The state of the
discrete input lines is transmitted back to the controller
unit. The discrete output signals represent the state of
discrete input signals at the wireless controller unit.
The WSNs incorporates transducers, signal conditioning,
an acquisition controller, a processor, power (battery), a
wireless radio, and an antenna into a sealed,
aerodynamically shaped, miniaturized package. It is
intended for external installation on a test vehicle via the
use of an electrocleavable adhesive [4].
The unit also includes onboard antenna tuned to the
Bluetooth band, I/O connector accessible through an
environmentally sealed cover provides access to the
discrete I/O signals and provides access to a switch that is
used to disconnect the battery in order to saves battery
when not in use or access for battery recharging.
4.1. Active wireless sensor with integrated
transducer and battery
MEMSbased accelerometer transducers with capacitors
is typically a structure that uses two capacitors formed by
a moveable plate held between two fixed plates. Under
zero net force the two capacitors are equal but a change in
For WSNs aircraft application can be use two types of
wireless sensors configurations:
148
SOMEASPECTSOFTHEDIFFERENTTYPESWIRELESSSENSORSIMPLEMENTATIONWITHINAIRBORNEFLIGHTTESTCONFIGURATION
force will cause the moveable plate to shift closer to one
of the fixed plates, increasing the capacitance, and further
away from the other fixed reducing that capacitance. This
difference in capacitance is detected and amplified to
produce a voltage proportional to the acceleration. The
dimensions of the structure are of the order of microns.
OTEH 2016
vapor, moisture, and temperature sensors. As a result of
their mass sensitivity, they can be used in numerous
physical, chemical, and biological applications [6], [7].
A basic SAW device consists of two Inter Digital
Transducer (IDT) on a piezoelectric substrate such as
quartz. The input IDT launches and the output IDT
receives the waves. The IDT consists of a series of
interleaved electrodes made of a metal film deposited on a
piezoelectric substrate. The width of the electrodes
usually equals the width of the interelectrode gaps giving
the maximal conversion of electrical to mechanical signal,
and vice versa. The minimal electrode width which is
obtained in industry is around 0.3 m, which determines
the highest frequency of around 3 GHz. If the electrodes
are uniformly spaced, the phase characteristic is a linear
function of frequency, the phase delay is constant in the
appropriate frequency range. This type of the SAW
device is than called delay line and can be used as a delay
line or as band pass filter. Load impedance causes SAW
reflection variations in magnitude and phase (ZL).
For some applications, wireless sensor unit may require
an external transducer and an external antenna
installation. In this application, the only element installed
on the skin of the test vehicle is the antenna. This
application requires the use of available power near the
sensor unit location, which eliminates the need of a
battery. This is very useful in the case where
instrumentation is installed within a pod and data is
transmitted back to the fuselage of the aircraft (Picture 4).
Picture 4. Wireless sensor unit with external transducer
Picture 5. SAW wireless delay line
This unit will be internally installed outside (missile
launcher), has an external triaxial sensor requiring
excitation power from the unit, uses 28 VDC power, and
uses an external antenna. All other building blocks used in
the unit with the internal transducer are used in this unit.
The transducer for use with this unit can be the triaxial
piezoelectric accelerometer. The WCM is installed as
part of host DAU OF airborne FTI a play the same role as
in WSN with builtin transducer and battery.
Sensing with acoustic waves is based on measuring
variations of acoustic propagation velocity of wave, or
wave attenuation. These variations imply changes in wave
properties (frequency for resonators, delay for delay lines)
which can be translated into the corresponding change of
the physical parameter measured.
In the second type of SAW devices SAW resonators,
IDTs are only used as converters of electrical to
mechanical signals, and vice versa, but the amplitude and
phase characteristics are obtained in different ways. In
resonators, the reflections of the wave from either metal
stripes or grooves of small depths are used (Picture 6).
4.2. Passive surface acoustic wave wireless
sensors
Surface acoustic wave sensors (SAW) are a class of
MEMS systems which rely on the modulation of surface
acoustic waves to sense a physical phenomenon. Surface
acoustic wave (SAW) sensors are used for identification
and measuring of physical quantities such as temperature,
pressure, torque, acceleration, tireroad friction, humidity,
etc. They do not need additional power supply for the
sensor elements and may be accessed wirelessly enabling
the use in harsh indoor/outdoor environments (Picture 5).
They have rugged compact structure, outstanding
stability, high sensitivity, low cost, fast real time response
and extremely small size (lightweight). SAW sensors are
used for identification of moving object and parts (so
called ID tags) and wireless measuring different
parameters. SAW wireless sensors are beneficial when
monitoring parameters on moving objects, such as tire
pressure on cars or torque on shafts. Sensors that require
no operating power are highly desirable in remote
locations, making these sensors ideal for remote chemical
Picture 6. SAW wireless resonators sensor
Typical saw wireless sensing systems include:
 a packaged saw sensor connected to an antenna,
 a separate transceiver connected to one or multiple
antennas or wireless control module (WCU) as a part
airborne FTI system.
The radio frequency transceiver (interrogator unit) sends
an electromagnetic pulse. The association of an antenna
149
SOMEASPECTSOFTHEDIFFERENTTYPESWIRELESSSENSORSIMPLEMENTATIONWITHINAIRBORNEFLIGHTTESTCONFIGURATION
with an IDT can be wirelessly interrogated. The pulse is
converted into a surface acoustic wave (SAW) on the
sensor (piezoelectric effect). Properties of the acoustic
wave will be modified under the effect of the physical
parameter which is sensed (e.g. temperature).The SAW
sensor response transmits these modifications back to the
RF transceiver. The measurement signals are contained in
the SAW transponders highfrequency response signal,
which the reader records and evaluates.
OTEH 2016
prototype of aircraft. Major challenges include
miniaturization of hardware, particularly the remote
sensors, power, and selection of the appropriate
communication protocol and operating frequency. The
challenges in the development of the wireless sensor
system for airborne applications are numerous and
required a systematic approach and careful study in many
areas of discipline. These areas include:
 Transducer Selection,
 Sensor power source related issues,
 Wireless Radio and Antenna (Interference with/from
other aircraft electronic systems),
 Line of sight requirements between sensor and
controller,
 Characterizing data latency in the wireless link,
 Frequency Study (The effect of S band telemetry
transmitters upon the wireless link),
 Network Throughput and Data Budget,
 Environmental qualification of hardware (MILSTD810 environmental tests including electrocleavable
adhesive for externally installed wireless sensor
modules),
 Ground tests EMC/EMI qualification of hardware
and software (MILSTD461/462) each airborne FTI
components and aircraft with integrated FTI
equipment in the Anechoic Shielded Chamber and
the Open Field Site,
 WSNs data encryption (security),
 Measurement uncertainty [1], [3], [5].
First, the interrogator generates a f0 carrier frequency RFI
Signal modulated by a short rectangular RF pulse during
aT=20 s time interval. If the interrogator frequency f0 is
equal to the resonance frequency of the resonator a
maximum of energy is stored in the SAW sensor.
When the interrogator stops the transmission after the
pulse of T=20s, it switches to the receive mode and then
is able to measure the received energy from the discharge
of the resonator into the antenna.
SAW wireless sensors have unique features and
unmatched benefits such as:
 operate in severe environments strong electromagnetic fields, high acceleration, high temperature,
explosive, corrosive, radiated environments,
 enable measurements on moving and rotating parts,
 need no battery,
 no cable or wire disposal and has feature infinite
autonomy,
 are small and light (e.g.: 5x5mm  2g),
 large measurement range (low to high temperature
and strain),
 multifunctional measurement (sensor for pressure
and temperature),
 high frequency (from 50 MHz to or 2.45GHz),
 readout distance of several meters (up to 15m),
 advanced digital transceiver easily connectible to
customer systems through standard interface [7], [8].
A series of ground (in the lab and flight line) and flight
tests after integration of airborne FTI subsystem must be
perform in order to evaluate the whole FTI system and
flight test process of concrete flying object start.
References
[1] Filipovic,Z., Stojic,R., Stojic,T., Vujic,D.: Design
and implementation of modern fligh test
instrumentation system for civilian and military
aplication, 4th International scientific conference on
defensive technologies, ISBN 9788681123409,
Belgrade, 67 October, pp. 440446, 2011.
[2] Filipovi,Z., Darag,S.A., Mohammed,D., Vujic,D.:
The sources of mesurement uncertainty related on
aircraft flight testing, 6th International scientific
conference on defensive technologies, ISBN 9788681123713, Belgrade, 910 October, pp. 107113,
2014.
[3] Diarmuid,C.: Wireless Data Acquisition in Flight
Test Networks, ITC 2015,
http://hdl.handle.net/10150/596417
[4] Vujic,D., Stojic,R., Filipovic,Z.: Wireless sensor
networks technology in aircraft structural health
monitoring, 5th International scientific conference on
defensive technologies, ISBN 9788681123584,
Belgrade, 1819, September, pp. 141147, 2012.
[5] Pellarin,S., Musteric,S.: Wireless sensor system for
airborne applications, International Telemetering
Conference Proceedings, 2007,
http://hdl.handle.net/10150/604468.
[6] Hribek,M., Risti,S., Radojkovi,B., Filipovi,Z.:
Typical aeronautics application of SAW WSNs depends
on the Program of testing or Usage Monitoring
programs some aircraft (prototype) sash as:
 structural health monitoring of aisles, slats, fuselage,
turbine blades,
 tire pressure monitoring of landing gear wheels,
 fuel/engine oil/hydraulics level monitoring ,
 fire/overheat detection,
 ice detection,
 temperature monitoring on brakes, inside engine, on
rotor,
 pressure monitoring inside engine, turbine,
compressor,
 critical equipment surveillance (on/off alarm) cockpit
door, cabin doors,
 emission monitoring,
 overload detection.
5. CONCLUSION
This paper describes some aspects of the different types
wireless sensors implementation within airborne FTI
subsystem. Design and implementation of the FTI System
is very important for successful evaluation of the testing
150
SOMEASPECTSOFTHEDIFFERENTTYPESWIRELESSSENSORSIMPLEMENTATIONWITHINAIRBORNEFLIGHTTESTCONFIGURATION
OTEH 2016
Wave Sensors in Mechanical Engineering, FME
Transactions, VOL. 38, No 1, (2010) 1118.
[8] Hribsek,M., Risti,S., Radojkovi,B., Filipovic,Z.:
Modelovanje i projektovanje filtara sa povrinskim
talasima i njihova primena u vojne svrhe,
VTG,VOL. 39, No 2, 2011.
Modelling of chemical surface acoustic wave (saw)
sensors and comparative analysis of new sensing
materials, International journal of numerical
modelling: electronic networks, devices and fields,
Int. J. Numer. Model. 26 (2013) 263274,
(wileyonlinelibrary.com). DOI: 10.1002/jnm.1870.
[7] Stojic,T., Hribsek,M., Filipovic,Z.: Surface Acoustic
151
UAS  FROM MINI TO TACTICAL
ADI COHEN
Aeronautics, NAHAL SNIR 10 ST. Yavne, Israel, +972545500972, adic@aeronauticssys.com
Abstract: Mini UAV's (unmanned aerial vehicles) are growing into scalable surveillance and reconnaissance platforms by
operational range and payloads capabilities. The Aeronautical Orbiter, a compact and lightweight Mini UAV, is presented
in this paper.
Keywords: lightweight mini UAV, Orbiter UAV system, mean features.
Highend payloads for day and night vision
1. INTRODUCTION
Exceptional navigability and safety in harsh
environmental conditions, specifically in strong winds
Aeronautics Ltd is a world leading manufacturer of
unmanned Intelligence, Surveillance, Target Acquisition
and Reconnaissance (ISTAR) solutions. Aeronautics
comprehensive systems provide real time decision making
capabilities such as border protection, maritime Surveillance
and more.
Growth potential allowing heavy payloads and new
extensions
Very low signatures: acoustic, thermal, visual, and radar
3. GRAPHIC OF THE ORBITER FAMILY
Deployed in dozens of countries on five continents, the
worldfamous Aeronautics UAS enable the military,
homeland security and law enforcement customers to make
their countries safer.
3.1. Orbiter 2 Mini UAS
The mature and combatproven Orbiter Mini UAS is part of
Aeronautics Orbiter UAS family. This compact and
lightweigt system is designed for ease of use by the
warfighters and security agents, providing efficient
operational solutions for tactical military and security
applications. The Orbiter 2 platform has proven top
performance and high reliability, providing lifesaving
support in conflict zones worldwide. In the maritime
configuration, Orbiter MUAS provides maritime
surveillance, reconnaissance and target acquisition solution
for small naval vessels operating maritime security and
naval warfare missions.
Since companys establishment in 1997, the Aeronautics
UAS have been the company's flagship products, and have
gained over 200.000 flight hours operational experience.
With the years, Aeronautics has diversified the product
portfolio to other ISTAR solutions, defense electronics and
integrated solutions.
Aeronautics inhouse vertical integration capabilities
facilitate rapid delivery of unique and tailored solutions to
customers, and provide excellent growth potential for
platforms and systems.
Fields of activity are: Unmanned aerial systems, aerial
special mission systems, ground & naval systems,
C4ISTAR solutions, HLS and civil applications.
The system and proprietary control software are compliant
with NATO STANAG such as 4586 and 4609.
2. COMPOSITION OF THE ORBITER
The Aeronautics Orbiter (henceforth the Orbiter) is a
compact and lightweight Mini UAV that provides a
complete Intelligence, Surveillance Target Acquisition and
Reconnaissance (ISTAR) for fielddeployed units. Designed
for use by infantry, artillery, marine and special forces, and
law enforcement organizations, Orbiter is a comprehensive
and flexible solution, ideally suited for covert military
operations, as well as for homeland security missions, peace
support and commercial operations.
The Orbiter MUAV system offers these capabilities:
High endurance
Long operational range
Figure 1. Orbiter 2 EO/IR
152
OTEH2016
UASFROMMINITOTACTICAL
EO camera with laser designation
Photogrammetring mapping
Main Features
Dual Payloads for day/night operation
Cameraguided flight
Rapid deployment, 7minutes to launch
Simple assembly, rapid turnaround
Electrical propulsion
Low silhouette, silent, covert operation
Automatic takeoff and recovery
Operational below cloud base, in harsh weather
conditions
Long endurance, extended operational range
Pneumatic Catapult launch, airbag ¶chute recovery
Mission autonomy, accurate navigation, with or without
GPS or datalink
Control and monitoring from vehicles
Encryped, digital datalink, frequency hopping
3.2. Orbiter 3 Small Tactical UAS
The mature and combatproven Orbiter 3 Small Tactical
Unmanned Aerial System (STUAS) is part of Aeronautics
Orbiter UAS family. Orbiter 3 is designed to deliver top
performance with the lightest and most advanced covert
platform available today. The Orbiter 3 platform has proven
robustness and high reliability, providing lifesaving support
in conflict zones worldwide.
Operational for up to 7 hours, over 100 km from its control
station carrying advanced, multisensor payloads, Orbiter 3
is designed for easeofuse at tactical operations.
The system and proprietary control interface are compliant
with NATO STANAG such as 4586 and 4609.
Figure 3. Orbiter 3 EO/IR cooled
Main Features
Highly transportable vehicle mounted system
Figure 2. Orbiter 2 COMM
Trisensor payloads for day, night operation under clouds
Orbiter 2 specifications
Wingspan: 3.00 m
MTOW: 10.300 kg
Payload weight: 1.50 kg
Maximal speed: 70 kt
Data link: LOS, up to 100 km
Unique SIGINTcapabilities
Rapid development, 7minute to launch
Simple assembly, rapid turnaround
Silent, electrical propulsion
Low silhouette, covert operation
Automatic takeoff and recovery
Advanced Image Processing Capabilities
Automatic video tracker
Video Motion Detector (VMD)
Video mosaic composition
DRoll and image stabilization
Digital zoom and super resolution
H.246 for video streaming
Operational below cloud base, in harsh weather
conditions
Long endurance, extended operational range
Catapult launch, net landing for maritime operation
Mission autonomy, accurate navigation, with or without
GPS or datalink
Control and monitoring from moving vehicles
Maritame
Orbiter 3 specifications
Ready to use kit used on any type of vessel
Wingspan: 4.40 m
No need for flight deck
MTOW: 30 kg
Net landing for maritime operation
Payload weight: 5.50 kg
Payloads
Maximal speed: 70 kt
Stabilized Dual EO/R
Continuous zoom
Data link: LOS, up to 150 km
153
OTEH2016
UASFROMMINITOTACTICAL
Payloads
Stabilized triple sensor: day, night (cooled FLIR) with
laser designation
Continuous zoom
SIGNIT capabilities
Photogrammetring mapping
Figure 5. Orbiter 1K
3.3. Orbiter 4 Small Tactical UAS
Aeronautics introduces the Orbiter 4 Small Tactical
Unmanned Aeral Tactical System (STUAS), part of the
Orbiter UAS family.
With long endurance capabilities Orbiter 4 is designed to
deliver top performance with the lightest, feedable, multimission and most advanced covert platform available today,
for both land and maritime operations.
Figure 4. Orbiter 4 EO/IR cooled
Orbiter 4 specifications
Wingspan: 5.00 m
MTOW: 50 kg
Carryng advanced multisensor payloads, Orbiter 4 is
designed for easeofuse to fit all operational needs.
Payload weight: 12 kg
Endurance: Up to 24 hrs
2 Payloads simultaneosly
Data link: LOS, up to 250 km
Service celling: 18.000 feet
Payloads
Stabilized triple sensor: day, night (cooled FLIR) with
laser designation
Continuous zoom
SIGNIT capabilities
Photogrammetric mapping
Synthetic Aperture Radar (SAR)
Automatic identification system
Figure 6. Orbiter family growth potential
References
[1]
[2]
[3]
[4]
[5]
154
www.aeronauticssys.com
www.commtact.co.il
Orbiter 2 Mini UAS Broshure
Orbiter 3 Small Tactical UAS Broshure
Orbiter 4 Small Tactical UAS Broshure
SECTION III
Weapon Systems and Combat Vehicles
CHAIRMAN
Professor Momilo Milinovi, PhD
Professor Dejan Mickovi, PhD
A PRELIMINARY DESIGN MODEL FOR EXPLOSIVELY FORMED
PROJECTILES
MOHAMMED AMINE BOULAHLIB
Military Academy, University of Defence, Belgrade, email: m.a.boulahlib@gmail.com
MILO MARKOVI
University of Belgrade, Faculty of Mechanical Engineering Weapon Systems Department, mdmarkovic@mas.bg.ac.rs
SLOBODAN JARAMAZ
University of Belgrade, Faculty of Mechanical Engineering Weapon Systems Department, sjaramaz@mas.bg.ac.rs.
MOMILO MILINOVI
University of Belgrade, Faculty of Mechanical Engineering Weapon Systems Department, mmilinovic@mas.bg.ac.rs.
MOURAD BENDJABALLAH
Military Academy, University of Defence, Belgrade, mourad.bendjaballa@hotmail.com
Abstract: The current paper proposes analytical approaches implemented in a performancecalculation program of
Explosively Formed Projectiles (EFP). The proposed analytical methods, mathematically describe the EFP forming
process aiming to optimize the initial phase of the EFP warhead design. A mathematical model, based on the wellknown theoretical approaches, is accomplished and implemented in a software. The developed software provides faster
analysis of EFP design process and the possibility to test new EFP configurations, in addition to the performances of
already existing ones. The adopted model is tested and validated for several types of EFP warheads according to
available experimental reports. Programs output results such as initial velocity, kinetic energy, axial and radial
deformation energies of liners, are compared with experimental data.
Keywords: explosively formed projectiles, analytical method, physics of explosion, performances software.
number of influencing parameters for this type of
warheads, as the explosive charge, liner form, materials
and cases, are adopted in the analytical method and
implemented in an algorithm for comparative analyses.
Otherwise, the algorithm offers the possibility to directly
export the adopted geometry of EFP warheads into
Autodyn numerical software, from the software package
Matlab, which considerably decreases preparation time.
1. INTRODUCTION
In the field of modeling and warheads' design, based on
EFP principles, only few papers have been written
containing analytical approaches to define the forming
process of projectiles [14]. Recently, most papers are
based on numerical approaches [510] which determine
the performances and gives detailed models of the
forming process. However, numerical software,
particularly Autodyn, which are often used for detailed
analyses in numerical approaches, require comprehensive
preparation of the expected warhead design, in addition to
their longlasting process of calculation.
This paper aims to present an improved analytical method
for the EFP performances, based on the integration of two
analytical models presented in papers [1, 2]. The first
model [1] is based on the active explosive charge masses,
which correspond to the charges in grid elements of
explosively propelled liner. The result of this model is to
integrate the explosively process in the initial EFP
velocity calculation. The second model [2] is based on
Gurneys method [11], for the final EFP velocity
estimations. This method particularly uses axial and radial
direction approach of the active explosive charge masses
for each element in liner's grid. These final estimations of
radial and axial plastic energy distributions, as stated in
the paper [2], present invariant expressions to be used for
the form estimations of EFP in the initial phase. These
expressions for radial and axial distributed energies from
the second method are implemented in the first one, in
order to provide two main performances of EFP, one is
Nowadays, the EFP warheads are present in many
systems that expect appropriate modernization and/or
optimization; as artillery submunition, antitank missiles,
mines etc. The current paper presents a software as a
solution to provide the ability to preliminarily design the
warheads, in addition to a further ability to analyze the
adopted design by more precise numerical software,
particularly Autodyn.
The research presented in this paper as well as in papers
[15], provide crucial information about the EFP
performances in a short time without the time requiring
comprehensive preparation. In the current paper, a large
157
OTEH2016
APRELIMINARYDESIGNMODELFOREXPLOSIVELYFORMEDPROJECTILES
the EFP velocity and second is the deformation energy
distribution, as the projectile form factor.
mai =
The results of these analytical methods and their
implementation in the algorithm contribute in improving
the accuracy of EFP velocity estimations. This is achieved
by an appropriate augmentation in the number of the grid
elements.
M Ki are the masses of metal cover segments. The idea in
this paper is to verify the analytical method implemented
in software for several types and geometries of EFP, and
compare velocity results with the available experimental
results. The data used for simulation software are
provided in papers [2, 8, 12, 13] and taken as
experimental base given in the appendix (table 3). The
geometry of the considered warheads is shown in fig.1.
The adopted analytical model is based on explosive
masses, energy and momentum balance equations to
estimate initial EFP velocity. By the liners partition in
elementary grid and by accepting that the detonation
pressure of explosive products attacks each particular
element on the grid, the impulses and momentum
exchanges, in addition to the final liners velocity can be
summarized [1]. The semispherical liner is divided into n
observed elements that start from its axis of symmetry. In
addition, full cylindrical explosive charge volume is also
divided in the same way. Each element of liner
corresponds to the amount of explosive segment
orientated normal to the surface of the liner and located
above it (figure 1). The initial velocities of these elements
depend not only on their position on the liner, but also on
liners geometry. For further analysis, the following
assumptions are accepted:
Detonation products attack metal liner immediately.
The motion of each discrete element of metal liner is
along the radius of liner and there are no colliding
effects between grid elements.
The constant compressive strain rate along axis is
0zi = const .
Figure 1. Adopted geometry of EFP warhead for
analytical analyses; 1 liner, 2 explosive charge, 3 case,
4 back plate; Input parameters: Dcaliber, Llength of
charge, lstarting cone position, 1thickness of liner
center, 2thickness of liner edge, 3thickness of cover,
4=5thickness of back plate, angle of cone, R1 outer
radius, R2inner radius
The final velocity of the EFP is performed by integrating
all absolute velocities of liner elements from the equation
(1), by momentum conservation law. It is given by the
following expression [1, 2 and 11]:
Using previous assumptions and energy balance equation
in detonation process, the velocity of a particular ejected
element on the liners grid, V0i [1] is equal to:
1 3i ; i = 1, 2,..., n
k 2 1 3 + i
(1)
0i M i
V0 E =
Where i = mai / M i , is the loading factor of ith liners
; i = 1, 2,..., n .
(4)
The differences between kinetic energies of elements
correspond to the plastic deformation work along zaxis
[1, 5, 11] as the relative motion towards liners mass
center. The expression that represents the energy of axial
deformation work is [2]:
The active mass of the explosive mai in the loading
factor, in equation (1), is a fictive explosive mass that
reproduces all effects of energies made by real explosive
masses and covers. Corresponding expression for this
mass is [1],
mi
M Ki M i
(1 +
) ; i = 1, 2,..., n ,
2
M i + M Ki + mi
i =1
n
i =1
element, mai  active mass of corresponding explosive
segment, M i  mass of liners ith segment, D
detonation velocity and k polytrophic coefficient of
detonation products, usually taken as k=3 [1].
mai =
(3)
used in the cases with (2) or without (3) warhead covers.
Values mi are the masses of explosive segments and
2. ADOPTED ANALYTICAL METHOD AND
ALGORITHM
V0i = D
mi 2
for M Ki = 0
2( M i + mi )
n
ADE = 1
M i (V0i cos ) 2
2
i =1
i = 1, 2,..., n
(2)
M V
i 0E
i =1
(5)
Besides, the second method considers the energy of radial
deformation work which corresponds to the part of kinetic
energy created by radial displacement of the elements.
This radial deformation work is presented by the radial
velocity values of each element, V0i sin i [1, 5, 11] and it
or
158
OTEH2016
APRELIMINARYDESIGNMODELFOREXPLOSIVELYFORMEDPROJECTILES
The algorithm offers the possibility to choose various
EFP configurations by varying the geometries and the
used materials, in addition to the number of segments to
be used, as presented by the block (a) fig. 2. After
approving the 3D visual model of the EFP by the designer
(block (b)), the algorithm computes the masses and the
volumes of each segment referring to the Guldins
theorem and based on the inputs.
is given in [2] as follows:
RDE = 1
2
M (V
i
0i
sin )2 ; i = 1, 2,..., n .
(6)
i =1
The algorithm, shown as a flow chart in figure 2,
significantly contributes in analyzing and evaluating the
affecting parameters on EFP main performances. Using
variations in inputs, the algorithm provides enough
precise output data such as axial initial velocity of EFP
liner as well as kinetic energies distributed in axial and
radial directions.
The volumes of segments for the liner the explosive as
well as the case are calculated in block (c) following the
next steps: firstly, the area of each segment is calculated
by double integral, where the intervals of integration are
defined by the segment boundaries. Secondly, the
distance traveled by each segment is computed in function
of the coordinate of the centroid and the angle of
revolution, which is equal to 2. Finally, the volume of
each segment, which is generated by its rotation about an
axis of revolution, is equal to the product of its area and
the distance traveled by its geometric centroid.
After that, the algorithm calculates the absolute velocity
of each segment using equation (1), in block (e) and
controlled by block (d). Then, the initial velocity of the
configuration as well as the kinetic energies distributed in
axial and radial directions are calculated by equations (4),
(5) and (6) respectively in block (f). Finally, the designer
has the ability to export the configuration into a numerical
software (Autodyn) for more precise study, in block (k).
In case that the results do not respect the system
requirements, the simulation process can be reinitiated
with new inputs.
4. RESULTS AND DISCUSSIONS
In order to investigate the model limitations and to
increase its precision, five simulations are carried on
according to configurations given in [5, 8 and 13], with a
variation in the number of segments.
The obtained results in the table 1 show that the initial
velocity for each configuration converges to a lower
asymptotic bound with the increase in the number of the
segments. The threshold number of segments used for
improving precision is about 70. The velocity beyond that
number reaches its asymptotic minimum and remains
unchangeable with the increasing number of grid
elements.
To validate the analytical model, the simulation results
are compared according to available experimental results
for known geometries of EFP warheads given in [2, 8, 12,
13]. The defined geometries and the characteristics of the
used materials are provided on Table 3 in the appendix.
The experimental values of the velocity of the EFP
warheads, together with the analytically calculated ones
are given on the figures 4 and 5, for each experiment.
Figure 4 represents a comparative analysis of
experimental and analytical results for 8 experiments. For
each experiment, the relative error of the analytical
method compared to the experimental data is expressed as
a percentage. From the presented results, obviously the
relative error is defined from 0.15% to 16.89%, and its
mean value is equal to 10.34%.
Figure 2. Algorithm for calculating EFP performances
159
OTEH2016
APRELIMINARYDESIGNMODELFOREXPLOSIVELYFORMEDPROJECTILES
Table 1. Convergence of selected results for adopted analytical method
N
10
30
50
70
90
110
Units
m/s
m/s
m/s
m/s
m/s
m/s
V0 E 1
2801.23
2787.01
2785.81
2785.48
2785.34
2785.3
V0 E 2
2168.17
2146.51
2143.11
2142.86
2142.58 2142.21
V0 E 3
2052.72
2032.52
2029.97
2029.14
2028.88 2028.55
V0 E 4
2079
2043.41
2042.15
2040.58
2040.29 2040.14
V0 E 5
2178.77
2155.06
2153.75
2152.03
2151.73 2151.73
On the figure 5, the comparative results of 6 experiments
derived with the same geometry but with different case
width, are presented. The results show quite small values
of the relative error that varies from 2.42% to 14.12%,
with a mean of errors equal to 8.2%.
The preliminary analytical results, shown in Figures 4 and 5,
are obtained based on the assumption that the process of
formation is completely ideal, whilst in reality, this is
unlikely to happen. For that reason, a correction factor Kc is
introduced to precisely take into account all the
imperfections in the experimental data. It is inserted as an
additional member in the equation (4), which has the task to
correct the value of the velocity calculated by the analytical
model. Then the equation (4) becomes as follows:
experiments vs. analytical 1
4000
experiment
analytical method
3500
14.97%
3000
0i M i
16.81%
velocity [m/s]
12.5%
2500
3.3%
6.88%
V0 E (c) =
11.21%
0.15%
K c i =1n
2000
; i = 1, 2,..., n .
(7)
i =1
16.89%
1500
With a specific correction factor, new obtained results are
presented in Figures 6 and 7. To review the result in a
compact way, the diagrams are organized so that the
previous preliminary results are presented together with
the new adjusted results given by equation (7).
1000
500
0
3
4
5
6
number of experiments
8
experiments vs. analytical 1
4000
Figure 4. Experiments vs. analytical results, where each
number of experiment is taken, for No.1, No.2, No.3,
No.4 and No.5 from [8], No.6 from [5], No.7 and No.8
from [14]
3500
experiment
18.37%
analytical
analytical with correction
3000
8.09%
12.14%
2500
2000
experiment
analytical method
5.74%
7.82%
9.78%
11.15%
velocity [m/s]
changing case experiments vs. analytical 2
12.14%
2.42%
2500
7.24%
2.6%
14.76%
4.14%
2000
1500
12.21%
velocity [m/s]
1000
1500
500
0
1000
3
4
5
6
number of experiments
Figure 6. Experiments vs. analytical results with
correction factor, where each number of experiment
corresponds to the one on figure 4
500
2
3
4
number of experiments
From Figure 6, seemingly and without doubt, the new
velocities are slightly lower compared to previously
obtained ones. The relative error of the adjusted analytical
model compared to the experimental data varies from
2.6% to 18.37%, with mean value of 9.94%.
Figure 5. Experiments vs. analytical results, where each
number of experiment is taken, for No.1, No.2, No.3,
No.4, No.5 and No.6 from [13]
160
OTEH2016
APRELIMINARYDESIGNMODELFOREXPLOSIVELYFORMEDPROJECTILES
show relationship between the errors in estimating the
velocities and coefficients kDE and Kc.
changing case experiments vs. analytical 2
2500
velocity [m/s]
2000
1.51%
3.51%
5.39%
6.7%
Table 2. Axial and radial deformation energy as well as
ratio coefficient
7.66%
1.67%
No.1
No.2
No.3
No.4
No.5
No.6
No.1
No.2
No.3
No.4
No.5
No.6
1500
1000
500
experiment
analytical method
analytical with correction
1
2
3
4
number of experiments
Figure 7. Experiments vs. analytical results with
correction factor, where each number of experiment
corresponds to the one on figure 5
Fig. 4.
Fig. 4.
Fig. 4.
Fig. 4.
Fig. 4.
Fig. 4.
Fig. 5.
Fig. 5.
Fig. 5.
Fig. 5.
Fig. 5.
Fig. 5.
ADE [J]
42259.62
33619.82
31495.86
31515.08
15157.18
49456.26
76846.92
77981.81
79818.56
81620.08
83251.63
84701.13
RDE [J]
30524.52
26849.15
24238.58
27256.37
25577.2
15676.72
46619.98
53368.5
57658.22
60652.4
62872.33
64588.93
kDE
1.38
1.25
1.30
1.17
0.60
3.15
1.65
1.46
1.38
1.35
1.32
1.31
5. CONCLUSION
The analytical method does not take into account the
imperfections in the formation process, thus referring only
to this fact lead the losses, so the analytical method
generally provides higher velocity values compared to the
experimental ones. Some experiments, though, as No.2,
No.5 and No.8 in Figure 4 show greater values of the
measured velocities than the analytically obtained ones,
which gives greater relative error as a percentage in figure
6. This illogical results are due to a lack of information on
the characteristics of the materials used, especially
explosives, as well as certain dimensions. In order to
make a comparison, the values that are missing have been
approximated which induces in some cases to illogical
results, which contributes to enlarge the relative error.
Although some results are illogical with greater error,
they do not exceed a reasonable limit, no more than 20%,
which is quite acceptable for analytical models.
Based on the research and the results obtained in the
current work, it can be concluded the following:
The analytical method implemented in the program
aims to increase the accuracy of the results by
enlarging the number of the used elements. It is
obtained that the convergence of the method is
achieved for n = 70 elements taking into account the
14 observed configuration, where for n> 70 no
considerable increase in the accuracy of the results can
be provided.
By using all the 14 experiments, the mean value of the
relative error of the analytical method without
correction factor is 9.41%, while using a correction
factor decreases the mean error to 7.57%.
Generally, the adopted analytical method gives
acceptable results referring to the experimental data.
Further research can be orientated to the influence of
the axial and radial energy distribution on the
projectile form which can help designing new semiempirical method for more precise preliminary
estimations of EFP parameters.
The corrected analytical model shown in Figure 7
provides a slightly better results of the velocities
compared to the previous model, in the range of 1.67% to
7.66%, with a mean value of errors equal to 4.1%.
To better understand the forming process, the energy of
axial and radial deformation work is calculated by eq. (5)
and eq. (6). The results are presented in Table 2, these are
the axial ADE and radial RDE deformation energy in
addition to their ratio coefficient kDE. This coefficient
determines the expected final projectile form; it varies
generally between 1 to 1.7. For materials as aluminum
this coefficient is less than 1, while for experiments
presenting lager liner radius (experiment No.6), kDE is
extremely large, more than 3. Further researches have to
Appendix
The presented results are obtained according to available
experimental results given in [2, 8, 12, 13], for known
geometries of EFP warheads. For each type of warheads,
the geometric parameters that present the input data for
the program are defined in Figure 1. In addition to the
defined geometry, the characteristics of the used materials
are also provided in table 3.
161
OTEH2016
APRELIMINARYDESIGNMODELFOREXPLOSIVELYFORMEDPROJECTILES
Table 3. Dimensions and materials of the selected experiments defined by figure 1
ACKNOWLEDGEMENTS
[7] Pappu, S, Murr, L.E.: 2002. Hydrocode and
microstructural analysis of explosively formed
penetrators, Journal of materials science, Vol.37, pp.
233248.
[8] Hussain, G., Hameed, A., Hetheringtom, J.G.: 2013.
Analytical performance study of explosively formed
projectile, Journal of applied mechanics and
technical physics, Vol.54, No.1, pp.1020.
[9] Fedorov, S.V., Bayanova Ya. M., Ladov, S.V.: 2015.
Numerical analysis of the effect of the geometric
parameters of a combined shapedcharge liner on
the mass and velocity of explosively formed compact
elements, Combustion, Explosion, and Shock Wave,
Vol. 51, No. 1, pp. 130142.
[10] Hussain, G., Malik, A.Q., Hameed, A.: 2011.
Gradient Valued Profiles and L/D Ratio of Al EFP
With Modified Johnson Cook Model, Journal of
Materials Science and Engineering 5, pp. 599604.
[11] Bender, D., Corleone, J. 1993. Tactical Missile
Warheads Explosively Formed Projectile,
American Institute of Aeronautic and Astronautic,
Washington.
[12] Teng, T.L., Chu, Y.A., Chang, f.a., Shen, B.C.:
2007. Design and Implementation of a High Velocity
Projectile Generator, Fizika Goreniya I Vzryva, Vol.
43, No. 2, pp. 233240.
[13] Weimann, K.: 1993. Research and development in
the area of explosively formed projectiles charge
technology, Propellants, Explosives, Pyrotechnics,
Vol. 18, pp. 294298.
[14] ivkov M.: 1984. Forming process of Explosively
Formed Projectile With Small Explosive Caliber and
Applications in Penetrating The Armor, Symposium
on Explosive Materials, pp.251262, Uice.
This research has been supported by the Ministry of
Education, Science and Technological Development of
the Republic of Serbia, through the project III47029, in
the 2016 year, which is gratefully acknowledged.
References
[1] Orlenko, L.P.: 2004. Fizika vzriva, (in Russian)
Glavnaja redakcija fizicesko matematiceskoi
literaturi, Moskva.
[2] Sharma, VK, Kishore, P., Bhattacharyya, AR,
Raychaudhuri, TK, Singh S.: An Analytical
Approach for Modeling EFP Formation and
Estimation of Confinement on Velocity, 16th
International Symposium of Ballistics, San
Francisco, 1996, pp.565574.
[3] Markovi, M.: 2011. Explosively Formed Projectiles,
MSc Thesis, (in Serbian), University in Belgrade,
Mechanical
Engineering,
Weapon
Systems
Department.
[4] Markovic, M., Milinovic, M., Elek, P., Jaramaz, S.,
Mickovic, D.: 2014. Comparative approaches to the
modeling of explosively formed projectiles,
Proceedings of Tomsk State University, Serie
Physics and Mathematics,V.293.
[5] Markovi, M., Elek, P., Jaramaz, S. Milinovi, M.,
Mickovi, D.: 2014. Numerical and analytical
approach to the modeling of explosively formed
projectiles, 6th International Scientific Conference
OTEH 2014, October Belgrade, pp. 910.
[6] Markovic M., Milinovic M., Jeremic O., Jaramaz S.:
2015. Numerical modeling of temperature field on
high velocitiy explosively formed projectile, 17th
Symposium On Thermal Science And Engineering
Of Serbia, October 2023, pp.175181.
162
TENDENCIES OF DEVELOPMENT OF AMPHIBIOUS ASSETS
IN ARMED FORCES OF NATO COUNTRIES
NENAD KOVAEVI
Military academy, Belgrade, www.inz.84kula@gmail.com
NENAD DIMITRIJEVI
Military academy, Belgrade, www.neshadim@mts.rs
Abstract: The paper presents a brief overview of modern developments and directions of further development of
amphibious means of overcoming water obstacles in armed forces in the countries of North Atlantic Treaty
Organization (NATO). In preparing the paper was a big problem later literature, and are therefore largely used data
from the Internet. Knowing amphibious assets of overcoming water obstacles in armed forces of NATO countries can be
seen in the purposeful way the effects of their use, and the need for innovation and investment in their resources.
Keywords: amphibious assets, armed forces, literature.
In Armed forces (AF) of NATO countries use two types
of resources for overcoming water obstacles, namely: an
establishment and meat assets. Under meat assets means
assets that are not in the formation of units (units nor
carry with them) are already in place or in the zone,
district, crossing water obstacles in social or private
property, and have the same purpose as the military posts
resources.
1. INTRODUCTION
The phrase "place crossing water obstacles", in principle,
we call a part of a certain type of water hazard, coast and
hinterland on its own and the opposite bank, to be used in
order to be able to provide a smooth and continuous
supply of people and moving vehicles of all categories,
over any kind of water barriers (rivers, canals, lakes). The
room size for a place crossing is different, considering the
type of crossing organized, conditions of access and exit
roads to the river, forestation and conditions coast
masking units, the number of funds to be crossing site to
cross the river and a number of other factors.
In an establishment of water assets for overcoming
obstacles to the AF of NATO countries are:
amphibious transporters,
landing crafts,
scaffolds (light scaffolding, heavy selfpropelled
amphibious staging, the staging of the pontoon parks),
In order to successfully overcoming water obstacles, they
are familiar with the characteristics that affect the
planning, organization and execution of the procedure.
These characteristics are: width, depth, flow velocity,
slope, coast, composition floor, height coast, access roads
and wooded shores. In tactical terms, the expression
crossing over the river means the overcoming of the river
by the units of the armed forces. For this, we should
distinguish between two ways of overcoming the river,
namely:
hovercrafts,
mechanized bridges (heavy and light),
sets pontoon parks,
folding bridges (,,Bailey") and
varnishes pedestrian bridges.
The work presents a systematization of existing, and the
tendency of further development of amphibious assets in
AF of NATO countries. The first part deals with the
historical genesis of the use of amphibious
assets. Systematization of amphibious assets in AF of
NATO countries is given in the second part, and the
directions of further development of amphibious assets in
AF of NATO countries discussed in the third part of the
paper.
crossing a river when the dam opposite shore is not the
enemy forces and
violent crossing rivers (forcing the river) when the
opposite shore defending enemy.
Throughout history, armed forces have crossed the water
hazard that could joined in battle or retreated.
Overcoming barriers to water as an element of engineer
operations and counteract in modern warfare has retained
the title of one of the most difficult tactical actions,
despite the fact that they invested enormous financial
resources in the development of equipment and resources
to perform the same.
2. HISTORICAL ORIGINS OF THE USE OF
AMPHIBIOUS ASSETS
The first known selfpropelled, amphibious vehicle
powered with steam, presented the American inventor
163
TENDENCIESOFDEVELOPMENTOFAMPHIBIOUSASSETSINARMEDFORCESOFNATOCOUNTRIES
Oliver Evans (Oliver Evans), 1805, under the name
,,Orukter Amphibolos". A year later the French combine
structural elements of the ship and passenger vehicles and
create a precursor to modern amphibious vehicles.
However, at that time, designers are not adequately
understood, however, for further development and
improvement of amphibious assets had to wait more than
a century.
OTEH2016
On the one hand, the countries of NATO (led by the
USA), the focus in the development of these resources to
focus on the development of amphibious characteristics of
military equipment and the development of assault
bridges.
On the other hand, the countries of the Warsaw Pact (led
by the USSR), the focus in the development of resources
for overcoming water obstacles are mainly focus on the
development of pontoon bridges and amphibious assets of
transport weapons and combat system.[2]
The importance of rapid and safe water obstacles to
overcome first realized the Germans, and for this purpose
to construct the first true amphibious vehicle in the world
in general, just before the Second World War. It was to
,,volkswagen" floating vehicle (VWSchwimmwagen),
made on the Volkswagen chassis civilian passenger
vehicles.
3. EXISTING AMPHIBIOUS ASSETS IN AF OF
NATO COUNTRIES
Vehicles of this type were equipped with SS troops and was
made entirely mechanical. Basic tactical and technical data
vehicle: weight: 1362 kg, length: 3,825 m, height: 1,615 m,
speed on the water: 10 km/h speed on dry: 80 km/h, the
crew: 4 soldiers, armed with machine gun 7.92 mm MG
34.[1] The vehicle is shown in Picture 1.
In order to better understanding of progress in terms of
development and innovation on the amphibious assets in
AF of NATO countries, first it is necessary to
systematization of the existing ones. Accordingly
structural solutions, amphibious assets are divided into:
amphibious combat vehicles,
engineering resources and amphibians
floater.
Combat amphibious vehicles are amphibious assets whose
primary purpose is to perform combat tasks, and with it
the amphibians, the most characteristic representative of
the amphibious infantry combat vehicles. These assets
have the advantage in a violent river crossing in the forces
of the first wave primarily due to the protection provided
by the units in a violent crossing the river and their
autonomy in movement and the possibility of taking the
fight out of the vehicle and create a bridgehead on
enemy`s coast.
Picture 1. ,,Volkswagen" floating vehicle
Engineering amphibious assets are whose main purpose is
to combat systems unit and a water obstacle to overcome
in a short period of time, the representatives of these types
of amphibians agents are: amphibious scaffolding and
conveyors. This type of funding is primarily intended for
the transport of all types of tanks and selfpropelled
weapons and vehicles, but also the people and personal
property in order to provide units that are solving certain
water hazard.[2]
In addition to the amphibious means the then Germany, in
order to prepare for the conduct ,,blitzkrieg" (a rapid pace
of war), the weapons introduced a whole series of new
means of overcoming water obstacles, starting with the
pneumatic and assault boats through to tanks, amphibious
transporter beams bridge and floating tanks. Faced with
the problem of coping with the large rivers in the USSR,
Wehrmacht forces have also attached great importance to
the modernization and development of pontoon bridges.
From sets of bridge load capacities from 3000 to 5000 kg,
with the beginning of the war, the weapons have already
been introduced in 1943, sets the bridge, whose capacity
enabled the successful transfer across water obstacles
almost all military equipment. [2]
Hovercraft are assets that are in their infancy, and they
actually reflected the future in terms of resources for
overcoming water obstacles. So far there are only arms
USA.
AF of NATO countries leading in the development and
specialization of research of all types of amphibious
assets. Primat in this sphere of development and
modernization of arms and military equipment are run by
the armed forces of the USA and the United Kingdom
(UK). Consequently all the other members of NATO
amphibious assets use models who are actually a modified
version of the funds that are used in the above mentioned
countries. [3]
The exponential development of science and technology
has always been subordinate to military objectives and
needs. By the end of the eighties of the 20th century was
dictated by the US and the USSR, as leaders of the two
militarypolitical blocs (NATO and the Warsaw Pact). As
a result, in this period, and runs into two different
approaches to solving the problem of overcoming water
obstacles.
164
TENDENCIESOFDEVELOPMENTOFAMPHIBIOUSASSETSINARMEDFORCESOFNATOCOUNTRIES
Specifications of the transporter: crew: 2 + 10 men, armed
with a machine gun 7.62 mm, weight: 19000 kg, the
relationship between engine power and vehicle weight
12.47 kW/t, length: 6.83 m, width 2.98 m, height: 2.30 m,
maximum speed: onshore 105 km/h, water 10 km/h.
The division of the types and designs of certain
amphibians agents that are in operational use in AF of
NATO countries, is given in the table 1.
Table 1. List of amphibious assets in NATO lands
No.
Armed
Forces
1.
USA
2.
UK
3.
French
4.
Germany
5.
Turkey
Types of amphibious assets
Combat
Engeener
Hovercraft
amphibious
amphibious
vehicles
assets
Models
Bradley M2/M3,
LAV 25, BV
206S,
AAV7A1,
LAV, LCACFuchs M93A1
LVT7
27, LCAC74
Fox,
Mowag Piranha I
8x8
FV 432 ARS, BV
206S, Fuchs
Viking BVS 10
LAV
M93A1 Fox
VAB ARS,
EFA,
LCAC27,
Panhard VBR,
GilloisEWK
LCAC74
AMX 10RC,
BV 206S, Luchs,
Fuchs M93A1
Fox, Fuchs
M2 i M3
M93A2, Fuchs
Panzer
AIFV, ACV S
GilloisEWK
OTEH2016
3.2. Amfibious armored vehicles ,,Viking
(BVS10)
,,Viking" is an amphibious all terrain vehicle. Is a
modified version of an armored personnel carrier ,,BV
206S. It's actually a kit consisting of two vehicles in
which is included a system for steering. It is located
within the Marine commandos UK. Shielded the steel
construction. The construction is made to reduce radar
reflection. Armor cant break a grain of 7.62 mm from a
direct hit, or shrapnel from artillery projectiles caliber 152
mm that fell no closer than 10 m from the vehicle. He
cant damage it by pressure on a mine, which in itself is
no more than 0.5 kg of explosives. [5]
In the following is presented by a representative of all
types of amphibious assets.
3.1. Armed transporter ,,Fuchs M93A1 Fox
German factory ,,Henschel Defense Systems", developed
in 1960 by ,,Fuchs for the purpose AF former West
Germany. Between 1979 and 1986 produced 996 of these
transporters. Since then, production has continued to
foreign markets, mainly in variants for special purposes,
such as for NBC reconnaissance. During the Gulf War
(1990  1991), the same version of the vehicle was
delivered to Israel, UK and USA. [4]
Picture 3. amphibious armored vehicle ,,Viking
(BVS10)
Of the weapons in possession of heavy machine gun 12.7
mm caliber machine gun 7.62 mm and antitank system
,,Milan". ,,Viking can overcome a water obstacle to the
depth of 1.5 m without preparation to sail. The crew
consists of a driver and 3 equipped marines, with a second
cabin can accommodate 8 fully equipped Marines. [1] A
vehicle is shown in Picture 3.
Transporter is sailing with two propellers. Standard
equipment includes a system for NBC protection and
passive equipment for night vision. A more modern
version of the vehicle ,,Fuchs is in almost all the units of
AF of the Federal Republic of Germany and is equipped
with stronger armor and a 7.62 mm machine gun mounted
on the front of the roof carriers. [1] Only the conveyor is
provided in picture 2.
Tactical and technical characteristics: combat weight:
10600 kg, the maximum speed on the mainland: 50 km/h,
the maximum speed for navigation: 5 km/h, driveline:
excavator, drive 4 tracks, maximum autonomy of 300 km.
3.3. Hovercraft
Hovercraft are assets of the latest technology and have
found wide application in overcoming water obstacles.
There were the seventies of the 20th century, during the
war in Southeast Asia. Then they are used only in the
coastal part of the sea and large river basins. The basic
principle that work is the movement of air cushion.
Tactical and technical features: displacement (full):
535000 kg, the crew of 27 people, length: 56.2 m, width:
22.3 m, the speed on the road: 85 km/h, the speed on the
water: 60 knots, engine: gas turbine [6]. Hovercraft set is
shown in Picture 4.
Picture 2. Armored Fighting Vehicle
,,Fuchs M93A1 Fox
165
TENDENCIESOFDEVELOPMENTOFAMPHIBIOUSASSETSINARMEDFORCESOFNATOCOUNTRIES
OTEH2016
on the development and a new type hovercraft  ,,armored
hovercraft" that are primarily intended for the transport of
the first combat echelon unit during execution desanta sa
broda na obalu. The planned maximum speed on land
would be around 100 km/h, and water up to 120 km/h.
Appearance hovercraft ,,LAV" is shown in Picture 5.
Picture 4. Hovercraft
4. DIRECTIONS OF FURTHER
DEVELOPMENT OF AMPHIBIOUS
ASSETS IN AF OF NATO COUNTRIES
Picture 5. USA and UK hovercraft
The asset would run strong motors that would ensure
floater elevation above the surface. This asset would have
another important characteristic, which is the ability to
overcome obstacles mines hovering above them, which is
very convenient, in terms of maintaining the pace of
descent.
In view of the further development of amphibious means,
OS NATO countries are working on two fronts:
modifications (modernization) of existing assets and
development of new types of hovercraft.
The amphibious assets, although proved to be very
efficient and reliable, modern defense industry the
majority of AF in NATO countries are planning to be in
the next 20 to 30 years in part (in some countries
associations, planned and completely) withdrawn from
operational use and to take their place floater. However,
some NATO countries, according to the characteristics of
their territory and waters, and still attach great importance
to the development of amphibious means, especially for
the transport of combat systems.
AF of NATO countries give great importance to the
development of resources for overcoming water obstacles.
The further development of these are aimed at the
construcamphibious assets with a focus on meeting the
following conditions:
Similarly, developing AF of Russian Federation with
amphibious vehicles from the eighties of the 20th century,
the BAZ ( )5921 i
BAZ5922, where the platforms of these vehicles placed
missile systems of high precision of target 9K79 ,,Point"
(constructor Sergey Pavlovich Nepobedimov) with
launchers to launch one and two rockets, according to
NATO codification system known as the SS21 ,,Scarab".
Modernisation of these types of amphibious assets started
in 2007, the most has been done in terms of making armor
amphibian plastic and composite materials with the intent
to significantly reduce the ingress of water into the body
of amphibians in case it gets affected. [7]
to meet the requirement in terms of maintenance and
operation.
the necessary resources that range great speeds,
they have low weight,
to have the chance to capacity,
they were shielded and
5. CONCLUSION
Based on the above findings, we conclude that the
tendency of further development of amphibious assets in
AF of NATO countries will be focused on the use of new
lightweight materials in the construction of amphibious
means of overcoming water obstacles. The materials must
be lightweight, but also sturdy and able to withstand high
loads, and to provide protection to crews, commonly used
composite materials. From amphibious assets primarily
seeks to be folded out of scaffolding and bridges different
capacity with two or more of crossing roads.
As we already said hovercraft are assets of the latest
technology, but the main problem of the first hovercraft
was that they were very slow and had the possibility of
small versus large capacity of its own weight. Then they
are used only in the coastal part of the sea and river basins
multiple.
However, the designers of amphibious assets in NATO
countries remains unresolved problem to find a way to
combat parts of joint units do independent of the
amphibious assets and pontoon materials and to move
under its own power over water obstacles. On this basis in
the coming years special place in the further development
of military equipment intended for innovation in terms of
constructing light and heavy armored amphibious
personnel carriers, selfpropelled antitank weapons and
amphibious tanks capable of underwater driving. This
Data on new types of hovercraft US Armed Forces have
not yet been published, so it is about them can be found
only images. It is only known to be already working on a
modernization of existing assets  hovercraft type
,,LAV. Also, AF of USA and UK, in addition to the
already existing cooperative model, is currently working
166
TENDENCIESOFDEVELOPMENTOFAMPHIBIOUSASSETSINARMEDFORCESOFNATOCOUNTRIES
funding will also allow: a fast passage across water
barriers, protect the people, but will also have the
possibility of an active and strong effect at the enemy.
OTEH2016
[3] Milojevi,D.: Tendecies of development of river
crossing assets in modern armies, New messanger of
Serbian Armed Forces, 1 (1) (2010) 125.
[4] http://desant.com.ua
[5] http://www.armytechnology.com
[6] http://worldweapon.com
[7] Kovaevi,N., Dimitrijevi,N., Babi,B.: Tendecies
of development of amphibious assets in armed forces
of the Russian Federation, ICDQM, Prijevoraak,
2016, Jun 2930, 552559
References
[1] Babi,B., Kovaevi,N.: Influence of amphibious
assets on environment, Risk and safety engineering,
Kopaonik, 2013, February 0206, 1522
[2] Kovaevi,N., Lazi,G.: Amphibious assets of armed
forces of NATO countries, Russian Federation and
Democratic Republic of Chine, Military Technical
Courier, 63 (1) (2015) 144168
167
ON ALGORITHM OF SYNCHRONIZED SWARMING AGAINST AN
ACTIVE THREAT SIMULATOR
RADOMIR JANKOVI
Union University School of Computing, Belgrade, rjankovic@raf.edu.rs
MOMILO MILINOVI
Faculty of Mechanical Engineering, Belgrade, mmilinovic@mas.bg.ac.rs
Abstract: The simulator algorithm of the system consisting of a group of armed mobile platforms, which in defence of a
territory against an active threat uses the synchronized swarming tactics, has been presented. The synchronized
swarming concept has been introduced in order to prevent the active threat to destroy the swarming platforms one by
one, reducing the probability of success of the swarm as a whole.
Keywords: swarming; armed mobile platform; synchronization; simulation; algorithm.
been developed. System activities have been presented by
time delays. The AMP group swarming in the model is
taking place in the battlefield represented by 2dimensional rectangle coordinate system (Figure 1). In the model
are moving:
1. INTRODUCTION
Swarming [1] is a tactics by which military forces attack
an adversary from many different directions, and then
regroup. Repeated actions of many small, manoeuvrable
units are going on, circling constantly through the
following four phases of swarming:
Disperse deployment of units in battlefield;
Gathering (concentration) of many units on
common target;
Action at a target from all directions;
Dispersion of units.
Units of the group (AMPi, i = 1, 2N);
Targets threat units (Pj, j = 1, 2M);
Command information system (C4I) messages.
In the time instant t = 0, AMPi are in the initial
deployment in the battlefield, and are characterized by
their maximal velocities (Vi), efficient ranges of their
weapons (RWi) and their individual possible effects on the
target threat unit Pj (Uij).
The swarming tactics is applied by much smaller units,
but their use is far more efficient, so in their actions as a
whole, they can often defeat many times superior
adversary [2].
In regular time intervals t, AMPs receive from C4I
system information on threat units (Pj) and other AMPi
movement, and at the same time give the reports of their
own current positions.
Although numerous examples of successful swarming
application have been recorded in history [3], the
significance of this tactics did not reach its full measure
until our days, due to brisk development of information
technologies and merging of computing and
telecommunications.More intensive military swarming
research began after 2000, and attained first results
mainly in the areas of unmanned vehicles (in the air,
underwater and on the ground), air force, navy and some
special ground forces units [4], [5], [6] and [7].
Y
NMP4(0)
4(2t)
Cilj(0)
4(3t)
C(t)
C(2t)
4(4t)
C(3t)
3(4t)
C(4t)
1(4t)
2(4t)
C(5t)
For small countries and their armies, as it is the case with
Serbia, one of the best investments could be the
adaptation of parts of their armed forces, especially
armoured and mechanized units (AMU) for swarming
tactics application [8]. That is the motivation for the
AMU swarming tactics research at Union University
School of Computing and at Belgrade Faculty of
Mechanical engineering, which is, according to our best
knowledge, the first research of the AMU swarming
tactics. So far, the discrete events simulation model of the
armoured mobile platforms (AMPs) group swarming has
1(3t)
V1(2t)
1(2t)
V1(t)
1(t)
2(3t)
3(3t)
3(2t)
3(t)
C(6t)
C(7t)
NMP3(0)
C(8t)
C(9t)
2(2t)
C(109t)
2(t)
V1(0)
NMP1(0)
4(t)
NMP2(0)
B
X
Figure 1. Swarming: group of 4 AMPs against 1 threat
168
ONALGORITHMOFSYNCHRONIZEDSWARMINGAGAINSTANACTIVETHREATHSIMULATOR
According to those information, AMPi are directed to
threat units and are moving towards them, with the aim to
as soon as possible take positions allowing a successful
swarming for destroying, disabling or hampering them in
accomplishing their mission.
OTEH2016
for dynamic checkout of swarming success criteria in the
course of simulation.
In order that AMPi participating in swarming could act
against threat Pj, the following conditions are to be
satisfied:
AMPi disposes the weapon Wi compatible with threat
Pj.
The distance between AMPi and threat Pj should be
within the effective range of the weapon Wi:
Dij = ( y j (t ) yi (t ))2 + ( x j (t ) xi (t ))2 DW i
(1)
The previous two conditions should be satisfied by
enough number of AMPs, so that their cumulative
effect, KUj, is greater or equal to critical threshold of
multiple AMPs effect, PKUj, which is characteristic of
the threat unit Pj:
N
KU j =
A K
ij
ij
U ij PKU j
(2)
Figure 2. Synchronization of swarming AMP1, 2, , 8
against threat Pj
i =1
Where:
They represent the basis for further development of this
class of simulators, including those for the synchronized
swarming of a group of AMPs against one or more active
threats.
Aij: is a coefficient (0/1) of AMPi to Pj previous
assignment in multiple target swarming models;
Kij: is a compatibility coefficient (0/1) of the weapon
Wi with the threat Pj.
In some real situations it is possible that a threat unit
ignores the swarmers activities and deals only with own
mission accomplishment. It happens in cases of a huge
supremacy of a threat over a number and type of
swarmers. However, a threat more often defends itself of
a swarm, trying to destroy AMPs one by one, so
hampering their gathering in a number which could
provide them successful swarming. One of possible
solutions for successful swarming of a group of AMPs
against an active threat is to include the synchronization
concept within the swarming tactics.
Uij: is a possible effect of the weapon Wi against the
threat Pj.
The relation of the threat Pj with the participants of the
swarm can be passive or active.
A passive threat is dedicated exclusively to its own
mission accomplishment and ignores the swarmers. If in
its movement such passive threat succeeds to reach its
assigned final position, and during that time the
cumulative effect of the swarming participants never
exceeds its critical threshold defined in expression (2), the
threat has accomplished its mission, and the swarming
tactics against it has failed.
In this paper the synchronization concept is given first,
and then the algorithm of the synchronized swarming of
an AMPs group against an active threat simulator is
presented.
An active threat, besides its own mission accomplishing,
is defending itself from the swarmers, in a way that, apart
from its moving towards its assigned final position, is
trying to destroy those APMs which in the meantime
could find themselves in the range of the threats own
weapons. In that way, the number of AMPs is decreasing,
as well as the probability of their swarming tactics
success.
2. SWARMING SYNCHRONIZATION
CONCEPT
If AMPs applying swarming against self an active threat
dont do that in a synchronized way, i.e. they dont reach
the distance from threat within the range of their main
weapons in approximately same time, there is a
probability of swarming failure, because the threat could
destroy AMPs one by one, as they appear within the range
of its own weapons. To solve that problem, the
synchronizing concept of the AMPs participating in
swarming has been introduced in the algorithm,
functioning as follows (Figure 2):
a. The synchronization zone (SZ) has been introduced
around the active threat, which is the circular ring
The algorithm and the results of the experiments with the
developed simulator of swarming of a group of AMPs
against a single passive threat have been given in [9], and
the algorithm of simulator of swarming against multiple
passive threats in [10]. During the development of
swarming simulator against passive threat, the
mechanisms for the simulation of movements of a threat
and swarmers have been elaborated, as well as the ones
169
ONALGORITHMOFSYNCHRONIZEDSWARMINGAGAINSTANACTIVETHREATHSIMULATOR
instance t = Tmj), functional dependence of coordinates
change in the simulated time.
represented by the expression:
Dzp = RP j (1 q)
OTEH2016
(3)
Where Dzp is the distance from the threat for which
has been defined that the AMP is in SZ, RPj is the
maximal efficient range of the threat weapon, and q is
a number between 0 and 1, defining the ring width,
which has been determined separately for every pair
of AMPi and threat Pj.
b. If there are enough AMPs at the distances which
enable their efficient actions, synchronization is no
longer needed and such AMP can apply successful
swarming.
c. If that is not the case, the AMPs which are further of
the SZ (AMP1, AMP2, AMP3, AMP=4 and AMP6
in Figure 2) will continue to approach the threat by the
shortest possible way.
d. AMP in the SZ are moving in parallel with threat,
based on so far known course and velocity of the
threat, obtained from the C4I system (AMP5 and
AMP7 in Figure 2).
e. AMPs situating between the threat and SZ, will move
away of the threat, directing their velocity vectors
contrary of the last known position of the threat
(AMP8 in Figure 2).
Such a mechanism enables that, while there are not
enough AMPs for successful swarming:
AMPs far from the threat continue to approach it by the
shortest possible way;
AMPs in the SZ, are moving in parallel with
target/threat at a relatively safe distance, ready to move
again towards the threat, when enough AMPs gather for
successful swarming cumulative effect;
Figure 3. Synchronized swarming simulator algorithm
AMPs near target/threat are moving away from it
towards SZ, in order to avoid individual destruction and
swarming failure.
Threats weapons characteristics: effective range RPj
(m), minimal time between repeated action and
probability of target hitting as function of distance.
3. SYNCHRONIZED SWARMING
SIMULATOR ALGORITHM
Characteristics of AMPs applying synchronized
swarming against an active threat: their total number N,
maximal velocities Vmi (m/s) weapons ranges DWi
(m), status Si (existing = 1, destroyed = 0),
compatibility of AMPi weapon Wi with a threat Kij
(0/1), possible effect of weapon Wi against threat Pj,
Uij.
The synchronized swarming of an AMPs group against an
active threat simulator is presented in figure 3. The
algorithm came from the basic algorithm for simulation of
swarming against a passive threat, which has been
upgraded by data structures and control mechanisms for
models that imply existence of an active threat, equipped
with resources for fighting against AMPs participating in
swarming.
Initial AMPs deployment, as a set of pairs of
coordinates xi (0) and yi (0) for t = 0. It can be random,
for simulation of the worst case when AMPs dont
expect appearance of a threat, or prescribed when some
information of appearance and mission of a threat are
known in advance. The latter case is used when the goal
of simulation is the research of the optimal initial AMPs
deployment for defence of such threat by means of
swarming tactics.
At the activation of the simulator, the initialization is done
first, comprising entering the following data:
C4I systems information renovation t (s)
Dimensions of simulated territory in which combat
activities take place: maximal values of coordinates xmax
(m) and ymax (m)
Upon the simulator activation, the status of every armed
mobile platform, AMPi, i = 1, , N, is checked up in the
main program loop.
Dynamic law of the threat movement: initial threats
coordinates (xp, yp at time instance t = 0), final threats
coordinates upon accomplished mission (xk, yk at time
If during the initialization the platform AMPi has not
170
ONALGORITHMOFSYNCHRONIZEDSWARMINGAGAINSTANACTIVETHREATHSIMULATOR
coordinates of its next position are calculated based on
its maximal velocity and the trajectory which it will
pass in time interval t. The goal of such motion of
AMPi is that it will find itself within the
synchronization zone SZ as soon as possible.
been assigned to participate in swarming against the threat
Pj, or its weapon is not compatible with the threat Pj, or
AMPi has already been destroyed by active threat (Si =
0), which is presented by the following logical
expression:
Aij Kij Si = 0
If the platform AMPi is out of SZ, but within the
effective range of the threats weapon, it is directed
opposite of the last known threats position provided by
the C4I system; the coordinates of its next position are
calculated based on its maximal velocity and the
trajectory which it will pass in time interval t. The
goal of such motion of AMPi is to try to avoid its
destruction by the active threats weapon and to find
itself within the synchronization zone SZ as soon as
possible.
(4)
the simulator proceeds to trial the next platform, AMP(i+1).
If condition (4) is not fulfilled, it is checked whether the
distance Dij, from AMPi to threat Pj, is less or equal to
threats weapon effective range, RPj.
If Dij RPj, the threats Pj fire action at the platform
AMPi is simulated by sampling the AMP destroying
probability distribution. That distribution depends on the
kind of threats weapon and the kind of AMP at which the
threat applies its fire action, and is entered at the
simulator initialization.
If the platform AMPi is within SZ, it continues motion
by the last course and velocity which the threat Pj had
until that moment, based on the last information
provided by the C4I system. The coordinates of the
platforms next position are calculated based on that
velocity, course and the trajectory which it will pass in
time interval t. The goal of such motion of AMPi is
to keep itself at safe distance, in the synchronization
zone SZ, and to follow the threat, moving parallel with
it and by its velocity, until the moment when enough
AMPs gather in SZ, which ensures that their cumulative
effect is greater of the critical threshold PKUij, i.e. the
condition (2) is satisfied and the success of the swarm is
provided.
In general case the probability of destroying AMPi is a
descending function of its distance to the threat Dij, with
the value of 100% for Dij = 0, the value defined by the
type of weapon at the distance equal to the weapon
effective range Dij = RPj and the value of 0 for distances
greater than RPj.
If the outcome of the threat unit Pj fire action is the
AMPi destruction, the value of Si = 0 is assigned to its
status and the simulator proceeds to trial the next platform
AMP(i+1).
The coordinates of the next position that the active threat
Pj should reach at the moment t = t + t, are calculated
based on its assigned law of the motion, which is entered
during the simulator initialization.
If AMPi has not been destroyed, it is checked whether it
is within the synchronization zone (SZ), i.e. whether the
following condition is fulfilled:
ROR j q Dij ROR j + q
All platforms AMPi, i =1, N, are analyzed in the
described way, and then the simulation clock is
incremented by 1 basic interval t. If a new value of
simulated time is greater than the threats mission
accomplishment time Tmj, it is concluded that the threat
has reached its final position before the fulfilment of the
condition (2), so the simulation of swarming against the
active threat is ended by failure.
(5)
If condition (5) is fulfilled, the value of possible
cumulative effect of all AMPs temporarily present in SZ
is updated, on the basis of the expression:
Kuj = Kuj + Ui
OTEH2016
(6)
If a new value of simulated time is still less than Tmj, the
threat Pj, and all platforms AMPi are moved to their
next positions, and the simulator begins new pass through
the main loop of checking the status of all swarming
platforms AMPi in the described way.
If the updated value of possible cumulative effect of all
AMPs temporarily present in SZ is greater than its critical
threshold PKUj, the condition defined by the expression
(2) is satisfied, and the swarm of AMPs has succeed to
neutralize the active threat Pj. If not, the simulator
proceeds to trial the next platform AMP(i+1).
5. CONCLUSION
If condition (5) is not satisfied, the platform AMPi is not
within SZ, the coordinates of its next position are
calculated, and then the simulator proceeds to trial the
next platform AMP(i+1).
The synchronization problem in swarming tactics is
significant, because the introduction of synchronization is
an attempt to reduce the failure probability in swarming
against an active threat.
The coordinates of the next positions, i.e. the new
deployment of the swarming AMPs in time t = t + t, are
calculated in the following way:
It comes to the fore in situations when, due to arriving of
armed mobile platforms at the distance of their possible
fire actions in different time instances, an active threat can
destroy them one by one, and question the successful
outcome of swarming tactics application.
If the platform AMPi is outside SZ and out of the
threats weapon effective range, it is directed to the last
known threats position provided by the C4I system; the
A new swarming synchronization concept has been
171
ONALGORITHMOFSYNCHRONIZEDSWARMINGAGAINSTANACTIVETHREATHSIMULATOR
presented in the paper, based on introduction of the
synchronization zone and various procedures of armed
mobile platforms, dependant on their temporary positions
related to active threat and synchronization zone.
OTEH2016
search and attack UAV swarms, Proceedings of the 2006
Winter Simulation Conference, Monterey, CA, USA,
2006. pp.13081315.
Nowak,D.J., Price,I., Lamont,G.B.: Self organized UAV
swarm planning optimization for search and destroy
using SWARMFARE simulation, Proceedings of 2007
Winter Simulation Conference, Washington, DC, USA,
2007. pp.13151323.
Pohl,A.J., Lamont,G.B.: Multiobjective UAV mission
planning using evolutionary computation. Proceedings of
the 2008 Winter Simulation Conference, Miami, FL,
USA, 2008. pp.12681279.
Singer,P.W.: Wired for war? Robots and military
doctrine, Joint Force Quarterly, 2009,52(1), pp.105110.
Jankovic,R., Milinovic,M., Jeremic,O., Nikolic,N.: On
Application of Discrete Event Simulation in Armoured
and Mechanized Units Research, Proceedings of the 1st
Internationa Symposium & 10th Balcan Conference on
Operational Research, Thessaloniki, Greece, 2011. Vol.2,
pp. 2835.
Jankovic,R.: Computer Simulation of an Armoured
Battalion Swarming, Defence Science Journal, vol. 61,
no.1, pp. 3643, January 2011.
Jankovic,R.: Data structures and control mechanisms for
multitarget swarming simulators, Electronic Letters, vol.
48, no. 16, pp. 997998, 2012.
Based on such synchronization concept, the new
algorithm of a group of armed mobile platforms
synchronized swarming against an active threat has been
developed and presented.
The described algorithm has been developed started from
the basic algorithm of a class of simulators of military
swarming systems, realized in so far research, especially
for the application of new tactical procedures in use of
armoured and mechanized units.
The basic algorithm, for unsynchronized swarming
against passive threats, has been improved by
mechanisms and data structures, which allow the
implementation of synchronization concept in swarming
against active threats.
References
Arquilla,J., Ronfeldt,D.: Swarming and the Future of
Conflict, Rand Corporation, 1999.
Edwards,S.J.A.: Swarming on the Battlefield Past,
Present and Future, Rand Corporation, 2000.
Edwards,S.J.A.: Swarming and the Future of Warfare,
Rand Corporation, 2005.
Price,I.C., Lamont,G.B.: GA directed selforganized
172
DETERMINING PROJECTILE CONSUMPTION DURING INDIRECT
MORTAR FIRE
ACA RANDJELOVI
University of Defense, Military Academy, Belgrade, aca.r0860.ar.@gmail.com
VLADO DJURKOVI
University of Defense, Military Academy, Belgrade, vlado.djurkovic@va.mod.gov.rs
PETAR REPI
Signal Brigade, Belgrade, repic92@gmail.com
Abstract: Determining projectile consumption is one of more important factors for achieving success when conducting
indirect fires tasks and depends on the accuracy of preparation of initial elements, mathematical expectation of number
of hits and probability of hitting the target. This paper shows the procedure for determining projectile consumption of a
120 mm mortar platoon during indirect engagement of predetermined target in accordance with the predefined number
of required direct hits, based on the developed model, and by applying appropriate formulas, coefficients and table
values.
Keywords: projectile consumption, probability of hitting the target, mathematical expectation of number of direct hits,
indirect fire.
determining elements of group engagement and projectile
consumption allowance. Also, projectile consumption is
greatly affected by the probability of hitting the target and
mathematical expectation of the number of hits that
combined with the response time represent basic factors
of efficiency [2].
1. INTRODUCTION
Mortar units are the main carriers of fire support for
infantry and mechanized battalions in the Serbian
Military. They execute their tasks by engaging the target
with accurate and precise fire of 120 mm and 82 mm
mortars.
By engaging their targets indirectly, mortar units achieve
better protection of their resources from the enemy fire
and take advantage of their maximum effective ranges in
a more effective way when executing their combat tasks.
However, indirect fire procedure requires more time, and
preparation for engagement is more complex with higher
projectile consumption.
Indirect mortar fire consists of two phases: preparation
and execution (Image 1.). Those two phases consist of a
number of activities that are executed step by step, the
right way and quickly. This provides right timing,
accuracy of fire and safety during the execution phase.
Group engagement is the final phase of indirect
engagement through which the main fire effect of combat
systems is achieved and the greatest material and/or moral
effects that are of huge importance for further operations.
By engaging the target in groups, mortar units can
neutralize, block, destroy or obstruct the enemy [1].
Image 1. Activities during indirect engagement
Efficiency of execution of listed tasks depends on
multiple interdependent factors out of which, determining
projectile consumption is the most important one.
2. EFFICIENCY OF MORTAR UNITS
Efficiency of mortar units represents the degree of success
of task execution expressed as impact of engagement and
projectile consumption or probability of successful task
Projectile consumption is determined after the assigned
task, area and/or target type, distance, ways of
173
OTEH2016
DETERMININGPROJECTILECONSUMPTIONDURINGINDIRECTMORTARFIRE
execution. It is described as the efficiency of engagement
as well as the capability to execute other tasks from the
domain of fire support, in various weather, spacial and
combat conditions [3]. Efficiency of group engagement is
achieved by:
 determining elements of group engagement as
accurately as possible,
 using appropriate number of weapon systems and
sufficient amount of projectiles,
 proper choice of projectiles, charges and aims, that is,
proper choice of engagement and type of fire,
 uniformly distributing fire in the target area,
 using the effect of surprise and
 timely corrections of elements during the execution of
group engagements.
formula [7]:
The effect of group engagement is the realistic indicator
of success of mortar fire and it is the result of indirect
fires. It is evaluated based on accomplished explosions on
the target and the degree of accomplished fire effect and it
is determined by: the size of deviation (error) of the
middle hit from the center of the target, by the number of
direct hits and the number of lethal pieces as well as
projectile consumption.
By applying given formula, for a 120mm mortar, we
calculated the probability of hitting a surface square
shaped target, with sides of length of 100m on the
distance of 4000m and assumption that the middle hit is
in the center of the target (Sp = Cc), and by different
accuracies of initial elements (Table 2).
Table 2. Comparative display of different probabilities of
hitting the target with a 120mm mortar
num
1.
2.
3.
Mortar units achieve efficient execution of the assigned
task by engaging the target in groups and by: using the
elements obtained through total preparation of initial
elements, elements obtained through fire correction and
by shifting fire on the topographicalgeodetic basis (TGB)
by using data from fire correction [4]. Listed ways are
followed by certain errors shown in Table 1.
16
3.
Fire correction on the target
0,5  2,2
0,9  9,0
4.
Shifting fire within allowed limits
from the target
1,3  1,5
2,5  3,5
120mm mortarplatoon
Demonstrated values maintain the degree of accuracy of
initial elements preparation which have a great influence
on group engagement and its efficiency for accomplishing
end (planned, desired) effects on target.
Accuracy of preparation of initial elements directly
influences the probability of hitting the target and
projectile consumption. Listed accuracy proportionally
affects the probability of hitting, and inversely
proportional on the consumption of projectiles. This
implies that the probability of hitting is greater and
consumption is smaller if group engagement is executed
with elements of complete preparation, which makes
indirect mortar fire more accurate and precise [5,6].
preparation
shortened
Number of hits
Mathematical
Expectation
4
Projectile
consumption
6
Unit
2.
Probability of hitting
(%)
70,35
33,38
18,17
Table 3. Mathematical expectation of the number of hits
and projectile consumption
Mean error
x Vd
y Vp
Number of
mo
1.
Elements of
engagement
table
Complete preparation
Shortened preparation
Comparative analysis of obtained results leads to the
conclusion that the value of probability of hitting the
target with elements obtained through complete
preparation is closer to probabilities of table values, and
that it is greater than probability of hitting the target with
elements obtained through shortened preparation of initial
elements.
Table 1. Numerical values of errors
The way of transition to group
engagement
Complete preparation of initial
elements
Shortened preparation of initial
elements
where:
v
 probability of hitting a surface target
x
 deviation from the middle hit to the center of the
target measured by distance,
y
 deviation from the middle hit to the center of the
target measured by direction,
a
 half of the length (depth) of the target,
B
 half of the width of the target,
Vd  probable deviation by distance,
Vp  probable deviation by direction.
Taking into account listed elements that determine the effect
of group engagement, it is possible to find out that it
increases with the decrease of error and increase of number
of direct hits with as low projectile consumption as possible.
Num
) (
V = 1 x + a x a x + b x b (1)
4
Vd
Vd Vp
Vp
Number of fired
projectiles (n)
Probability of
hitting(p)
Number of direct
hits
(=nxp)
Number of desired
hits()
Probability of
hitting (p)
Projectile
consumption
(n=a/p)
complete
40
0,1817
0,3338
7,26 7
13,35 13
20
0,1817
0,3338
110,07 110
59,91 60
Influence of accuracy of preparation on projectile
consumption has been analyzed for a 120mm mortar
platoon for achieving 20 direct hits via mathematical
expectation of the number of hits when firing 40
projectiles (10 per mortar) and by applying the calculated
probability of hitting with complete and shortened
Probability of hitting a surface target (square or
rectangle shaped) is calculated using the following
174
OTEH2016
DETERMININGPROJECTILECONSUMPTIONDURINGINDIRECTMORTARFIRE
following model:
preparation of initial elements in accordance with the
above stated conditions of firing (Table 3).
Up = 11501,30111 = 195 projectiles
Comparative analysis of obtained data leads to conclusion
that by using complete preparation of initial elements we
can expect greater number of direct hits and lower
projectile consumption, which makes the unit more
efficient in solving assigned tasks for up to 45%.
Dividing calculated result by the number of weapon
systems in a mortar platoon gives us projectile
consumption per system:
Up =
3. FIRING TASK MODEL
Up
= 195 = 48, 75 49 projektila (3)
br. orudja u vodu
4
Multiplying projectile consumption per system (Up') by
the number of mortars in a mortar platoon gives us total
projectile consumption for given task execution, which
is 196 projectiles in this case (49 projectiles x 4 mortars).
Firing task model based on which projectile consumption
is determined is founded on following assumptions:
task is conducted by a 120mm mortar platoon
consisting of 4 mortars with a light TF M62P3 mine,
with a UTU M62 fuse, set to immediate,
target has area of 1ha  sides 100 x 100 m (air defense
artillery battery),
group engagement is conducted with elements of
complete preparation,
distance to target is 4000 m,
the goal is to neutralize the enemy with the degree of
30%,
probability of hitting the target (p) with the first
projectile is 33,38%,
number of expected/desired hits () is 40 projectiles,
allowed number of projectiles (n) for combat task
execution (firing) is 160 projectiles.
In accordance with the allowed number of projectiles for
task execution, mathematical expectation (ME) was
calculated in accordance with the probability of hitting the
target (p):
MO = n x p = 160 x 0,3338 = 53,41 53 hits
(4)
In accordance with the developed model and conditions
for task execution, projectile consumption was calculated
(n) for getting expected/desired hits () on target
depending on the probability of hitting the target (p):
n = a = 40 = 119,83 120 projektila
p 0,3338
In accordance with the given conditions for task execution
we determined: total number of projectiles for a unit
(necessary for task execution), number of projectiles for
each mortar, mathematical expectation of the number of
hits with allowed number of projectiles, as well as the
necessary number of projectiles in order to accomplish
desired hits.
(5)
Based on calculated results, we can get the picture of the
effect of projectile consumption on task execution and
ability to execute indirect fire tasks in accordance with the
commander's intent.
In case of the model mentioned above, it can be
concluded that the allowed amount of projectiles (n) does
not enable target neutralization with the degree of 30%,
however, it provides 40 direct hits on target with fewer
projectiles than allowed.
4. PROJECTILE CONSUMPTION FOR
FIRING TASK EXECUTION
With this conclusion, the commander and superior
command get a realistic picture of possibilities of mortar
systems under their command, and as a result of that a
proposal to increase the number of projectiles or reduce
the degree of neutralizing the target under 30% can be
made.
Calculations of projectile consumption provide realistic
image of assigned tasks of mortar units for commanders
and superior command.
Projectile consumption necessary for firing tasks
execution is calculated using the following formula [8]:
Up = Pc(ha) N kp kdg kpr kvgu (2)
5. CONCLUSION
where:
 Pc(ha), target area = 1,
 N, table consumption for target neutralization(5E) on Dg
10 km and En = 25 % = 150 projectiles,
 kp, transition coefficient for En = 30 % = 1,30,
 kdg, distance coefficient for Dg = 5 km (Dg 10 km)
= 1,
 kpr, preparation coefficient (complete preparation of
initial elements) = 1,
 kvgu, type of fire and fuse coefficient (fuse with current
effect) = 1.
The importance of determining projectile consumption
was highlighted in this paper as well as its impact on
solving indirect firing tasks in accordance with
commander's intent.
Critical analysis of the effect of accuracy of preparation
of initial elements on the probability of hitting the target
was conducted, as well as its effect on mathematical
expectation of the number of hits and projectile
consumption and their causal relationship with the
efficiency of mortar units.
All in all, a valid calculation of projectile consumption
was conducted in accordance with conditions given by the
fire task model, based on whose results appropriate
By inputting table values into the formula, we get the
estimate of projectile consumption for a given task by the
175
DETERMININGPROJECTILECONSUMPTIONDURINGINDIRECTMORTARFIRE
conclusions were made and a realistic possibility for task
execution was estimated for a 120mm mortar platoon.
[2] Kokelj,T.: Zbirka resenih zadataka iz teorije
artiljerijskog gadjanja, VIZ, Beograd 1999.
[3] Kova,M.:
Metod
odredjivanja
efikasnosti
artiljerijskih i raketnih jedinica za podrsku, Sektor
za skolstvo, obuku i NID, VIZ, Beograd, 2000.
[4] Randjelovi,A., Djurkovi,V., Kokelj,T.: Uticaj
tacnosti pripreme pocetnih elemenata posrednog
gadjanja
na
izvrsenje
grupnog
gadjanja
minobacacem 120 mm, 6. Medjunarodna naucna
konferencija OTEH, Beograd, 2014.
[5] Kokelj,T., Regodi,D.: Tacnost potpune pripreme
pocetnih
elemenata
posrednog
gadjanja,
Vojnotehnicki glasnik, br. 2, 2005.
[6] Randjelovi,A., Kokelj,T., Komazec,N.: Tacnost
skracene pripreme pocetnih elemenata posrednog
gadjanja minobacacem 120 mm, 17. medjunarodna
DQM konferencija, Beograd, 2014.
[7] Kokelj,T., Randjelovic A.: Zbirka resenih zadataka iz
pravila gadjanja naoruzanjem pesadije, VIZ,
Beograd 2010.
[8] ivanov,Z.: Teorija gadjanja, VIZ, Beograd 1979.
Listed calculations were done by applying appropriate
formulas and table values of coefficients that enable us to
obtain usable data which represent the base for making
proposals about more efficient engagement of mortar
units and possible demands to change the type of
engagement on the given target.
Based on that, it can be concluded that projectile
consumption represents an important factor of task
execution and together with the probability of hitting the
target and accuracy of preparation of initial elements and
group engagement, influences the degree of
accomplishment of fire effects on target during task
execution.
Also, determining projectile consumption defines the
possibility of achieving desired number of direct hits and
decision making and setting realistic tasks by the
commander and the superior command.
References
[1] Artiljerijsko pravilo gadjanja, SSNO,
artiljerije/, VINC, Beograd, 1991.
OTEH2016
/Uprava
176
PROPELLER AND SHIP MAIN EGINE SELECTION IN CORRELATION
WITH OVERALL EFFICIENCY PROPULSION COEFFICIENT
IMPROVEMENT
JOVO DAUTOVI
Technical Test Center, Vojvode Stepe br. 445, Belgrade; email jovodaut@gmail.com
VOJKAN MADI
Technical Test Center, Vojvode Stepe br. 445, Belgrade; email vmadic@gmail.com
SONJA URKOVI
Military Technical Institute, Bulevar vojvode Bojovica br.2, Belgrade; email sonja.djurkovic@yahoo.com
Abstract: This paper presents propeller selection method when the power of the ship engine is known and also influence
of correct propeller selection on a significant improvement in overall efficiency propulsion coefficient. After that it
shows the method of main engine selection in case of vessel modernization, i.e. replacement of one type with another
type the drive when retain the old propellers. Main engine selection, in that case, is shown in the case of selection
electric motor drive of the ship KOZARA in the process of its modernization during which diesel engine propulsion is
replaced by diesel electric propulsion. In the end, power shaft measurement results on the shaft lines is shown which
confirms main engine correct selection.
Keywords: propeller selection, electric motor selection, propeller curve, shaft power measurement, performance
recording.
and drive engine should be such that the operating point
of the diesel engine always is inside the diesel work areas
for which the engine is designed, in order to increase the
overall efficiency coefficient [1].
1. INTRODUCTION
During designing of the ships special attention should be
paid to the appropriate propeller and drive engines
selection and their mutual pairing.
For proper utilization of the possibilities for the engine, it
should be possible to achieve maximum revolution per
minute (rpm) speed at which the engine has maximum
power.
During operation of the ship it is usually to do
modernization after years of use, which often includes
replacement of the one type drive engines with the other
one. In this case, usually retain the old propellers, and
attention should be paid to the proper selection of the new
drive engines and they must be matched with the
propellers.
If propellers match with maximum rpm in reducing rpm
of a percentage, engine power will be reduced in
accordance with the propeller curve. In other words, the
shaft torque curve will not be in accordance with the
engine torque curve than the lower the full power of the
engine would never be used.
An example for this case is the modernization of ship
"Kozara", which was carried out during 2011 and 2012.
During this period diesel engine propulsion was replaced
with diesel electric propulsion (DEP). DEP is consisting
of two diesel engines, two electric drive motors with
gearboxes and bow thrusters. That type of modernization
was performed for the first time on a Serbian warship.
The main difference between diesel propulsion and DEP
is that in a diesel propulsion system propeller is driven by
diesel engine and in DEP system propeller is driven by
electric motor. Electric motor power supply is realized
from the generators which are driven by diesel engines.
Figure 1 shows the corresponding characteristic of diesel
engine and propeller curves [2].
For constant torque (ie. max torque):
2
(2.1)
which indicates that the power is proportional to the rpm
(figure 1). If the torque is not constant, it is
assumed that the power changes per cubic parabola curve
as a function of number of revolutions
2
, which means that the torque is proportional to the
square of rpm.
2. PROPELLER SELECTION
Significant improvements in overall efficiency coefficient
on ship standard drive systems mainly achieved by proper
selection of the propeller. Mutual working of propeller
(2.2)
177
PROPELLERANDSHIPMAINEGINESELECTIONINCORRELATIONWITHOVERALLEFFICIENCYPROPULSIONCOEFFICIENTIMPROVEMENT
which means that the torque coefficient
is constant
(
). If it is known that the fixed blade
), then
propellers advance ratio is constant (
the propeller thrust coefficient
is also constant
is constant then
, but if
). If
(
and
(or
for
J is constant then
is
permanently reduce thrust force
).
assumption in the normal speed range of displacement
ships, and power formula is:
OTEH2016
Note also the upper rpm limit on the (design) line 1 say
point D at 105% maximum rpm. At this point the full
available power will not be absorbed in the clean hullcalm water trials, and the full design speed will not be
achieved. In a similar manner, if the ship has a light load,
such as in ballast, the design curve 1 will move to the
right (e.g. towards line 1a ) and, again, the full power
will not be available due to the rpm limit and speed will
be reduced.
These features must be allowed for when drawing up
contractual ship design speeds (load and ballast) and trial
speeds.
(2.3)
is defined by
The basic assumption is made that
the engine manufacturers as the Propeller Law and they
design their engines for best efficiency (e.g. fuel
consumption) about this line. It does not necessarily mean
that the propeller actually operates on this line, as
discussed earlier.
It is important to match the propeller revolutions, torque
and developed power to the safe operating limits of the
installed propulsion engine. Typical power, torque and
revolutions limits for a diesel engine are shown in Figure
2, within ABCDE [2].
Figure 1. Matching of propeller to diesel engine
It should be noted, however, that the speed index does
change with different craft (e.g. highspeed craft or very
, where x
slow displacement ships), in which case
will not necessarily be 3, although it will generally lie
between 2.5 and 3.5.
Typically it is assumed that the operator will not run the
engine to higher than 90% of its continuous service rating
(CSR).
The marine diesel engine characteristics are usually
,
based on the propeller law which assumes
which is acceptable for most displacement ships. The
basic design power curve passes through the point A.
Figure 2. Typical diesel engine limits (within ABCDE)
It should be noted that the propeller pitch determines at
what revolutions the propeller, and hence engine, will run.
Consequently, the propeller design (pitch and revolutions)
must be such that it is suitably matched to the installed
engine.
When designing the propeller for say clean hull and calm
water, it is usual to keep the actual propeller curve to the
right of the engine ( line), such as on line 1 , to allow
for the effects of future fouling and bad weather. In the
case of line 1 the pitch is said to be light and if the
design pitch is decreased further, the line will move to
line 1a , etc.
3. ELECTRIC ENGINES SELECTIONS ON
THE SHIP KOZARA
For the correct selection of electric drive engines on the
ship Kozara it was necessary to calculate the power
which engines deliver to the propellers. The first step was
to calculate power which was delivered to the propeller
by old diesel engines. The transmission of power from the
diesel engine to the propeller produces shaft transmission
losses which are expressed by shaft transmission
. For calculation of shaft transmission
efficiency
efficiency next coefficients are taken into account: thrust
As the ship fouls, or encounters heavy weather, the design
line and then
curve 1 will move to the left, first to the
to line 2 . It must be noted that, in the case of line 2 , the
maximum available power at B is now not available due
to the torque limit, and the maximum operating point is at
C, with a consequent decrease in power and, hence, ship
speed.
178
PROPELLERAND
DSHIPMAINEGINESELECTIONINCO
ORRELATIONWITHOVERALLEFFICIENCYPROPULSIONCOEFFICIENTIM
MPROVEMENT
bearing (0.999), steady beearing  9 pieces
p
(0.995)) and
bulkhead stuuffing box  3 pieces (0.995). In this way shaft
transmission efficiency is
.
(3.4))
hich power iss
EM calculated poower is 287.66 kW. EM wh
w estimatedd
250 kW is chosenn for ship proopulsion. It was
that ship with 2550 kW EM will achieve the requiredd
speed. In the asseessment was ttaken into acccount the factt
that ship with thhe new drivve engines has a smallerr
n of the drivee
displlacement. Thee influence onn the selection
EM has had a huuge differencee in the purchase price off
E higher poower. If selected power EM
M
seleccted EM and EM
enterred in the aboove formula itt can be obtaiined power off
the engine
e
whichh can be perm
manently delivered to thee
prop
peller and that equals
=238 kW.
g
Delivered poower in case of classic dieesel drive is given
by:
(3.1)
where
OTEH2016
wer.
nominal dieseel engine pow
Figure 3 shows an old shiip propulsion system with diesel
d
drive engine.. Left and righht drive system
ms are identicaal.
After calculating the electric m
motor power itt is necessaryy
to seelect the type of electric mootor. Asynchrronous motorss
havee been selectedd because of iits advantagess compared too
the other
o
electric motors.
m
Figure 3. Diesel engiine drive proppulsion system
m
o the
1 diesel enngine; 2 thruust bearing; 3 , 5 first part of
propeller shaaft steady bearrings; 4 bulkkhead stuffingg box
I;6, 8 seconnd part of the propeller shaaft steady bearrings;
7  bulkheead stuffing box II; 9, 10  third
t
part of thhe
propeller shaaft steady bearrings; 11  bullkhead stuffingg box
III; 12 sterrn tube;13 stern boss; 14 bracket bearring.
Two
o electric motoors, each havinng a power off 250 kW typee
B6A
AZJ 35404 prroduced by K
Koncar GIM, are installedd
in th
he ship Kozzara. The traansfer of pow
wer from thee
electtric motor to the propeller is carried ou
ut through thee
gearb
box. An elecctric motor drives a propeeller. Electricc
moto
ors speed and
a
torque control is carried outt
autom
matically via voltage frequeency converteers.
After calcullating of pow
wer which old
o diesel enngines
delivered to the propelleers it was staarted with electric
mation.
drive enginess power estim
Two
o gearboxes ZF W3501 produced by ZF
F
Fried
drichshafen AG
A of Germanny, the transmiission ratio off
1: 3,,968, are instaalled in the shhip in order to
o transmit thee
poweer from the EM
M to the propeller.
Figure 4 shoows a diesel electric
e
propuulsion system. The
new propulsion system allso consists of
o an identicaal left
and right system. In addiition, the figuure 4 also shoows a
bow thruster.
Com
mpared to thee other variannts of power transmissionn
from
m the drive engine
e
to proopellers, electric drive hass
several advantagees that can bee seen from th
he diagram inn
ure 5 [3].
Figu
Figure 4.
4 Diesel electtric drive proppulsion system
m
1 dieseel engine; 2 generator; 3 electric drivee
engines; 4 gear box; 5 bulkhead stuuffing box I; 6,
6 7
propeller shaaft steady bearrings; 8 bulkkhead stuffingg box
II; 9 stern tube;
t
10 sterrn boss; 11 bracket
b
bearinng; 12
boow thruster
At the beginnning, the lossses in the trannsmission of power
p
from the eleectric motor to the propelller are estim
mated.
Gear box efficiency (ttogether witth thrust beearing
(0.98) andd shaft transm
mission efficiency
efficiency)
(0.97) aree taken into acccount.
Delivered power
p
in case of DEP is givven by:
(3.2)
which indicaates that electrric engine pow
wer is calculated by
the formula:
Figure 5. Propeller, diesel enngine and elecctric motor
curvees
 electric motoor curve;
 diesel motor curve;
overloaded propeller
p
curvee;
 propelller curve
(3.3)
mula is entereed
instead of
w be
If in the form
it will
calculated thhe electric mottor (EM) pow
wer
, in ordder to
DEP gave thhe same deliveered power too the propeller like
diesel enginee:
It caan be seen on diagram that drive diesel engine,
e
whichh
is most
m commonlyy used on inland waters caan change thee
speed in the rangee from about 440%
to 100%
(from
m
ng regime to 100% of nominal rpm). Speeed above thiss
idlin
179
PROPELLERANDSHIPMAINEGINESELECTIONINCORRELATIONWITHOVERALLEFFICIENCYPROPULSIONCOEFFICIENTIMPROVEMENT
OTEH2016
values of power
range can be increased only in the short term, usually up
to 110%
. It is not possible even to reduce the rpm
below this range, because it can cause the engine stop.
The ship load curves are often very different. The
example is loaded or unloaded push boat or a tugboat in
free run regime or in tow regime. The figure 5 shows two
cases of propeller curve, one when propeller is overloaded
and the other when the propeller curve has shape of cubic
parabola as a function of rpm.
Number
1.
2.
3.
4.
5.
Torque
(Nm)
510,68
1925,27
4123,54
6425,7
6758,64
Shaft speed
(min1)
101
202
301
375
393
PowerP
(kW)
5,4
40,7
129,9
252,2
278
Figure 7. shows diagrams of torque and shaft power as
function of the rpm.
At the electric engine transmission power increases
linearly with the propeller speed of rotation up to a
nominal value, but from that point it is possible to
increase the speed of rotation up to 60% above nominal
value, with power at a nominal value. As a result, the
electric motor can operate at 100% power for all cases
propeller curve, unlike the diesel engine in case of
propeller curve for overloaded propeller can be operated
at speeds from idling to speed n1. This is the biggest
advantage of electric motor as the drive unit.
4. PROPELLER SHAFT TORQUE
MEASUREMENT AND PERFORMACE
RECORDING
Figure 7. Diagram of shaft power and torque
measurement
By measuring propeller shaft power can be checked is
propeller accurate chosen according to engine. Figure 6
shows an acquisition system for the shaft power
measurement.
Shaft power measurement confirm electric motor power
and the possibility that the motors run in overload of 10%
over a period of one hour.
Based on the results of measuring the torque and power to
the propeller shaft in the expression for calculating the
propeller characteristic:
it is determined coefficient K = 0.00476. The expression
for the ship Kozara propeller characteristic is[7]:
Figure 6. Propeller shaft torque and rpm measurement
It is a multichannel mobile measuring system SPIDER 8
manufacturer HBM (HBM  Hottinger Baldwin
Messtechnik) SPIDER 8 is connected with the
corresponding sensors strain gages type XY21 6/120 [4]
and optical rpm measuring system AO1 type [5]. By this
measuring system propeller shaft torque and rpm are
measured. Based on the measured propeller shaft power it
could be drawn propeller curve and find a constant
that depends on the characteristics of the propeller.
0,00476
(4.3)
Table 2. Value of power to the propeller shaft , calculated
on the basis of the measured torque P and power
calculated on the basis of the propeller characteristic .
Shaft torque and rpm measurement is made on the left
intermediate shaft behind gear box.
Shaft power, i.e. propeller absorbed power is calculated
by the formula:
Table 2 shows the value of power to the propeller shaft,
calculated on the basis of the measured torque P and
power calculated on the basis of the propeller
characteristic .
Propeller shaft torque measurement on dieselelectric
drive ship Kozara was performed 09.05.2012. during
test trailing at the river Danube.
2
60
(4.2)
(4.1)
Number
Shaft speed
(min1)
P (kW)
1.
101
5,4
4,9
2.
202
40,7
39,2
3.
301
129,9
129,8
4.
375
252,2
251
5.
393
278
288,9
(kW)
Figure 8 shows diagram power to the propeller shaft,
calculated on the basis of the measured torque P and
power calculated on the basis of the propeller
characteristic , as a function of the propeller speed.
Shaft torque measurement results at different propeller
rpm are shown in table 1 [6].
Table 1. Shaft torque measurement results and calculated
180
PROPELLERANDSHIPMAINEGINESELECTIONINCORRELATIONWITHOVERALLEFFICIENCYPROPULSIONCOEFFICIENTIMPROVEMENT
OTEH2016
1. Results of shaft power measurements confirm good
choice of electric drive engine and the possibility that
the motors run in overload of 10% over a period of
one hour.
2. Good matching of measured and calculated shaft
power results fully confirms the results validity and
measurement power methods
3. Shaft power result measurements also confirm correct
operation of frequency converters. Although there
were possibility to be problems with THD voltage
distortion it turns out that it is not necessary to apply
any filters to reduce voltage high harmonics in the
network[8].
References
[1] V. J. Gitis, V. L. Bondarenko, T. P. Jefimov, J. G.
Poljakov, B. M. urbanov, Theoretical basis of the
exploitation of marine diesel engines (translated
from Russian ) SSNO , Belgrade, 1973.
[2] A. F. Molland_ S. R. Turnock, D. A. Hudson, Ship
resistance and propulsion _ practical estimation of
ship propulsive powerCambridge University Press
(2011).
[3] B. Bilen, Z. Nikoli, Z. ovagovi, D. Bulovan,
Improvement of the driving characteristics of the
ships with electrical transmission, Institute of
Technical Sciences SANU, Belgrade.
[4] HBM  An Introduction to Measurements using
Strain Gages.
[5] Instructions for measuring the torque on the propeller
shaft of the ship, TOC, C.33.003.
[6] The results of the test vessel, Handover record, the
Ministry of Defence , Belgrade 2012.
[7] J. Dautovi, Diesel electric drive of river military
ships as a method of improving ship maneuvering
characteristics, PhD thesis, Belgrade 2016.
[8] D. Vueti, I. Vlahini, The impact of the serial
inductance to reduce harmonic distortion of power
line commutated inverter electric propulsion system
of the ship , Maritime. god 19. 2005, p. 6575.
Figure 8. Diagram power to the propeller shaft,
calculated on the basis of the measured torque P and
power calculated on the basis of the propeller
characteristic , as a function of the propeller speed
5. CONCLUSION
It can be concluded that for the proper functioning of the
ship drive system it is important to match the propeller
revolution speed with torque and power of the drive
engine. The propeller pitch determines propeller
revolution speed and also engine revolution speed.
According to this, the propeller (pitch and rpm) must be
designed to be matched with the drive engine.
In the case of ship "Kozara" modernization is shown the
way of drive engine type changing and retain the old
propellers. Drive engine calculating power method is also
shown and acc. to this electric motor type is chosen.
In order to check the accuracy drive engines selection
during the ship trial shaft torque measurement was
performed. Based on the results of torque measurements
electric engine power delivered to the propeller was
calculated.
Based on the shaft torque (power) measurement results on
the ship "Kozara" it can be concluded the following:
181
OPTIMIZATION OF PLANETARY GEARS AND EFFECTS OF THE THINRIMED GEAR ON FILLET STRESS
MILO SEDAK
Belgrade University, Faculty of Mechanical Engineering, Belgrade, msedak@mas.bg.ac.rs
TATJANA M. LAZOVI KAPOR
Belgrade University, Faculty of Mechanical Engineering, Belgrade, tlazovic@mas.bg.ac.rs
BOIDAR ROSI
Belgrade University, Faculty of Mechanical Engineering, Belgrade, brosic@mas.bg.ac.rs
Abstract: Planetary gears take a very significant place among the gear transmissions, and they are widely used in
military and civil industry applications such as marine vehicles, aircraft engines, helicopters and heavy machinery.
Planetary gears are complex mechanisms which can be decomposed into external and internal gears with the
corresponding interaction, which requires geometrical conditions in order to perform the mounting and an appropriate
meshing of the gears during their work. Planetary gears have a number of advantages as compared to the transmission
with fixed shafts such as a compact design, with coaxial shafts, high power density and higher efficiency, which is
achieved by reducing gear weight using thinrimed gears. The purpose of this paper is to present the optimization
model for the planetary gears, where the objective function is the weight of gears, and functional constraints imposed
upon their respective structural design. Hence, the objective is to minimize rim thickness of the gear in order to achieve
highperformance power transmission and minimize weight. This paper presents the results of an investigation with
finite element analysis (FEM) into the effects of thinrimmed gear geometry on the root fillet stress distribution.
Keywords: Planetary Gear, Thinrim Gear, Finite Element Analysis, Internal Gear, Root Stress.
in [5]. Therefore it is necessary to include the bending
fatigue failure and crack propagation that is above all
influenced by rim thickness, into the design phase to
avoid component failure mode. Moreover, it has been
shown that in addition to mass reduction, the rim
thickness introduces gear flexibility that decreases the
influence of internal gear and carrier errors [6]. The
purpose of this paper is to present a generalized
optimization procedure to the minimum volume design of
the planetary gearbox which takes into account the
bending fatigue life of the thinrim gears.
1. INTRODUCTION
Planetary gear sets are of fundamental importance in
many applications that require high performance
mechanical energy transmission systems and are widely
used in automotive, military and aerospace industries.
Because of their larger torquetoweight ratio, compared
with fixed shaft transmission systems, and other
numerous advantages, planetary gears found their
application in a rotorcraft transmissions, vehicle
transmissions systems and jet propulsion systems [12]
etc.
The usual methods of gear design and analysis are based
on the current standards published by the International
Organization for Standardization, German Institute for
Standardization and American Gear Manufacturers
Association. Although, the design and analysis of the
gears has been described and organized systematically in
the existing literature and standards, the procedures for
the design of the planetary gear set still require
specialized design knowledge, thus presenting a good
source for further research. This is mainly due to the
complex geometric and kinematics relations of the
planetary gears that provide mounting and appropriate
meshing of the planetary gears during their work. As a
result the conventional trial and error type of calculation
is timeconsuming and the complexity of the considered
problem as well as the accuracy of the final construction
is limited by the tentative choices and designers intuition.
The weight minimization problem has been of
considerable interest, especially in high power density
transmission applications. A number of wellestablished
works concerning the minimummass design problem of
the gear reduction systems include the minimization of
the center distance as well as the volume of the internal
gears by Tong and Walton [3]. More recently, the
application of evolutionary algorithms on minimum mass
gear design problem was considered in [4]. To ensure low
weight of the planetary gear reduction unit the rim of the
internal gear, as well as the other thinrimed gears in the
planetary set, must be as thin as possible.
However, the rims that are too thin can affects the root
stress distribution resulting in the tooth bending fatigue
failure and cracks. The investigation of the rim thickness
on the fatigue life by finite element method was presented
182
OTEH2016
OPTIMIZATIONOFPLANETARYGEARSANDEFECTSOFTHETHINRIMMEDGEARONFILLETSTRESS
The potential improvement of such design process can be
made employing the computer based optimization process
where the immediate benefits includes the applicability to
the gear design of arbitrary complexity. Using the finite
element (FE) method to calculate stress and deflection in
the case where no analytical of empirical relation is
known, broadens the scope of applicability of
optimization based design.
2. FORMULATION OF THE
OPTIMIZATION PROBLEM
The problem of minimum mass design of spur gears has
been of a considerable interest for a number of
researchers, primarily due to the requirement of low mass
design in highperformance power transmission
applications. Weight minimization is the most frequent
structural optimization problem, generally formulated as a
constrained optimization problem which can be stated as
follows
Optimization process of complex mechanical design
problems often leads to complicated objective functions
with a large number of design variables and highly
nonlinear constrains. Thus, it is not rare that such
optimization problems are multimodal with the discrete
search space. Therefore, conventional gradientbased
optimization methods are ineffective for the solving of
such complex nondifferentiable optimization problems
due to the high number of design variables and the
methods disadvantage of being trapped into local minima.
In the past decades, metaheuristic optimization methods
have proven effective against such drawbacks, and have
become powerful methods for solving tough optimization
problems in many fields of science and engineering. The
metaheuristics are able to provide global solution to the
optimization problem with higher number of design
variables.
minW ( x ) ,
(1)
xD
subject to the n equality and m inequality constrains
gi ( x ) 0, i = 1, , m
h j ( x ) = 0, j = 1, , n.
(2)
The objective function of such problem for homogenous
material can be defined as
W (x) =
V ,
i
(3)
i =1
where is the weight density and Vi is the volume of the
i th structural element. For the purpose of the
considerations in this paper the volume of the internal thin
rimmed gear is approximated using the rim thickness as a
main parameter. Therefore, the objective function of the
thinrim internal gear is given as follows:
Metaheuristics such as Simulated Annealing (SA),
evolutionary algorithms (EAs) and Swarm optimization
algorithms have been a subject of considerable interest in
the solving of complex mechanical power transmission
design problems. Recent work related to the minimum
mass design of spur gears include the application of
genetic algorithm on the mass minimization of single
stage helical gear unit, including the sizing of the shafts
and gear housing [78].
W ( x ) = V = b ( mn z + hF )
4
In this paper, an optimization procedure based on
simulated annealing method has been used in combination
with finite element method to generate optimal thin rim
gear configuration subjected to constraints on maximum
stress value. SA models the physical process of heating a
material and then slowly lowering the temperature to
decrease defects, thus minimizing the system energy.
Simulated annealing has the ability to avoid being trapped
in local minima. Therefore, it can provide global optimal
solutions even with discontinuous and nondifferentiable
objective functions. The Stressstrain calculations of the
root fillet stress of thinrim gear are performed on the
model based on FE method.
( H r + 1)2 mn2 z 2 (4)
where H r is the rim thickness, mn is the normal gear
modulus, hF is tooth height and b is the width of the
gear. The rim thickness can be defined as a ratio of the
rim thickness to the root radius
Hr =
Rs rf
rf
(5)
where Rs is the outer and rf is the root radius of an
internal gear.
The corresponding constraints of the minimum mass
optimization problem, take into considerations the
strength of the gear, trough the factor of safety, and
achievable size constraints. Functional constraints in the
form of inequality are given as
The paper is organized as follows. In section 2 the
formulation of the constrained thin rimed gear
optimization problem with appropriate constraints and
feasible domain is given. Moreover, section 2.1 presents
penalty method to convert formulated optimization task
into the unconstrained optimization problem. Sections 2.2
and 2.3 summarize the application of the Conjugate
gradient and Simulated annealing algorithms to the given
optimization problem, respectively. The formulation of
the computational finite element model is given in section
3. The obtained simulation results are presented and
analyzed in section 4. Finally, the future work and
conclusions are drawn in section 5.
g (x) =
[ F ]
F
SF > 0
(6)
where F is the working bending stress, [ F ] is the
allowable bending stress and S F is bending stress factor
of safety. The corresponding size constraints
min max
183
(7)
OTEH2016
OPTIMIZATIONOFPLANETARYGEARSANDEFECTSOFTHETHINRIMMEDGEARONFILLETSTRESS
are included in optimization model. Finally, taking into
consideration the given optimization problem (4) and
corresponding constraints (6) and (7) the nonempty
feasible domain is defined as follows
D = x \n
g (x) 0 h (x) = 0
2.2. Conjugate gradient
Conjugate gradient (CG) algorithm was first proposed as
the algorithm for the numerical solution of a system of
linear equations, and later generalized for the
unconstrained optimization of the nonquadratic functions
by Fletcher and Reeves [10]. The main advantage of CG
algorithm is its simplicity and a rapid convergence,
characterized by the quadratic convergence rate.
Moreover the algorithm does not require any explicit
second derivatives of the objective function making it less
computationally expensive.
(8)
Thus, the vector of design variables has the following
form
x = [ H r , z, mn , hF , b ] .
T
(9)
The descent direction in the initial iteration is chosen as a
direction of the negative gradient of the objective function
at initial point. Thus, the initial iteration of the conjugate
gradient method is the same as of the Steepest Descend
algorithm. Therefore, the descend direction in the initial
iteration is given as
2.1. Penalty method
Penalty method is a popular and intuitive constrainthandling method that successfully approximates the
constrained optimization problem with the unconstrained
optimization problem using the penalty function [9]. The
main idea of the method is to modify the objective
function to a following form
FW ( x ) = W ( x ) +
( )
d ( 0 ) = W x ( 0 ) ,
( )
P (x) ,
i
where W x( 0) is the gradient of the objective function
(10)
calculated at the initial point x( 0) .
i =1
The sequence of the approximations of optimal solution
where Pi ( x ) is the penalty function corresponding to the
{x( ) , k = 0,1,} is generated using the iterative form of
k
i th constraint. The penalty function is introduced to
eliminate any solution that leads to the constraint
violation. Therefore, the constrained weight minimization
problem (1) can be converted into an unconstrained
minimization problem that has the following form
min FW ( x ) = min W ( x ) +
x\ n
x\ n
the Conjugate Gradient method given as
x( k +1) = x( k ) + k d( k ) ,
(11)
i =1
{()
0,
xD
si x Ri x
/ D,
i = 1, , m
>0
>0
(16)
and d( k ) the descent direction of the algorithm in the k th
iteration given by
(12)
where Ri is the penalty factor that measures the
importance of the i th penalty function and si ( x ) is a
continuous function that considers the equality and
inequality constrains, defined as
si ( x ) = gi2 ( x ) , i = 1,, m .
min k ( ) = min W x( k ) + d( k ) ,
The penalty function can be defined as
Pi ( x ) =
(15)
where k > 0 is the step length computed such that the
function of one variable has a local minimum at = k ,
that is
P (x) .
(14)
(k )
( )
( )
W x( k ) ,
k =0
=
.
(k )
( k 1)
+ k d
, k 1
W x
(17)
There are several formulas for the calculation of the
conjugacy parameter k , among them the bestknown is
the FletcherReeves [9] formula that quickly finds the
optimum of the convex quadratic function, defined as
(13)
Based on the definition of the penalty function given in
(12) it can be seen that the penalty method directly
evaluates every feasible solution based on their objective
function values, while the penalty function is applied to
the infeasible solutions thus decreasing their fitness value.
A major disadvantage of the penalty method is the
adequate choice of the penalty factor, which largely
affects the efficiency of the search process. Penalty factor
primarily depends on the level of violation, and varies
accordingly.
k =
where
( )
W ( x ( ) )
W x ( k )
k 1
(18)
denotes the Euclidian norm. The key
advantage of the CG algorithm lies in the fact that the
descent direction in the current iteration is constructed
based on the gradient of the objective function in the
current and previous iteration. Thereby the convergence
speed of the algorithm can be improved with the careful
184
OTEH2016
OPTIMIZATIONOFPLANETARYGEARSANDEFECTSOFTHETHINRIMMEDGEARONFILLETSTRESS
control of the parameter k .If the solution in the previous
iteration leads closer towards the optimal solution
parameter k takes higher values, and the decent
direction in the previous iteration prevails in the search
process. However, if the solution leads away from the
optimum the gradient of the objective function in the
current iteration takes the crucial part in the construction
of the descend direction.
The process of metal annealing requires a precise control
of the temperature and the cooling rate which is specified
by the annealing schedule. From (20) is clear that as the
temperature decreases the probability of accepting worse
solutions decreases exponentially. Rapid cooling results
in the SA algorithm failing to explore entire solution
space and converge towards the global optimal
configuration. On the contrary, extremely slow cooling
will produce slow convergence and long execution time
of the SA algorithm. Therefore, to avoid such undesirable
effects the geometric cooling schedule is used, as the most
commonly used cooling schedule in SA literature,
described by the following temperatureupdate relation
2.3. Simulated annealing
Simulated annealing (SA) is a metaheuristic technique
designed to mimic the annealing process in material
processing which is successfully applied to global
optimization problems. It is first presented by Kirkpatrick
et al. [11] for a combinatorial optimization problems as a
generalization of the Metropolis Monte Carlo integration
algorithm. The fundamental idea of the simulated
annealing algorithm is to use random search which not
only accepts changes that minimize the objective
function, but also retain some of the solutions that lead
away from the local minimum. The SA algorithm uses the
Boltzmann probability distribution to keep the worse
solutions, and avoid being trapped in local minima.
T ( k +1) = T ( k ) ,
where is a constant cooling factor such that ( 0,1) .
2.4. Termination criteria
The adequate choice of the termination criterion could
determine whether or not the optimization algorithm is
efficient and successful in the search of the optimal
solution. To terminate the iterative procedures of the
proposed optimization algorithms the following
termination criteria are used
The algorithm will run until a predetermined number
of iterations is reached and
The relative change in objective function value
between two consecutive iterations is less than the
predefined positive number, mathematically written
Algorithm starts from an initial solution vector x( 0) , with
an initial temperature parameter T ( 0 ) and the temperature
cooling factor . A candidate solution x c , is generated
randomly applying a small perturbation to the solution
vector x( 0) , and objective function values are evaluated.
Then, the objective function difference can be expressed
as:
( ) f (x
f = f x
( 0)
).
( )
(
W (x )
W x( k ) W x( k +1)
(k )
(19)
If the change in the objective function values between
current and candidate solution is positive, i.e. f > 0 , the
candidate solution improves the objective function, and
the current solution is replaced by the candidate solution.
Otherwise, the objective function difference is negative,
i.e. f < 0 and the candidate solution leads away from
local minimum. Therefore, the candidate solution x c is
accepted with the transition probability determined by:
p=e
f
kT
( f < 0 ) ( p > r )
otherwise
(23)
3. COMPUTATIONAL FE MODEL
There are major difficulties with the traditional method
used in gear design regarding the computation of the
relative deformations and stresses. The model for
determining the bending stress is derived from the
cantilever beam and presents an approximation to the
exact stress distribution within the gear tooth. To correct
the beam model additional semiempirical stress
concentration factors are added to the model.
where k is the Boltzmanns constant and T is the
temperature parameter which controls the annealing
process. Such solution is accepted to explore other
potential regions of the solution space and converge
towards the global minima. Therefore, the solution for the
( i + 1) th iteration is accepted according to:
xc ,
x( i +1) = ( i )
x ,
) .
When the relation (23) is satisfied the search procedure
has reached the region of the optimal solution and further
iterations cannot lead to the significant improvement of
the objective function.
(20)
(22)
The model employed in this paper is based on the finite
element method to compute relative deformations and
stresses and provides an alternative to the traditional
method for the calculation of the bending stress. This
model does not require any empirical factors to compute
bending stress which can be predicted accurately, thus
allowing fully autonomous optimization process with a
high number of design variables as well as satisfying
complex constraints.
(21)
Where r is a uniform random number in the range of
[ 0,1] .
185
OTEH2016
OPTIMIZATIONOFPLANETARYGEARSANDEFECTSOFTHETHINRIMMEDGEARONFILLETSTRESS
The contact surfaces between the teeth of the pinion and
internal gear are discretized with a large number of finite
elements in order to accurately represent the complex
involute shape of the teeth profile. Moreover, the regions
of possible stress concentration, such as the trohoide arc,
are modeled by finer FE mesh.
different rim thickness to evaluate how this geometric
property may affect the bending fatigue failure and crack
propagation.
Table 1. Parameters of the planetary gear set used in this
paper (all dimensions are in mm , unless specified
differently)
The FE model for calculating stressstrain field at tooth
root of the planetary gear set used in this paper is shown
in Picture 1.
Sun
Planet
Internal
No. of teeth
34
18
70
Module
1.5
1.5
1.5
Pressure angle [ ]
21.3
Outer diameter
53.893
31.594
105.408
Root diameter
47.019
24.774
112.382
51
27
105
Base circle
Youngs
N/mm 2
modulus
Poisons ratio
2.07 105
0.3
Density kg/m3
7860
It is assumed that the critical working bending stress is at
the point where maximum equivalent stress of the FE
model is achieved.
Picture 1. FE computational model of the planetary gear
set used in the analysis
The proposed optimization methods, i.e. simulated
annealing and conjugate gradient optimization algorithms
are applied to the constrained thin rimed gear weight
minimization problem using the described penalty
method. Therefore the constrained optimization problem
is relaxed to unconstrained optimization problem using
the modified objective function, according to the (10).
The solutions of the FE model are obtained using Gauss
elimination nonlinear equation solver [12]. Due to the
high degrees of freedom of the FE model and contact
conditions between teeth profiles along the line of action,
the stiffness matrices of the system are often singular. To
avoid this numerical problem, the FE mesh is carefully
discretized in the regions of possible singularities. The
calculations of the FE model are performed assuming
homogeneous and isotropic material, without any
imperfections or damages.
The convergence of the optimum internal gear design
parameter values obtained using the proposed
optimization procedure based on SA and CG optimization
algorithms are presented in Picture 2. Moreover, the
variation of the objective function for the rim thickens
parameter for different number of evaluations using SA
and CG algorithms are shown in Picture 3.
4. SIMULATION RESULTS
In this section the results of the numerical simulation are
presented to verify the improvements in the optimal
design solution using the proposed SA based optimization
procedure compared to the results obtained by
conventional CG algorithm.
2.6
Simulated annealing
Conjugate gradient
2.4
Objective function value [kg]
2.2
The optimization procedure, that takes into
considerations the effects of the rim thickness on the
internal gear root fillet stress, is applied to an example
planetary gear set. The example planetary gear set, whose
corresponding FE model is given in Picture 1, consists of
floating sun gear and three equally spaced planetary
gears, with the internal gear rigidly connected to the
housing. The torque of Tin = 1500Nm is applied to the
input of the system. The basic design parameters of the
considered planetary gearbox are listed in Table 1.
2
1.8
1.6
1.4
1.2
1
0.8
The root fillet stress has been obtained from the FE model
of internal gear with the constant geometric properties and
10
15
20
Number of Iterations
25
30
35
Picture 2. FE computational model of the planetary gear
set used in the analysis
186
OPTIMIZATIONOFPLANETARYGEARSANDEFECTSOFTHETHINRIMMEDGEARONFILLETSTRESS
number of iterations, although the SA algorithm can
efficiently solve the optimization problem and produces
the solution with the minimal weight that satisfies the all
constrains. It is observed that the SA algorithm
outperforms gradient based algorithm, and achieves
global optimal design configuration.
14
Simulated annealing
Conjugate gradient
13
12
Rim thickness [mm]
11
10
Future work may include a multiobjective optimization
problem that includes additional technoeconomic
objective function. Moreover, the design parameters of
the optimization problem can be further extended to
include additional parameters.
9
8
7
6
5
4
OTEH2016
References
0
10
15
20
Number of Iterations
25
30
35
[1] Bill,R.C.: Advanced Rotorcraft Transmission
Program, 46th Annual American Helicopter Society
Forum, (1990).
[2] Lancaster,O.E.: Jet propulsion engines, Princeton
University Press, 2015.
[3] Tong,B., Walton,D.: The optimisation of internal
gears", International Journal of Machine Tools and
Manufacture, 27(4) (1987) 491504.
[4] Yokota,T., Taguchi,T., Gen,M.: A solution method
for optimal weight design problem of the gear using
genetic algorithms, Computers & industrial
engineering, 35(3) (1998) 523526.
[5] Kramberger,J., Sraml,M., Potrc,I., Flasker,J.:
Numerical calculation of bending fatigue life of thinrim spur gears, Engineering fracture mechanics,
71(4) (2004) 647656.
[6] Bodas,A., Kahraman,A.: Influence of carrier and
gear manufacturing errors on the static load sharing
behavior of planetary gear sets, JSME International
Journal Series C, 47(3) (2004) 908915.
[7] Chandrasekaran,M.. Padmanabhan.S.: Single speed
gear box optimization using genetic algorithm",
ARPN Journal of Engineering and Applied Sciences,
10(13) (2015) 55065511.
[8] Wang,H., Wang,H.P.: Optimal engineering design of
spur gear sets, mechanism and Machine Theory, 29
(7) (1994) 10711080.
[9] Byrd,R.H., LopezCalva,G., Nocedal,J., A line search
exact penalty method using steering rules",
Mathematical Programming, 133(1) (2012) 3973.
[10] Fletcher,R., Reeves,C.M.: Function minimization by
conjugate gradients, The computer journal, 7(2)
(1964) 149154.
[11] Kirkpatrick,S.,
Gelatt,C.D.,
Vecchi,M.P.:
Optimization by simulated annealing, Science,
220(4598) (1983) 671680.
[12] Hughes,T.J.: The finite element method: linear static
and dynamic finite element analysis, Courier
Corporation, New York, 2012.
Picture 3. FE computational model of the planetary gear
set used in the analysis
From Picture 2 it can be seen significant improvement in
optimal design solution that the SA based optimization
procedure achieves, which represents global optimal
solution with minimum mass design. The CG based
optimization procedure converges towards the local
solution therefore achieving worse optimal design
solution. However, concerning the convergence properties
of the algorithm, it is observed that the CG algorithm
converges towards optimal solution in significantly fewer
number of iterations, where the SA algorithm requires
three times more number of iterations to converge. This is
primarily due to the global search characteristics of the
SA algorithm, which requires more iterations to
thoroughly explore feasible domain and find a global
optimal solution.
Comparing the data presented on Picture 2 and Picture 3 it
is observed that as the rim thickness parameter changes,
the mass of the internal thin rimed gear changes
proportionally. This conclusion is in agreement with the
assumptions made in FE model, and the material
homogeneous and isotropic assumptions.
The observations made on the numerical simulations
show that the proposed thin rimed gear optimization
procedure based on SA algorithm can achieve better
design solutions, compared to the traditional gradient
based algorithms, however with the cost of higher number
of iterations.
5. CONCLUSION
In this paper, the optimization procedure of the thin
rimmed gear design in relations to the root fillet stress is
developed and analyzed, employing the FE model to
compute deformations and root fillet stress in the
optimization process.
The presented simulation results show that the CG
algorithm converges towards the solution in fewer
187
PROJECTION OF QUALITY A COMPLEX TECHNICAL SYSTEM
LJUBIA TANI
Project Management College, Belgrade, ljtancic@gmail.com
PETAR JOVANOVI
Project Management College, Belgrade
SAMED KAROVI
Project Management College, Belgrade
Abstract: The paper presents solution of task interior ballistic projection weapons for concrete caliber small arms
weapons. Optimal solution is looked in selection the best physical  chemical and ballistic characteristics powder, mass
powder, powder chamber and other characteristics with him is execution interior ballistics calculation on the computer.
Solution is compared with existence solutions and conclusions are given.
Keywords: projection small arms weapons, optimization.
Quality management project is one of the subprocesses
of the global project management concept. The main
objective of a particular project management, in addition
to minimizing wasted time, resources and costs, is the
completion of the project within the required or necessary
quality [6].
1. INTRODUCTION
The implementation of modern business and other
activities, ventures and projects is burdened with
extraordinary complexity and uncertainty, which are
caused primarily by an increasing and growing
complexity of the projects themselves and the
environment in which they work and the extremely rapid
pace of the development of science, technology and
civilization as a whole.
This paper discusses the design of complex technical
systems, and the design of the optimum solution to a
ballistic design of a complex technical system [7] and [8].
A specific caliber weapon as a complex technical system,
structurally composed of aggregates, components, subassemblies and parts and is functionally designed for
firing shots in small arms is described.
This complexity most often leads to serious problems in
the implementation of various ventures and projects
which is reflected in substantial delays and increased total
cost of implementation and ineffective implementation on
the whole [1] and [2].
Firing bullets is a complex thermodynamic and
gasdynamic process of an extremely fast, almost
immediate, converting the chemical energy of the powder
first into the heat and then into the kinetic energy of
powder gases that cause the movement of the missile
system  filling  tube movement [9] and [10].
In order to realize complex projects effectively various
management disciplines, methods and techniques are used
today. The best of them appears to be project
management, an operational management discipline that
enables effective management of the implementation of
various projects and programs [3].
For the illustration: The process of firing takes a few
hundredths of a second, the maximum pressure of powder
gases reaches a value of order 100 to 600 MPa, the
temperature of powder gases at the moment of creation is
21003800 K and in the moment of leaving the missile
tubes it is 1500 to 2000 K; the maximum velocity of the
projectile at the moment it leaves the pipe is 70 to 1500 m
/ s and the acceleration is of order 150000 to 600000 m / s
** 2.
For an efficient project management of complex technical
systems (military, industrial, construction and other
projects) it is necessary that special attention should be
paid to the management of risk and the quality of the
project as part of a comprehensive project management
[4], in addition to technical, time, cost constraints and
elements.
Risk management involves a set of management methods
and techniques used to reduce the possibility of realizing
unwanted and harmful events and consequences, thus
enhancing the ability to achieve the planned results [5].
So it is a set of methods that enable minimization of
losses and reduce the probability of losses as well as the
costs incurred by this reduction.
Order of magnitude indicates that this is a very dynamic
phenomenon of impuls character and that there is a high
risk in the theoretical and experimental phases of the
design.
The aim of this paper is to specify, based on the defined
required data such as muzzle velocity and maximum
188
OTEH2016
PROJECTIONOFQUALITYACOMPLEXTECHNICALSYSTEM
pressure of powder gases, only a certain number of
optimum ballistic parameters [8]: the most suitable model
and type of gunpowder, the mass of propellant and
propellant chamber volume based on the criteria of the
quality of the complex technical system and highlight
possible risks that may accompany the execution of this
complex project.
conservation of mass, momentum and energy using
GaussOstrogradski transformation a mathematical model
is developed for an arbitrary moment of time during the
combustion of gunpowder. The process of firing is
observed from the time of the burning propellant behind
the projectile create enough pressure at which the sleeve
missiles will be etched into the grooves of pipes and start
up the missiles. It is assumed that at this time all the
initial and boundary conditions are known. After
completion of the powder combustion, the twophase flow
becomes a single flow, the flow of a propellant gas. The
established mathematical model then becomes a classic
gas dynamic model [9].
2. BASICS OF BALISTIC PROJECT DESIGN
AND OF MATHEMATICAL MODEL OF
TWO PHASE FLOW
The process of firing in the complex chamber systems is a
gasdynamic process that is, in the space of the stationary
bottom of the tube and the moving projectile,
characterized by streaming of the two phases: solid
from burning gunpowder grains and gaseous from
powder gases as a propellant combustion product [10].
The ballistic design involves determining the optimum
basic structural characteristics of tube filling conditions
and energy characteristics of firing rules based on the
quality of the products that are put before a system is
described, and is defined by three main objectives [8]:
The system of equations is derived in Eulerian
coordinates t (time) and x (any position in the tube of the
breech face to the bottom of the projectile) [10], and then
transformed into a system of Lagrangeian coordinates t
and s (the mixture of gunpowder and propellant gases in
the individual points behind the projectile). The initial
assumptions and the overall performance, with additional
equations, initial and boundary conditions are given in
[12] and [13]; here, only the general form of partial
differential equations is described:
 the classic task of defining the barrel caliber, projectile
weight and muzzle velocity and determining the
characteristics of the channel tube, the terms of system
charging, the basic characteristics of charge and all the
data necessary for the calculation of pipe systems and the
projectiles;
Yk
+
t
k ,l
l =1
Yl
= bk
s
k =1, ... , m
(1)
where: m  number of equations
(m = 5 while powder combustion lasts and
m = 3 after the combustion is finished) and
the general task of adopting the maximum range, the
penetration on a given target range for subcaliber
projectiles and armor projectiles or the horizontal
distance and height of the target for antiaircraft
systems, selecting the system caliber, the projectile
weight, the muzzle velocity and all the other
parameters, in the same way as the classical task does.
It is obvious that the classical task is part of the
generalized task. The general task generally consists of
two parts. The first part is the process of selecting an
optimal complex technical system, the other is the
classic problem of ballistic design.
ak,l  variable coefficients, some of which are equal zero.
A given system of equations is nonlinear, nonhomogeneous and nonstationary, and is solved
numerically using the theory of finite differences.
Numerous numerical schemes can be set upon the system.
When choosing a scheme three conditions are to be me
[14]:
1. Condition of Consistency the scheme cannot have a
tendency to be a solution to any other systems
2. Condition of Convergence the errors in the
calculation must be limited
the simplified task of, compared to the classic, defining
some additional constraints. Thus, for example, a
complex technical system is designed to use either the
same ammunition as any already existing system or the
same pipe. This task is worked out in the same way as
the classic task, however, some of the parameters are
already given in advance and need not be analyzed.
This task involves the problem of modification of some
part of the existing complex system on the basis of
change in the internal ballistic parameters.
3. Condition of Stability the errors tend to zero value.
The condition of stability and convergence of numerical
schemes is developed [15,16] and a program for a
personal computer is devised that is used to analyze the
influence of the initial parameters [17].
Stability condition of the numerical scheme is defined by:
max 1
The basic elements of the internal ballistic design are the
pipe and the bullet. They are a necessary condition for the
process of firing bullet in the complex chamber system to
happen. The today's theory of war anticipates as large a
range of complex systems as possible, so there is a
tendency in the development to modify the existing or
construct complex technical systems to achieve a better
performance [11].
(2)
where is : max  largest eigenvalues matrix [A].
Let is the increase by time and h is the increase by s
and = rh , then value r is solved from the convergence
condition of the total system error, [15]:
Based on the initial assumptions and the law of
189
OTEH2016
PROJECTIONOFQUALITYACOMPLEXTECHNICALSYSTEM
m
Al =
ak ,l
k =1
rl ,i 1
Al i
r = min min ( rl ,i )
i l
The parameters evidently have different influence
gradient on results, so, for example, the unit grain burning
speed has the smallest percentage of change and the
highest influence on the results, while pressure engraving
has a reversed situation.
(3)
3. THE RESULTS OF THE FACTOR PLAN OF
EXPERIMENT 2n
After the influence parameters selection is finished, all
parameters which increase the results is taken and with
these parameters values the output limit is received (table
2.). The initial parameters, based on (QPI) have high and
low border allowances. After that, all initial parameters
which decrease exit results are taken and so the other limit
output is received.
The program solution to the theoreticalnumerical model
can be tested with different values of the initial data
which influence the outputs differently. All initial
parameters must satisfy the Quality Product Instruction
QPI and the Instruction of Commission International
Permanente CIP [18] which define the allowed tolerances
for some parameters. However, initial parameters can
have higher and lower value as allowance values with the
same reliability.
Table 2. Initial parameters (maximum, medium and
minimum)
To estimate the initial parameters influence on the
mathematical model results and to execute the parameters
rank, the experiment factor plan 2n, based on the [19], is
carried out. The factor plan results are:
There are initial parameters which have not significant
influence on the model either alone or in combination
with other initial parameters. They are: conduct heat
coefficient (0), preliminary period time (t0), projectile
beginning position (X0), powder gases density at the
beginning of first period (0), dynamical viscosity
coefficient (), transition heat coefficient (b) and initial
powder grain temperature (Tb0).
There are initial parameters which have an essential
influence on the model either alone or in combination
with other initial parameters. They are: pressure
engraving (p0), projectile mass (m), powder mass (mb),
powder gases covolume (), propellant grain area (Szo)
and unit grain burning speed (uzo).
1.
p0
2.
3.
m b0
4.
5.
Sz0
6.
u z0
+
+
+
+
+
+

Percent % pm, bar
20
20
1.56
1.56
1.55
1.55
1.46
1.46
1.3
1.3
1.3
1.3
3009
3009
3055
2964
3059
2959
3032
2990
3075
2945
3112
2906
V0, m/s
712
710
709.7
710
717
702
713
709
715
704.8
721
700
ULAP762.min
7.9 10 10
7.8 1010
m b0 ,kg
0.001645
0.00162
0.0015948
m, kg
0.0080232
0.0079
0.0077767
Sz0 , m 2
4.0733 10 07
4.127 1007
4.1806 1007
,m3 / kg
0.0009243
0.00091
0.00089769
p0 ,Pa
1.5 10
+07
1.25 10
p m ,Pa
333.8 10 +06
311 10 +06
279.5 10+06
V0 ,m / s
(+9.3%)
736.7 (+3.1%)
720.8
(8.45%)
689.8 (3.474%)
+07
1 10
+07
4. RESULTS OF CALCULATION OF TWOPHASE FLOW AND PROJECTION
BARREL
Model (chemical composition) and the type of powder
(shape and size) are usually selected on the basis of the
total pulse pressure of powder gases. If it is necessary to
design a new powder first, then the desired chemical
composition of gunpowder is defined. Based on the
chemical composition of the powder the thermo chemistry
of gunpowder [13] is defined, that is, the mathematical
model for calculating the specific characteristics of
gunpowder. Based on the chemical processes occurring
during the combustion of gunpowder, the relations of the
products of combustion (major and minor) are set as well
as the energy balance and the relations of these equations
are developed to perform the calculation of physical,
chemical and ballistic characteristics. For a selected
complex system of a particular caliber, based on the
analysis conducted in the literature [21] the gunpowder
NC  08 is adopted, with all the physical  chemical and
ballistic characteristics relevant to internal ballistic
calculation.
Table 1. Initial parameters influence on calculations
results
Level
ULAP762.srd
8 1010
The execution of the program with the border values of
initial parameters yields the highest and lowest flow
characteristics values. Out of all the flow characteristics,
this paper describes and analyses only the powder gases
pressure and the projectile speed. If the mathematical
model is defined correctly, the experimental results must
be within or near those limits output calculations.
Table 1. shows the allowed tolerance measures in
percentages for parameters with significant influence on
the results for automatic rifle 7.62 mm and ammunition it
uses, based on QPI and CIP [18]. The parameters
influence upon the output is received by computer
program realization for experiment factor plan 2n in which
the relative connection criterion of exit and initial
parameters influence are defined.
Entrance
Parameter
rank
ULAP762.max
u z0 ,m / sPa
Exit
rank
6.
4.
3.
5.
2.
1.
190
OTEH2016
PROJECTIONOFQUALITYACOMPLEXTECHNICALSYSTEM
The calculations are performed on the basis of the
FORTRAN language program on the PC computer [13].
The paper does not offer the program solution but only
the calculation results with a certain comment presented
in Descartes coordinate system and in a form of family
curved lines. Picture 1. presents flow characteristic
propellant gas pressure (p) as functions of time (t) and
situation in the tube (x), the function of propellant gas
density is similar to that of pressure.
Risk management is somewhat different from other
management processes, as in the management of risk it is
difficult to exert the risk event control, but a prior
preparation and response to possible future risk events is
made, in order to reduce the probability of their
occurrence and to increase the probability of achieving
the expected results of the project.
Since the experiments were carried out on a real system
that captures a very dynamic character and appearance of
the pole to the probability of the risk to the greatest extent
possible a detailed analysis of risk factors (highrisk
event, risk probability and the size rolls) is conducted. For
a reliable risk management all the subprocesses such as
risk identification, risk analysis and risk assessment, risk
response planning, and control of the application of the
risk are carried out [24].
Analyzing risks in the project measurement and data
processing of complex technical systems in Table 3 we
conclude that there is a risk that has a critical importance
for the realization of the whole project. The event which
carries the greatest risk of contracting the project is to
supply the devices. Otherwise, the risk event causes other
risk events too. To prevent the occurrence of the risk we
have to reach the agreement with the supplier of
equipment to pay large fines for equipment delay and
oblige him to train the staff.
Picture 1. Pressure propellant gas
Obviously, the internal ballistic characteristics dependant
diagrams fully meet the goal defined in the introduction.
After this phase, the design follows the design of the inner
tube route with all the details such as rifling, transition
cone, gunpowder chamber, etc. After designing the inner
tube of the route the route of external design and
conceptual design of the bullet follows. The designing of
the external route pipes can be realized in classical
modern analytical methods and numerical methods [22]
and [23].
Measuring the propellant gas pressure in the small arms
barrel caliber 7.62 mm is executed with for the purpose of
comparing the calculation data and the real data in the
small arms. For the experiment execution, the trunk,
barrel, wrapping and pressure transducer piesoelectrical
measurer were provided. A measurement chain for
measuring pressure propellant gas by block scheme
(Picture 3.) is connected to the complex technical system
barrel.
Using a matrix to draw diagrams in the previous program
for engineering drawing and design "Katia" is modeled
conceptual design of pipes:
Picture 2. Modeled tube firing weapons 7.62 mm
Picture 3. Block scheme for data measuring and
processing
5. A COMPARATIVE ANALYSIS OF
EXPERIMENTAL DATA AND RESULTS OF
CALCULATION
MTP  piesoelectrical measure transformer pressure, CC coaxial cable, AS amplifier signal, RS  registrar signal
(measure tape recorder), DO  digital oscilloscope, C computer, Pl  plotter and Pr  printer
Project management includes the management of project
risk [4], in order to increase the probability of achieving
the desired objectives of the project and reduce the
possibilities of adverse events and adverse outcomes [5].
191
OTEH2016
PROJECTIONOFQUALITYACOMPLEXTECHNICALSYSTEM
model and type of powder, powder mass and volume of
the powder chamber. Then the design of the inner tube
route with all the details such as the cartridge chamber,
the transition cone and the missile lead is made. After
designing the inner tube of the route follows the route of
the design of the external tube and the conceptual design
of the bullet. This projected system, compared to the same
caliber specific system has optimum properties, which
justifies the application and verifies the chosen
calculation method.
This paper gives a theoretical  experimental analysis of
the process of discharging the tube complex systems
based on experimental research and numerical modeling
on a computer. The analysis revealed that the medium
terms of the calculation give the optimum inputoutput
parameters and acceptable results of the calculations for
the specific complex technical system [21].
Picture 4. Diagrams model and central experiments value
The comparative analysis included the pressure of powder
gases and projectile velocity in the complex system pipe
over time. The results of experiments and models confirm
the nature of the change of pressure and velocity, and the
basic assumption of the mathematical model that the
pressure of powder gases and projectile velocity function
of time and the projectile path. Comparing the
experimental and calculated results of the propellant gases
pressure and the projectile velocity their good mutual
compliance is recognized, which confirms the correctness
of the mathematical model.
Pressure curves on all the measure points and curves of
the model on the measure point 1 are presented as
maximum  (curve h  higher), medium (curve m)
and minimum  (curve l  lower) obtained with the
initial parameters. The mathematical model and computer
program are corrected with the conditions for
experimental barrel and those calculations are executed
with such conditions, because the firing conditions in the
experimental barrel are different from the conditions in
the fighting barrel. Fig. 1. shows that medium
experimental results are in calculation results zone limited
with higher and lower curves. For this reason, at the end,
medium calculation results are compared with medium
experimental values for position 1.
References
[1] PMI PMBOK Guide, A Guide to the Project
Managment Body of Knowledge, 4th Edition,
Project Managment Institute, Newtown Square, The
USA, 2008.
[2] Kerzner,H.: Project Management, A Systems
Approach to Planning, Scheduling and Controlling,
10th ed. John Wiley&Sons, Inc. New Jersey, The
USA, 2009.
[3] Turner,J.R.: The handbook of projectbased
managmentLeading strategic change in organizations, McGrawHill, The USA, 2009.
[4] Jovanovi,P.: Project Management, Faculty of
Organizational Sciences, Belgrade, 2006.
[5] Smith,P.: Merritt G.: Proactive Risk Management:
Controlling Uncertainty in Product Development,
Productivity Press, New York, 2004.
[6] Barkley,B.T.: Project Management in New Product
Development, New York, McGrawHill, 2008.
[7] Tancic,Lj.: Classic Interior Ballistics, Ministry of
Defence, Human Resources Sector, Military
Academy, Belgrade, 2006.
[8] Tancic,Lj.: Interior Ballistic Design, Ministry of
Defence, Human Resources Sector, Military
Academy, Belgrade, 2014.
[9] Jaramaz,S., Mickovic,D.: Internal Ballistics,
University of Belgrade, Faculty of Mechanical
Engineering, Belgrade, 2002.
Interior
ballistics,
Military
[10] Cvetkovic,M.:
The average experimental value for maximum pressure
(294.77 MPa) practically is identical with the model
(295.1 MPa) for 20 elements for variable s, and all
experimental curves are compatible with the calculations
so that the correctness of the given theory is validated.
The experimental results prove the character of pressure
in smallarms barrel as function of time and position in
the barrel.
6. CONCLUSION
This paper presents the design choice of the optimum
solution to the ballistic design complex systems for a
particular caliber. For an efficient project management in
the complex technical system implementation, special
attention is paid to risk and project quality management.
The probability of risk is reduced to a minimum by the
analysis of risk factors and a strategy is defined to reduce
and respond to the risk event. What specifically should be
done to minimize the impact of risk events is described in
detail.
The quality of the project is defined by the regulations of
permanent international commission on the quality of
products that define the allowed tolerance for all inputs of
crucial influence upon the outputs of specific complex
systems.
The main goal of the work is reflected in the choice of the
optimum ballistic parameters, primarily: the most suitable
192
OTEH2016
PROJECTIONOFQUALITYACOMPLEXTECHNICALSYSTEM
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Technical Academy of Yugoslav Army, Belgrade,
1998.
Jovanovic,P., Tancic,Lj., Lajsic,Dj., Vukovic,M.:
Innovation Management and Project Management,
College of Project Management, Belgrade, 2012.
Cvetkovic,M.: Application insteady gasdynamic on
the interior ballistics problem to small arms, Ph. D.
dissertation, High Military Technical School,
Zagreb, 1984.
Tancic,Lj.: Numerical computation of insteady
models in interior ballistics to small arms, Ph. D.
dissertation, Military Technical Academy of
Yugoslav Army, Belgrade, 1997.
Smith,G.D.: Numerical Solution of Partial
Differential Equations: Finite Difference Methods,
Oxford University Press, 1985.
Cvetkovic,M., Tancic,Lj.: A comparisons analysis
experimental and calculations results for twophase
flow in the small arms, II International Symposium
Contemporary Problems of Fluid Mechanics,
University of Belgrade, Faculty of Mechanical
Engineering, Chair of Fluid Mechanics, Belgrade,
1996.
Cvetkovic,M., Tancic,Lj.: Analysis of conditions
numerical modeling of twophase flow in small arms,
XXII Yugoslav congress of applied and theoretical
mechanics , JUMEH 97, Vrnjaka banja, 1997.
Cvetkovic,M., Tancic,Lj.: Initial parameters
[18]
[19]
[20]
[21]
[22]
[23]
[24]
193
influence on two phase flow model in the small arms
barrel, 6th Symposium On Theoretical and Applied
Mechanics, Republic of Macedonia, Struga, 1998.
October 13
***, Commission Internationale Permanente (CIP),
Geneve, 1995.
Pantelic,I.: Introduction in theory engineer's
experiments, People's University  Radivoj Cirpanov,
Novi Sad, 1986.
Tancic,Lj.: The experimental research two phase
flow parametars in the small arms barrell, Review
Technical Research of Yugoslav Army, Belgrade,
1999., No 2, P. 3 9
Cvetkovic,M., Tancic,Lj.: Two phase flow models
sensibility in small arms on initial parameters,
Review Technical Research of Yugoslav Army,
Belgrade, 1999., No 3, P. 3 7
Tancic,Lj.: Optimization of internal ballistic
parameters at the design small arms, Second
Scientific Conference of the defense industry (I8I13), Belgrade, 2007.
Jovanovic,P., Tancic,Lj.: Project quality menagment
of complex interior ballistic systems, 5th International
Scientific Conference on Defensive Technologies,
OTEH 2012, Belgrade, 2012., 1819 September
Jovanovic,P.: How to become a good manager, College
of Project Management, Belgrade, 2010.
STRESS ANALYSIS OF INTEGRATED 12.7 mm MACHINE GUN MOUNT
ALEKSANDAR KARI
Military Academy, University of Defense in Belgrade, aleksandarkari@gmail.com
DUAN JOVANOVI
Faculty of Engineering, University of Kragujevac, djovanovic.jovanovic7@gmail.com
DAMIR JERKOVI
Military Academy, University of Defense in Belgrade, damir.jerkovic@va.mod.gov.rs
NEBOJA HRISTOV
Military Academy, University of Defense in Belgrade, nebojsahristov@gmail.com
Abstract: The paper describes the problem of integration 12.7 millimeters machine gun on the mobile platform. Based on
the dimensions of the machine gun modeling of machine guns mount with cradle is done. The completed model is fully
functional and realistic. Using optimized internal ballistic parameters the calculations of recoil forces and loading of mount
and rotating bearing are executed. The loading calculation of bearing was made in two ways. In the first case finite element
method is applied and software package FEMAP was used. The second method is based on calculating the resistance
components of bearing from the equilibrium condition. At the end the comparative analysis of data obtained from these two
methods was done.
Keywords: integrated machine gun, stress analysis, recoil force, bearing.
characteristics that affect the efficiency of weapons,
primarily projectile velocity at the muzzle, and the required
tactical and technical characteristics of the machine guns
presented in Table 1 [2].
1. INTRODUCTION
The resistance of weapon and its firing stability depend on
the value of recoil forces [1]. In the case of the integration
of weapons, if the barrel is rigidly connected to the mount,
the recoil force is fully transmitted to the mount. In order to
avoid inconveniences of such connections, or extended
firing time load mount and reduce its intensity, barrel is
elastic connecting to the mount which enables the
movement of the barrel or whole weapon during firing in
the direction of the axis of the barrel.
Table 1. Tactical  technical characteristics of machine
gun M87
Characteristic
Caliber, d
Barrel length, l
Total length of machine guns, L
Weight of machine guns, m
Rate of fire, n
Effective range, Dmax
The study of conditions to be satisfied by stand of machine
gun to be installed on a mobile platform is the main goal of
this paper. This involved the construction of the rotating
mount that could withstand the loads created by recoil force,
as well as calculation of the forces that are transmitted on
the mobile platform. As an example, for the integration was
selected machine gun 12.7 mm M87.
Value
12,7 x 108 mm
1100 mm
1560 mm
24,8 kg
700 bullet/min.
1500 m
2. RECOIL FORCE
For the selection the most rational construction of mount, it
is necessary to define the acting forces on the mount as a
whole, but also on its individual parts. In the process of
firing of the machine gun, force of powder gases acting on
the bolt. After unlocking, on the bolt acts the force of the
return spring that opposes the force of the pressure powder
gases.
In order to reach the optimal solution, the conditions of
minimal force and minimal recoil mass of mount are set,
taking into account the boundary allowable conditions of
materials resistance, as well as, the conditions in tactical
use. The research was carried out through four main
phases:1) Optimization of internal ballistic parameters and
their influence on the recoil force of the machine gun; 2)
Determination of machine gun recoil force; 3) Designing the
appropriate mount based on the construction of the machine
gun; 4) Determination of the mount and bearing loading.
If the machine gun is on the mount and there is an elastic
connection between gun and mount, there is the resistive
force of recoil (force of shock absorber, buffer, hydraulic
brake, etc.). There is also a frictional force on the guide rails
of machine guns which opposes to movement of a machine
gun during recoil.
The goal is to minimize the pressure of powder gases in the
barrel, but with as less as possible reduction of
194
OTEH2016
STRESSANALYSISOFINTEGRATED12.7MMMACHINEGUNMOUNT
Finally, the total recoil force is:
JJJG JJJG JJJG JJJG JJG
Pkn = Fbg + Fop + Fbf + Ftr
(1)
where: Fbg  force of gas pressure, Fop  force of spring,
Fbf  force of buffer and Ftr  frictional force.
Activity of these forces is shown in Picture 1.
Picture 3. The optimal calculated velocity of the
projectile in the barrel
When the bottom of the projectile passed over the orifice,
gun powder gases enter the returnees chamber and acting on
the piston, creating a new force.
Picture 1. The forces acting on the weapon
The listed forces are parallel to the axis of the barrel, so that
the intensity of the total recoil force can be expressed as:
The pressure in the gas cylinder is calculated based on the
model by E. L. Bravin [5]:
Pkn ( x1 , x2 , t ) = Fbg ( t ) Fop ( x1 ) Fbf ( x2 ) Ftr ( x2 ) (2)
where: t time from the start firing, x1 movement of bolt
and x2 movement of machine gun during recoil.
pk ( t ) = p e
Some forces were calculated by applying the mathematical,
mechanical and ballistic methods, while others are measured
on real model of machine guns.
t
b
1 e b
(3)
where: p  pressure of propellant gas in the barrel at the
orifice for the evacuation of powder gases at the moment
when the projectile passes the orifice, t duration of action
of gas on the piston, coefficient of powder gases effect
and b  the ratio of unit impulse and pressure.
By internal ballistic calculation [3] the obtained maximum
pressure in the barrel of machine guns 12,7 mm 87 is pm=
328,98 Pa.
With all of the calculated values that exist in the expression
(3), a function of pressure changes in the gas cylinder can be
represented by:
pk = 120 106 e
t
0.001343
t
31.257
0.001343
1 e
.
(4)
The force of the bolt return spring acting after unlocking of
bolt. Change of the force of the bolt return spring,
depending on the bolt traveled distance x1, is obtained
experimentally:
Picture 2. The optimal internal ballistic curve of pressure
in the barrel
Fop = 157 + 396.08 x1 .
From a total of 15 parameters that are entered in the input
file to the internal ballistic calculation the three basic
ballistic characteristics of gunpowder are varied: covolume
of powder gases, the specific energy of powder gases and
unit burning rate of gunpowder.
For the purpose of realization elastic connection of machine
guns and gunmount is designed buffer. The experimentally
determined spring force of the buffer is:
Fbf = 760 + 102142.9 x2 .
The parameters are changed so that avoid the case of
incomplete combustion of gunpowder [4]. By varying these
parameters within the range of 10% and by selecting the
optimum value thereof, reduced the maximum pressure in
the barrel pm = 308.25 Pa, which represents a reduction of
the maximum pressure in the barrel of 6.3% (Picture 2). The
calculated velocity of the projectile at the muzzle is V0 =
841 m/s (Picture 3).
where x displacement of machine gun.
Total recoil force in the phases is shown in Picture 4.
195
(5)
(6)
OTEH2016
STRESSANALYSISOFINTEGRATED12.7MMMACHINEGUNMOUNT
Picture 4. The recoil force change in phases
Picture 6. Field stress of gunmount (MPa)
3. LOADING OF DESIGNED GUNMOUNT
AND BEARING
In order to calculate the forces acting on the bearing and on
the basis of technical documentation from the factory
"ZastavaArms", by software CATIA V5 R18 model of
machine gun assembly with a cradle and gunmount is made
(Picture 5). On the basis of this 3D model, the required
positions of the barrel axis and the mass of projected cradle
and the gunmount are obtained [6].
a) in the xdirection
b) in the ydirection
c) in the zdirection
Picture 7. Loading of bearing in MPa
By simulating loading machine gun assembly, cradle and
gunmount were obtained the following values of forces that
load bearing: Fx = 1949 N, Fy = 1238 N and Fz =3284 N [6].
Picture 5. Isometric view of the assembly machine guns
on the cradle and gunmount
The calculation was performed taking into account the
following assumptions [6, 7]:
The effect of force of gravity is entered as a negative
acceleration acting on the whole body, while the effect of
the recoil force entered as the effect of surface pressure (the
bottom of the barrel). As constrains were adopted to the
lower surface of the bearing is fixed. Thanks to the
simplified geometry, finite element mesh can be generated
automatically. The final element is the shape of a
tetrahedron. After completed analysis, required stresses and
forces are reading. In Picture 6 field stress in MPa is
represented in isometric view of the assembly.
Mobile platform to which is mounted a machine gun is
located on a horizontal surface;
Distribution of additional vertical load on the ball of
bearing is sinusoidal;
Horizontal forces distributed per balls as on radial ball
bearings.
To determine the strain of roller tracks it is necessary to
determine the calculation load on the balls and the roller
track, and then check the strain on the surface pressure.
For the calculation of bearing is necessary to know the
intensity of the forces acting on the bearing. The stress of
bearing caused by recoil force and force of gravity. The
simulated effects of the forces on the bearing in the
direction of x, y and z axes are shown in Picture 7.
Picture 8 shows the forces that overload the rotating bearing
in the vertical plane. Markings used in the figure are the
following: Pkn recoil force, Fz vertical reaction of
bearing, G the total weight of the assembly machine gun
and gunmount, D mean diameter of the bearing, G, G
polar coordinates of weight force, z, z polar coordinates
of vertical reaction [7].
196
OTEH2016
STRESSANALYSISOFINTEGRATED12.7MMMACHINEGUNMOUNT
Picture 9. Scheme of rotating bearing loading in the
horizontal plane xy
Coordinates of resulting horizontal reactions Fh are
generally (D/2, h). Coordinates of reaction mechanism for
gunmount rotation Fmk are designing determined (a = D/2,
mk = 45). Unknown parameters (Fh, Fmk and h) are
determined from equilibrium conditions:
Picture 8. Scheme of rotating bearing loading in the
vertical plane xz
Vertical reaction of the bearing is obtained from the force
equilibrium condition (Picture 8):
= Fz G FR sin = 0 .
(7)
(9)
= 0 FR a cos Fmk D = 0
2
On the basis of system of equilibrium conditions equations
(eq. 9) and condition that the axis of machine guns at the
geometric center of the assembly, next parameters are
obtained:
Fmk =
M y = 0 , and
geometric sizes are taken from CAD model (b = 135.7 mm,
h = 609.2 mm, z = 28.747 mm).
2 FR a cos
=0
D
4a
4a 2
sin mk + 2 = 2292.14 N . (10)
D
D
2a cos mk
h = 0
tg h =
2a
D 1 sin mk
D
Fh = FR cos 1
Because a = 0 (axis of machine guns is in the plane of the
axis of rotation of bearing) and y = 0 (force of gravity is in
longitudinal plane of gunmount), complete assembly is
symmetrical in relation to the plane xz, angle coordinate is
G = 0. Then the polar coordinate z is:
G G cos G + FR b sin FR h cos
z =
.
G + FR sin
= 0 Fh sin h Fmk cos mk = 0
Polar coordinates of vertical reactions were determined from
Force FR in this case is the maximum force of buffer spring
(FR = 2292.14 N). For minimal elevation angle =0, value
of vertical reaction is minimal Fz min = 1358.7 N , whereas
for a maximum elevation angle =75 the value of vertical
reactions of bearing is the maximum Fz max = 3572.7 N .
M x = 0,
= 0 Fmk sin mk + Fh cos h FR cos = 0
where elevation angle, FR resistance force of recoil.
the momentum equation
The effect of calculated forces and loading of bearing shown
in Picture 10 [7].
(8)
For the most favorable case momentum loading of bearing
(=0), according to Picture 8, we get that z = 999 mm.
If z > D/2, this means that the load tends to overturns
assembly and must operate rollover protection. In the
specific case (D = 415 mm) is obtained that the structure is
not stable in terms of rolling over, and it is necessary to use
the rollover protection.
The effect of force on the bearing in the horizontal plane (xy) is shown in Picture 9. Marks used in the figure are the
following: Fh horizontal reaction of bearing, Fmk reaction
mechanism for rotation, h angle coordinate of horizontal
reaction and mk angle coordinate of reaction mechanism
for rotation [7].
Picture 10. The loading of rotating bearing
197
OTEH2016
STRESSANALYSISOFINTEGRATED12.7MMMACHINEGUNMOUNT
By comparing the values of reaction forces of rotating
bearing, obtained using these two methods, it can be seen
that they are very close and that in the vertical direction are
different by about 8%, and in the horizontal direction by
about 1%.
5. CONCLUSION
In this paper are compared results of two methodology for
loadings calculation of heavy machine guns mount.
The paper was performed calculation of machine gun 12.7
mm force recoil, discussed the impact of internal ballistic
parameters to it, as well as its effect on the assembly of
machine guns, the cradle and the gunmount.
Regardless of which way to calculate reaction forces of
rotating bearing, they can further used for the calculation of
the elements of a machine gun automatic control, as well as
for calculations of mobile platform loading during firing.
Based on the design parameters of the machine guns, a gunmount was designed together with the cradle and the
rotating bearing. Modeling was done in the software
package CATIA V5R18 with the condition of minimum
recoil force and minimum allowable mass of gunmount.
Since the selected machine gun is gas operated, apart from
the basic force of pressure of powder gases at the head of a
bolt which causes recoil of machine gun, should be taken
into account the force of pressure of powder gases on the
forehead of the piston in the gas cylinder. All other forces
that occur during firing, are the resistance forces of recoil
and have the opposite direction.
ACKNOWLEDGMENT
This work is part of the project III47029 in 2016, funded by
the Ministry of Education, Science and Technological
Development of Republic of Serbia.
References
[1] Ristic Z. Mechanic of Artillery Weapons. (In Serbian),
Odbrana Media Center, Belgrade, 2016. (in print)
[2] Babak F. K. Machine guns. (In Russian), Polygon,
Sankt Petersburg, 2005.
[3] Cvetkovic M. Internal Ballistic. (In Serbian), Military
Academy, Belgrade, 1998.
[4] Tancic LJ. Internal Ballistics Design. (In Serbian),
Odbrana Media Center, Belgrade, 2014.
[5] Petrovic M. Mechanics of Automatic Weapons. (In
Serbian), Odbrana Media Center, Belgrade, 2007.
[6] Chobitok V. A. Construction and Calculation of Tanks
and Infantry Combat Vehicles. (In Russian), Voennoe
izdatelstvo, Moscow, 1984.
[7] Jovanovic D. Integration of machine gun 12.7
millimeters on mobile platform master thesis, Faculty
of engineering, Kragujevac, 2015.
By internal ballistic calculation the maximum pressure in
the barrel of 12.7 mm machine gun which obtained is pm =
328.98 Pa. By varying appropriate ballistic parameters of
powder, an optimal solution is obtained (pm = 308.25 Pa)
that provided minimal recoil force, with the least possible
disruption of output values as initial velocity of the
projectile.
With such obtained recoil force, the calculation of load and
forces acting on the rotating bearing was made. The
calculation is done in two ways. The first used the method
finite element by program FEMAP. The most loaded parts
of the gunmount are under stress from 19.43 to 38.83 MPa.
The second method is based on calculation of bearing
components load from the equilibrium condition.
198
STRATEGY IMPLEMENTATION OF DUALSEMIACTIVE RADAR
HOMING GUIDANCE WITH COUPLING OF TANDEM GUIDED AND
LEADING MISSILE OF AIR DEFENCE MISSILE SYSTEM ON REAL
MANEUVERING TARGET
MARKOVI STOJAN
University of defense, Military academy, Pavla Jurisica Sturma 33, 11000 Belgrade, Serbia, stomarkovic@yahoo.com
MILINOVI MOMILO
Faculty of Mechanical Engineering, University of Belgrade Kraljice Marije 16, 11120 Beograd 35, Sertbia,
mmilinovic@mas.bg.ac.rs
NENAD SAKAN
Institute of Physics, University of Belgrade, Pregrevica 118, Zemun, 11080 Belgrade, Serbia, nsakan972@gmail.com
Abstract: The paper investigated the possibilities of newly formed dual system SARH (Semiactive radar homing)
missiles SA6 midrange category, on standard manoeuvrable targets.The system was set up as a dualcooperative
guidance of the first launched (guidedslave) missile from second launched missile (leadingmaster), with a time delay
of 3s. Angle position of target is measuring using two existing RHH, which together with additional illuminators on
missiles, represent two active seekers, with a period of mutual scanning and with changing the rolesstatus of 1s. Also,
missiles synchronously communicate between each other through free caudal antenna, when it is situation based on the
smaller current error, determines the status of a missile. Proces of coupled guidance is modeled with specially designed
program in MATLAB with appropriate adjustment AD, AL and effective radius of action Warhead. Missiles are guiding
in calculated point of overtaking with clasical method PN with changing the angle of launch SPLV, by changing the
distance, altitude and speed targets and and other given parameters of flight. Simulations were examined specific
conditions and regimes of shooting with one or two missiles, the correlation results and made conclusions of the most
efficient mode and optimal performances of dual system homing missiles AAD.
Keywords: slave and master missile guidance, control system simulation MATLAB, dual navigation missile, Semi Active
Radar Homing, Proportional Navigation, Miss distance missiletarget, Radar lighter.
Disadvantage of this system is that it binds to the resource
on the ground and thereby prevents its engagement to
launched and guidance other missiles upon targets in AS
(AirSpace). Also, drawback of this system that he may
warn opponents, respectively aircraft as a target, that is
launched missiles on it and that the same guiding from the
ground. Semiactive RHH has potential possibility to
measure the relative direction to the missile as well as
angle speed LOS (Line of Sight) missiletarget, but there
is no possibility of measurement distance to target In this
paper, missile has been analyzed with SARH way of
guidance which has two rear tail antenna, operable to
receive a facing direct signal illumination target, which
carries a reference frequency, and thus enabling the
measurement of Doppler frequency and estimate III MF
(Mixed Frequency) however, the change of distance to
target irrelevant and it has no need to be separately
measured by. This channel is desirable, and his role is
not to guide the missile, but to allow her to distinguish
target from the cluster and interference signal and CEW
(Countermeasure Electronic Warfare). There is a second
parallel channel, and that is angle grip and guidance upon
target through the front and face of the antenna, regulated
AOT (Automat Operation target) who later become the
1. INTRODUCTION
Classical artillery, with multi tools, relies upon prediction
trajectory of target in open loop control, with the help of
Optical or radar devices for control, tracking and action,
respectively fire control, primarily through LFT (Low
Flying Target). In order to achieve the required the
accuracy of the action, either by LFT as well as on the
medium altitudes targets, used in missile systems AAD
(AntiAir Defence) that have automatic control closed
loop for continuous reading errors in order to provide
meeting missile with the target and its translation in
corrective
maneuver,
with
additional
required
acceleration. Designing law of guidance, that the distance
to the goal (ultimate current miss) decreases to the lowest
possible value, in real world resulting in a final breach by
objective, which compensates with the action of remote
lighter missile. Representative and reliable way of
guidance missile in terms middle range AAD is Semi
Active Radar Homing which provide that the source of
radiation missile, which illuminates the target, located on
the ground, both at the beginning and below, in the next
phase of guidance, provides at the end reliable work or
RHH (Radar Homing Head) missile.
199
STRATEGYIMPLEMENTATIONOFDUALSEMIACTIVERADARHOMINGGUIDANCEWITHCOUPLINGOFTANDEMGUIDEDAND
dominant in missile guidance, and can be stored with the
services of the other radar and optical devices (TOVTelevision Optical Viewfinder) in the AIR. Many years of
observation AAD system, with completely pros and cons
imposed the idea of specific research methods of fighting
with modern targets in AS, with decreases the intense of
probability of their hit and destruction, relative to the
initial set requirements. Otherwise, effectively reducing
and zeroing error homing missile, with increased range
and speed of target, his CEW and maneuver, intensively
thinkinging world in last ten years with the development
of modern theories of guidance and applying the most
modern mathematical and software tools.
PdualM1,M 2 = 1 /(1 Psin glM ) 2
OTEH2016
before launching missiles, which continues to fly ballistic
in POT (Point of OverTaking) with the accuracy of (+4o),
and of itself doesnt see the target.
With the work off start phase of missile propulsion,
achieves the necessary speed of missile, and at the same
time, protection of other missiles and resources, with the
delay activation of RAMJET engine. After that
stabilization of missile occur and followup of its stable
and manageable flight, engagements, the already
calculated point of overtaking, but now with the classical
method PN (Proportional Navigation) with default AL
(Area Launch) with their borders. This means that the part
of radarcomputer system on the ground did all the
procedures provided for
complex procedures (46
equations), actions and modes of work, from
observations, expectations, appearance and allocation
target in the launch zone, to give accurate coordinates and
creating conditions for launch. It is calculated and the
time entered reinforcement RL (RadarLighter) the 0.6s
distance till encounter with target with a given AD (Area
of Destruction) from near and far, upper and lower limit,
dictated the height of target, dead zone, its speed,
parameter target and type of aircraft as a target, figure 1.
(1)
Second phase of flight missiles is marching mode of work
when RAMJET propulsion engine is activated. Missile is
included in the loop of guidance, ate target is still
continuously illuminates CW illuminator with
1S91(RSSGRadar Station Surveillance and Guidance),
and at same time monitors TR (Tracking Radar) ie pulse
radar. Necessary presence included continuous pulse
transmitter RG (Radar Guidance 1S31) represents, with
observation RS (Surveillance Radar 1S11) all the while
guiding missile to hit and destruction of the target, great
technical and tactical problem of hitting the target and
same time surviving crew and resources. It shows practice
and out dated system, which makes it inferior, in relation
to the modern offensive resources from AS Also,
possibility of hitting modern targets is drastically reduced,
and to achieve the required real norms its necessary a far
greater number consumption of missiles by one target
(from 18 to 22 missiles) in a short period of time, which is
economically unsustainable
Figure 1. Starting kinematic show dualcooperative
homing two missiles on one target
With the work off start phase of missile propulsion,
achieves the necessary speed of missile, and at the same
and impossible in our conditions short period of time,
which is economically unsustainable and impossible in
our conditions. So, exclusion RG from system and
substation on missile (placing additional illuminator on
missiles) and modernization of the existing receivers
(digitalization, silent in SSTSolid State Techniques) in
the existing RHH, with innovated roles of tail antena.
Figure 2. Area of bursting fragments WARHED and
optimal position of antenna activation RL at the midline
One of the ways, or strategies in real application and
improving efficiency, probability of hitting and destruction
of target (manufacturers information: p1M= PsinglM = 0,80 and
p2M = 0,96) for MIG19 with out considering CEW, of the
existing system of miderange, will be analyzed and exposed
on the basis of the initial equation.
Missiles get divided but each other interchangeable roles
(guided  slave missile M1  that excels 3s and she is
always in role that attack target and leadingmaster
missile M2with illuminates target T and flight behind
guided M1 and support it) is the core cooperatively
coupled homing two missiles to one target, SARHARH
(Active Radar Homing), figure 4[2].
2. FORMULATION OF THE PROBLEM
OF DUAL HOMING MISSILE
The initial phase of flight AAMS (AntiAircraft Missile
System) is the command radioguidance, whose
realization in system initially begins on the launcher,
200
STRATEGYIMPLEMENTATIONOFDUALSEMIACTIVERADARHOMINGGUIDANCEWITHCOUPLINGOFTANDEMGUIDEDAND
OTEH2016
Picture 3. Starting conception strategies simulation of dual homing AAD missiles on one target
relative to SNR for method ARH through the square of the
distance relationships RM1T and RM2T , eq 2 [1].
Figure 4. Proposal algorithm decision  making in tandem
homing missiles
Missiles are changing roles changing the specified status,
at defined intervals of time and thus obtaining SARH in
the air, with all know pros and cons. Algorithm applied in
strategy dualcoupled homing is shown in figure 4. It can
be seen that guided missile M1 travels in PO (Point of
Overtkong), optional guided missile M1, which
illuminates target, with criteria change of roles of
missiles. Algorithm decision, running parameters, which
are in accordance with the decisions state (missiles),
expressed as control functions of time in respective
sequences, processed in [2], from the time of operation
target to its cooperative destruction.
Figure 5. Matlab block scheme simulation
SNRSARH = SNRAHR / [( RM 1T / RM 2T ) 2 ]
(2)
Missiles fly based on criteria set current failures m (md miss distance) and achieving the lowest endflops in
relation to the position and maneuver of the target in area.
Starting pictures kinematics, as the basis of simulation,
shown in figures 4 and 5.
Missiles are changing roles SARH  ARH with an interval of
1[s]. Their effectiveness is, in general, can be expressed
through SNR (Signal Noise Ratio) for method SARH
201
STRATEGYIMPLEMENTATIONOFDUALSEMIACTIVERADARHOMINGGUIDANCEWITHCOUPLINGOFTANDEMGUIDEDAND
Unfavorable position, with structural constraints and
performance of missiles (max 4s flight toward as per
predictable path) at one time for a second missile, that
was with big fault, becomes a priority. So, there is a
possibility of additional processing through GP
(Generator of Searching) in Homing Head, respectively
additional grip (if target was lost for flight in
remembered TP (Point of overtaking). Also, she has
enough time and benignity of movement, so that in the
area, guided missile its longitudinal axis and with angle of
beam transmitter RL, placed in the optimal position
parallel to the longitudinal axis of aircraft (decreases the
tilt radiation pattern in relation to longitudinal axis with
command changing antenna) for action remote radar
lighter and destruction of target, figure 6.
OTEH2016
autopilot and executive actuators of missile. At the
entrance, a module target formed and module of
comparing errors of missile M1 relative to M2 i module
of additional calculation, with respect to the effective
radius of action W (Warhead). System has possibility to
of giving the stroke duration (clock) roles of missiles
(guided, leading), with the necessary delay missile launch
M2 than 3 s. Matlab model simulation, picture 5, done on
the basis of kinematic relations of picture 1, includes:
model of gravity and shape of Earth, kinematics and law
of guidance, dynamics and 6DOF, disruption and settings,
aerodynamics (parameters and coefficients), missile
launcher, geometry of missile, radar lighter i warhead,
masspropulsioninertia characteristics, run the simulation
with default parameters, reading file, select a time
changes status of missile ending at the closest point of the
encounter and presentation of data and results (diagrams)
given simulation. In simulations, with solutions ballistic
flight, ARH mode and SARH mode, were used real
parameters of missiles: navigation constante [N=3.5],
timing constant HHM [Tg= 004s], memory Point of
Overtaking the Target (POT) from Generator Search (GS)
RHHM tmemory = [2,54s], radius of efficient action
warhead [refW = 25m], time reinforcement RL [tearmRL =
0,6DPOT], coefficient effectiveness radar lighter Kru>
0.90, fugure 6, side overload [ab=15g], longitudinal
overload [au=23g], loss of guiding until meeting with
starting maneuver of flight [tvgsmanevart = 68 s], clearance
failure of missile
[rpr= 320m].
Figure 6. Dependence Kru of speed targets (bomber,
fighter) and failures of AADS missile
4. SHOVING RESULTS MATLAB
SIMULATION AND RESEARCH
Based on the collected data in the real world, setting
model SARH as bistatic radar in the air (alternating shifts,
giving  active radiation and receptionsemi active)
obtained are the result of the launch of two tandem, the
two missiles (guided and leading) on real targets (strategic
bombers and attack fighter). Also, experiential data,
which provides original rule shooting system, were
amendments of the input data and parameters, on what
basis are made numerous simulations using the Matlab
2R2011a [4], whose Simulink model custombuilt
analysis and set up for this type of missile, in figure 5.
The results are shown in Table 1. First, it provides a case
in shooting case of forced release of missiles with
SPLV1,2.
Figure 7. Integral law probability failure missile SA6 at
various speeds target less than 12m for a given probability
P
Then it examined individually shooting with one missile
on modern target flying to meet in standardized height
7km and then dual shooting on target which are flow with
zero radial velocity (helicopter) on 3km. Then engage
targets outgoing which is defined by the manufacturer in
the development system (strategic bombers and attack
fighter MIG19).
3. REALIZATION OF SIMULATION
TANDEM DUAL COUPLED MISSILES
Conceptual scheme of simulation coupled guidance
tandem missiles AAD, in software program MATLAB,
has shown in picture 5. We see that they are
mathematically coupled two module 6.DOF missile,
202
STRATEGYIMPLEMENTATIONOFDUALSEMIACTIVERADARHOMINGGUIDANCEWITHCOUPLINGOFTANDEMGUIDEDAND
OTEH2016
Table 1. Showing results of Matlab simulation for default parameters of target and and avariety of shooting modes and
launching missiles AADS
It is taken case of shooting in the autonomous mode
2K12M and centralized (K1M) management of fire from
the upperlevel. In the end results were compared for
various modes of launching missiles (shooting with one
missile and shooting with two coupled missiles, on
different specific altitudes of 0.011km, 3712km, and set
speed targets in and checkout) as well as other
parameters of the target (helicopter, drone, bomber, attack
fighter, modern fighter to meet and outgoing), testing
time shift status missiles M1 and M2 and its impact on the
ultimate failure, table 1. For that purpose given algorithm
was used in pictures 3 and 4 for autonomous launching
two coupled missiles with limited time simulation
propulsion missile, with allow overload and parameters of
target, etc. Used as real technical parameters of geometry
missiles, its aerodynamic parameters, blend together in
two coupled 6DOF dynamics model and and the
kinematics associated with the module objective (uniform
movement or maneuver as sinusoidal movement, or by
changing the position and target vectors). In table 1, has
shown results of simulation, for various modes of
launching and variable of parameter target (elevation,
distance, altitude, speed, maneuver, maneuver frequency,
parameter). In the analysis was tested changed duration
status of missile M1 and M2. For specific targets,
criterion manufactures SA6 missile system they were:
203
STRATEGYIMPLEMENTATIONOFDUALSEMIACTIVERADARHOMINGGUIDANCEWITHCOUPLINGOFTANDEMGUIDEDAND
OTEH2016
Table 2. Research optimal time simulation shift missiles
M1,M2
Figure 8. Tipical engagement geometry AM
Figure 9.
strategic bombers and attack fighter, serial number 8. and tandem rocket launching, the best solution for LF targets,
9. Table 1, was adopted as the optimal value time duration the distance from the launch 9km, height 3km, target
status of missile of t =1s.
speed 1.2Ma, side overload of 8g and with parameter
0km. The ultimate failure of the rocket is upgraded from
Appearance representative diagram of radiation, with two
48.80m to 11.83m, which is within the limits of efficient
SPLV and tandem dual coupled AAD missiles, shown as
operations radar lighters, and it is 412%, in relation to
a variable kinematic size and trajectory of target and
launch of a single missile on the same target. Due to the
missile, figure 9. Given failures PAAD missile, very well
requirements set SA6, which date from 55 years ago, use
agree with a given dependence Kru = f(r[m]), figure 6,
of dual launching of two missiles represent a good basis
referred to in [3], the original instructions rules producer
for observation and modification SAM in midrange
system. Pictures 1,2,3 show (al) the end results and given
category. Further directions of research should be directed
failures. Comparing the results, with the results of the
on expansion mutually coupled guidance on modern
launch of a single missile on same target, result in an
target with complex maneuver, improving the algorithm
improvement of about 3.9 8.02 %.
by the expanded method (PPN) and salvo launch of
homing missiles SAM on one target.
5. CONCLUSION
References
The obtained representative results show that the strategy
of dual SARH, coupled guiding and leading missile, with
optimal interval of time changed status of 1s improved
around 8.02%, for almost all targets and modes of
shooting on average. Thereby were analyzed strict
requirements, dictated by modern high performance
targets (attack fighters) with curvilinear maneuver and in
encounter and in leaving phase. Realistic performance of
SA6 were taken into account with from near and far
boundary zone of launches and destruction, on various at
different altitudes. It turned out that the results of the dual
[1] Mullins,J.: AMRDEC, Radar TechnologyRadar
Guidance Techniques, AIAA Symposium Huntsville,
Alabama, 2005,
[2] Markovic,S., Milinovic,M., Sakan,N.: Test simulator
for the discrete multiparametric decision flight
system, OTEH2014, Beograd, VTI.
[3] UARJ PVO, SSRP KUBM  Teorija i Pravilo
gaanja, ubenik 122, VIZ Beograd, 1977.
[4] MATLAB 2R2011a, Aerospacemodul, Simulink.
204
EXPERIMENTAL INVESTIGATION OF OILS IN FOURSTROKE
ENGINES
SRETEN PERI
University of Defence, Military Academy, Pavla Juriia turma Str. No. 33, 11000 Belgrade, sretenperic@yahoo.com
BOGDAN NEDI
University of Kragujevac, Faculty of Engineering, 6 Sestre Janjic Street, 34 000 Kragujevac, nedic@kg.ac.rs
Abstract: Confirming the basic causes of failures and their elimination, control of certain phenomena, is defining
proactive maintenance, as a new method that reduces maintenance costs and prolongs the life of assets. Determination
of tribomehanical systems condition has very important role in development of friction theory and practice, wear and
lubrication. There are used today different physical and chemical methods and tribology methods for tribomehanical
system diagnosis. Experience in technical systems exploitation shoved that the most effective failure prognosis is
according to parameters, particles created as result of wear, which are reliable indicators of wear. Analysis of oil
samples which contain particles, created as results of wear, enable evaluation of system tribology condition in different
phases of system exploitation.
The paper presents the physical chemical tests in the analysis of oils that are used for the assessment of his condition.
Furthermore, the results of experimental research on the physical chemical characteristics of the oil sampled from
engines of the vehicles (PUCH 300GD, PINZGAUER 710M and IK 104) are shown. All of these vehicles were in
regular use by the Serbian armed forces.
The performed research has revealed some significant changes in the physical chemical characteristics of oil for engine
lubrication. These changes directly depend on the condition of the entire engine elements, i.e. depend on their
functional characteristics. The research results are originating from the research of the paper authors.
Keywords: Monitoring, Analysis oils, Lubrication systems, Proactive maintenance, Diagnosis.
detergents and dispersants, preventing
antioxidants, preventing corrosion inhibitors.
1. INTRODUCTION
Using Oil Analysis programs for engine oils has several
benefits: reduction of unscheduled vehicle downtime,
improvement of vehicle reliability, help in organizing
effectiveness of maintenance schedules, extension of
engine life, optimization of oil change intervals and
reduction of cost of vehicle maintenance.
oxidation
To know analytical properties of lubricants is the base to
make a decision in development, production and
application of lubricants. The lubricant classifications and
approved
system
specify
many
performance
characteristics and analytical tests. The analytical tests are
classical and instrumental. Instrumental technical have the
advantages in small quantity of the sample and rapid
analyze. As a part of the common proactive strategy of
the hydraulic systems maintenance, concept of online
monitoring is introduce in practice, recently [26,9]. It is a
combination of the measurement procedures, by which
sample of fluid is to be analyzed is taken directly from the
system and the results of the measurements are
continuously. Online monitoring considering, first of all,
control of cleanliness classes (according to ISO, NAS,
SAE), control of humidity, viscosity, permittivity (acid),
temperature
In application, oils change their properties through [1]:
contamination by combustion products and metal wear
particles, consumption of additives which is chemical and
bears impact on important oil functions and base oil
oxidation.
The primary role of engine oil is the lubrication of
moving engine parts and reducing friction and wear of
metal surfaces which provides the good engine
performance and its long life. In order to provide a
defined quality of engine oils during production and for
final products to meet the product specifications we need
to know the physical chemical characteristics of engine
oils.
The following tests are the most used in condition
monitoring: Viscosity, Total Acid Number (TAN), Total
Base Number, Water, Spectrometric analysis and Particle
Count.
Certain physicalchemical characteristics which are
significant for the quality of engine oils are achieved by
adding additives to base oils. The most frequent additives
are for: improving of viscosity indeximprovers, reducing
pour point depressants, maintaining engine cleanness
Viscosity is the resistance of a fluid to flow and the most
important lubricant physical property. The fluid is placed
in a "viscometer" (a calibrated capillary tube for precise
205
OTEH2016
EXPERIMENTALINVESTIGATIONOFOILSINFOURSTROKEENGINES
flow measurement between two premarked points on the
tube) and preheated to a given temperature in a "viscosity
bath" (which is usually oilfilled). After the oil reaches
the desired viscosity temperature, gravityinfluenced flow
of the oil is initiated in the viscometer and timed between
two calibrated points. This time becomes the determinant
for the result.
of engine condition monitoring by oil analysis. This part
presents the results of experimental research of physicchemical characteristics of engines oil which was sampled
from engines of PUCH 300GD, Pinzgauer 710 and
IKARBUS IK 104P vehicles [10], [11].
The research was carried out in two vehicles PUCH
300GD (PUCH1, PUCH2), two vehicles PINZGAUER
710M (PINZ1, PINZ2) and two vehicles IKARBUS IK
104P (IK104P1, IK104P2).
Total Acid Number (TAN) is a neutralization number
intended for measuring all acidic and acidacting
materials in the lubricant, including strong and weak
acids. It is a titration method designed to indicate the
relative acidity in a lubricant. The TAN is calculated from
the amount of KOH consumed. The acid number is used
as a guide to follow the oxidative degeneration of oil in
service.
The research was conducted through periodic sampling
oil from engine vehicles listed above. Apart from the
fresh oil (zero sample), samples are taken after 1.000
km, 2.000 km, 3.000 km, 4.000 km and 5.000 km for
vehicles.
The physicalchemical characteristics of oil in accordance
with standard methods are examined, shown in Table 1.
Total Base Number (TBN) is a neutralization number
intended for measuring all basic (alkaline) materials in the
lube (acidneutralizing components in the lubricant
additive package). The converse of the TAN, this titration
is used to determine the reserve alkalinity of a lubricant.
The TBN is highest when oil is new and decreases with
its use. Low TBN normally indicates that the oil has
reached the end of its useful life.
Table 1. Implemented tests and methods for examining
the physicchemical characteristics of oil
Characteristic
Kinematic viscosity, mm2/s
Viscosity Index
Flash Point (C)
Pour Point (C)
Water Content, mas.%
Total Base Number (TBN),
mgKOH/g
Insoluble substances in
pentane, %
Insoluble substances in
benzene, %
Fe Content, %
Cu Content, %
Water can be detected visually if gross contamination is
present. Excessive water in a system destroys a lubricant's
ability to separate opposing moving parts, allowing severe
wear to occur with resulting high frictional heat. There are
several methods used for testing the moisture
contamination (crackle, FTIR water, centrifuge, Karl
Fischer) each with a different level of detection (1000
ppm or 0.1 % for first three methods and 10 ppm or 0.001
% for Karl Fischer method).
Spectrometric analysis is a technique for detecting and
quantifying metallic particulates in used oil arising from
wear, contamination and additive packages. The oil
sample is energized to make each element emit or absorb
a quantifiable amount of energy, which indicates the
elements concentration in the oil. The results represent
the concentration of all dissolved metals and particles.
The equipment for spectrometric analysis is the standard
equipment for oil analysis laboratories today. It provides
information on technical system, contamination and wears
condition relatively quickly and accurately. Spectroscopy
is moreorless blind to the larger particles in an oil
sample, more precisely, to particles greater than 10 m in
diameter, which are more indicative of an abnormal wear
mode [7].
Method
SRPS B.H8.022
SRPS B.H8.024
ISO 2592, ASTM D 92
ISO 3016
ASTM D 95
ASTM D 2896
ASTM D 893
ASTM D 4055
ASS
ASS
The analysis was done on the fresh (new) oils and oils
that are used in the engines of vehicles. During the
sampling of oil choice of the sampling were conducted
carefully according to the actual oil usage, which enabled
each sample as representative one.
The wear mechanism of a tribological lubrication system
consists in the wear of contact surfaces, and lubricant
consumption. If there is wear of the contact surfaces,
there are wear particles present.
Regardless of the availability of numerous methods for
diagnosing the physicchemical changes of lubricants, in
order to create a true picture of the condition of lubricants
from the user system, it is of importance to satisfy the
precondition of the possibility to obtain a representative
sample. That is why it is extremely important to take the
sample in a proper way.
Particle Count is a method used to count and classify
particulate in a fluid according to accepted size ranges,
usually to ISO 4406 and NAS 1638 [8]. There are several
different types of instrumentation on the market, utilizing
a variety of measurement mechanisms, from optical laser
counters to pore blockage monitors.
Allowable values of deviation limits of individual
characteristics of the oil are conditioned by the type of oil,
working conditions and internal recommendations of the
manufacturer of lubricants and users. Limited value
characteristics of oils that condition the change of oil
charging from engine are given in Table 2. They represent
2. THE RESULTS OF OIL ANALYSIS AND
DISCUSSION
In this part are presented the results of oil analysis
examination during application in fourstroke engines by
physicchemical methods in order to evaluate possibilities
206
OTEH2016
EXPERIMENTALINVESTIGATIONOFOILSINFOURSTROKEENGINES
Table 5. The results of testing samples of used oil from
engines examined vehicles [10]
the criteria for the change of oil charge. Deviation of only
one source changes characteristics of oil charge, no matter
of what a characteristic is about.
Sample
Viscosity at
100C, mm2/s
100 ppm
50 ppm
Flash
Point,C
Used engine oil in examined vehicles are shown in Table
3. Characteristics of zero samples of engine oil are shown
in Table 4, and the results used oil samples in Table 5.
Index
Viscosity at 40 C and 100C,mm2/s
Viscosity Index, %
Total Base Number (TBN), mg
KOH/gr
Flash Point, C
Water Content, %
Products wear Content Fe,
ppm(g/gr)
Products wear Content Cu,
ppm(g/gr)
Max. allowed
variation
Engine oil
20 %
5%
The fall to 50
%
20 %
0,2 %
Viscosity
Physicalchemical characteristics oil
and products wear
Viscosity at
40C, mm2/s
Table 2. Allowed values deviation of physicochemical
characteristics of new and used oil
Table 3. Used engine oil in examined vehicles [10]
Cu Content
(ppm)
Fe Content
(ppm)
TBN,
mgKOH/g
Engine oil from engine of PUCH 300 GD vehicles
SAE
API
Manufacturer
classification
classification
SAE 15W40
API SG/CE
FAM Krusevac
Engine oil from engine of PINZGAUER 710 M vehicles
SAE
API
Manufacturer
classification
classification
SAE 30/S3
GALAX Beograd
Engine oil from engine of IKARBUS 104 P vehicles
SAE
API
Manufacturer
classification
classification
SAE 15W40
API SG/CE
FAM Krusevac
Table 4. Results of zero samples of oil from the engine
[10]
Characteristic
Color
Density, gr/cm3
Viscosity at 40 C,
mm2/s
Viscosity at 100
C, mm2/s
Viscosity Index
Flash Point, C
TBN,
mg KOH/g
104,63
14,12
11,67
230
240
10,5
9,8
PUCH
2
104,8
110,4
111,8
113,8
115,9
127,5
14,12
14,23
15,03
15,65
16,12
17,03
135
131
126
123
120
115
230
215
210
204
202
188
10,5
9,4
8,9
8,7
8,1
7,6
27,4
59,8
71,2
71,4
86,8
27,4
2
3,4
3,7
3,9
5,4
2
IK104
1
104,8
96,9
96,2
92,3
90,8
90,2
14,12
13,74
12,86
12,46
12,32
12,29
135
132
130
125
122
119
230
217
214
213
210
189
10,5
8,8
8,7
8,4
7,9
7,3
30,1
32,5
35,6
37,5
38,5
30,1
1,5
1,9
3,2
4,4
4,9
1,5
IK104
2
104,8
104,4
101,9
97,1
94,8
93,1
14,12
13,62
13,55
13,25
12,95
12,61
135
133
131
127
124
121
230
212
210
202
193
184
10,5
8,1
7,7
7,2
6,8
6,4
20,5
46,3
57,6
62,8
69,6
20,5
3,2
5,1
6,3
7,7
9,1
3,2
PINZ
1
104,6
100,4
94,4
86,3
79,1
75,9
11,67
10,9
10,3
9,96
9,37
8,74
100
96
93
89
84
82
240
196
186
168
154
136
9,8
9,6
9,1
8,3
7,6
7,1
19
19,8
38,3
54,3
105,4
19
3,5
4,1
5,3
6,9
8,7
3,5
PINZ
2
104,6
100,9
96,1
88,6
82,2
76,9
11,67
10,55
10,4
10,15
9,64
9,07
100
97
95
91
87
84
240
193
177
159
143
128
9,8
9,4
8,4
7,8
6,6
6,2
17,9
40,9
86,7
132,8
261
17,9
3,3
3,8
6
8,1
9,7
3,3
The viscosity index is an empirical number which shows
how the viscosity of some oils changes by increasing or
reducing the temperature. High viscosity index shows
relatively small tendency of viscosity to change upon
influence of certain temperature, as oppose of low
viscosity index which shows greater viscosity change
with temperature.
Type of motor oil
FAM
Galax
SAE 15W40
SAE 30/S3
3,0
3,0
0,881
0,902
104,81
0
1
2
3
4
5
0
1
2
3
4
5
0
1
2
3
4
5
0
1
2
3
4
5
0
1
2
3
4
5
0
1
2
3
4
5
0
1
2
3
4
5
PUCH
1
104,8
111,0
113,5
119,4
126,4
132,7
14,12
14,64
15,47
16,05
16,63
17,56
135
129
122
119
116
112
230
220
208
205
197
192
10,5
9,1
7,2
6,5
6,1
5,2
98,4
123
137,1
149,4
165,3
98,4
4,9
5,9
6,7
7,3
7,9
4,9
During the exploitation it is desired that the viscosity
changes as lesser as possible with the change of
temperature. If during work temperature modes are
changeable and cause major changes of viscosity that may
cause disruptions in the functioning of the system, which
is a manifestation of increased friction, wear and damage.
Change of engine oil Viscosity Index is shown in the
Picture 1. The decrease in the Viscosity Index oil is evident
for all vehicles, exceeding the limit of 5 % (Table 2).
207
OTEH2016
EXPERIMENTALINVESTIGATIONOFOILSINFOURSTROKEENGINES
Picture 2 shows the changes viscosity at 40 C engine oils
during exploitation.
Index viscosity "0" sample:
135 (SAE 15W40)
100 (SAE 30)
140
PUCH1
PUCH2
135
PINZ1
125
Max allowed increase 20%
125,7 (SAE 15W40 )
120
Max. allowed decrease 5%
128,25
( SAE 15W40)
110
Viscosity at 40C, mm 2/s
Viscosity Index
130
PINZ2
100
IK104P1
90
80
Max. allowed decrease 5 %
95 (SAE 30)
0
1000
2000
3000
IK104P2
4000
5000
PUCH1
PUCH2
115
Viscosity of "0" sample
104,8 (SAE 15W40)
104,6 (SAE 30)
105
PINZ1
PINZ2
95
IK104P1
85
Kilometrage, km
75
0
Picture 1. The change of Viscosity Index [10]
Max allowed decrease
20% 83,8 (SAE 30)
1000
2000
3000
IK104P2
4000
5000
Kilometrage, km
The most important engine oils characteristic is the
viscosity defined as a measure of inner friction which
works as a resistance to the change of molecule positions
in fluid flows when they are under the impact of shear
force, or in other words, it is the resistance of fluid
particles to shear.
Picture 2. The change of viscosity at 40 C [10]
The increase viscosity at 40 oC engine oil is evident for
PUCH1 and PUCH2 vehicles, exceeding the limit of 20
%. The decrease viscosity at 40 oC engine oil is evident
for PINZ (exceeding the limit of 20 %) and IK104P
vehicles, Picture 3.
The viscosity is a changeable category and it depends on
the change of temperature and pressure. A higher
temperature reduces the viscosity and makes a fluid
thinner.
Viscosity "0" sample:
14,12 (SAE 15W40)
11,67 (SAE 30)
20
Viscosity at 100C, mm 2/s
Multigrade engine oils among numerous additives always
contain also viscosity index improvers. These additives
are special types of polymers, which in small
concentration significantly improve engine oils
rheological properties, especially viscosity and viscosity
index.
However, during engine oils utilization, degradation of
viscosity index improvers i.e. Break down of polymeric
molecules occurs. It results in reduction of their molecular
weight what leads to viscosity loss and oil film thickness
decrease, which causes undesirable phenomena of friction
and wear.
Max. allowed increase 20%
17,04 (SAE 15W40)
18
PUCH1
PUCH2
16
Max. allowed decrease 20%
11,29 (SAE 15W40)
14
PINZ1
PINZ2
12
Max. allowed decrease 20%
9,33 (SAE 30 )
IK104P1
10
IK104P2
8
0
1000
2000
3000
4000
5000
Kilometrage, km
Picture 3. The change of viscosity at 100C [10]
Reasons for the increase of viscosity lubricants are as
follows: oxidation of lubricants, cavitations due to
foaming lubricants, dissolution of lubricants with water,
pouring and charging system viscosity fat greater than
recommended and contamination of solid particles and
products wear lubricants.
TBN of "0" sample
10,5 (SAE 15W40)
9,8 (SAE 30)
11
Total Base Number, mg KOH/g
PUCH1
Max. allowed decrease 50%
5,25
( SAE 15W40)
10
The reasons for the reduction of lubricants viscosity are:
lubricants contamination of fuel (for motor oil), shearing
additive for reclamation viscosity, drop point of flash,
grinding molecules, lubricants contamination without
solubility with water, pouring and charging system
viscosity less fat than recommended, and the impact of
liquid cooling. Also, the causes may be high temperature,
load, uncontrolled long interval use, insufficient amount
of oil in the oil system, inefficient cooling systems and
the like.
PUCH2
9
PINZ1
8
PINZ2
7
IK104P1
6
Max. allowed decrease 50%
4,9 (SAE 30)
IK104P2
5
0
1000
2000
3000
4000
5000
Kilometrage, km
Picture 4. The change of TBN [10]
As expected, kinematic viscosity usually decreases in
time due to fuel penetration, or in well maintained
engines, there occurs a slight increase as a result of the
increase of the oil insoluble, without fuel penetration.
TBN is a neutralization number intended for measuring
all basic (alkaline) materials in the lube (acidneutralizing
components in the lubricant additive package). The TBN
208
OTEH2016
EXPERIMENTALINVESTIGATIONOFOILSINFOURSTROKEENGINES
is generally accepted as an indicator of the ability of the
oil to neutralize harmful acidic byproducts of engine
combustion. The TBN is highest when oil is new and
decreases with its use. Low TBN normally indicates that
the oil has reached the end of its useful life. TBN is a
measure of the lubricant's alkaline reserve, and mostly
applies to motor lubricants. If a lube contains no alkaline
additives, there is little use to determine a TBN, as there
will likely be none. Combustion acids attack TBN, e.g.,
sulfuric acid, decreasing as it consumes.
(100 ppm, Table 2) for PUCH1 and PINZ2 vehicles.
Content of cooper is significantly below the allowable
limits (50 ppm, Table 2) for all vehicles.
300
Max content of Fe 100 ppm
Content of Fe, ppm
Figure 4 shows the changes of total base number (TBN)
engine oils. The decrease TBN engine oil is evident for all
vehicles. Until 5.000 km TBN value does not exceed the
allowed limit, except for PUCH1 vehicle.
120
0
1000
2000
3000
4000
3000
4000
5000
12
PUCH1
Max . content of Cu 50 ppm
10
PUCH2
8
PINZ1
6
PINZ2
4
IK104P1
PINZ1
PINZ2
IK104P2
0
1000
2000
3000
4000
5000
Kilometrage, km
IK104P1
Max. allowed decrease 20%
192 (SAE 30)
Max. allowed decrease 20%
184 (SAE 15W40)
2000
Picture 6. The change of content Fe [10]
160
140
1000
Kilometrage, km
Content of Cu, ppm
Flash point, C
180
IK 104P1
PUCH2
200
PINZ2
100
IK104P2
PUCH1
220
PINZ1
150
Picture 5 shows the change of flash point for engine oils.
The decrease in the flash point is noticeable, and by the
end of exploitation testing exceeds the allowed limits (20
%, Table 2) for PINZ vehicles.
Flash point "0" sample:
230 (SAE 15W40)
240 (SAE 30)
PUCH2
200
50
Flash point represents data that shows what temperature
leads to open fire ignition by the steam created by oil
heating. In engine oil analysis the flash point determines
the presence of fuel oil, which is a consequence of poor
motor (bad work injectors). The reduction of flash point is
due to the penetration of fuel.
240
PUCH1
250
Picture 7. The change of content Cu [10]
IK104P2
5000
3. CONCLUSION
Kilometrage, km
The interpretation of used oils analysis is very complex,
because the individual analyses are interdependent. That
is the reason why it is necessary to know the entire oil
analysis, and not bring conclusions based on individual
analysis results. It is also necessary to establish both
normal and critical quality levels for specific oils in given
engines and under specific application conditions.
Picture 5. The change of flash point [10]
Analysis of the contents of different metals that are in the
lubricant is very important. Metal particles are abrasive,
and act as catalysts in the oxidation of oils. In motor oils,
the origin of the elements may be from the additives, the
wear, the fuel, air and liquid for cooling. Metals from the
additives can be Zn, Ca, Ba, or Mg and that indicates the
change of additives. Metals originating from wear are: Fe,
Pb, Cu, Cr, Al, Mn, Ag, Sn, and they point to the
increased wear in these systems. Elements originating
from the liquid for cooling are Na and B, and their
increased content indicates the penetration of cooling
liquid in the lubricant. Increased content of Si or Ca,
which originate from the air, points to a malfunction of
the air filter.
The lubricant, being an inevitable factor in the
tribomechanical system of engine has apart from the
usual lubricating role, also an important role in detecting
the engine operation efficiency and condition. This is
achieved through a systematic monitoring of oil in
application and a permanent contact between the motor
oil manufacturer and user.
Analyses from used oil sample should always be
compared with previous samples and final conclusions
should be based on trend analysis and has two closely
related objectives: to obtain information on the lubricant
drain intervals and preventive maintenance of the
machine.
Iron and copper content (Pictures 6 and 7), as a product of
wear, in the oil charge to the end of exploitation testing
has a growing trend.
Content of iron is significantly above the allowable limits
209
OTEH2016
EXPERIMENTALINVESTIGATIONOFOILSINFOURSTROKEENGINES
Investigations it was realized that there is a change of
physicalchemical characteristics of oil for lubrication in
the engines vehicle. These changes are in direct
dependence on the state of all elements tribomechanical
engines system, and depending on their functional
characteristics.
[6] V. Macian, R. Payri, B. Tormos, L. Montoro:
Applying analytical ferrography as a technique to
detect failures in diesel engine fuel injection systems,
Wear, Vol. 260, No. 45, pp. 562566, 2006.
[7] R.I. Taylor, R.C. Coy: Improved fuel efficiency by
lubricant design: a review, Proceedings of the
Institution of Mechanical Engineers, Part J: Journal
of Engineering Tribology, Vol. 2014, No. 1, pp. 115, 2000.
[8] M. Piest, C.M. Taylor: Automobile engine tribologyapproaching the surface, Wear, Vol. 241, No. 2, pp.
193203, 2000.
[9] S.A. Adnani, S.J. Hashemi, A. Shooshtari, M.M.
Attar:The Initial Estimate of the Useful Lifetime of
the Oil in Diesel Engines Using Oil Analysis,
Tribology in Industry, Vol. 35, No. 1, pp. 6168,
2013.
[10] S. Peric: The development of a method of diagnosis
the condition from the aspect of physicalchemical
and tribological characteristics of lubricating oils of
vehicles, PhD thesis, Military Academy, Belgrade,
2009.
[11] S. Peri, B. Nedi: Monitoring oil for lubrication of
tribomechanical engine assemblies, Journal of the
Balkan tribological association, Vol. 20, No 4, pp.
646 664, 2014.
References
[1] J. Denis: Lubricant properties analyses and testing,
Editions Tehniq, Paris, 1997.
[2] D. Grgi: Online monitoring of oil quality and
conditioning in hydraulics and lubrications systems,
in: Proceedings of 10th SERBIATRIB '07,
Kragujevac, Serbia, pp. 305309.
[3] I. Maui, P. Todorovi, A. Brkovi, U. Proso, M.
apan, B. Jeremi: Development Of Mobile Device
For Oil Analysis, Tribology in Industry, Vol. 32, No.
3, pp. 2632, 2010.
[4] V. Macian, B. Tormos, P. Olmeda, L. Montoro:
Analytical approach to wear rate determination for
internal combustion engine condition monitoring
based on oil analysis, Tribology International, Vol.
36, No. 10, pp. 771776, 2003.
[5] L. Guan, X. L. Feng, G. Xiong, J. A. Xie: Application
of dielectric spectroscopy for engine lubricating oil
degradation monitoring, Sensors and Actuators A:
Physical, Vol. 168, No.1, pp. 2229, 2011.
210
OPTIMIZATION OF THE BOX SECTION OF THE SINGLEGIRDER
BRIDGE CRANE BY GRG ALGORITHM ACCORDING TO DOMESTIC
STANDARDS AND EUROCODES
GORAN PAVLOVI
Lola Institut, Belgrade, goran.pavlovic@li.rs
VLADIMIR KVRGI
Lola Institut, Belgrade, vladimir.kvrgic@li.rs
STEFAN MITROVI
Lola Institut, Belgrade, stefan.mitrovic@li.rs
MILE SAVKOVI
Faculty of Mechanical and Civil Engineering in Kraljevo, Kraljevo, savkovic.m@mfkv.kg.ac.rs
NEBOJA ZDRAVKOVI
Faculty of Mechanical and Civil Engineering in Kraljevo, Kraljevo, zdravkovic.n@mfkv.kg.ac.rs
Abstract: The paper considers the problem of optimization of the box section of the main girder of the singlegirder
bridge crane. Reduction of the area of the box cross section is set as the objective function. The algorithm of
generalized reduced gradient (GRG2 algorithm) was used as the methodology for determination of optimum
geometrical parameters of the box section. The criteria of permissible stresses, local stability of plates, lateral stability
of the girder, static deflection, dynamic stiffness, minimum plate thickness and production feasibility (distance between
the webs) were applied as the constraint functions. Verification of the used methodology was carried out through
numerical examples and the comparison with some existing solutions of cranes was made. The comparative
optimization results show changes of the box section optimum geometric values due to domestic standards or
eurocodes.
Keywords: singlegirder bridge crane,box section, optimization, GRG2 algorithm, eurocodes.
In most cases, the optimization of the girders is performed
by FEA. In paper [02] design optimization is performed
using principal calculations and detailed 3D FEA by
changing primarily thickness of the box girder plates.
Principal static and d ynamic calculations for two models
of double box girder, are given in the paper. As a result of
optimization a reduction of mass of 38% achieved and
stressdeformation characteristics considering yield
stranght and stability of construction was not endangered.
In paper [03] the main aim is to reduce the structural mass
of a realworld doublegirder overhead crane, through the
use of modern computer modelling and simulation
methods and applications by CATIA software. The
structural mass reduction are designed and verified by
structural static stress simulations. The numerical analysis
of stress and strain field for double beam bridge crane is
done under maximum load conditions by using ABAQUS
software in [04], the distribution and largest area under
load of bridge is obtained. At the same time three natural
frequencies and mode shapes of double beam bridge crane
is analyzed, which the results provide a theoretical basis
and reference for double beam bridge crane designer.
Similarly, in paper [05] FEA of main girder of bridge
1. INTRODUCTION
Singlegirder bridge cranes are widely applied in
industrial plants. The box section of the main girder is
most often used for medium and high carrying capacities
of these cranes. The mass of this girder has the largest
share in the total mass of the singlegirder bridge crane,
and that is the reason why it is very important to reduce it
in order to obtain a lighter structure, which also reduces
the market price of the crane.
Most papers treat the problem of optimization or stress
analysis of doublegirder bridge cranes. It is known that
doublegirder cranes are intended for lifting and
transportation of large loads and for larger spans than
singlegirder cranes. However, the number of singlegirder cranes installed in plants is significant so that
optimization of their main girders is justified. In some
cases it is more economical to use the singlegirder bridge
crane in relation to the doublegirder bridge crane from
reduction of mass of the girders point of view, [1]. This is
good reason to be pay more attention for these types of
cranes.
211
OTEH2016
OPTIMIZATIONOFTHEBOXSECTIONOFTHESINGLEGIRDERBRIDGECRANEBYGRGALGORITHMACCORDING
crane is conducted by ANSYS software. The natural
frequencies and the vibrationmade vectors of the first 6
orders are obtained, and the vibration hazardous areas are
found. Rational optimization of girder is able to avoid the
resonance frequency region, which will help to improve
the reliability and lifespan of the girder.
2. MATEMATHICAL FORMULATION OF
THE OPTIMIZATION PROBLEM
As the optimization task represents mass minimization, it
is necessatry to determine the values of geometrical
parameters of the crosssection of the girder which define
its minimum area.
In addition to the application FEM, is increasingly being
applied various analytical and numerical methods for
optimization process. In paper [06] the optimum design of
boxtype of crane girders is considered by using nonlinear
programming techniques. The limitations on the stresses
and the deflections induced in the girder in different load
conditions are stated in the form of inequality constraints.
The effects of collision, wind loading and damping time
of vibration are also considered in the problem
formulation. The resulting nonlinear programming
problem is solved by using an interior penalty function
method. Similarly, in paper [07], by using the same
method, the mass of the girder was reduced up to 17,11%
for medium rank, and 11,56% for heavy rank. In paper
[08] was made parametric synthesis of the span of beams
bridge cranes on the basis of multivariate analysis of their
geometrical parameters and combining design and test
calculations. Developed the methodical and software for
calculation of metal structures of bridge cranes with
optimal weight and size characteristics. In paper [09] the
method of Lagrange multipliers was used as the
methodology for approximate determination of optimum
dependences of geometrical parameters of the box
section. The criteria of lateral stability and local stabiity
of plates were applied as the contraint functions. The
obtained results of optimization of geometrical parameters
were verified on numerical examples. In order to solve the
optimization of the bridge crane box girder, on the basis
of the polyclonal selection algorithm, an improved
polyclonal selection algorithm based on negative selection
is proposed in paper [10]. By taking advantages of the
clone deletion and supply operation based on negative
selection, the optimization ability of the proposed
algorithm is improved. The experiment results of the
typical function test and the crane box girder optimization
verify the effectiveness of this algorithm.
Minimization of the mass corresponds to minimization of
the volume, i.e. the area of the cross section of the girder,
where the given boundary conditions must be satisfied.
The area of the cross section primarily depends on: height
and width of the girder and thickness of plates.
The optimization problem defined in this way can be
given the following general mathematical formulation:
minimize f ( X )
subject to: gi ( X ) 0, i = 1,..., m
and li X i ui , i = 1,..., n
where:
f ( x )  the objective function,
gi ( X ) 0  the constraint function,
m  the number of constraints,
li , ui  lower, i.e. upper boundary, where ui > li .
X = { x1 ,..., xn }  represents the design vector made of
T
n design variables.
3. OBJECTIVE FUNCTION
The model is presented in Picture 1.:
y
b
x1
2
x1
x
Having in mind that there are a large number of singlegirder bridge cranes in plants, this paper deals with the
investigation into optimization of the box crosssection of
singlegirder bridge cranes. As it can be seen in the
mentioned papers, there are different constraint functions
so that it can be concluded that a better objective function,
i.e. smaller girder mass is obtained for a larger number of
constraints.
y1
t2
yc
t2
yB = y C
The mentioned papers point to the importance of
optimization of the main girder of the crane and creation
of the model which can allow a more real description of
the crane behaviour in operation.
y2
3xt2
x2
yA
t1
b1
P
t3
A
C
xA
xC
Having in mind previously stated results and conclusions,
the objective of this paper is to define optimum values for
box section geometric parameters of singlegirder bridge
crane, which would lead to mass reduction.
xB
b2
Picture 1. The box section of the singlegirder
bridge crane
212
OPTIMIZATIONOFTHEBOXSECTIONOFTHESINGLEGIRDERBRIDGECRANEBYGRGALGORITHMACCORDING
The highest stresses occur at the middle of the span. The
values of the bending moments in the corresponding
planes are:
The objective function is represented by the area of the
crosssection of the box girder.
The vector of the given parameters is:
G
x = ( Q, L, M cv , M ch , ka , mk , d ,...)
M VI = R ( L e1 ) 2 / 4 L + q L2 / 8
(8)
M HI = Rh ( L e1 ) 2 / 4 L + q L2 ka / 8
(9)
(1)
where:
Q the carrying capacity of the crane,
where:
L the span of the crane,
= 1, 05  the coefficient which depends on the
classification class (1,05 for classification class 2), [15],
mk the mass of the trolley,
d the distance between the wheels of the trolley,
= 1,15  the dynamic coefficient of the influence of
load oscillation in the vertical plane, [11],
Mcv and Mch the bending moments in the vertical and
horizontal planes,
q specifically weight per unit of length of the girder,
ka = 0,1  the dynamic coefficient of crane load in the
horizontal plane, [11].
e1 = F2 d / R ,
while the values of the corresponding forces are:
The area of the cross section (2), i.e. the objective
function, is:
A = 2 h t2 + b t1 + b2 t3
F1 = F2 = ( Q + mk ) / 2 g , R = F1 + F2
(10)
F1, st = F2, st = (Q + mk ) / 2 g = Fst
(11)
F1, h = F2, h = Fst ka , Rh = F1, h + F2, h
(12)
Ft = R ( L e1 ) / 2 L + q L / 2
(13)
P = R / nk
(14)
(2)
The geometrical properties in the specific points of box
section (Picture 1) shall be determined by wellknown
expressions (Wx1, Wy1, Wx2, Wy2, Sx2, Ix, WxB, WyB, WxA,
WyA, SxA, WxC, WyC).
4. CONSTRAINT FUNCTIONS
4.1. The criterion of of strength
where:
Strength check is conducted in specific points of girder.
Actual stress r has to be lower than critical stress k
which depends on load case.
r k
(3)
k = k 1 = f y / 1
(4)
k = k 2 = f y / 2
(5)
P  the vertical crane wheel load,
nk  number of the wheels of trolley.
1. The strength in specific points of the girder:
Point 1:
Maximum normal stress at point 1:
1,u = M VI / Wx1 + M HI / Wy1 k
according to domestic standards, or:
k = f y / m /1
OTEH2016
(15)
Point 2:
(6)
Maximum normal stress at point 2:
z 2 = M VI / Wx 2 + M HI / Wy 2
according to eurocodes,
where:
(16)
Maximum tangential stress at point 2:
f y the minimum yield stress of the plate material,
= Ft S x 2 / (2 t2 I x ) k / 3
1 the factored load coefficient for load case 1,
Maximum equivalent stress at point 2:
2 the factored load coefficient for load case 2,
m = 1,1 general resistance factor, [14].
2,u = z 2 2 + 3 2 k
The constraint function has the following form:
g1 = r k 0
(17)
(18)
2. The strength in the bottom flange of the girder in
specific points:
(7)
213
OPTIMIZATIONOFTHEBOXSECTIONOFTHESINGLEGIRDERBRIDGECRANEBYGRGALGORITHMACCORDING
Point B:
OTEH2016
A,u = zA2 + oyA , Ed 2 zA oyA , Ed + 3 A 2 k (28)
I According to domestic standards:
Point C:
Maximum equivalent stress at point B:
B ,u = M VI / WxB + M HI / WyB k1
I According to domestic standards:
(19)
The normal stresses due to the local pressure of the trolley
wheel at point C, [15]:
II According to eurocodes:
The normal stresses due to the local pressure of the trolley
wheel at point B, [13]:
xP,C = K Cx P / t32 k1 , zP,C = K Cz P / t32 k1 (29)
oxB , Ed = cxB P / t32 k , oyB , Ed = c yB P / t32 k (20)
K Cx , K Cz  the corresponding coefficients for calculating
stresses at point C, [15].
cxB , c yB = 0
the
corresponding
coefficients
for
Maximum normal stress in the direction of the z axis at
point C:
calculating stresses at point B, [13].
zC = K Cz P / t32 + M VI / WxC + M HI / WyC
Maximum equivalent stress at point B:
B ,u = M VI / WxB + M HI / WyB + cxB P / t32 k
(21)
(30)
Maximum equivalent normal stress at point C:
Point A:
C ,u = zC 2 + xP,C 2 zC xP,C k 2
(31)
I According to domestic standards:
II According to eurocodes:
The normal stresses due to the local pressure of the trolley
wheel at point A, [15]:
P
x, A
= K Ax P / t k1 ,
2
3
P
z, A
The normal stresses due to the local pressure of the trolley
wheel at point C, [13]:
= K Az P / t k1 (22)
2
3
oxC , Ed = cxC P / t32 k , oyC , Ed = cCy P / t32 k (32)
K Ax , K Az  the corresponding coefficients for calculating
stresses at point A, [15].
cxC , cCy  the corresponding coefficients for calculating
stresses at point C, [13].
Maximum normal stress in the direction of the z axis at
point A:
zA = K Az P / t + M VI / WxA + M HI / WyA
2
3
Maximum normal stress in the direction of the z axis at
point C:
(23)
zC = cxC P / t32 + M VI / WxC + M HI / WyC
Maximum tangential stress at point A:
A = Ft S xA /(2 t2 I x ) k1 / 3
Maximum equivalent normal stress at point taki C:
(24)
C ,u = zC 2 + oyC , Ed 2 zC oyC , Ed k
Maximum equivalent stress at point A:
A,u = zA2 + xP, A2 zA xP, A + 3 A2 k1
(33)
(25)
(34)
4.2. The criterion of local stability of plates
II According to eurocodes:
1. Local stability according to domestic standards:
The normal stresses due to the local pressure of the trolley
wheel at point A, [13]:
Local stability check is done in the same manner for both
the flange and web plates. The paper considers the case
with one row of horizontal stiffeners, along with vertical
stiffeners being placed at the distance 2h (Picture 2).
oxA , Ed = cxA P / t32 k , oyA , Ed = c yA P / t32 k (26)
cxA , c yA  the corresponding coefficients for calculating
stresses at point A, [13].
Maximum normal stress in the direction of the z axis at
point A:
zA = cxA P / t32 + M VI / WxA + M HI / WyA
(27)
Picture 2. Stiffeners of the box girder
Maximum equivalent normal stress at point A:
214
OTEH2016
OPTIMIZATIONOFTHEBOXSECTIONOFTHESINGLEGIRDERBRIDGECRANEBYGRGALGORITHMACCORDING
To satisfy the stability condition it must be:
4.4. The criterion of girder stiffness
max ux v
(35)
In order to satisfy this criterion, it is necessary that the
deflection in vertical plane have the value smaller than the
permissible one:
where:
max  maximum pressure stress,
f = F1, st L3 1 + (1 6 2 ) / 48 E I x f dop (45)
ux  buckling stress limit, [12],
where:
v  yield stress, [12].
F1,st  force acting upon girder beneath the trolley wheel,
The constraint function has the form:
f dop  permissible deflection in vertical plane,
g 2 = max min( ux , v ) 0
(36)
= 1, =d/L, [15].
The constraint function has the form:
Maximum pressure stress of the top flange:
p = 1 ( M VI / Wx1 + M HI / Wy 2 )
g 4 = f f dop 0
(37)
4.5. The criterion of dynamic stiffness
Maximum pressure stresses for the webs in zone 1 and
zone 2, respectively:
1 = 1 ( M VI / Wx 2 + M HI / Wy 2 )
M VI y1 t1 h1 M HI
+
y1 t1
Wy 2
Wx 2
2 = 1
To determine dynamic stiffness, it is necessary to analyse
the vertical oscillation of main girder with the payload
(Picture 3).
(38)
1,0
(39)
11
2. Local stability according to eurocodes:
m1
The criterion of local stability of the top flange plate is:
p x fy / m
(46)
Picture 3. The model of oscillation of the main girder
with concentrated mass
(40)
The mass m1 is determined according to the expression,
[15]:
where:
x  a reduction factor, [14].
m1 = Q + mk + 0, 486 mm
The criterion of local stability of the web in zone 1 and
zone 2, respectively:
(47)
where:
1 x1 f y / m
(41)
mm  the mass of the girder.
2 x2 f y / m
(42)
The time of damping of oscillation is determined from the
expression, [15]:
T = 3 / d Td
4.3. The criterion of lateral stability
where:
Safety check for lateral stability of the girder is done in
compliance with [12]. So, it has to be fulfilled:
zV 1 = M VI / Wx1 k 3 = f y / I
(48)
= 2 11 m1  the period of oscillation,
11  the deflection of the girder caused by the action of
(43)
the unit force,
where:
Td  the permissible time of damping of oscillation, [15],
 the buckling coefficient, [12].
d  the logarithmic decrement which shows the rate of
damping of oscillation, [15].
Using the above mentioned relations, the constraint
function reads:
g3 = M VI / Wx1 f y / I 0
The constraint function for the criterion of dynamic
stiffness is:
(44)
g5 = T Td 0
215
(49)
OPTIMIZATIONOFTHEBOXSECTIONOFTHESINGLEGIRDERBRIDGECRANEBYGRGALGORITHMACCORDING
The justification of application GRG2 algorithm was
checked on five solutions of singlegirder bridge cranes
which are in operation. It can be observed that greater
area of box girder cross section are obtained according to
eurocodes in comparison with those which are calculated
due to domestic standarads.
5. NUMERICAL PRESENTATION OF
OBTAINED RESULTS
The optimization is done by generalized gradient method
(GRG2 algorithm) and using Solver tool from Analysis
module of EXCEL software.
Variable parameters are section height and width and
plates thicknesses, (50). All constraint functions stated in
previous chapters are taken into analysis.
( h,b,b1 ,b2 ,t1 ,t2 ,t3 )
OTEH2016
The conclusion is that further research should be directed
toward a multicriteria analysis where it is necessary to
include additional constraint functions, such as: material
fatigue, influence of manufacturing technology,
optimization of the ratio of plate thicknesses, types of
material, conditions of crane operation and economy.
(50)
The case with one row of longitudinal stiffeners is
analysed. Also, the longitudinal stiffeners are adopted to
be placed at the distance of h/4.
References
[1] Babin,N., Georgijevi,M., ostakov,R. Izbor tipa
mostovske dizalice, sanduaste konstrikcije u funkciji
kriterijuma minimalne teine nosaa, SMEITS,
Beograd, 1979.
[2] Beirovi,A.,
Vukojevi,D.,
Hadikaduni,F.:
Optimization of double box girder overhead crane in
function of cross section parameter of main girders,
15th International Research/Expert Conference
Trends in the Development of Machinery and
Associated Technology TMT 2011, 1218
September 2011, Prague, Czech Republic.
[3] Sankar,A., Vijayan,V., Ashraf,I.: Reducing the
structural mass of a realworld double girder
overhead crane, International Journal of Advances in
Engineering & Technology IJAET, Vol. 8, Issue 2,
pp. 150162, april 2015.
[4] Yu,Q., Mao,X.: The Performance Analysis of Double
Beam Bridge Crane Based on Computer Simulation
Technology, Key Engineering Materials, Vol. 584, pp.
107111, Trans Tech Publications, Switzerland, 2014.
[5] Peng,R., Qin,X.: Modal analysis of crane girder
based on ANSYS Workbench, Advanced Materials
Research, Vol. 951, pp. 5861, Trans Tech
Publications, Switzerland, 2014.
[6] Rao,S.S.: Optimum Design of Bridge Girders for
Electric Overhead Traveling Cranes, Journal of
Engineering for Industry, Vol. 100, pp. 375382,
august 1978.
[7] Tian,G., Zhang,S., Sun,S.: The Optimization Design of
Overhead Traveling Cranes Box Girder, Advanced
Materials Research, Vols. 538541, pp. 28502855,
Trans Tech Publications, Switzerland, 2012.
[8] Anzev,V.Y.,
Tolokonnikov,A.S.,
Potapov,S.A.,
Kalabin,P.Y.: Improvement of the method of
calculation of span of beams load lifting machines of
bridge crane type, 
, ,
, 2013. . 7, .1, pp. 144153, 2013.
[9] Savkovi,M., Gai,M., Pavlovi,G., Bulatovi,R.,
Zdravkovi,N.: Optimization of the box section of the
main girder of the bridge crane according to the
criteria of lateral and local stability of plates, The
7th International Symposium KOD, p.p. 113120,
Balatonfred, Hungary, 2012.
Minimum thickness of the web plates is adopted to be 5
mm and minimum thickness of the bottom and top flange
is adopted to be 6 mm, which are also the constraint
functions. In addition, as one more constraint function,
minimum distance between the web plates is taken to be
25 cm.
Table 1. Opimization results according to domestic
standards
Saving
Q (t) L (m) A1 (cm2) A2 (cm2)
(%)
Amiga 8t
8
15,1
208,8 151,8
27,30
Amiga 5t
5
15,2
177,2 138,3
21,95
Radijator
10
16,8
248,8 161,8
34,97
10t
JEEP
16
15
237,6 192,87
18,83
16t
Lola 5t
5
16,5
256,6 172,8
32,66
Table 1 and Table 2 show the results of the optimization
for five solutions of singlegirder bridge cranes, according
to domestic standards and eurocodes, respectively. A1 and
A2 are the values of the area of the box crosssection
before and after optimization, respectively.
Table 2. Opimization results according to eurocodes
Saving
Q (t) L (m) A (cm2) A (cm2)
(%)
Amiga 8t
8
15,1
208,8 180,24 13,68
Amiga 5t
5
15,2
177,2 159,3
10,10
Radijator
10
16,8
248,8 182,95 26,47
10t
JEEP
16
15
237,6 231,15
2,71
16t
Lola 5t
5
16,5
256,6 180,24 13,68
6. CONCLUSION
The paper presented optimum dimensions of box section
of singlegirder bridge crane by GRG2 optimization
method. The criteria of permissible stresses, local stability
of plates, lateral stability of the girder, static deflection,
dynamic stiffness, production feasibility (distance
between the webs) and minimum thicknes of the plates
were applied as the constraint functions. The objective
function was minimum crosssectional area, whereby
given constraint conditions were satisfied.
216
OPTIMIZATIONOFTHEBOXSECTIONOFTHESINGLEGIRDERBRIDGECRANEBYGRGALGORITHMACCORDING
OTEH2016
standardizaciju, Beograd, 1986.
[13] EN 19936, Eurocode 3  Design of steel structures Part 6: Crane supporting structures, 2007.
[14] prEN 1300131, Cranes  General Design  Part 31:
Limit States and proof competence of steel structure,
2010.
[15] Ostri,D., Toi,S.: Dizalice, Institut za mehanizaciju
Mainskog fakulteta Univerziteta u Beogradu, 2005.
[10] Chen,C., Shen,Y., Yuan,M.: Optimization Design of
Crane Box Girder Based on the Improved Polyclonal
Selection Algorithm, IOSR Journal of Mechanical
and Civil Engineering (IOSRJMCE), Volume 11,
Issue 4 Ver. II (Jul Aug. 2014), pp 120125, 2014.
[11] SRPS M.D1.050, Standardi za dizalice, Jugoslovenski
zavod za standardizaciju, Beograd, 1968.
[12] SRPS U.E7, Standardi za proraun noseih elinih
konstrukcija,
Jugoslovenski
zavod
za
217
MATHEMATICAL MODELING DYNAMIC PERFORMANCE OF
ARTILLERY FIRE SUPPORT IN THE OFFENSIVE OPERATION
DAMIR PROJOVI
Military Academy, Belgrade, damirpro@yahoo.com
ZORAN KARAVIDI
Military Academy, Belgrade, zkaravidic@gmail.com
MIROSLAV OSTOJI
Military Academy, Belgrade
Abstract: This paper presents a mathematical model of a artillery fire support as a tool for the analysis course of action
in the decisionmaking process. Defined as mathematical dependence of the input model with indicators of the situation
in a certain moment of combat. The model is based on indicators predict the state of the artillery unit during combat.
Results of the predictions provide more efficient evaluation course of action and faster transition to the next phase of
the decisionmaking process.
Keywords: Mathematical modeling, the dynamics of combat operations, artillery fire support, offensive operation.
Situation models treat some operational problems of the
general model. In current practice, it is known as
operational or tactical mission.
1. INTRODUCTION
Model is often defined as a simplified picture of reality,
and the analogy between the model and the real system is
the basic assumption underlying the method of modeling.
Given that the real systems are extremely complex,
simplifying is required in the modeling process. This
especially applies to organizational systems, such as
military organizations, and even more so when that
organization system performs combat.
According to the generality criteria, considered model
rocket artillery fire support is special. According to the
completeness criteria it is partial. Relationships within the
model, between the input and output variables are defined
and mathematically supported.
2. ROCKET ARTILLERY FIRE SUPPORT
IN THE OFFENSIVE OPERATION
Besides in combat there are a number of phenomena that
are not repeated in the form in which they were before
happened, and are very difficult to predict. From this
point it is possible to make models of differentiation and
performance dependent and independent variables and
analyze it or to experiment. Experimenting on real
systems, when it comes to combat is almost impossible.
Artillery support has a great tactical mobility and constant
as soon as possible readiness for action on the request. In
addition, Artillery support is completely automated in
area of preparation and firing procedures. It is able to
simultaneously, accurately and efficiently act on a very
large number of targets. It is permanent included in digital
environment, updating the combat situations in real time,
so that after carrying out a combat order, immediately
ready to execute a new task.
Models of war games could be with different level of
generality and can be divided into: general, special and
singly. The most general are those that can be applied to a
number of different situations. In contrast to the general,
the special models are made for solving a part of the
general problem. Individual models treat a problem from
the framework of a particular model.
Artillery is grouped and developed to engage on main
direction of operations. The reason for it is a largescale
and more efficient participation in the decisive stages of
the operation. Grouping artillery units includes deliver
artillery and missile units fire support artillery into
appropriate groups. Artillery unit of the fire support
forming artillery and rocket groups.
According to the criteria of the completeness, models
could be: complex, the partial and situational. Complex
model contains a generic model, one or more of the
partial model and the one or more situational model.
Complex models display the operation as a war
phenomenon in a particular space or time from the
beginning to the end, including preparations. Partial
models are related to some content or security of
operations. It is known in the practice as ''Appendix''.
Artillery fire support is action of artillery fire at the
enemy in support of its own forces in combat operations.
It realizes with various types of artillery fire and missile
strikes to destroy and neutralize combat units and
opponents resources.
218
MATHEMATICALMODELINGDYNAMICPERFORMANCEOFARTILLERYFIRESUPPORTINTHEOFFENSIVEOPERATION
Artillery fire support in the offensive operation is divided
into two phases:
preparation of artillery attacks,
artillery support of attack. [1]
3. MODEL ARTILLERY FIRE SUPPORT
The mathematical model artillery support offensive
operations are determined by two groups of conditions
that more closely define the actual conditions of current
operations. They are related to the preparation and the
execution phase. Considering the artillery firing,
according to phases of the operation, there are two phases
of artillery firing. The preparation phase of operation
corresponding to the preparation of initial elements for
firing stage while execute the operation corresponding to
the execution phase of group shooting.
OTEH2016
Errors in determining the target position (Exc, Eyc),
Errors in topographic and surveying preparation
(Extgr, Eytgr),
Errors determining ballistic corrections (Exb),
Errors in determining meteorological corrections
(Exm, Eym),
Errors in technical preparation (Extp, Eytp),
Errors in firing tables (Extg, Eytg),
Errors of rounding elements (Exc, Eyc),
Errors of data processing (Exop, Eyop).
The first group of conditions defining the technical
characteristics of the unit or the technical characteristics
of the tools and resources that are used to prepare initial
elements for a group shooting.
The second group relates to the definition of situational
parameters in the operation, which is operationalized and
quantified with the determination of their mutual
correlation.
Picture 1. Summary error of initial elements of complete
preparation for firing
Summary error of complete preparation of initial elements
for firing is:
According to the distance: (1)
3.1. Model preparation artillery fire support
The phase of determining the initial elements is the most
complex functions of artillery firing, because it tested all
restrictions on which depends whether the shooting is not
possible or it is not in the chosen sector with defined
restricted areas, reefs shelter, with selected set of
meteorological and ballistic data and the specific
relationship firing position and target.
2
2
2
2
2
2
2
Ex = Exc
+ E xtgr
+ Exb
+ Exm
+ E xtp
+ Extg
+ Exz2 + Exop
(1)
According to the direction: (2)
2
2
2
2
2
2
E y = E yc
+ E ytgr
+ E ym
+ E ytp
+ E ytg
+ E 2yz + E yop
The application of initial elements for the shooting
demands: the exact coordinates of the units and targets;
accurate data on meteorological and ballistic conditions of
firing; oriented instruments and artillery pieces directed to
the main direction according to the instructions for
topographic and geodetic support of artillery; possession
of appropriate firing tables depending on the altitude.
These characteristics are defined through the probable
errors. Their combined effect is expressed as the summary
middle value of probable error of initial elements
preparation for firing. It will later define probability of
hitting a certain area or destination. [2]
(2)
For further consideration it is important to answer the
question, what is the size of possible target area after the
completion of initial elements preparation. Or determine
the size of land, at the distance and direction, which the
probability is 100% of hitting first projectile fired with
elements of complete initial elements preparation.[3]
Theoretically speaking, errors can be extends from minus
infinity to plus infinity. It has been proved that
satisfactory accuracy is meet if errors are smaller in
absolute value than 4 probably errors in both direction.
This means that the probability of hitting target at the
distance and direction, after the completion of initial
elements preparation, ranges from 4 PE to + 4 PE,
picture 2.
Complete preparation initial elements of indirect firing
task is to determine, as accurately as possible, the actual
conditions for fire, comparison with the tables
conditions, calculating corrections and eventually
determine the best initial elements for firing. The actual
firing conditions are variable for each firing. Normal
firing conditions are a constant value for all firing and
include conditions for which firing tables is calculated.
Area possible target position after the complete
preparation of the initial elements, by distance and
direction, and the probability that the target is at the some
point of that area is represented by ordinates of the
equation possible position of the target.
Complete preparation indirect firing initial elements
consists of the following independent group sources of
errors:
219
MATHEMATICALMODELINGDYNAMICPERFORMANCEOFARTILLERYFIRESUPPORTINTHEOFFENSIVEOPERATION
OTEH2016
tables for a certain distance. Normal law that determines
the probability of finding the target at the distance and
direction, as follows:
E +a
E a
p = 1 x
x
4 Vd
Vd
E y +b
Ey b
V p
Vp
(3)
Determination of fire regimes
In the approach of a mathematical model forming it is
shall be determined the fire regime as an average rate of
fire as a parameter that defines the artillery unit at all. Fire
regime data are known and can be found in the technical
instruction or in firing tables.
Determining the size of the singleshot destroyed surface
of the target
Picture 2. Possible position of the target area
The probability of hits is a complex event that is equally
probability by the distance multiply probability by
direction.
Singleshot destroyed target surface with appropriate
weapons is equal to surface of successful action of
fragmentation projectile. The effect of a singleshot firing
of all weapons artillery units obtained by multiplying the
size 1 and 2 and the total number of artillery weapons.
This size is important for determining the initial firepower
of artillery unit.
3.2. Model dynamics performance artillery fire
support
To form the mathematical model it is necessary to
determine the derived parameters that quantitatively
express the initial firepower opposing forces in combat. It
is a fundamental part of this paper that is dedicated to the
methodology of determining these parameters for a
specifically problem at dynamics of opposed artillery
combat. The task consists in determining the
mathematical expectation unruined number of combat
units of the two groups at any point of time.
The starting dependence
Based on the defined baseline data term can be imported:
Starting firepower first group
U1 =
1 p1 1 N1
S2
(4)
Starting firepower of the second group
The mathematical model based on the setting makes the
initial determination of firepower, taking into
consideration the following parameters:
N1 and N2, the total number of weapons translated
into the basic caliber
p1 and p2, hit probability
1 and 2, rate of fire
1 and 2, singleshot destroyed surface
S1 and S2, surface on which are deployed two artillery
gropes
Vm1 and Vm2, fire capability of both groups
U2 =
2 p2 2 N 2
S1
(5)
At the start of combat t initially firepower be value U1
t and U2t
Determination fire capability of artillery units
Operation is planned and in plane exists table of targets
which will be shot. Based on the above it counts the total
expenditure of the projectiles:
Pb k = in=1 (U p )i
For each of these parameters will be present the procedure
for calculating, according to the basic concept of
mathematical model.
(6)
It shall be put the total consumption of projectiles in the
equation, where we use the concept of number approved
combat kit of ammunition operational model.[4] Based on
it we get the possibility of fire :
Determination the probability of target hitting
The repairs of the first missile in the group firing after the
initial elements complete preparation for shooting are
summary probably errors Ex and Ey. Repairs are
calculated from center of target. Value of probably error
of distance Vd and direction Vp are random errors that are
under the normal law of errors and taken from the firing
Vm =
Ob / k
Pb / k
(7)
On firepower, during the operation, proportionally affects
the possibilities of fire respectively the number of
220
OTEH2016
MATHEMATICALMODELINGDYNAMICPERFORMANCEOFARTILLERYFIRESUPPORTINTHEOFFENSIVEOPERATION
undestroyed number of artillery pieces which uses the
following terms:
approved combat kit of ammunition:
U1 =
N11 p1 1 Vm
S2
(8)
MO1 ( t ) = V1 ( t ) N1
(17)
Defined firepower relates situational parameters specified
operating model with the mathematical model of the
dynamic combat operations artillery units.
MO2 ( t ) = V2 ( t ) N 2
(18)
where is:
Furthermore, at any point of time firepower is
proportional to surface (S1 and S2) as the percentage of
undestroyed areas.
Values
V1 ( t )
and
V1 ( t ) =
S1 ( t )
S1
(9)
V2 ( t ) =
S2 ( t )
S2
(10)
V2 ( t )
accurately
defines
MO1  mathematical expectation number of undestroyed
weapons in the first group at the time (t)
MO2  mathematical expectation number of undestroyed
weapons in the second group at the time (t)
4. IMPLEMENTATION OF THE MODEL
IN THE OFFENSIVE OPERATION
To illustrate the model to get solutions, it takes the
example offensive operation [4]. In the model of
offensive operation, division LRSV 128mm M77 formed
first artillery group (AG1), until the enemy selfpropelled
122mm howitzer battery 2s1 formed second artillery
group (AG2).
the
development dynamics of the combat between the two
sides, because it allows to determine the mean value
undestroyed units of the first and second group involved
in the combat.
For offensive operation in table 1 provides basic
information defining the problem. The main characteristic
of the problem is that the forces both of sides deployed in
defined areas, and the losses in the battle were
proportionate to the percentage of area neutralization.
It is obvious that the value of V1 ( t ) and V2 ( t ) decrease
in the course of time, so the increment is:
for the first group
V1 ( t ) = U 2V2V1t
(11)
Table 1. Situation parameters
No.
Input
1.
Target range
The probability of
2.
hitting
3.
Rate of fire
4.
Number of weapons
Singleshot destroyed
5.
surface
The surface of enemy
6.
deployed artillery unit
for the second group
V2 ( t ) = U1V1V2 t
(12)
From the previous term, when allowed to t0 obtained
Diners differential equations:
V1
= U 2V2V1
dt
V2
= U1V1V2
dt
(13)
(14)
Group 1
12106m
Group 2
11550m
0.16087
0.195
5 min
18
1
5 min1
6
0.1 ha
0.1 ha
3 ha
9 ha
Starting firepower of units is calculated on the basis of
initial data from table 1. Based on the above analytical
expressions obtained numerical values are: U1 = 0.48261
and U2 = 0.065107. Based on the data from table 1, values
U1 and U2 follows values shown in Table 2.
The initial conditions of Dinners equations are defined
for t = 0 where: V1 (0) = 1; V2 (0) =1. Substituting values
for the second differential equation obtained terms:
Based on the obtained model, it could be calculated the
maths expected value of the percentage undestroyed area
for all the time duration of the combat.
Table 2. Results of counting
Time of artillery
V(t)a
combat [min]
1
0.95
2
0.92
3
0.90
4
0.89
5
0.88
6
0.87
7
0.87
8
0.87
9
0.87
Based on the calculated values for size V1, V2, it is
possible to determine the mathematical expectation
From the data in Table 2. it could be conclude the
following:
V1 = eU1t
U1 U 2
U1eU1t U 2 eU 2t
(15)
V2 = eU 2t
U1 U 2
U1t
U1e U 2 eU 2t
(16)
which constitute simultaneously an analytical model of
artillery and counter artillery combat. [5]
221
V(t)b
MOa(t)
MOb(t)
0.63
0.40
0.26
0.17
0.11
0.07
0.05
0.03
0.02
17.09
16.54
16.20
15.98
15.84
15.75
15.69
15.65
15.62
3.75
2.39
1.54
1.00
0.65
0.43
0.28
0.18
0.12
MATHEMATICALMODELINGDYNAMICPERFORMANCEOFARTILLERYFIRESUPPORTINTHEOFFENSIVEOPERATION
Percentage of undestroyed surface is going to change in
a function of time. For example: for the first artillery
group after 2 minutes of combat this percentage is 92%
and for the second artillery group value is 40%.
The mathematical expectation undestroyed number of
weapons differs analog to upper conclusion.
OTEH2016
increase the total number of weapons involved in
combat.
Using the Diners equations in the preparation of combat
operations, headquarters of artillery groups can
quantitatively determine the balance of power and to
predict the artillery combat results. Adjusting certain
parameters can assure success in their own unit.
5. CONCLUSION
References
Based on the results obtained by applying the analytical
model in the offensive operation can be concluded that
such an approach in the analytical modeling provides
great opportunities in the simulation and research of real
problems of the use of artillery units in combat.
[1] Borbeno pravilo artiljerije, GVS KoV, 2013,
Beograd.
[2] ivanov,.: Teorija gaanja, udbenik za vojne kole
(smer artiljerije) i artiljerijske jedinice, SSNO, G
JNA, UA216, VIZ, Beograd, 1979, str. 262.
[3] Kokelj,T.: Uticaj tanosti pripreme poetnih
elemenata posrednog gaanja na preciznost
artiljerijske vatre, magistarski rad, 2004, Vojna
akademija.
[4] Projovi,D.: Optimizacija artiljerijske vatrene
podrke u operacijama Vojske Srbije primenom
globalnog pozicionog sistema, 2016, PhD, Vojna
akademija, Beograd.
[5] Petri,J., Koji,Z.: Operaciona istraivanja, 1990,
Nauna knjiga Beograd.
The first part has general importance. In the second part,
these parameters are applied in specific conditions that
make it possible to describe the combat process with
Diners system of differential equations. Based on
equations, it performs the analytical model of combat
operations for specific conditions.
Based on the analysis of mathematical expectation
undestroyed number of weapons, it can be concluded that
the success of the mutual artillery combats it:
increasing the probability of hitting;
increasing the rate of fire (reaction time);
increasing the surface on which they placed their own
artillery units;
222
MODELING AND MULTIBODY SIMULATION OF LAND ROVER DEFENDER 110
RIDE AND HANDLING DYNAMICS
NABIL KHETTOU
Military Academy, University of Defense in Belgrade, Serbia, nabil.khettou@gmail.com
DRAGAN TRIFKOVI
Military Academy, University of Defense in Belgrade, Serbia, dragan.trifkovic@va.mod.gov.rs
SLAVKO MUDEKA
Military Academy, University of Defense in Belgrade, Serbia, slavko.muzdeka@gmail.com
Abstract: This paper deals with the implementation of a multibody simulation model of the offroad ground vehicle Land
Rover Defender 110 in order to simulate its handling and ride behavior. Laboratory measurements on the vehicle were
carried out in order to determine component and subsystem parameters. The required measurements are then used to build
the multibody model using the software MSC.ADAMS/Car. The ride and handling dynamics of the obtained 94degrees of
freedom model are assessed for both the vehicle negotiating a discrete speed bump and performing a double lane change
manoeuver. The obtained results suggest that model responses are in accordance with the nature of the simulation tests.
However, filed measurement should be carried out to correlate the obtained simulation results.
Keywords: computer modelling, model validation, field testing, vehicle dynamic, vehicle conversion.
Thus, the simulation model should be kept as simple as
possible, but good enough to accurately represent the
dynamic behavior to be investigated [6]. According to the
author, ground vehicle dynamics can be classified into three
major components: vertical dynamics (vehicle response to
road disturbances), lateral dynamics (vehicle response to
driver steering), longitudinal dynamics (vehicle response to
engine throttle and braking). Depending on the application,
many scientific papers investigate one aspect of vehicle
dynamics. In [7], the author has suggested a method to
identify lateral tire forces using a simple vehicle model and
can be applied to the analysis of vehicle handling
performance. Pazooki [4] has developed a comprehensive
offroad vehicle model for ride analysis using a 3D tireterrain interaction model. The author has investigated both
suspended and unsuspended vehicle model responses arising
from road roughness profile. A high resolution computer
based simulation model was developed by Leatherwood [8]
which aims to emulate the ride and handling performance of
a ground military vehicle. Even though the author has tried
to establish highresolution models of all major subsystems,
the assembled full vehicle model seems to describe more
accurately the ride dynamics of the actual physical vehicle
than the handling behavior.
1. INTRODUCTION
Simulation and computer modeling have transformed the
unidirectional process of vehicle design and development
into a next level strategy which allows engineers to
reproduce manoeuvers and tests on virtual models to assess
the dynamic behavior of vehicles at different design levels
(complete redesign, derivative design, variant design, model
update, etc.) [1]. Nevertheless, vehicle simulations are
intended to reduce the cost and the duration of the vehicle
development process and help identifying errors and
deficiencies at early stages of the design process. Ride
comfort and handling properties are one of the major key
features to be investigated using vehicle simulation models.
Ride comfort is expressed as the level of discomfort
experienced by the passenger in terms of frequency and
amplitudes of mostly vertical oscillations induced by road
geometry and engine vibration. Meanwhile, handling
properties are related to the response and the stability of the
vehicle to driver and environmental inputs such as gust,
wind and road disturbances [2,3]. In the field of vehicle
dynamics, these two features can be evaluated by
performing predefined transient and steady state
manoeuvers according to internationally accepted
procedures (ISO 38881, STANAG4357ED). Generally,
ride comfort can be assessed performing straight line
driving maneuvers over standardized road obstacles and
geometries (potholes, planks, ramps, sinus waves, stochastic
uneven etc.) [4]. On the other hand, handling qualities can
be evaluated during standardized openloop steering
maneuvers and cornering events (impulse steer, fish hook,
double lane change etc.) [5]. Literature suggests that the
requirements for vehicle simulation models should be in
accordance with the considered dynamic characteristics.
1.1. Background and motivation
The need for developing vehicle simulation model comes to
the fore especially when considering the possibility of
changing the use of an existing vehicle concept or installing
multipurpose equipment. For small production series,
generally there are no financial means to develop a new
vehicle concept, yet developers seek the best available
223
MODELINGANDMULTIBODYSIMULATIONOFLANDROVERDEFENDER110RIDEANDHANDLINGDYNAMICS
platform whose dynamic behavior should be examined
before and after the reconstruction or equipment installation.
An example of this can be found in military vehicles which,
depending on the application, need to be improved,
armored, weaponized or equipped with tactical surveillance
and reconnaissance equipment.
3.
4.
5.
Modern threats and challenges require prompt reaction to
face the changing character of modern warfare and the use
of new technologies. Thus a respectful army should quickly
and efficiently equip its military units with appropriate
equipment. Here comes the advantage of developing
military mobile solutions by converting existing vehicles.
Furthermore, army can achieve significant savings in terms
of maintenance when using one platform for different
purposes. Through the last 20 years it has been noticed an
increasing need to outfit land forces with lightweight ground
vehicle for surveillance and reconnaissance of battlefield,
borders and military facilities. While leading military
industries develop new concepts of these vehicles, others
upgrade an existing vehicle by integrating the required
equipment. The most common equipment for the above
mentioned vehicles consists of combination of different
sensors, transducers and telecommunication devices such as
TV and thermal cameras, ground surveillance radars, laser
devices, acoustic sensors and radio stations, all integrated in
one block mounted on a telescopic mast. The analysis of the
integration possibility of this equipment on wheeled vehicle
should answer questions about arrangement, positioning,
additional mass, mechanical stress of vehicle structure,
dynamic stability, ride and maneuverability.
OTEH2016
assemble the full vehicle model;
Design appropriate manoeuvers to assess the full
vehicle response;
Perform maneuvers on the model;
Discuss the results and draw conclusions for further
researches.
All tests and measurements were carried out at the premises
of the sponsor institution with the use of the available
resources according to a predefined research plan.
Dimensions of characteristic components of the suspension
system were manually measured and then drawn using a
CAD software in order to determine their centers of gravity
as well as moments of inertia. The coordinates of the vehicle
CoG and the moments of inertia of the suspended mass have
been determined by Uys et al. [9] and are used in this work.
Suspension system topology was obtained by determining
the coordinates of the suspension hard points. Some
measurements were carried out in the laboratory to
determine spring stiffness, shock absorber characteristic, tire
radial stiffness, steering ratio and antiroll bar torsional
stiffness. Based on the above mentioned, each major
subsystem (Front suspension, rear suspension, front wheels,
rear wheels, steering subsystem and chassis) was built in the
software package MSC.ADAMS/Car using the
corresponding template which contains information about
the topology of the subsystem and the way it communicates
with other subsystems. The assembled full vehicle model
was then adjusted so that the total weight, weight
distribution and the inertia of the chassis match up with the
actual physical model.
1.2. Research goal and approach
The behavior of the model was analyzed on the base of the
model responses performing straight line bump course at
different speeds to evaluate the vertical dynamics, and
double lane change course to evaluate vehicle stability and
handling performance.
The main objective of this paper is to contribute to the
preliminary analysis of the possibility to change the purpose
of an existing military wheeled vehicle by integrating
special equipment for border surveillance and
reconnaissance. While mounting this equipment on a
commercial vehicle is less expensive and timesaving, the
degradations of the ride and handling performance are
considerable and have to be investigated. This is mainly
because the chassis and suspension components as well as
tires have not been modified to match up with the added
weight and the modification of the vehicles center of
gravity (CoG). The aim of this work is to create and validate
a multibody model of the Land Rover Defender 110 that is
intended to predict the ride and handling limits of the
upgraded version. First, a multibody simulation model of
the vehicle is built using available documentation and
measurement of component parameters. The detailed
simulation procedure, including driver inputs and the
maneuvers performed on the model, are presented.
2. FULL VEHICLE SIMULATION MODEL
The aim of this work is to build and validate a multibody
model of the wheeled road vehicle Land Rover Defender
110 through series of carefully selected ride and handling
tests. The main reason for this choice is the availability of
the vehicle and the fact that the sponsoring institution has
chosen this platform to integrate the surveillance and
reconnaissance equipment. The principal technical
characteristics of the vehicle which are relevant to the
model, as well as moments of inertia about the CoG, are
presented in Table 1. Picture 1 illustrates the position of the
vehicles CoG [9].
In order to achieve this goal, a stepbystep chronological
methodology was performed as follows:
1. Identify the relevant subsystems, parameters and
measures required for the model implementation in
ADAMS/Car and the behavior to be assessed
(subsystem topology, total weight, weight distribution,
moments of inertia, spring and damper characteristics,
antiroll bar characteristic, steering ratio);
2. Create templates for all considered subsystems and
Picture 1. Defenders CoG position
224
MODELINGANDMULTIBODYSIMULATIONOFLANDROVERDEFENDER110RIDEANDHANDLINGDYNAMICS
Table 1. technical characteristics of the vehicle
Model
DEFENDER 110
Wheel base
Track width
Unloaded Weight
Front suspension
Rear suspension
Axle ground clearance
Tire size
Roll moment of inertia
Pitch moment of inertia
Yaw moment of inertia
2794 mm
1486 mm
2125 kg
Live beam axle
Live beam axle
250 mm
235/85R16
744 kg.m2
2440 kg.m2
2478 kg.m2
OTEH2016
correspondent forceangle curves. Shock absorber forcevelocity curves were determined using a hydraulic press.
The tire radial stiffness was measured with the wheel fixed
at the center and the load was applied on one side of the tire.
Transmission ratio was calculated using a displacement
sensor to measure wheel deviation along with an angle
transducer to record steering wheel angle. All the
aforementioned characteristics were then processed and
stored in an appropriate form to be later used in
MSC.ADAMS/Car.
2.2. Suspension system configuration
The Defenders front suspension system shown on Picture
2a [10] consists of a live rigid axle connected to vehicle
body by means of two radius arms (1) to provide
longitudinal guidance of the axle and to react to longitudinal
forces. A Panhard rod (2) attaches the axle to the vehicle
body to prevent lateral axle displacement. This
configuration allows the axle to move only up and down
along zaxis, as well as rotation about the xaxis. A pair of
coaxial coil spring (3) and shock absorber (4) are mounted
vertically on each end of the axle. The rear suspension
shown on Picture 2b has the same configuration. However,
the Panhard rod is replaced with a triangular linkage (5).
This combination forms Roberts mechanism which
provides lateral axle positioning and makes the connection
point between the axle and the linkage (coupler point) to
move on a straight line vertically. Springs (6) are mounted
vertically at each end of the axle, while the shock absorbers
(7) are tilted slightly at an angle about yaxis to introduce
some longitudinal damping. Both front and rear suspension
systems are fitted with antiroll bars to help reduce body roll
when the vehicle corners.
2.1. Component parameters identification
In order to quantify all the required parameters for the
multibody simulation model, it was necessary to carry out a
number of measurements on the target vehicle. The
suspension topology was determined by estimating the
coordinates of the most representative bushing centers of the
front and rear suspensions in reference to a coordinate
system located at the mid contact point between the wheel
and the ground. The dimensions of the linkages are
measured and drawn into ADAMS/Car to estimate their
mass as well as their inertia properties. Individual spring
stiffness characteristics were measured using a displacement
sensor attached to an axial force transducer to obtain the
corresponding forcedisplacement curves. Compression
speed was kept small enough to consider quasistatic
loadings and the hysteresis effect was neglected. Front and
rear antiroll bars were tested in the same way to obtain the
Picture 1. Land Rover Defenders suspension systems. a front suspension, b rear suspension. 1 radius arm, 2Panhard rod, 3 front coil spring, 4 front shock absorber, 5 triangular linkage, 6 rear coil spring, 7 rear shock
absorber, 8 radius arm.
of the vehicle body;
2.3. Model implementation
3. Default properties are used for all bushings within the
In order to adapt the complexity of the model to the
model;
behavior being assessed, a number of assumptions and
simplifications were adapted and used within the model, 4. Suspension components are considered to be rigid
bodies;
those are:
5. Because of lack of technical documentation, the
1. The vehicle chassis is modelled as a rigid body;
steering subsystem was modelled using a predefined
2. Default values are used for the aerodynamic properties
MSC.ADAMS template.
225
MODELINGANDMULTIBODYSIMULATIONOFLANDROVERDEFENDER110RIDEANDHANDLINGDYNAMICS
The Land Rover Defender is modelled as a multibody
dynamic system using MSC.ADAMS/Car, comprising beam
axle front and rear suspension, steering subsystem, front and
rear wheels and vehicle chassis. The primary purpose of this
model is to simulate the ride and handling behavior of the
vehicle. Therefore, having accurate representation of the
brake system and the powertrain is not meaningful, apart
from considering the effect of their mass properties on the
vehicle behavior. A nonlinear MSC.ADAMS Pacejka 2002
tire model [11] was used and fitted with experimental data
to consider vertical dynamic of the tire under normal loads.
The front suspension is modelled as a rigid axle attached to
the vehicle body by means of two longitudinal rods
connected to the chassis by hook joints at the rear end and to
the axle by spherical joint at the front end. The Panhard rod
is modelled as a lateral rod attached to the chassis by hook
joint and to the axle by spherical joint. The steering system
is modelled as a rack pinion system using a predefined
MSC.ADAMS/CAR template and fitted with measured
steering ratio. Each end of the rack is connected to the tie
rod by constant velocity joint (convel joint). The tie rod is
connected to the wheel hub through the steering arm by a
spherical joint. The antiroll bar is modelled as two separate
bars attached to the axle by spherical joints at the lower ends
and supported by bushings fixed to the vehicle chassis. The
other ends are connected to each other by a torsional spring
which applies moment around the axis joining the two bars.
This moment is function of the angular displacement of the
two bars around the aforementioned axis. The rear
suspension is modelled in a similar way. However, the
Panhard rod was replaced by a flexible square section beam
to be representative of the triangular linkage on the real
vehicle. The beam is attached at the rear end to the middle
of the axle by a spherical joint while the front end is fixed to
the chassis. Coil springs and dampers are modelled with
OTEH2016
nonlinear curves using measured experimental data. The 3D
models of front and rear suspension and the steering system
as well as their kinematic schemes are shown in pictures 3
and 4. In addition, symbols in Picture 4 are painted in the
same colors as their corresponding elements in Picture 3.
The full model has 94 degrees of freedom, 42 moving parts,
18 spherical joints, 12 revolute joints, 9 Hooke joints, 3
translational joints, 2 convel joints, 1 cylindrical joint, 8
fixed joints and one motion defined by the rotation of the
steering wheel.
Picture 2. MSC.ADAMS/Car full vehicle model. a front
suspension and steering, b rear suspension.
2.4. Driver model
The vehicle is controlled using ADAMS/Car Driving
Machine. Vehicle longitudinal velocity is controlled so that
the actual velocity matches the desired input. The required
acceleration or deceleration to keep the vehicle velocity at
the desired value is expressed as wheel torque or brake that
must be applied to the wheels. Errors in longitudinal
velocity are compensated using controller that calculates a
feedback torque depending on the difference between actual
and the desired velocity.
Picture 4. Kinematic schemes of the suspension systems. a front suspension, b rear suspension.
226
MODELINGANDMULTIBODYSIMULATIONOFLANDROVERDEFENDER110RIDEANDHANDLINGDYNAMICS
The vehicle lateral control is achieved considering a
simplified bicycle model of the actual vehicle model in
order to identify the relationship between the path curvature
and the necessary steer angle input. Whenever the
simulation model deviates from the target path because of
the bicycle model simplification, a connecting contour is
built to bring back the vehicle to the target path. The vehicle
is steered with the parameters from the bicycle model and
the yaw rate of the simulation model is corrected using a
feedback controller. The latter calculates the error between
the yaw rates of the simulation model and the bicycle
model, so that the simulation model follows the contour as
closely as possible.
OTEH2016
2.5. Bump test
This test is used to evaluate vertical and pitch dynamic of
the vehicle negotiating a discrete obstacle (Picture 5) for
speeds ranging from 15 to 50 km/h. The test consists of
driving the vehicle straight ahead on a flat road at the
desired longitudinal speed in order to reach a steady state
before negotiating the obstacle. The vehicle is carefully
tuned so that the transient event begins at 0degree pitch
angle. The vehicle is set so that the front wheels
simultaneously make first contact with the obstacle. No
brake is applied during the test. The bump test finishes
when the vehicle regains its initial steady state.
2.4. Driver model
The vehicle is controlled using ADAMS/Car Driving
Machine. Vehicle longitudinal velocity is controlled so that
the actual velocity matches the desired input. The required
acceleration or deceleration to keep the vehicle velocity at
the desired value is expressed as wheel torque or brake that
must be applied to the wheels. Errors in longitudinal
velocity are compensated using controller that calculates a
feedback torque depending on the difference between actual
and the desired velocity. The vehicle lateral control is
achieved considering a simplified bicycle model of the
actual vehicle model in order to identify the relationship
between the path curvature and the necessary steer angle
input. Whenever the simulation model deviates from the
target path because of the bicycle model simplification, a
connecting contour is built to bring back the vehicle to the
target path. The vehicle is steered with the parameters from
the bicycle model and the yaw rate of the simulation model
is corrected using a feedback controller. The latter calculates
the error between the yaw rates of the simulation model and
the bicycle model, so that the simulation model follows the
contour as closely as possible.
Picture 5. Discrete obstacle profile.
2.5. Double Lane Change test
The double lane change manoeuver is used to evaluate
handling and stability of the vehicle in lateral dynamics. The
manoeuver consists of switching the vehicle from one lane
to another and back to the initial lane without hitting any of
the cones defining the track (Picture 6). Depending on
vehicle width, the different dimensions of the double lane
change track used in this paper are defined by STANAG4357 [12]. Steady state should be reached before the vehicle
crosses the first section of the test. The vehicle is driven
along a straight line at the desired speed. Pitch and roll
angles are carefully controlled so that the vehicle begins the
test at 0degree pitch and roll angles.
Picture 6. Double lane change test track.
crossing a bump obstacle at 30 km/h. From the diagrams, it
After entering the first section of the test track, throttle is
is evident that vehicle model reproduces the expected
maintained and the test is accomplished at constant velocity.
behavior in term of curve shape, especially the suspension
The test finishes when the vehicle regains its initial steady state.
deformations and vertical chassis acceleration. Picture 8
indicates the simulation results obtained for the vehicle
3. SIMULATION RESULTS
performing a double lane change maneuver at 70 km/h. As
can be seen from the diagrams, the simulation model truly
The MSC.ADAMS model was tested by analyzing the
predicts the behavior of the vehicle with regards to the
relevant parameters of the real vehicle negotiating a bump
nature of the considered test, especially simulated roll
obstacle at 20, 30 and 40 km/h and performing a double lane
velocity and lateral acceleration, which are significant for
change manoeuver at 50, 60 and 70 km/h. Picture 7
handling assessment.
represents the obtained results for the vehicle model
227
MODELINGANDMULTIBODYSIMULATIONOFLANDROVERDEFENDER110RIDEANDHANDLINGDYNAMICS
Picture 7. Simulation results for a bump test at 30 km/h.
228
OTEH2016
MODELINGANDMULTIBODYSIMULATIONOFLANDROVERDEFENDER110RIDEANDHANDLINGDYNAMICS
Picture 8. Simulation results for a double lane change manoeuver at 70 km/h.
229
OTEH2016
MODELINGANDMULTIBODYSIMULATIONOFLANDROVERDEFENDER110RIDEANDHANDLINGDYNAMICS
OTEH2016
nonlinear vehicle dynamics simulation for limit
handling, Proc. Inst. Mech. Eng. Part J. Automob.
Eng., vol. 216, pp. 319327, 2002.
[6] Allen, R., Rosenthal, T., Requirements for vehicle
dynamics simulation models, SAE Tech. Pap. Ser. No
940175, 1994.
[7] Kim, J., Effect of vehicle model on the estimation of
lateral vehicle dynamics, Int. J. Automot. Technol.,
vol. 11, no. 3, pp. 331337, 2010.
[8] Letherwood, M., Gunter, D., Ground vehicle modeling
and simulation of military vehicles using high
performance computing, Parallel Comput., vol. 27, no.
12, pp. 109140, 2001.
[9] Uys, P. E., Els, P. S., Thoresson, M. J., Voigt, K. G.,
Combrinck, W. C., Experimental determination of
moments of inertia for an offroad vehicle in a regular
engineering laboratory, Int. J. Mech. Eng. Educ., vol.
34, no. 4, pp. 291314, 2006.
[10] Workshop manual supplement and body repair manual.
Publication No. LRL 0410ENG, 2nd Edition. Land
Rover, 2001.
[11] Pacejka, H. B., Tire and Vehicle Dynamics, 2nd ed.
ButterworthHeinemann, 2002.
[12] NATO, STANAG 4357ALLIED VEHICLE
TESTING PUBLICATIONS (AVTPS). 1991.
5. CONCLUSION
This paper deals with the development of a multibody
vehicle model that can be used to predict vehicle dynamic
performance of a real vehicle.
A multibody simulation model of the Land Rover Defender
was developed in MSC.ADAMS/Car using vehicle
parameter measurement. The 94degree of freedom
simulation model was tested when performing bump test
and double lane change maneuver at various speeds. The
obtained results truly reproduce the expected behavior. The
developed model will be validated later against
experimental data using an instrumented vehicle.
References
[1] Weber, J., Automotive Development Processes. Berlin:
Springer, 2009.
[2] Wong, J. Y., Theory of ground vehicle, 4th ed. New
Jersey: Jon Wiley & Sons, 2008.
[3] Gillespie, T. D., Fundamentals of vehicle dynamics.
Warrendale, PA 150960001: Society of Automotive
Engineers, Inc., 1992.
[4] Pazooki, A., Rakheja, S., Cao, D., Modeling and
validation of offroad vehicle ride dynamics, Mech.
Syst. Signal Process., vol. 28, pp. 679695, 2012.
[5] Allen, W. R., Chrstos, J., Howe, G., Validation of a
230
PERSPECTIVES OF USE OF SWITCHED RELUCTANCE MOTORS IN
COMBAT VEHICLES
RADOSLAV RUSINOV
Technical University, Sofia, rusinovr@abv.bg
ILIYAN HUTOV
Joint Forces Command, Sofia, iliyan.hutov@gmail.com
RADI GANEV
University of Structural Engineering & Architecture, Sofia, radiganev@abv.bg
KOSTADIN LAZAROV
Technical University, Sofia, klazarov@tuplovdiv.bg
Abstract: The article makes a proposal for the use of switched reluctance motors for military transport and combat vehicles.
They are compared in terms of reliability, energy source, energy consumption, size and weight. The advantages of switched
reluctance motors and the prospects for their use are exhibited. A principle solution for drive operation is proposed.
Keywords: switched reluctance motors, electric vehicle, motor control, combat vehicle.
for electrical power for future military systems onboard a
vehicle. The power management and distribution system can
supply continuous power to meet such loads as propulsion,
thermal management and other small power consumers and
can also be used to supply the intermittent power to
drive/charge a pulsed power system for electric weapons
(ETC gun and DEW) or EM armor. Furthermore, the
availability of these high levels of electrical power onboard,
may be used to reduce the logistical burden of the fleet by
eliminating, in certain instances, the towed generators,
necessary to provide electric power in the field.
1. INTRODUCTION
There are several potential benefits of electric drives that are
moving the technology advancement to civil and military
applications. While some of the payoffs are common to civil
and military markets, there are some unique to both
applications as shown in Table 1.
Table 1. Electric Vehicle Benefits
Military
Civil
Vehicle Packaging
Flexibility
Fuel Economy
Fuel Economy
Onboard Power Generation
Reduced Emissions
Stealth Potential (Silent
Improved Drivability
Movement)
Improved Accelerations
Improved Accelerations
Reduced Maintenance Cost
Reduced Maintenance
Increased Silent Watch
Period
Stealth Potential (Silent Movement): The significant
onboard energy storage system can be used to meet silent
watch requirements for extended periods of time for various
missions. Depending on the power requirements of the
silent watch, a mission can be extended over a few hours,
far exceeding the silent watch capability of the current
fleets. Silent mobility over a limited distance is also
achievable where the vehicle can move in or out of a hostile
territory with a reduced chance of being detected.
For military applications the most tangible benefits are:
1) Fuel economy,
2) Available power onboard,
3) Silent watch and silent mobility,
4) Flexibility of component packaging and integration,
5) Enhanced diagnostic and prognostics.
Flexibility: An electric drives system consists of modular
components connected by cables thus giving the vehicle
designers more designing freedom. This avoids the
constraints of conventional mechanical drive systems, which
require the engine to be connected to the wheels via gear
boxes and rigid shafts. This means the components can be
arranged and integrated in the vehicle for the optimum
utilization of the available space.
Available power onboard: The electrical drive system
consists of two sources of power: the enginegenerator and
the energy storage system. The main power management
and distribution system can be designed and sized to meet
the demand of all electric power consumers in the vehicle.
This is extremely beneficial due to the increasing demand
A source of electrical powered unit with electric motors,
compared to internal combustion engines, is more reliable,
faulttolerant, easier to repair and with lower operating costs.
The electric motors have a higher coefficient of efficiency, a
much better performance, are easier to manage [1]. They have
231
OTEH2016
PERSPECTIVESOFUSEOFSWITCHEDRELUCTANCEMOTORSINCOMBATVEHICLES
Typical for this type of engine is the phase switching
necessity. The control unit could be electronic (Fig. 2) [2] or
mechanical.
also, greater flexibility in the deployment source of energy
compared to the fuel tanks. The drive with electric motors
saves a number of facilities such as transmissions, decreases
significantly the requirements to the cooling and lubrication
systems.
In case of failure of the electronic unit, phases switching
could be realized by mechanical commands controller and
mechanical contacts. This enables an emergency travel on
short distances.
Because of number of features, the machines center of
gravity is high, which leads to some limitations in
maneuverability and their resistance. The electric motors are
suitable for incorporation in the machine wheels. This
determines a lower center of gravity, greater stability and
maneuverability.
An advantage of the topology of the inverter (Fig. 3) for
feeding the SRM is that it is impossible to obtain a short
circuit.
2. MAIN
The switched reluctance motors (SRM) are characterized by
simple construction, lack of coils and magnets in the rotor,
lack of mechanical switches, maintainability, reliability,
wide range of speed control, high speeds, high starting
torque, overload capacity. They are suitable for work in
conditions of high temperatures, dust, explosive
environments, and military applications. They also have
some disadvantages, such as pulsations of the moment.
Figure 3. Inverter for supply voltage of the stator
windings.
In military vehicles the reliability is crucial. In case of
electronic switching block failure, the stator voltage is
supplied by a mechanical switch (command controller)
driven by an operator. In emergency mode, the engine,
should deliver 80% of the nominal power.
Advantages concerning the reliability  no mechanical
switch, the stator and the rotor have simple construction,
which increases the meantime before failure (MBF).
Figure 1. Switched reluctance motor
Such type motors are very suitable for gearless drives. They
have high torque at low speeds. Apparent from the
expression (1), the motor shaft torque depends on the square
of the speed:
The number of poles of the SRM may vary. Fig. 1 shows an
engine with configuration 64 (six poles in the stator and
four poles in the rotor). Pulsation of the motor torque
depends on the number of poles.
M =
SRM can work in the absence of phase (phase winding
interruption or failure in one of the power switches).
dL ( .i ) i 2
d 2
(1)
From the expression of the torque some conclusions could
be made [3]:
Torque is proportional to the square of the current, so it
does not depend on the current direction. This determines
the use of the unipolar power source, which allows
reducing the number of electronic switches in the inverter.
Due to the fact that the torque of SRM depends on the
square of the current, the mechanical characteristic
resembles that of a DC motor with excitation in series, i.
e. has a large starting torque.
The change of rotation direction is carried out by
changing the sequence of the feeding pulses to the motor
phases, which is simple for implementation.
Figure 2. SRM 6/4 an electronic switch of the stator
voltages.
SRM can operate in four quadrants of the mechanical
232
PERSPECTIVESOFUSEOFSWITCHEDRELUCTANCEMOTORSINCOMBATVEHICLES
characteristics, i. e. to realize braking modes by returning
energy to the source.
OTEH2016
3. CONCLUSION
The electric vehicles (EV) are existing technologies but
many development efforts are still needed to put valuable
products on the commercial market. The demand of
development and design effort in the field of drives, energy
sources and charging infrastructure is becoming enormous
and a challenging field for the European Community.
Easy implementation of rotation speed management.
The future of the access to energy, the environmental and
climate problems and the need to solve the mobility
problems in the urban warfare are fields in which the
electric vehicles offer a large number of interesting and
adequate solutions.
References
[1] Rusinov,E.: Application of Adaptive Digital Filter for
Rotor Position Determination of Switched Reluctance
Motor, Automatics and Informatics, Sofia, 2014
[2] http://machinedesign.com/
[3] Krishnan,R.: Switched Reluctance Motor Drives, CRC
Press, 2001
[4] All Electric Combat Vehicles (AECV) for Future
Applications, RTO TECHNICAL REPORT, TRAVT047, 2004.
Figure 4. Graphic of SRM mechanical characteristic.
Motor power: the drive system should be approximately 1 to
1.5 MW for a 50/60ton battle tank and 200/300 kW for a
10/20ton combatsupport vehicle [4].
The addition of power will allow:
Increase of acceleration at all speed ranges;
Increase of the maximum speed in all kinds of terrain and
grades;
Stealth operations (limited silent mobility and silent
watch).
233
SECTION IV
Ammunition and Energetic Materials
CHAIRMAN
Slobodan Jaramaz, PhD
Nikola Gligorijevi, PhD
PHYSICOCHEMICAL PROPERTIES AND THERMAL STABILITY OF
MICROCRYSTALLINE NITROCELLULOSE ISOLATED FROM WOOD
FIBER
MOHAMMED AMIN DALI
Military Academy, University of Defence, Belgrade, PhD researcher, dalima380@gmail.com
Abstract: The nitrocellulose (NC), which was discovered in 1832, has a great importance in the military field. In fact, it is
considered as the quasiuniversal material for gunpowder, double base propellants and composite double base propellants.
The purpose of this study is divided in two parts. The first part is to improve the performances and properties of
microcrystalline nitrocellulose samples that are prepared in the laboratory and to compare them with conventional
nitrocellulose during the artificial ageing: using FTIR and XRD. In the second part, the aim is to improve the thermal
stability of nitrocellulose samples with two kinds of stabilizers (ethyl centralite and N(2acetoxyethyl)pnitroaniline which
was tested via DSC, using 3 % by weight of the stabilizer for each preparation. All samples are subjected to artificial
ageing. Then, the thermal degradation is evaluated using non isothermal kinetic models in order to investigate the thermal
degradation energy variation (activation energy, Ea). In addition, the present work has demonstrated that DSC
measurements allow to evaluate the actual state of the nitrocellulose and to prove that this analytical technique is capable to
distinguish the differences in thermal decomposition processes in nitrocellulose during ageing as well as to determine the
storage time.
Keywords: wood, cellulose, nitrocellulose, microcrystallinity, ageing.
1. INTRODUCTION
On the 19th century, the black powder was the only
explosive substance used. However, the invention of
nitrocellulose by HenriBraconnot in 1832 terminated the
use of black powder. This invention was the cause to start a
renovation in the field of energetic materials. This polymer
synthesized by nitration of the cellulose became popular not
only by its military applications as an ingredient in
formulations of powder, propellants and pyrotechnic
compositions, but also in the civilian sector like paintings,
photography and cosmetics. The importance and the
frequent use of nitrocellulose in the formulation of energy
products caused to start many studies on the phenomena of
degradation of the substance, which made the attribution of
causes it to nitrogen oxides freed from the nitrate esters
during storage at room temperature, that is why, Alfred
Nobel used the DPA as a stabilizer in the nitrocellulose.
In this study, the nitrocellulose sample was obtained from
laboratory by nitration of the cellulose which is itself
isolated from the wood fibers (the pinewood).
2. EXPERIMENTAL PART
2.1. Obtaining the native cellulose and micro
crystalline from wood
Picture 1. Protocol for obtaining the native cellulose and
microcrystalline from wood fibers
Obtaining cellulose fibers and microcrystalline cellulose
from wood requires a set of treatments and pretreatments as
it is given in Picture 1, [1, 2, 3]
237
PHYSICOCHEMICALPROPERTIESANDTHERMALSTABILITYOFMICROCRYSTALLINENITROCELLULOSEISOLATEDFROMWOODFIBER
OTEH2016
The FTIR spectrum interpretation of: softwood (BR), pin wood
after soxlhet extraction (BRES), hemicellulose (HCR), native
cellulose (CR) and commercial cellulose (CRO) (Picture 5),
shows the existence of bands of absorbance hemicellulose at
1732 cm1, the two peaks, appearing at 1610 cm1 and 1512 cm1
, correspond respectively to carbonyl groups (C=O) and to
stretching vibrations of (C=C) which are special groups of
lignins. The comparison of this spectrum with the spectrum of
the intermediate treatments shows the complete elimination of
hemicelluloses by alkaline treatment. This disposal is justified
by the disappearance of the peak at 1732 cm1. [4, 5]
Picture 2. Wood blocks after grinding
Picture 2. Pure native cellulose (a) from pine wood (b)
commercial cellulose
Picture 5. FTIR specters of commercial and synthesized
cellulose
The spectrum superposition of the commercial cellulose
with the synthesized cellulose enables the identification and
confirms that they have the same nature of the commercial
cellulose. [4, 5]
2.2. Obtaining the original nitrocellulose and
microcrystalline nitrocellulose
Nitration samples of native and microcrystalline commercial
and synthesized cellulose, are produced by a sulfonitric
mixed acid by a weight / volume ratio (w/v)=(1g/30ml)
[79]. After treatment of stabilization during 24 hours, the
nitrogen content of each sample is given in the table [8]:
Picture 3. Pure native microcrystalline cellulose (a) from
pine wood (b) commercial microcrystalline cellulose
In picture (3,4) it can be seen that there is no difference
between synthesized samples and commercial samples.
Table 1. Nitrogen content of nitrocellulose sample (NCR)
softwood nitrocellulose, (NCMCR) softwood microcrystalline nitrocellulose, (NCCOM) commercial nitrocellulose native made in china, (NCMCCOM) commercial
microcrystalline nitrocellulose.
Picture 4. IR specters of intermediate pretreatment of the
pin wood.
238
Sample
Nitrogen content (%)
NCR
12,24
NCMCR
12,84
NCCOM
13,42
NCMCCOM
12,98
PHYSICOCHEMICALPROPERTIESANDTHERMALSTABILITYOFMICROCRYSTALLINENITROCELLULOSEISOLATEDFROMWOODFIBER
OTEH2016
Picture 11. Oven used for the aging
Picture 8. Xray diffractograms of the
nitrocellulose samples
The Previous diffractograms show that the synthesized
nitrocellulose of both type of cellulose (native and
microcrystalline) has an identical structure to commercial
nitrocellulose. [10]
Samples are taken for three times (0 day 8 days and 15
days) to be analyzed by FTIR, XRD and DSC.
3. RESULTS AND DISCUSSION
3.1. Aging effect on chemical functions
2.3. Kinetics Study of the degradation synthesized
and stabilized nitrocellulose
The evolution of the chemical function during accelerated
aging is followed by FTIR.
In order to illustrate the role of stabilizers, (NCStab) thin
films have incurred accelerated aging at a temperature of
80 C for a period of 15 days.
The preparation of (NCStab) thin films is performed using
the following proportions of the four varieties nitrocellulose
(NCCOM, NCMCCOM, NCR, NCMCR), with 3% of
stabilizer (ethyl centralite EC and N(2acetoxyethyl)pnitroaniline ANA). Each mixture is solubilized in in
acetone. [12]
Picture 12. NC COM infrared specters during aging.
Picture 9. Chemical structure using the stabilizers
Picture 10. Thin films of the mixtures NC+stab,
(a) NC+ANA, (b).NC+EC.
All samples are cut and placed in tubes in order to realize an
accelerated aging in a thermostatically bath, controlled for
15 days at 80C. [12]
Picture 13. NCR infrared specters during aging
239
PHYSICOCHEMICALPROPERTIESANDTHERMALSTABILITYOFMICROCRYSTALLINENITROCELLULOSEISOLATEDFROMWOODFIBER
OTEH2016
Picture 14. NCMC COM infrared specters during aging
Picture 17. NCMCCOM XRD diffractograms
during aging
Picture 15. NCMCR infrared specters during aging.
According to infrared specters during aging, we can see an
attenuation of the intensities of the (ONO2) function for all
nitrocelluloses samples, and the appearance of (C=O)
functions at 1714 cm1, it is logical saw thermal degradation
of nitrocellulose, according to the reaction [6]:
Picture 18. NCR XRD diffractograms during aging
> CHONO 2 > C = O + HNO 2
3.2. Aging effect on the crystallinity:
For understanding the effect of aging on crystallinity
nitrocellulose samples, they have experienced by XRD
analysis during aging every 5 days.
Picture 19. NCMCR XRD diffractograms during aging
The diffractograms of the nitrocellulose samples (NCR,
NCMCCOM, and NCMCR) indicate a disappearance of the
principal peak of the crystallinity. However, in the
commercial nitrocellulose (NCCOM), the crystallinity is
partially preserved at the end of aging. This disappearance
of the peaks of synthesized samples has the same source. It
is faster in the case of conventional (NCR) compared to
microcrystalline (NCMCCOM and NCMCR). This
difference appears more clearly in the diffractograms after
Picture 16. NCCOM XRD Diffractograms during aging
240
PHYSICOCHEMICALPROPERTIESANDTHERMALSTABILITYOFMICROCRYSTALLINENITROCELLULOSEISOLATEDFROMWOODFIBER
10 days. The evolution of the crystallinity of synthesized
nitrocellulose samples differs from the commercial one due
to processing and stabilization conditions. [10]
According to Table 4, the difference between the native and
microcrystalline nitrocellulose, stabilized and not stabilized,
can be noticed clearly. After 15 days of accelerated aging,
stabilized nitrocellulose samples are more stable than nonstabilized sample saw degradation temperatures, the
decrease in enthalpies of degradation of the nonstabilized
samples is more important than in the stabilized samples,
which confirms the importance of using a stabilizer during
the storage period. it can be seen also that the sample of
NCMCCOM has a decrease in temperature and enthalpie of
degradation, less than the sample of NCR nostabilized after
15 days of aging (see Table 3). This indicates that the
microcrystalline nitrocelluloses are more stable than
conventional nitrocelluloses [10, 11].
3.3. Aging effect on the enthalpy of degradation
and degradation temperature:
Using the DSC method, variations in enthalpy of
degradation and degradation temperature are followed. The
obtained results are summarized in Tables 2 and 3.
Table 2. Temperatures and degradation enthalpies of NC
unaged samples (=5C/min)
Sample
NCR
NCMCR
NCCOM
NCMCCOM
Enthalpy of
degradation (J/g)
1184,281
1296,352
1563,603
1296,508
Degradation
temperature (C)
203,90
203,64
203,19
203,49
3.5. Determination of kinetic parameters and
storage time
The determination of activation energies and preexponential factors for aged samples are accomplished by
Kissinger's (1) and Ozawa's (2) method (see Table 5). [11]
Table 3. Temperatures and degradation enthalpies of NC
samples after 15 days of aging (=5C/min)
Sample
NCR
NCMCR
NCCOM
NCMCCOM
Enthalpy of
degradation (J/g)
236,856
907,447
938,162
848,254
OTEH2016
Degradation
temperature (C)
183,96
180,7
189,77
182,24
Ea
ln 2 = ln A R
E
R
Tm
a
Tm
(1)
Ea
R Tm
(2)
ln ( ) = 2,315 9, 4567
The values of activation energies (Ea), pre exponential
factors (Log (A)), the rate constants (Log(K)) and critical
temperatures of explosion (Tb) for nitrocellulose samples,
stabilized and nonstabilized, are obtained by the linear
leastsquares and iterative methods using MATLAB and
mentioned in the Table 5. [1012, 14]
From the Table (1.2), it can be clearly seen that both
thermodynamic parameters (enthalpy of degradation and
degradation temperature) decrease during aging. This is
justified by accelerating the degradation phenomenon due to
the nitrogen oxides accumulation which makes the
nitrocellulose unstable (low temperature of degradation) and
less energetic because of the disappearance of the nitro
groups (NO2). The decrease in enthalpies degradation is
important in the case of conventional nitrocellulose relative
to those microcrystalline, it means that for a long period, the
crystallinity and the reduction of polymerization degree,
affect the stability and the energy properties of
nitrocellulose. [11]
Table 5. Kinetic parameters of nitrocellulose samples with
and without stabilizer after 15 days aging via Kissinger's
method
Ea
(KJ/mol)
NCCOM
104,16
NCCOM+EC
127,97
NCCOM+ANA 135,17
NCR
89,55
NCMCCOM
95,55
NCMCR
106,04
NCMCR+ANA 135,17
NCMCR+EC
138,54
Sample
3.4. Stabilizer effect on aging
Via the DSC method, the effect of stabilizer is studied on
nitrocellulose aging. The variation in enthalpy of
degradation and degradation temperature was followed for
NC samples with and without stabilizer. The obtained
results are summarized in Tables 4.
Log(A)
log K
Tb (K)
9,05
12,13
12,39
7,98
8,74
10,07
13,39
13,62
8,74
10,30
10,75
7,71
8,00
8,56
10,29
10,65
477,16
483,04
481,87
481,27
477,67
466,52
470,54
472,52
Table 6. Kinetic parameters of nitrocellulose samples with
and without stabilizer after 15 days aging via Ozawa's
method
Table 4. Temperatures and enthalpies of degradation of
NC+stab samples after 15 days of aging (=5C/min)
Ea
Log(A)
(KJ/mol)
NCCOM
106,55
//
NCCOM+EC
129,24
//
NCCOM+ANA 136,35
//
NCR
92,56
//
NCMCCOM
98,23
//
NCMCR
108,13
//
NCMCR+ANA 135,88
//
NCMCR+EC
139,15
//
Enthalpy of
Degradation
degradation (J/g) temperature (C)
NCMCR+EC
1102,096
187.09
NCMCR+ANA
1223,096
183.07
NCCOM+EC
1012,305
195,23
Enthalpy of
Degradation
Sample
degradation (J/g) temperature (C)
NCCOM+ANA
999,327
195,02
NCMCCOM
848,254
182,24
Sample
Sample
241
log