Fluent Spring 2004

VOL XIII ISSUE 1 SPRING 2004
APPLIED
COMPUTATIONAL
FLUID
DYNAMICS
FLUENT Makes
Splash Olympics
a
Automotive
In-Cylinder Power
at the
Environmental
Nuclear Power Plant Thermal Discharge
HVAC
The Heat is On But Not at Fluents Data Center
Materials
No Sweat: Modeling Clothing and Fabrics with FLUENT
Healthcare Industry Supplement Inside!
Editors Note
On the Cover: Contours of shear stress (frictional drag) on the body of an elite swimmer; the CFD results contributed to the design of competitive swimwear
Courtesy of Speedo International Ltd.
On the Supplement Cover: Wall pressure in the vicinity of an abdominal aortic aneurysm fitted with a stent-graft
Courtesy of the Institute of Diagnostic Radiology, University Hospital of Zrich
Fluent News is published by
10 Cavendish Court Lebanon, NH 03766 USA 1-800-445-4454 2004 Fluent Inc. All rights reserved. Editor: Liz Marshall Contributing Writer: Keith Hanna Contributing Editor: Erik Fergusen Project Manager: Kat Barrell Design: Lufkin Graphic Designs FLUENT, FiDAP, GAMBIT, POLYFLOW, G/Turbo, MixSim, FlowLab, Icepak, Airpak, and FloWizard are trademarks of Fluent Inc. Icepak and Airpak are joint developments of Fluent Inc. and ICEM-CFD Engineering. All other products or name brands are trademarks of their respective holders.
hen we leave our day jobs and venture out into the world, familiar fluid flow patterns are recognizable everywhere. In gazing at them, many of us often wonder what it would be like to model something entirely different from the systems we focus on at work. For those of us with an interest in sports, the opportunities to do this are endless. For Barry Bixler at Honeywell, these mental meanderings became a reality when he first simulated the flow around a swimmers arm (Fluent News, Spring 2000). That project led to a much larger role in recent years that focused on the development of new competitive swimsuits for Speedo International, a leading performance swimwear brand. Our cover story describes how he used CFD in this novel application to help design a new swimsuit that will hopefully lead to gold winning performances in Athens. Other uses of CFD for sporting activities are also featured, ranging from the design of handlebars for racing bicycles to wheelchairs for paraylmpians. For those who enjoy sporting activities in the mountains, trips to remote huts are challenging and memorable. At one such hut in Sweden, excessive snowdrifts have become an annual problem, in part because they can compromise fire escape routes. On page 14, an article describes a clever use of FLUENTs multiphase model to simulate the snowdrifts around this hut. The results will be used to help guide the design of future mountain huts in the region. Other simulations of both outdoor and indoor airflow are presented as well, including one that is particularly close to home. As described in an article on page 10, cooling problems in Fluents Lebanon, NH data center were completely eliminated thanks to simulations using Airpak. Cooling was the focus of many projects reported on in this issue, ranging from notebook battery packs (p. 21) to Ducati heat exchangers (p. 28). Heating, however, was of interest in an
unusual project carried out at the University of Exeter, reported on page 36. Archeological digs in Sri Lanka unearthed dozens of ancient blast furnaces built near the crests of hills, and researchers questioned whether or not the furnaces could reach temperatures hot enough to smelt iron. CFD simulations of a typical furnace demonstrated that the prevailing up-hill breezes stoked the fires to temperatures that could easily perform this process. Using CFD to understand archeological findings is contrasted in this issue with examples of work being carried out at the technological front lines, as reported in the Healthcare Supplement. The healthcare market has seen rapid growth during the past twenty years, and CFD is playing an ever-expanding role in all of the participating industries, as evidenced in recent issues of Fluent News and in the current supplement. Topics such as blood flow in patientspecific vessels, novel drug delivery systems, and advances in modeling liquid reactions for pharmaceutical applications are just a few of the topics explored. During the past year, Fluent engineers have been busy performing an extensive matrix of tests for the upcoming release of FLUENT 6.2. On page 34, a detailed account of one of these tests, a turbomachinery validation, is described. More details on the new features in the release appear in the FLUENT product report on page 42. Other product reports are also presented. The Support Corner on page 44 provides an overview of Fluents User Services Center, a wealth of CFD information that is widely used by analysts around the world. Whether through the USC or through personal contact with Fluent staff and distributors, please keep in touch with your questions, comments, and success stories. s
Liz Marshall fluentnews@fluent.com
feature stories
sports
Speedo Goes for Gold with CFD Alive & Kicking Underwater On the Pulse of the Olympics From Bronze to Gold
Contents
23
departments
36
academic news
CFD in the 1st Millennium AD Southampton Partnership Optimizes Collaboration Cavitation and Hydraulic Flip CFD in the Politecnico di Milano Classroom King Fahd University of Petroleum & Minerals Adopts Fluent Software
applications
10 14
hvac
The Heat is On But Not at Fluents Data Center Shooting the Breeze in Pittsburgh Feeling the Breeze on Arabian Nights
40 44 46
product news
CFD Without Borders Using FloWizard Taking Meshing to a New Level Whats New in FLUENT 6.2
environmental
Snow Drift Modeling Reaches New Heights Nuclear Power Plant Thermal Discharge Dispersing Urban Pollutants in Budapest NanoMist Suppresses Fire
support corner
The Fluent User Services Center Maximize the Return on Your Fluent Software Investment
19 20 16 21 22 23
paper
From Pulp Fiction to Pulp Fact in Paper Recycling
materials
No Sweat: Modeling Clothing and Fabrics with FLUENT
around Fluent
Fluent India Moves to Pune Infotech Park Upcoming UGM Dates
electronics
Notebook Computer Batteries Meet the Thermal Grade
healthcare industry supplement
semiconductors
Preventing Wafer Contamination
S2 S3
overview
A Clear Vision for the Healthcare Industry
automotive
In-Cylinder Power The Shape of Things to Come with Automotive Gas & Water Tanks Windshield Wipers at High Speed Cool Ducati Heat Exchangers
blood flow
CFD Helps Interventional Radiologists Heart Valve Dynamics During a Cardiac Cycle Hemodynamics in a Carotid Bifurcation FSI Through Partnerships
S2
29 30
Marine
Modeling Marine Propellers
S6 S8 S9 S10
drug delivery
Drug Delivery from Coronary Stents Needle-Free Drug Injections
chemical
Realistic Distillation Modeling Size Distribution Modeling in FLUENT Predicting Residence Time Distribution with FLUENT Taking the Pressure Off in Outer Space
pharmaceuticals
Improving Reaction Selectivity by Clever Mixing
medical equipment
Virtual Prototyping System for Blood Pump Design
33 19 34
power generation
ATEX for Explosive Atmospheres
respiration
FSI Produces a Breath of Fresh Air Breathing Easy with the Dynamic Mesh Model
turbomachinery
Mapping the Eckardt Centrifugal Compressor
sports
Oilflow lines over the body colored by dynamic pressure
Fluent NEWS spring 2004
Speedo Goes for Gold with CFD
sports
n 1896, when the French Baron, Pierre De Coubertin lobbied to resurrect the ancient Olympic Games as a modern amateur athletic event of sporting excellence, little did he know that one day it would become the biggest event in the world, with an estimated viewing audience of over half the Earths population. For two weeks in Athens this August, the Games will come to Greece once again for the XXVIII Summer Olympiad, and attention will turn to the sportsfest where the elite athletes of the worlds nations compete for personal and national glory. For most events today, the difference between winning and not winning a medal is usually a fraction of a percentage point in overall performance. Given such a landscape, any legal and costeffective competitive advantage will be sought out by national teams, individual athletes, and equipment manufacturers to gain that extra edge. Two sports that tend to dominate the viewing coverage throughout the Summer Games are swimming and track and field, and it is in these sports that many of the closest performances occur. Speedo International Ltd, the worlds leading swimwear brand, identified the potential for bringing CFD technology to the swimming pool after reading an article in Fluent News (Spring 2000) about Barry Bixler, a principal aerospace CFD engineer working for Honeywell in Arizona. At the time, Barry had spent much of his spare time working in his garage on a CFD analysis of the hydrodynamics of swimming strokes for USA Swimming. Speedos Research & Development team in Nottingham, England contacted Bixler and Fluent Europe Ltd in Sheffield to see if it would be possible to harness CFD technology to enhance their world-leading Speedo Fastskin suit, which made headlines in 2000. At the time, it was the worlds fastest swimsuit, in part because it mimicked the skin of a shark. Swimmers from over 130 countries wore the swimsuit in Sydney, where it revolutionized the swimming world by contributing to over 80% of all the swimming medals, and swimmers wearing the suit broke 13 out of 15 world records. Speedo saw the potential of CFD-based fluid flow analysis to understand swimming hydrodynamics, even though CFD had never been used in swimwear development before. The wide use of FLUENT in other sports such as Formula One motor racing and Americas Cup Yachting convinced Speedo that it had the proper qualifications for their sporting CFD challenge. In March of this year, Speedo launched their new fast suit, Fastskin FSII, at several high profile media events in Sydney, London, and New York. The development of the suit, which they claim to be the fastest in the world, has involved a collaborative team of experts. CyberFX, the Hollywood-based body scanning and special effects company behind the recent Spiderman, Matrix, and Charlies Angels movies, was involved in the swimsuits development process from the beginning. Its unique body scanning technology allowed Speedo to create anatomically correct virtual and actual models of the average male and female swimmer to provide the perfect base upon which to test the suit. With the help of CFD modeling work
Pulsed water flow pathlines over the body colored by local static pressure
Visualization by EnSight
The CFD modeling we have done has

been both challenging and extensive. Using FLUENT, we have been able to mathematically model a water flume on our computers. We analyzed the flow and obtained incredibly consistent and accurate results. The CFD analyses, combined with flume testing, enabled Speedo to evaluate a huge number of swimsuit design variables, and eventually, we believe, obtain the worlds fastest swimsuit.
Barry Bixler CFD Consultant to Speedo
sports
Speedo Fastskin FSII swimsuit in action
CFD simulation of water flow over typical female (top) and male (bottom) elite swimmers in the glide position showing contours of shear stress (frictional drag) on the swimsuit region
Speedos customized swimsuit colored by CFD-predicted frictional drag is modeled at the New York launch
by Bixler and the University of Otago Water Flume in New Zealand (the lowest turbulence intensity flume in the world, capable of producing very strong currents), Speedo R&D engineers were able to show that the Fastskin FSII swimsuit they have developed increases underwater speed by reducing passive drag by up to 4% compared to the next best suit. Based on these findings, Speedo expects the swimsuit to help elite swimmers (who all have access to Fastskin FSII under FINA rules) to break even more world records in Athens. Once the swimsuits were validated with the water flume trials, the CFD software was incorporated into Speedos design process to evaluate the drag and fluid flow characteristics around the male and female swimmers for various flow scenarios. Bixlers CFD work allowed Speedo to simulate the flow around the virtual swimmer bodies, thus making Fastskin FSII as hydrodynamic as possible. During the swimsuit development process, the CFD models helped to eliminate variability in the results so that Speedo could conduct a greater number of tests and achieve more accurate results than ever before. Moreover, CFDs ability to visualize the details of the water flow around the athletes revealed information that had not been known previously. The results were used to guide the position of the seams, gripper panels on the underside of the forearms, and vortex riblets on the chest, shoulders, and back of the suit. In addition, the CFD predictions of water flow over male and female bodies helped to guide the positioning of different patches of fabric and fabric coatings available to Speedo to optimize the suits created for the different swimming events, such as breaststroke and backstroke. The actual scanned models that had been made into precision mannequins were used in over 1,000 flume tests carried out in Otago. The results supported the CFD findings that the new Fastskin FSII full bodyskin is the fastest and most advanced performance swimwear in the world. Engineers at Speedo believe that they have now achieved a transfer of motor racing technical design methodology to the hydrodynamics of swimming and swimsuits. This is probably a first for a sportswear brand, and a milestone of applied engineering to the complexities of ergonomic sports systems. In the wake of Speedos new CFD design process, this integrated technical design approach will almost certainly pave the way for many other sports in the future, especially those dominated by equipment aerodynamics and hydrodynamics. Ultimately, however, the success of any swimsuit is in its acceptance by the worlds leading swimmers and coaches. Speedo has not been neglecting the interface with elite athletes, and has been working with a group of swimmers from around the world to develop Fastskin FSII. They have selected eleven elite swimmers, known as the Speedo FSII Team, who are at the forefront of Speedos challenge in Athens to evaluate the suit in pool performance tests. The group includes the latest phenomenon from the US, 18 yearold Michael Phelps. Phelps, who broke three world records and won three gold and two silver medals at the 2003 World Championships in Barcelona, will be wearing a Fastskin FSII in Athens in a bid to match Mark Spitzs famous record of seven gold medals at the 1972 Summer Games. If successful, Speedo will pay Phelps a cool one million dollar bonus. Performance tests by elite swimmers on the new suit are taking place this spring, and will provide the ultimate feedback in terms of performance increases to be gained from this high-tech innovative swimming product. Certainly, the temperature in the Athens swimming pool is heating up already thanks to this exciting and pioneering use of CFD by Barry Bixler and the Speedo R&D team. s
sports
Alive & Kicking
Underwater
By Andrew Lyttle and Matt Keys, Western Australian Institute of Sport, Mt. Claremont, Australia
Digitizing the underwater video footage (left) and deriving the kinematic variables (right)
he underwater phases of modern swimming form a large component of the total event time, and can often lead to the difference between earning a medal and finishing with the rest of the pack. Optimal underwater starting and turning techniques require maximizing the distance achieved by minimizing the deceleration rate caused by hydrodynamic drag forces. Currently in elite competition there is a wide range of underwater technique strategies used by swimmers with very little scientific rationale applied in their selection. Previous empirical testing conducted by Lyttle et al.1 examined the net force produced during underwater kicking due to the propulsive force and active drag. Results were compared to prone streamlined gliding in order to prescribe an approximate velocity at which to initiate underwater kicking. The study assumed the steady state (constant velocity) condition at which the testing was conducted, so it limited the applicability to real swimming where the body is continually accelerated and decelerated. A study now underway at the Western Australian Institute of Sport (WAIS) seeks to discriminate between the active drag and propul-
sion generated in underwater kicking with the goal of prescribing an optimal kick profile in swim starts and turns. The use of FLUENT CFD will allow variables such as the amplitude and frequency of the kicking movement, as well as the effects at different velocities and depths, to be examined from a fluid dynamics perspective. The objective information gained can be used in technique analysis. An elite swimmer from the WAIS with a proficient underwater kicking ability has been filmed underwater, performing both maximal high amplitude, low frequency dolphin kicks, and low amplitude, high frequency dolphin kicks. A CFD model is being developed that is based on an accurate 3D mapping of the swimmer using a Cyberware WBX whole body laser scanner. The 3D mapping will be combined with 2D joint kinematic data recorded from digitized underwater video footage of dolphin kicks from the same swimmer. Once the detailed model is imported into FLUENT, it will be used to model the fluid flow around the upper body as well as the animated lower limb in order to differentiate flow lines and calculate propulsive and resistive coefficients. The dynamic mesh model in FLUENT will be used to simulate the movement
of the lower limbs. The base of the model will be set up by defining the individual components as solid bodies and creating a user-defined function (UDF) to match the output from the digitized kinematics data. During the next phase of the study, alterations to the inputs and model constraints will be investigated to examine the effects of variations in underwater kicking form and technique. The objective information gained from this type of analysis will equip the sports scientists with the tools to more accurately provide advice on technique modifications in order to gain the extra edge at the elite level. s
The authors would like to acknowledge the financial support of the Western Australian Institute of Sport and Australian National Elite Sports Council Sports Science/Medicine subcommittee for funding this project.
reference:
1 A.D. Lyttle, B.A.B. Blanksby, B.C. Elliott, and D.G. Lloyd, Net forces during tethered simulation of underwater streamlined gliding and kicking techniques of the freestyle turn, Journal of Sports Science, 18, p.801-807, 2000.
sports
On the
By Keith Hanna, Fluent News
Pulse of the Olympics

he sports and leisure industry today is a global $500 billion business. The industry has seen an increased benefit from CFD modeling over the years, with applications that range from Formula 1 racing cars to golf club design, from ski jumping postures to Olympic swimsuit materials. Sporting events are increasingly competitive, both on and off the field. To ensure that UK athletes have the best possible edge, a 2M EU Objective 1 Grant was obtained in October 2003 to launch SportsPulse in Sheffield, England. Sheffield is in the South Yorkshire region of the UK, and has a unique cluster of world-class sports research groups and leading sports industry companies. By drawing together this geographical cluster of sports science, sports medicine, sports engineering, and sports industry economics through a hothouse environment, SportsPulse is aiming to incubate, energize, fund, develop, and commercialize new and innovative ideas for athletes and sports of all kinds. SportsPulse is directed by Dr. Steve Haake, head of the Sports Engineering Research Group (SERG) at the University of Sheffield, and pioneering founder of the International Sports Engineering Association. The SportsPulse project harnesses the synergies of a cluster of eight partner organizations in the region, including Fluent Europe Ltd. in Sheffield. A state-ofthe-art 3D scanner and FLUENT software are components of a leading sports CFD center
SportsPulse 3D Scanner being used on Olympic cyclist Bradley Wiggins by Olympic triple jump world record holder, Jonathan Edwards
The front fork design (left) currently being optimized for British Cycling for use with a disc front wheel; a handlebar analysis (above) has also been conducted by SportsPulse for British Cycling
under development. SportsPulse hosts sports forums three times each year in South Yorkshire to promote cross-fertilization of ideas and technology innovations. In December 2003, the organization was represented at an Engineering in the Olympics seminar in London, hosted by the UK Institute of Mechanical Engineers. Special guest attendee at the event was Jonathan Edwards, a physicist by training and the current Olympic triple jump world champion and world record holder. He was intrigued by the combination of modern scanning technology and CFD software to help improve the performance of the British Olympic Cycling Team, represented at the event by Bradley Wiggins, current world individual pursuit champion. Dr. Haakes team is also working with the British Olympic Sailing, Cycling and Paralympic teams for the Athens Olympiad. The team, with its resident CFD expert, Dr. John Hart, is involved in a wide range of CFD applications in support of several British Olympic Performance Directors and is aiming to have a medal-winning impact in these non-traditional CFD application areas in Athens. Fundamental PhD fluid flow research studies relating to sports using CFD are also planned for the years ahead. s
more.info@
www.sportspulse.org or contact: info@sportspulse.org
Contours of surface pressure and water pathlines for a Yngling Olympic class hull and keel sailing at a characteristic incidence angle
Water pathlines colored by velocity magnitude for a Yngling Olympic class hull and keel, sailing at high incidence angle (tacking)
sports
From Bronze
to
By Steve Haake, University of Sheffield, Sheffield, UK; and Leigh Bramall, Fluent Europe Ltd.
Gold
Surface pressure and pathlines for the unoptimized chair geometry
Visualization done using EnSight from CEI
Surface pressure on the athlete before (left) and after (right) optimization
Recirculation zone behind the athlete after optimization of the wheelchair; the athlete is in an un-optimized position
ue to its successful use by the Americas Cup and Formula 1 series teams around the world, CFD has become widely recognized as an effective tool in yacht and automotive racing design. One outgrowth of this success has been the crossover of CFD technology to other areas of the sporting world. While swimming, cycling, and skiing have already benefited from CFD analysis, the sport of wheelchair racing in Great Britain has recently begun to make use of fluid flow modeling to help its athletes in their quest for Paralympic glory in the 2004 Athens games and beyond. Toward this goal, in August 2001 UK Athletics commissioned a feasibility study from the Sports Engineering Research Group (SERG), a leading center for research in sports technology based at the University of Sheffield, and the Center for Sports & Exercise Science at Sheffield Hallam University. Together, the two institutions have applied a range of sports science and technology resources, including CFD, to the project. The CFD focus of the project has been to provide visual and numerical data on the aerodynamic performance of Dave Holding, a British athlete competing in the 100-meter and 200meter wheelchair sprints, and whose physical shape was considered the least aerodynamic among the five athletes in the overall study. Assessing the performance of Holding and his chair required mechanical studies to be initiated, in which a series of resistant forces were analyzed to measure their impact on his performance. The wheel alignment of the chair was tested to identify its effect on rolling resistance, and the effect of the chairs mass was also thoroughly investigated. Eventually, a final optimized design was achieved. The mechanical modifications were further supplemented by conceptual ideas for specialized sportswear, such as an aerodynamically optimized helmet, to ensure minimum resistance for Holding and his chair in racing position.
To compare the aerodynamics of each chair design, two geometries were created one representing Holding in the chair prior to any design changes, and the other incorporating the suggested modifications. Solid CAD models of the wheelchair were built using Pro/ENGINEER, while two human forms representing Holding were constructed using SAMMIE CAD Ergonomics Design software. The geometries were imported into GAMBIT, and meshes were generated for each model of Holding in position in the chair. The geometries were complex, requiring tetrahedral meshes of several hundred thousand cells to ensure the accuracy of the results. The meshes were read directly into FLUENT, where inlet boundary conditions were used to simulate wheelchair speeds of approximately 7 m/s. The initial results showed the existence of a recirculation zone behind the wheelchair and a large wake effect, caused by separation over the shoulders of the athlete. Design modifications to the clothing and wheelchair structure, however, were found to minimize the impact of these phenomena, making a 3% reduction in drag possible equivalent to a reduction of 0.03 seconds over 100 meters. Other tests showed that these drag reductions, combined with modifications that resulted in a total mass reduction of 2.2% and improvements to the alignment of the rear wheels, could deliver a further improvement of 0.16 seconds over 100 meters. For the 100-meter race at the Sydney Paralympics in 2000, a time reduction of this magnitude would have propelled the thirdplace competitor into first place! These findings reinforced the potential significance of the research. The success of the project has prompted further study across an expanded range of sports with a view toward boosting the success of UK athletes in forthcoming Olympic and Paralympic games. s
hvac
The Heat is On But Not at Fluents

Data Center
By Walter R. Schwarz, HVAC&R Industry Manager, Fluent Inc.
10
hvac
he thermal cooling of high-density computer data centers is a challenging yet mission-critical task in many companies today. The complex nature of airflow and the frequency at which equipment is added, changed, or relocated only add to the challenge. Fluent Inc. is not immune to these types of problems since its continued growth in business has been accompanied by a continued growth in computing resources. Some Fluent office locations around the world have data centers that are cooled with underfloor air supply, and have equipment arranged in hot-aisle/cold-aisle configurations. At the corporate headquarters in Lebanon, NH, however, the central data center is cooled using overhead air supply. As is true for many such facilities, it was probably designed for occupant comfort rather than for the cooling needs of a data processing environment. Until recently, not much attention was given to the thermal cooling of the room, since most of the computing resources were deployed in individual offices. With the changing trend toward clustering computer resources in a central location, however, cooling has become an important concern. During the past year, the Director of Computer Services and Building Manager were experiencing difficulties in maintaining optimal cooling conditions in the data center. According to ASHRAEs Thermal Guidelines for Data Processing Environments (2004), the recommended equipment environment temperatures are 20C to 25C (68F to 77F). Compounding the situation in Lebanon were plans to bring more computer horsepower and its associated thermal load into the existing data center. Initial thoughts were to spend considerable money to bolster the cooling capacity with more HVAC capital equipment. Before proceeding with such a large investment, both managers approached the HVAC industry team to see
if Fluent software could be used to analyze the problem. Airpak airflow modeling software was used to quickly create a virtual model of the data center, which is approximately 67 m2 (725 ft2) in area with a power density of about 394 W/m2 (37 W/ft2). The actual power of the equipment was directly measured and used in the simulation, instead of the equipment nameplate ratings, which can lead to over design of the HVAC system. The first analysis was of the data center with the existing equipment and air distribution system that made use of four-way ceiling diffusers. The simulation results reflected much of what the managers already knew that there was a wide range of rack inlet temperatures, and several locations where the equipment was subjected to higher than desired temperatures. The managers were impressed by the close agreement between the equipment temperature readouts and the temperatures predicted by Airpak. Moreover, the simulation revealed some short-circuiting of supply air into the returns before it had a chance to cool the equipment. Learning from the initial analysis, the equipment was repositioned in the room to develop a hot aisle/cold aisle arrangement, and the four-way diffusers were replaced with one-way diffusers that discharged cool supply air vertically down. Simulation of this improved design showed far better cooling of the equipment. Other analyses were performed to assess the vulnerability of critical servers in the room to hypothetical failure scenarios, where a refrigerant compressor might fail in one of the HVAC systems. Future simulations will be done to help design the cooling system to accommodate an expected doubling of computer resources in the room. The goal is to make sure that the data center is adequately protected and capital investments in HVAC system upgrades are wisely planned. s
It was remarkable to identify local cold spots predicted by the simulation and to then go to the data center and verify the results with your hands. The simulation was able to precisely pinpoint the temperature variations in the room.
Alister Cate Building Manager, Fluent Inc.
I knew I had a problem. The HVAC system in my data center was not keeping up with the existing thermal load and I was about to add a lot more computing power. I was almost convinced by several contractors to spend $40,000 on more AC, when I turned to Fluents own HVAC team for recommendations. Following their advice, we changed and relocated the diffusers, and rearranged the existing equipment. Afterwards, I was able to add all the planned new hardware, without any additional AC. And the temperatures in the room have never been cooler or more uniform. I am very impressed!
Earl Zmijewski Director of Computer Services, Fluent Inc.
Pathlines colored by temperature in the data center at Fluent Inc. illustrate the airflow distribution from the new vertical diffusers
Plan view of temperature contours two feet above the floor
11
hvac
Shooting the Breeze

in
Pittsburgh
A schematic of the proposed addition
By Erik Olsen, Andrew Gilles, and Jessica Wojtas, OWP/P, Chicago, IL
Temperature contours on selected planes in the proposed addition show the range of temperatures that can be expected with natural ventilation
Velocity vectors show the airflow through the building, driven by natural convection
multidisciplinary team of architects and engineers applied Airpak airflow modeling software to develop their awardwinning entry in the first annual U.S. Green Building Council (USGBC) Design Competition. The team from Chicagobased OWP/P was the third-place winner in this competition for emerging green builders. Entrants designed a theoretical addition to the existing facility for The Pittsburgh Project, an urban, neighborhoodbased, Christian community development organization serving the north side of Pittsburgh. Submissions were required to meet the stringent requirements needed to earn a Platinum rating under the USGBCs LEED (Leadership in Energy and Environmental Design) green building rating system, while also meeting the aggressive target budget of $100 per square foot. To enhance occupant comfort, minimize the buildings energy use, and keep the project cost low, the team designed a naturally ventilated building with no mechanical ventilation or cooling. The design concept relied on buoyancy-driven flow developing within a solar-heated cavity created between the brick exterior of the existing building and a new, three-story plane of glass located at one end of the new building. This solar tower pulls air through the operable windows of each individual room, down the corridors, and out of an exhaust stack at the top of the building. The thermal mass of the brick wall maintains warm temperatures in the solar tower through the night, driving cool night air through the building to remove heat built up during the day. The team used Airpak to verify and develop the design concept. The Airpak model was developed from very early hand sketches and computer-based architectural models in order to confirm that sufficient ventilation would be provided by this design. Early models showed that openings between the three levels outside of the solar tower allowed too much air to rise from floor to floor, bypassing the tower and creating uneven temperatures and air quality in the building. These openings were closed off from airflow with suspended pieces of glass, allowing light transmission and visual communication between levels. The final model shows large amounts of air flowing in through all windows and out through the solar tower, in keeping with the design intent. This type of early design stage modeling exercise allows concepts that might be considered unrealizable to be explored and tested. Unviable concepts can be rejected, while promising concepts can be refined. s
more.info@
A section view of the proposed addition, showing the original building at left, the glass tower in the middle, and the proposed addition at right
www.usgbc.org
12
hvac
Feeling the Breeze on Arabian Nights

By A.K. Ahmed, Fluid Codes Ltd. (Fluent Distributor for the Middle East), Greenford, Middlesex, UK
eople living in countries with a hot climate and high humidity during the summer have, for centuries, adopted natural ventilation systems for their traditional houses. One such approach, used along both sides of the Gulf coast, is the wind tower house. These houses make use of one or more towers to cool the internal premises. To better understand the cooling process, the Al-Bastakia tower house in the old section of Dubai was studied. The house contains a rectangular bedroom (3 x 10 x 4 m3) and a square cross-section wind tower (3 x 3 x 15 m3), located at the east side of the room. One of the tower walls extends 1m down from the ceiling to create a partition between the tower and bedroom. The room has three horizontal windows halfway up the west side wall of the room. The cooling performance of a tower house is affected by two main factors: the tower design, and the strength and direction of the wind. The tower design details, such as the height, cross sectional area, location, and orientation have a significant effect on its performance. The direction of the wind outside the house governs the general flow pattern inside the bedroom. The main stream blowing in through the tower, which is higher in both flow rate and temperature than that blowing in through the window, mixes with the weaker flow inside the room to govern the cooling process. Full-scale 3D models of the tower and its attached bedroom were simulated using FLUENT. Four scenarios of wind blowing from the north, south, east, and west were solved to determine the steady-state internal flow for each. Wind speeds of 10 m/s and a temperature of 311K were used at the tower inlet on the windward side, with pressure outlets on the three remaining sides. Speeds of 7 m/s with a temperature of 301K were assigned to the open windows. The doors to the room were assumed closed. For the west wind scenario, two main flow circulation loops were observed. A primary circulation loop was established between the west and east channels of the tower. A secondary loop was established inside the bedroom. Beneath the wall partition, a one-dimensional downward flow was generated. This air curtain played a significant role in the entrainment and mixing of the two circulating streams. The air curtain helped force cold air from the windows downward, strengthening the circulation pattern in the room. The height of the wall partition was found to play a significant role on the strength of the air curtain, so modifications to the partition could provide an easy and low cost solution for improving the cooling performance. While the west wind was found to represent the best scenario for cooling, the south wind represented the worst. No air curtain developed, so a single circulation loop became established, with the coolest air pooling above the windows. The CFD results have shown that the placement of the windows and tower can be chosen so that the cooling of these houses best utilizes the renewable wind energy in an efficient manner. The tower house concept has been used for many types of buildings, including markets, mosques, and traditional cafes. Hopefully, modern houses and buildings in the Middle East will continue to adopt this green architecture approach that has played such an important role in its heritage. s
A typical wind tower house in Dubai

Photograph 2004, www.datadubai.com
Pathlines colored by temperature illustrate recirculation zones in the tower and room when the wind blows from the west
Iso-surfaces of temperature show that for a westerly wind, cooler air collects below the windows and circulates in the room
13
environmental
Snow Drift Modeling

Reaches New
By Per Olowson, F-Energi&Milj AB, Gteborg, Sweden
Courtesy of the Swedish Touring Club (STF)
Heights
14
environmental
The CFD model of the mountain station illustrates the surrounding hilly terrain in the north of Sweden
rifting snow can cause significant problems in open terrain. Snowdrifts usually develop behind the crests of mountains or hills, in surface depressions, or around obstacles such as buildings, fences, or trees. In exposed, windy locations, the amount of snow that accumulates due to snowdrifts can be hundreds of times greater than the snow that falls directly on the ground. Strong winds and low visibility due to drifting snow are often reasons for closing mountain roads. Around buildings, snowdrifts can cause serious problems with accessibility and the blockage of emergency exits. Snowdrifts also increase the wear and tear on buildings, as well as the maintenance costs. The snow deposited around a building in a windy place is usually more a result of the wind transporting and redistributing the snowflakes than of free falling snow. The wind patterns around a building therefore play a strong role in determining where and how much snow will accumulate. The Swedish Touring Club (STF) has a number of mountain stations that are open for the public throughout the year. One of the stations, Sylarnas mountain station, is situated in the north of Sweden, 1000 meters above sea level, close to the Syl massif. Since the station is positioned in open terrain in a valley, it is exposed to hard winds and snowdrifts. In the early 1980s, a new station was built after a fire destroyed the old station. When designing the new building, wind loads and snowdrift problems were taken into consideration and wind tunnel tests were made of the construction. Despite the tests performed, however, the problems with snowdrifts around the building have been so severe that STF now is planning to re-build the station. At the time when the new station was built, only experimental methods were available to assess
the snowdrift potential. Today, increased computing power and advanced numerical methods have made it possible to use CFD as a powerful yet cost-effective tool to solve the complex twophase problem that snowdrift represents. STF therefore asked F to evaluate different retrofit proposals with the aid of CFD. Based on drawings, topological data, and wind statistics, a computational model was created. In order to validate the methods used, the existing building was simulated first and comparisons were made with observations of the snowdrift around the building. Agreement with these observations was good, so simulations of new building designs were made. The air flow and transport of snow particles were simulated using the mixture model in FLUENT. By applying an apparent viscosity to the snow phase, the particles were slowed down in regions with high snow concentration and/or low shear stress. When the snow phase is slowed, the density of the blowing snow increases, and this in turn slows the wind speed. Since accumulated snow is a result of a history of varying wind directions and speeds, a number of representative wind conditions were included in transient simulations performed for each geometry. The calculations showed that it is better to make the building higher than lower. It is also important to avoid steps in the construction, such as exterior stairwells. The roof angle was also found to have a significant influence on the amount of drifted snow. With the aid of the CFD study, valuable information regarding the snowdrift at the mountain station was achieved. Moreover, the increased knowledge about the nature of snowdrifts will very likely lead to major cost reductions for STF in the future. s
Strong winds cause recirculating flow and subsequent snow drifts near the house
A surface of constant snow volume fraction is used to illustrate the drifts that form around the building
Early in the transient simulation, snow drifts begin to form in expected locations in the vicinity of the house
15
environmental
Nuclear Power Plant

T
Thermal Discharge
By Fangbiao Lin and George E. Hecker, Alden Research Laboratory Inc., Holden, MA; and Brennan T. Smith and Paul N. Hopping, Tennessee Valley Authority, Knoxville, TN
Study area of thermal plume from Browns Ferry Nuclear Power Plant
Map and computational domain of the overall river model
Temperature contours on the water surface
0 depth (ft) 5 10 15 20 70 75 80 temperature (F) 85 90 measured FLUENT prediction
he TVA Browns Ferry Nuclear Power Plant (BFNPP) withdraws its condenser cooling water from Wheeler Reservoir using a shoreline intake. Waste heat from the plant increases the cooling water temperature before it is returned to the reservoir through three multiport diffusers containing a total of about 22,500 downstream-facing ports. The ports are 2-inch diameter holes situated 6 inches apart both vertically and horizontally. The numerous small diffuser ports, the buoyancy of the heated effluent, and the large scale of the discharge area present major challenges to developing a reliable three-dimensional CFD model for predicting the temperature distribution and flow patterns in Wheeler Reservoir. When different treatment alternatives for plant waste heat are considered, the change each creates in the temperature distribution in the reservoir is crucial information for assessing the environmental impact. FLUENT was recently used to develop an innovative CFD model of the Wheeler Reservoir area. To simulate the discharge from thousands of diffuser ports, a two-zone modeling approach was used that consists of a multiple jet sectional model and an overall river model. The river model contained about 2 million computational cells for simulating the cooling water discharged from the thousands of diffuser ports. The multiple jet sectional model, which simulated the flow from the individual discharge ports over a one foot slice of the diffuser pipe in great detail, provided the information for developing sectional boundary conditions for the diffuser effluent in the river model. The realizable k- turbulence model was used and a second-order accurate solution was obtained. CFD was used to predict the surface water temperature distribution as well as the stratification resulting from temperature differences. Validation of the CFD models was performed using data from field measurements, hydraulic model tests, and other experiments. For the sectional diffuser model, the results of a vertical section of diffuser ports were compared with data from a hydraulic model test of a small segment of the BFNPP diffuser1. These comparisons verified that the sectional model simulations can predict the thermal stratification, expansion, dilution, and near-field behavior of the multiport jets. For the full-scale river model, comparisons between the temperatures predicted by FLUENT and measured along the centerlines of the operating diffusers and at downstream monitoring stations were found to be in close agreement. Indeed, the validations showed that the overall river model, based on the two-zone approach, can reproduce the major features of temperature and flow in the diffuser mixing zone, including the mixing and temperature rise patterns of thermal plumes. s
reference:
1. D.R.F. Harleman, L.C. Hall, and T.G. Curtis, Thermal Diffusion of Condenser Water in A River During Steady and Unsteady Flows with Application to the T.V.A. Browns Ferry Nuclear Power Plant, Hydrodynamics Laboratory Report No. 111, Massachusetts Institute of Technology, Cambridge, MA, September 1968.
Temperature comparison at one of the downstream monitoring stations
16
environmental
Dispersing
Urban Pollutants
in
ir quality in the urban environment has become one of the most important issues of environmental protection. The most significant urban pollutants, such as the emissions from vehicles, are influenced by architectural structures, city planning, and traffic control measures. To examine the environmental impact of new construction and its associated traffic patterns, atmospheric pollutant transport needs to be predicted, since field measurements cannot be made. Wind tunnel investigations and numerical simulation are the tools that can be applied during the planning and licensing phases of the project. A new Millennium City Center is being planned in the southern part of Budapest, close to the bank of the Danube River, between two bridges. The City Center will include a row of relatively large buildings: a conference center, museums, a concert hall, hotels, and residences. Since these buildings will be built adjacent to a busy roadway (carrying 60,000 vehicles/day) between the Danube and a neighboring district of Budapest, the possibility of an adverse effect from the new buildings and increased traffic on the air quality in the neighboring district has been considered. To respond to the concerns, the Department of Fluid Mechanics at the Budapest University of Technology and Economics was commissioned to perform experimental and numerical analyses of the dispersion process in order to predict the effect of the planned City Center. A model of the site was constructed for analysis in the test section of a wind tunnel and FLUENT simulations of the area were performed for a variety of wind conditions. A detailed numerical model of the existing buildings in the area, including the new construction, was prepared
Budapest
By Tams Rgert, Tams Lajos, kos Cscs, Istvn Goricsn, Mrton Balcz, Department of Fluid Mechanics, Budapest University of Technology and Economics, Budapest, Hungary
for the FLUENT simulations. The geometry is complicated. Several of the buildings are characterized by critical forms and inter-building connections. Because of the special geometrical constraints, an unstructured mesh of 978,000 tetrahedral elements was created. The wind tunnel and numerical models were both made in 1:500 scale. The computational domain included an empty space around the outermost buildings corresponding to one kilometer at full scale to avoid the direct forcing effects of the boundary conditions. The height of the computational domain was 0.6m, corresponding to 300m at full scale. The realizable k- model was used for turbulence closure. The CFD models illustrated many well-known effects, such as street channeling and the formation of vortices behind buildings. The concentration of a tracer gas at 24 sampling points, 1.5m above the street level, was measured and calculated for five dominant wind directions, with and without the City Center buildings in place. Considering the probabilities of wind velocities and directions, a dimensionless annual mean concentration was calculated from the CFD results for every sampling point and compared with values calculated similarly from the results of wind tunnel measurements. While the trends are the same, CFD predictions at only about two-thirds of the sampling points were in good agreement with the measured values. The discrepancy in the low concentration regions was suspected to be the result of measuring errors. Most importantly, both the numerical simulation and the experiment showed that because of the lifting effect of the new buildings, a slight decrease in pollution in the neighboring district would occur. s
The model of buildings prepared for FLUENT
Concentration distribution and the effect on the City Center in a SW-SSW wind direction
100 normalized mean concentration wind tunnel FLUENT prediction 10
0.1
R1
R3
R5
R7
R9 R11 R13 R15 U1 measurement points
U3
U5
U7
Comparison of FLUENT and wind tunnel data: annual mean concentrations with the City Center in place
17
environmental
NanoMist
By K.C. Adiga, NanoMist Systems LLC, Warner Robins, GA
Suppresses Fire
fire, and would vaporize before contact. This notion is changing, however, as a patented ultra-fine water mist technology with the trade name NanoMist has recently attracted a great deal of attention. Using droplets less than 10 m in diameter, the technology has demonstrated efficient fire suppression capability in various configurations. With the help of CFD and laboratory tests, scientists from NanoMist Systems have demonstrated that this ultra-fine mist may be a potential alternative to HFC-227ea for meeting both government and industrial fire protection needs, pending full-scale testing. The CFD results have yielded an improved understanding of how to generate, scale, and deliver fine mist clouds into a fire location, while addressing concerns about the premature loss of liquid droplets. The new technology has some very attractive features, including 1) near selfentrainment of gas-like mist into the firebase; 2) considerably less water required compared to conventional water sprays; and 3) the relatively non-wetting nature of the mist on nearby surfaces because of fast vaporization. The amount of water required to put out a fire is far less compared to larger droplets because of the increased total surface area of the smaller droplets, the rapid increase in heat removal from the combustion zone, and the dilution of oxygen due to expansion as vaporization occurs. While the process technology required to produce such a fine mist and deliver it to the fire area on a commercial scale has been difficult to develop, FLUENT has helped make this aspect of the development and evaluation quick and affordable. In the CFD study, a medium scale fire was generated using a volumetric heat generation source term. The discrete-phase model (DPM) was used to simulate the mist. Stochastic particle trajectories, influenced by turbulent fluctuations, were computed. The coupled solution predicted the rate of vaporization of water droplets subjected to the fire, and the subsequent cooling of the local gas field. The predicted centerline peak temperatures were used to judge the fire cooling capacity of the mist. CFD results vividly show the entrainment of water mist droplets into the firebase as seen from DPM droplet trajectories. The firebase pulls the mist from its surroundings. The behavior predicted by the CFD simulations was reproduced closely by tests of mist deployed to the base of heptane pool fires. These fires were put out in 10 seconds or less. NanoMist Systems is now collaborating with the U.S. Naval Research Laboratory (NRL) and Hughes Associates, Inc. (HAI) in evaluating NanoMist fire suppression technology for electronics applications. Preliminary results are very promising. s
Using the DPM, the entrainment of surrounding water mist into the firebase is shown by stochastic droplet trajectories
n the 1960s, Halon 1301 was hailed as an effective fire suppression chemical. Shortly thereafter, it was identified as an ozone-depleting environmental hazard, and a search for newer and safer fire suppression agents was renewed. The chemical HFC-227ea, a safer alternative to Halon 1301, is the most popular agent in use today. It has some limitations, however, because of by-products formed during its use. Water in some form would be an ideal permanent solution. Throughout the years, the concept of using an ultra-fine water mist (with droplet diameters < 10 m) for fire suppression has been generally ignored, based on the notion that such extremely small droplets would not have enough momentum to reach the
NanoMist local flooding experiments show the entrainment of surrounding mist into the firebase, extinguishing the heptane pool fire
18
paper
Pulp Fiction to Pulp Fact in Paper Recycling

From
By John Egan, Kadant Black Clawson, Mason, OH
n the paper industry, pulpers are often used in the stock prep, recycle, and paper machine processes. Hydrapulpers, which hydrate and defiber (break up into individual fibers) dry paper stock, are used to re-pulp bales of paper, re-slurry waste from the paper producing process, and recycle old corrugated cardboard and paper discarded at the mill. For all of these applications, a rotor is used to circulate the pulper slurry and provide the shear force required to fully defiber it. The rotor typically interacts with a perforated bedplate, and together the two determine the level of pulper circulation and shear force imparted to the material for a given pulper tub design. Because the fluid flow in a pulper is difficult to visualize, engineers do not fully understand how the pulpers mechanical components contribute to the process of fully defibering a slurry. Most of todays rotors were designed decades ago, but because they offer stable performance and reliability in a critical productiondriven application, these rotors remain in service with little demand placed on suppliers to improve them. At Kadant Black Clawson, CFD has played an integral role in recent efforts to improve the efficiency of their Hydrapulper rotor design. Engineers began by analyzing existing designs to understand their flow characteristics. They then evaluated a number of alternative designs (using a proprietary simplified rotor performance program) in hopes of identifying improvements that would result in more efficient operation. Several promising new first cut designs were developed, and detailed solid geometries of them were created and meshed using I-DEAS from SDRC. Models with 1.5 to 2 million cells were imported into FIDAP, where solutions were carried out in the fixed frame of reference. Flow in the rotating frame
of reference was viewed in Fieldview (a product of Intelligent Light). The ability to view the results in both the rotating and stationary frames provided unique insights that led to a number of useful rotor performance improvements. A CFD analysis was performed for each of the first-cut rotor designs. Using a stereolithography rapid prototyping process, a series of modular rotors was built and tested in order to evaluate the accuracy of the CFD predictions for these designs. Accurate calculation of power was critical to the success of the modeling efforts. In particular, the net power (equal to the total pulper power minus the mechanical power loss, rotor underside power, and stock shredding/defibering power) was computed for each case. Predictions for net power were found to be within 3% of measured values. The next step was building a fullsize version of the best design and testing it in a production 20-foot diameter, Kadant Black Clawson 10D Hydrapulper. A new 82-inch diameter rotor was built using the design developed with the aid of CFD to replace a conventional rotor in one of two identical pulpers running in parallel. Trials were run in batch mode, and performance (under stable operating conditions) was compared for the new and conventional rotors. Scaled up model results predicted a net power draw that was within 3% of the actual power consumption. The actual power savings based on a series of continuous trials, comparing the performance of both the new and conventional rotors installed in parallel pulpers, was found to be 25%. Just as important, the client has reported that defibering and circulation performance using the new rotor is at least equal to that of the conventional rotor. s
Computer model of an 8-ft pulper with a new low-power rotor
Transformation of velocity vectors to the rotating frame of reference (using Fieldview) was critical to refining the rotor design
An 81.6-inch-diameter Vortech rotor installed in a 10D Hydrapulper
19
materials
No Sweat:
By Jim Barry and Roger Hill, Creare Inc., Hanover, NH
Modeling Clothing and Fabrics with FLUENT

dvances in textile technology are expanding the performance boundaries for clothing and technical fabrics. Outdoor and athletic clothing provide weather and thermal protection and manage moisture. Protective clothing controls the exposure of laboratory and hazardous materials workers, heathcare providers, fire fighters, and military personnel to chemicals, biological materials, and heat sources. Depending on the specific application, characteristics such as impermeability to hazardous materials, breathability, rapid moisture transport, insulation value, weight, cost, and ruggedness must be balanced to deliver comfort and performance. Recent innovations now enable CFD to play a significant role in the development of technical fabrics and clothing systems. In clothing development, a macroscopic CFD approach has been used that looks at the interactions of the fabric with the body and its surrounding environment. From this macroscopic perspective, each fabric layer is treated as a porous medium with spatial- and time-varying properties. Multiple fabric layers and air gaps between layers can be addressed. Microscopic analyses of fabrics where individual fibers are modeled form a separate area for CFD applications. FLUENTs porous media model provides a starting point for
the fabric model, and substantial new capabilities for transport processes in the fabrics can be incorporated via user-defined functions (UDFs). Important transport processes in textile fabrics include the convection and diffusion of vapor and liquid, as well as the adsorption and desorption of fluids from solid fibers. Using UDFs, engineers at Creare have implemented models for variable permeability (dependent on local moisture content in the fibers and liquid blockage), condensation/evaporation in the fabric, vapor/liquid sorption to fibers, and capillary transport of free liquid. A special sorption model has also been developed for fabrics incorporating activated carbon. Using the extensibility built into FLUENTs graphical user interface, the setup of the properties and modeling options for the fabric zones has been automated. In addition, visualization capabilities have been customized using FLUENTs standard tools. Applications of the textile fabric models include predicting the penetration of chemical agents through protective materials. Protective fabrics often are tested as swatches in small, enclosed cells. The CFD models of the swatch tests not only provide detailed insights into the transport processes within the fabric, but enable a better understanding of the effects of convective flow above the swatch on test results. To predict the
Circular pattern of chemical agent droplets on a fabric swatch in a test cell
performance of protective fabrics on humans, simplified 2D models of clothed body sections and more complex 3D models have been developed. Starting from laser body scan data, a combination of computer-aided design software and GAMBIT has been used to create a meshed model of a human with one or more layers of clothing. For both protective and outdoor clothing, thermal comfort is a critical factor. A sweating skin model has also been implemented in FLUENT using UDFs. Using 3D models of clothed humans, the effects of varying fabric properties, weather conditions (including temperature, wind, and humidity), metabolic output, and clothing ventilation features have been assessed. s
Humidity levels under ventilated single-layer clothing; the clothing fit is not symmetrical on the body
Clothed human model shows contours of sweat evaporation, assisted by a slight headwind
20
electronics
Notebook Computer Batteries Meet the

Thermal Grade
By Hossein Maleki and John Johnson, Motorola Energy Systems Group, Lawrenceville, GA
or the last few years, the notebook computer business has grown at an annual rate of about 25%, and the same growth rate is predicted to continue for the next few years. Such growth is attributed to the development of small, yet sophisticated notebook computers that have met both the users ordinary and complex application requirements. Todays notebook computers may include 1.2 - 2.0 GHz CPUs (e.g. Pentium-IV), CD/DVD players, high resolution LCDs, ample disk space (100s of MB), wireless connectivity, and remote information sensing capabilities. Unfortunately, these components are less than 100% efficient, so considerable heat is generated during operation. In addition, rechargeable batteries self-heat while charging and discharging because of their inherent electrical, thermodynamic, and electrochemical impedances. Furthermore, Li-ion cells should be operated and/or maintained at temperatures lower than their chemical stability limit. Extended exposure of Li-ion cells to temperatures above 60C will degrade their performance. A combination of notebook computer features compact size, high functionality, fast operating speeds, long run time, and high-resolution visuals has created new challenges in notebook battery thermal management. Cell performance differs from manufacturer to manufacturer, heat generation varies with operating conditions, and products must be cost effective. These requirements have added a great deal of complexity to the thermal management issue. For practical reasons, thermal simulation plays a critical role in the design of highly reliable notebook computer batteries. Prediction of the thermal responses of a battery pack based on its real-time application requirements can save engineering resources and product design costs
while ensuring quality and reliability. Motorola Energy Systems Group uses Icepak software to complete a full, problem-to-solution analysis for the development of batteries with optimum cooling capability. As an example, a simulation of a notebook computer battery was conducted to determine the real-time thermal performance of the battery while charging at 2.2A, and discharging at a constant power of 55W, both at 42C ambient. Results showed that the battery had a reasonable level of heat dissipation capability since the temperatures of its cells rose above 60C for only a short time. However, the temperature rise of the charge control integrated circuit (IC) increased to near its critical limit of 73C. Autopsy results showed that this particular IC had a small die-to-pad size ratio and a large volume of molding compound, both of which caused increased heat accumulation during the charge control process. Overall results indicated that the battery temperature rise during charge is dominated by the power dissipation from the control electronics, and that the temperature rise during discharge is dominated by heat dissipation from the cells. The results from this modeling exercise provided an understanding of the thermal response of the battery that helped to reduce the product design cycle-time. Note that, during discharge, the Icepak results show cell temperatures that are higher than the experimental values. This is due to the fact that CFD captured the temperature over the entire body of the cells, whereas the measured value is the cell skin temperature, which is generally lower than the temperature at the center of cell, especially during discharge. CFD can therefore provide important information that could have been very difficult to obtain experimentally. s
Thermal profile of the electronics and cells inside of the notebook battery discharged at 55W constant power at 42C ambient
3.2 2.8 heat generation (W) 2.4 2.0 1.6 1.2 0.8 0.4 0 500
discharge power 10.0 W
1000
1500 time (sec)
2000
2500
3000
Heat generation profiles of 2.2-A cells from four different manufacturers at 40C ambient; the variation in heat released increases as the cells are discharged at higher power
Temperature (C) after Charge Component Capacitor Cells Charge Control IC Charge FET Discharge FET Exp. 63.6 51.9 73.1 66.9 63.4 CFD 61 45-51 67 69 62
Temperature (C) after Discharge Exp. 68.8 57.5 68.0 68.4 71.8 CFD 66 62-70 67 68 67
Temperatures of control electronics and cells of the battery at the end of charging and discharging
21
semiconductors
Preventing
Wafer Contamination
ing requires a factor of three more time than ideal purging to reach the same condition. While classical and ideal purging begin with comparable discharge rates for air, the classical method drops off considerably after the initial 25% of air is removed from the device. The FLUENT results also showed that with a fully loaded FOUP, air flow between the wafers was significantly less than that along the wafer edges. This was attributed to the fact that the classical inlet position directs nitrogen upwards toward the base of the bottom wafer, where it is diverted to flow around the entire stack. Furthermore, the narrow spacing between wafers discourages circulation, and the slightly denser humid air causes the nitrogen to rise above, rather than mix with the air. To improve the purging performance, a number of modifications were tested. A plenum purging system was introduced that injects purge gas from the sides of the FOUP. Different purge gases were tested, such as argon and dry air, and the temperature of the purging gas was altered to promote
By Roland Bernard and Hisanori Kambara, Alcatel Vacuum Technology, Annecy, France; and Arnaud Favre, INOPRO, Grenoble, France
he front-opening unified pod (FOUP) is a 300mm wafer handling and contamination control device, used in the semiconductor manufacturing industry. The pods are used to transport wafers between processes, and care must be taken when using the pods to avoid wafer contamination. For example, the presence of humidity in the enclosed environment of the pods can cause a variety of phenomena, such as native oxide growth, corrosion, and film cracking. The presence of emitted organic compounds can lead to degradation of the electrical properties of integrated circuits fabricated from the wafer material. Simple intuitive purging with an inert gas such as nitrogen, a technique used to keep the pods free of contamination, does not always maintain wafer cleanliness compatible with high yields. At Alcatel, rigorous numerical simulation with CFD has been used to show how pod purging can be improved. The potential inefficiency of purging a FOUP with an inlet and outlet located on the bottom of the pod the classical purge configuration has been examined using FLUENT. The FOUP was initially filled with ambient air with 40% relative humidity, a typical value for a wafer fabrication cleanroom. A transient analysis was done as a steady flow of nitrogen was introduced through the inlet. The calculation was continued until the air mass fraction reached 1% (with 0.4% relative humidity). To test the effect of the nitrogen gas flow rate, a range of inlet velocities was considered. For each case studied, the time required to reach the above conditions was recorded and compared to a computed value for the ideal purge time, defined as the ratio of the FOUP volume to the volumetric flow rate. For the base case, using a 2L/min gas flow rate, results showed that classical purg-
The original humid air mass fraction after about 5 minutes of purging has taken place
natural convection between the wafers. These and other modifications have reduced the purge time to a point that is closer to the ideal case. s
This work was supported by the European MEDEA+ T301 project and by the MINEFI (French Ministry of Finances & Industry). The authors also wish to thank ST Microelectronics and Alcatel Vacuum Technology for providing the experimental data and organization.
100 mass fraction of air left in FOUP (%)
10
ideal purge 1
Alcatel purge
plenum inlet purge 600 time (sec) 900
classical purge
300
1200
1500
Comparison of different FOUP nitrogen purge times
Mass fraction of the residual gas using plenum injector purging after a) 1.5 minutes, b) 5 minutes, and c) 22 minutes
22
FOCUS on CFD
In The Healthcare Industry
Newsletter Supplement
S2 S3
overview
A Clear Vision for the Healthcare Industry
blood flow
CFD Helps Interventional Radiologists Heart Valve Dynamics During a Cardiac Cycle Hemodynamics in a Carotid Bifurcation FSI Through Partnerships
S6 S8 S9 S10
drug delivery
Drug Delivery from Coronary Stents Needle-Free Drug Injections
pharmaceuticals
Improving Reaction Selectivity by Clever Mixing
medical equipment
Virtual Prototyping System for Blood Pump Design
respiration
FSI Produces a Breath of Fresh Air Breathing Easy with the Dynamic Mesh Model
overview
healthcare industry
A Clear Vision
for the Healthcare
By Marc Horner and Ahmad Haidari, Fluent Inc.
Industry
he healthcare industry is facing a dynamic future with many challenges stemming from an expanding elderly population in the developed world, higher levels of care, and a constant push for new, more effective pharmaceuticals and biopharmaceuticals. From the discovery, manufacturing, and delivery of new drugs to biomedical device design and disease research, CFD is a key tool that will be at the forefront of this growing industry. Fluent has been providing biomedical modeling solutions for over a decade. It has a dedicated and knowledgeable staff, and has academic and industrial partners to meet the needs of many far-reaching applications. For the pharmaceutical and medical products industries, the many capabilities in Fluent software offer solutions for both drug manufacturing and drug delivery. FLUENT is used in applications such as spray dryers, scale-up, mixing, and chemical reaction. Devices such as inhalers, blood pumps, artificial heart valves, stents, and catheters have been successfully modeled with FLUENT and FIDAP to better understand their performance and interaction with the human body. FIDAPs fluid-structure interaction (FSI) model has been widely used to capture the deformation of physiologic structures as a result of air, liquid, or blood flow. POLYFLOW is ideal for extrusion, blow molding, and thermoforming, processes that often involve viscoelastic materials, and which are used for the manufacture of pharmaceutical packaging and medical devices. Bioengineering has experienced rapid growth in recent years, thanks to the ability to create 3D meshes from MRI and CT scan data of blood vessels, the digestive tract, or air passageways. Simulations of these systems are geared towards understanding the underlying causes of disease or analyzing device performance. The import of medical scan data into GAMBIT has been assisted by recently developed filters, and through partnerships with industry leaders who specialize in this capability. Microfluidic applications are at the forefront of many industries, including healthcare. In the biomedical area, so-called lab-on-achip applications include chemical/biological agent detection, DNA sequencing/analysis, and drug discovery. The various analytical tasks are performed in a series of microchannels whose characteristic width and height are between 10 and 1000 m. Two critical applications for this type of analysis are the separation (isolation) and detection of a target species. Both processes rely on differences in the mobility, or migration speed of the species when subjected to an applied electric field. The electrohydrodynamics (EHD) modeling capability in FIDAP has been successfully used to predict the separation and pre-concentration of chemical species in this manner. The collection of stories in this supplement is representative of the wide range of applications of interest in the healthcare industry today. Articles that focus on blood flow, breathing, drug delivery, medical equipment, and pharmaceuticals highlight the efforts of some of our customers. Indeed, the work of these and other analysts, who have adopted CFD for biomedical applications, has led to the enormous growth of flow modeling in the healthcare industry during the past few years. CFD capabilities for biomedical applications will continue to expand, and will hopefully help us all lead longer and healthier lives! s
The geometry of the eye (left), including the sclera (black), iris (blue), cornea (gray), lens (pink), and choroid and ciliary body (red). A continuous, room temperature air stream cools the cornea surface for 30 seconds. After the air stream is stopped, the cornea warms up again. Natural convection currents develop in the anterior chamber, and these are illustrated by contours of velocity magnitude (right). By studying the rate at which the eye warms up, researchers can determine the rate of blood flow to the eye through the choroid and ciliary body.
Courtesy of Dr. S. Orgl and Dr. T. Picornell, Augenklinik Basel, Switzerland
Velocity magnitude displayed on a number of slice planes through the nasal cavity; the nostrils are at the lower left
high mobility ions
low mobility ions Separation of two species in a solution exposed to an applied electric field, resulting from the different mobilities of each; the width of the microchannel containing them is expanded by a factor of 10 for visualization purposes
S2
blood flow
CFD Helps
Interventional
Radiologists
By Dr. Thomas Frauenfelder, Institute of Diagnostic Radiology, University Hospital of Zrich, Zrich, Switzerland
n abdominal aortic aneurysm (AAA) is a pathologic dilatation of the abdominal aorta reaching a diameter of more than 3cm. Two different types of AAA are known: the suprarenal and the infrarenal, situated above and below the renal artery, respectively. Nine percent of all people above the age of 65 suffer from an AAA. Clinical studies reveal a very high risk for rupture of the more frequent infrarenal AAAs when they are larger than 5cm. In clinical practice, a threshold of 5cm for the maximum transverse diameter of an AAA is typically used to recommend interventional treatment. Recent studies have found, however, that calculations of flow velocity and pressure using CFD seem to be more conclusive for assessing the risk of rupture than the 5cm rule. Today, endoluminal stent-graft implantations, which use fabric-covered stents, have gained increased acceptance as a less invasive alternative to the surgical treatment of AAAs, with fewer complications. This method requires a very accurate imaging standard before the procedure, since the stentgraft dimensions must be chosen prior to implantation. Most imaging techniques result in only static anatomical models, however. Thus, while they allow for optimal and precise stent planning, they do not take into consideration the effects of the temporally-varying flow and pressure that are always present in the vessel. To better address these concerns, FIDAP has been used to investigate the pressure and flow patterns in patientspecific models of AAAs before and after stent-graft implantation. Three major steps are involved in the process: (1) data acquisition by CT angiography; (2) 3D solid model generation; and (3) numerical simulation. Once the CT data becomes available, Amira, visualization and volume modeling software from TGS, is used to construct the 3D geometry and delineate structures such as the vessel lumen, vessel
wall, thrombus, calcifications, and post-operatively, the stent. A triangular surface mesh is generated and imported into GAMBIT, where a volumetric mesh is built and the properties of objects and boundaries are identified. A turbulent fluid-structure interaction (FSI) simulation is then run in FIDAP, where the blood is treated as a non-Newtonian fluid. The transient velocity profile at the inlet boundary is taken from Doppler-ultrasound measurements. Other values, like density, Youngs modulus, and Poisson coefficient, are based on material-specific data that have been adapted during test simulations. Many important observations have been made from the FIDAP simulations to date. Prior to implantation, the flow patterns show major vortices in the concavities of the aneurysm, which, in the context of blood rheology, can explain the formation of intraluminal thrombi. After stent-graft implantation, the flow velocity increases and the vortices disappear inside the stent. New eddies appear at the stent-end, however, which can lead to the need for a second intervention in some cases. The wall pressure is very high before implantation at locations where the blood hits the vessel. This is a possible explanation for the growth of the aneurysm and the formation of aortal kinking. After implantation, the high pressure areas are mostly found adjacent to the stent bifurcation. This is probably due to the funnel-like stent geometry. When the pressure drop is found to be asymmetric inside the stent limbs, asymmetric wall shear stresses can develop, which can lead to stent migration or rupture. After implantation, the increase in blood volume can be assessed in the leg that, prior to intervention, had a weaker blood flow. The reason for this increased flow is due to the more symmetric geometry of the stent. In the future, CFD, in combination with virtual stent placement, can help choose the optimal stent. s
Pathlines (above) and wall pressure (right) suggest problematic conditions for an abdominal aortic aneurysm
After deploying a stent-graft, pathlines through the artery show improved flow
S3
healthcare industry
blood flow
healthcare industry
Heart Valve Dynamics

During a Cardiac
1.0 inlet velocity (m/s) 0.5 0.0 -.05 4.0 pressure drop (mm Hg) 2.0 0.0 -2.0 90 leaflet angle (degrees) 60 30 0 Inlet velocity, pressure drop, and leaflet angle during the cardiac cycle (left); the resulting flow through the valve is shown below at three times:
Cycle
he interaction between blood and the structures that transport it plays an important role in many biofluid dynamic flow problems, such as blood flow in the heart and vessels, through heart valves, in cardiac assist devices, and in artificial organ design. Different numerical techniques have been used to tackle this fluid-structure interaction (FSI) problem. At Ghent University, a new fluid-structure interaction model has been developed that is based on the dynamic mesh model in FLUENT. The model has been used to simulate a prosthetic aortic valve, and the results have been helpful for visualizing the blood flow in the region of the valve throughout a cardiac cycle. The FSI technique was implemented in FLUENT using journal files and user-defined functions (UDFs). An external FSI code (written in C++) drives the transient calculation. This code runs a subiteration loop for every timestep in order to solve the fluid-structure interaction problem. During every subiteration a journal file with FLUENT commands is produced by the FSI code and executed by FLUENT. In this journal file an estimation of the position of the valve is calculated in a Define on Demand UDF. This valve position is used in the Define Grid Motion UDF for the dynamic mesh model, and the corresponding flow field is computed. Once the solution converges, the forces on the leaflet are calculated using another Define on Demand UDF. When the first subiteration finishes, the external FSI code checks the convergence criteria for the fluid-structure interaction problem. If certain criteria are not met, a new subitertion is started. The valve position is adjusted using a stabilizing subiteration scheme that uses a numerical derivative of the moment on the leaflet1, and a new flow solution is computed, starting from the results at the previous time. This process continues until the prescribed conditions are satisfied, at which point the time is incremented by the FSI code, and a new valve position for a new time is computed. Using between three and four subiteration loops, the FSI problem typically converges to within 3 or 4 orders of magnitude before the next time step is started. A fully implicit coupling procedure is therefore achieved by using a separate solver for the fluid problem (FLUENT) and for the structural problem (the FSI code). Initial tests of the model were performed on a 2D case that illustrated the dramatic change in the flow in the aortic sinus during a cardiac cycle (http://navier.ugent.be/~kris). More recently, a 3D simulation involving 500,000 cells has been performed to track the blood flow and motion of one leaflet of a bi-leaflet heart valve for a complete cardiac cycle. Clinical and in vitro studies show different dynamic behavior under different physiological conditions, such as aortic versus mitral position, and the expected flow patterns and leaflet movement can be predicted by the FSI results. The FSI model could therefore be an important means of better understanding the phenomena that drive the coupled behavior of blood flow and artificial heart valve leaflets. This new FSI algorithm has promise as a major engineering tool for unraveling the hemodynamics associated with thrombolitic and hemolytic events of existing and new mechanical heart valves. s
By Kris Dumont and Pascal Verdonck, Institute of Biomedical Technology (IbiTech); and Jan Vierendeels, Dept. of Flow, Heat, and Combustion Mechanics; Ghent University, Ghent, Belgium
1 time (sec)
0.1 sec
0.6 sec
reference:
1.025 sec
1 J. Vierendeels, K. Dumont, and P. Verdonck, 33rd AIAA Fluid Dynamics Conference and Exhibit AIAA20033720, pp.23-26, June 2003.
S4
blood flow
Hemodynamics
in a
Carotid Bifurcation
mation on the flow and shear stress fields near the apex (flow divider), a representative bifurcation geometry was adapted for this study to extract the common features of hemodynamics and wall motion, and to illuminate the time- and spatially-varying characteristics of wall shear stress and circumferential strain. For the outlet flow boundary conditions of the ICA and ECA, waveforms obtained from phase-contrast magnetic resonance velocity measurements were used to construct the time variation of fully developed parabolic velocity profiles. An isobaric timevarying pressure was applied at the inlet of the CCA. A modified waveform of the CCA wall distention was applied for the inlet CCA pressure transient. Results illustrate that the maximum displacement occurs at the side wall of the bifurcation region. The wall stretching that occurs suggests that the outer side of the ICA wall may experience more cyclic strain than other locations. The wall shear stress distribution at the time of peak pressure shows that the carotid sinus experiences very low wall shear stress, whereas much higher wall shear stress occurs in the ECA. s
The authors are grateful to Dr. Peter J. Yim from the Radiology Department of the University of Medicine and Dentistry of New Jersey for providing the carotid velocity measurements, and to Ms. Midori Ohki, Fluent Asia Pacific Co., Ltd., for her support.
By Tada Shigeru, Department of Mechanical Engineering and Science, Tokyo Institute of Technology, Tokyo, Japan; and John M. Tarbell, Department of Biomedical Engineering, The City College of New York / CUNY, New York, NY
he common carotid artery (CCA) divides in the neck into the external and internal carotid arteries. The external carotid (ECA) supplies the tissues of the face and neck with oxygenated blood from the heart. The internal carotid (ICA) supplies the anterior portion of the brain. The carotid artery bifurcation is one of the common sites where atherosclerotic plaque and stenosis are prone to develop. Immediately above the bifurcation, there is a sinus, or bulb, in the internal carotid, which plays a role in the progression of atherosclerosis, depending on a complex spatiotemporal pattern of mechanical forces. Although recent advances in computational modeling techniques have allowed the incorporation of realistic vascular geometries, providing information that is otherwise difficult to get, most of these studies have focused on hemodynamics alone. At the Tokyo Institute of Technology, a fluid-wall coupled modeling technique incorporated in the latest FIDAP solver has been used to elucidate the relationship between wall shear stress and mechanical stress in the arterial wall. These fluid-structure interaction (FSI) simulations have demonstrated how local mechanical factors involved in the fluid-wall coupling participate in the pathogenesis of atherosclerosis. Transient simulations that track the blood flow through a complete cardiac cycle have been performed. Although a cast model from a specific individual may provide detailed infor-
Contours (grayscale) of the displacement magnitudes of the thick-walled carotid bifurcation, along with line contours of the velocity (color) at the time of peak pressure in the cardiac cycle; the ECA branch is below and the ICA branch, containing the carotid sinus, is above
Contour plot of the wall shear stress at the time of peak pressure in the cardiac cycle showing a significant difference in the level of wall shear stress between the ICA (upper) and ECA (lower) branches
FSI Through Partnerships
luid-structure interaction (FSI) is the focus of a new technology initiative involving Fluent and ABAQUS, Inc. Announced in December 2003, this collaboration will leverage the best-in-class tools of each company to deliver co-simulation capability for FSI. The joint initiative will provide engineers with easy-to-use tools that will take into account the important coupling effects between fluid flow phenomena and structural behavior, including sophisticated simulation of nonlinear structural responses. Coupled fluid-structure simulation capability is important for many applications in the automotive, aerospace, biomedical, and process industries. The planned co-sim-
ulation capability between FLUENT and ABAQUS is expected to address applications such as fan blade flutter, microfluidics, in-vivo vascular flows, and flow-induced vibrations in structures and equipment, as well as many consumer goods design and packaging applications. Implementation of co-simulation using FLUENT and ABAQUS will provide our customers with an outstanding coupled fluid-structure solution, notes Paul Bemis, Fluents VP of Product Management. Many of our customers most challenging design problems require assessment of interactions between fluid and structure. By teaming with ABAQUS, Inc. we will provide a solution based on industry-leading best-in-class technologies. s
ABAQUS is extremely pleased to be working closely with Fluent, who are a recognized leader in CFD, notes Ken Short, VP of Marketing at ABAQUS, Inc. This initiative is a significant step toward a practical and useable FSI capability from two of the most respected solution providers in the software industry.
S5
healthcare industry
drug delivery
healthcare industry
Drug Delivery
from
Coronary Stents
By Neil Bulman-Fleming, Rosaire Mongrain, and Isam Faik, McGill University, Montral, Canada; and Olivier Bertrand, Universit Laval, Qubec City, Canada
inimally invasive catheter-based interventions have revolutionized the world of heart surgery. Lengthy recovery times associated with open-heart surgeries are routinely avoided through the use of balloon angioplasty and coronary stenting. Stenting is a process involving the deployment of a metal scaffolding inside a blocked artery to maintain blood flow to the heart muscle. Though these techniques have achieved a high degree of success, they are not by any means perfect. In a large percentage of cases, the patients body responds to the implantation of a stent by growing many layers of smooth muscle cells over the implant. This response can often lead to blockage of the coronary artery the exact problem that the stent is intended to solve initially. One promising solution to this problem is the deployment of drug-loaded stents in diseased arteries. By coating the stent structure with a chemical-infused polymer, drugs that limit smooth muscle cell growth can be delivered directly to the area responsible. Over a critical therapeutic period of about one month, uniform drug distribution can effectively eliminate the proliferation of smooth muscle cells. The challenge is to achieve a uniform drug distribution in the region of the stent, and to verify that an appropriate dose is provided. Animal trials investigating drug-loaded stents can give a general indication of an implants success, but do not provide accurate dose distributions in the arterial tissue. One practical, inexpensive and efficient means of studying these dose distributions is through numerical modeling. Although real tissues are complex, irregular, and widely varied among patients, appropriate simplifications to the coronary artery tissues can yield meaningful general dose distribution results. In a study carried out at McGill University in Montral, Canada, FIDAP was used to model the pharmacokinetics of drug-eluting stents. The goal of the study was to develop a tool for evaluating the dose delivery characteristics of real 3D stent geometries. Such a tool could then be used by stent designers to modify geometries in order to ensure optimal dose delivery outcomes. Three common stent geometries were modeled, and a simulation was carried out in a quarter-artery domain. By examining mass concentration over time at field points throughout the arterial wall, an estimate for how the drug moves through the solution domain was obtained. Dose homogeneity, final concentration, and stent contact area values were combined into a single parameter, the Local Delivery Effectiveness Score (LDES). This score, out of ten, combines key factors that may contribute to the success of partic-
One example of the quarter-domain stent geometries used in the numerical model
(a) Longitudinal (constant ) and (b) cylindrical (constant radius) sections of the therapeutic region after 7 days; the flow travels from bottom to top in (a), and the thickness of the vessel wall is exaggerated to illustrate the spread of the drug
ular stent designs, and provides a simple, onestep evaluation criterion. In order to increase the accuracy and realism of this basic model, work is now underway to use actual patient-specific artery geometries in a numerical model. Images obtained using Intravascular Ultrasound (IVUS) can be reconstructed from individual transverse slices to build a 3D model. Such models incorporate arterial tissues as well as calcified deposits and other inclusions known to be present in diseased coronary arteries. These geometries, along with accurate diffusion information for the tissues represented in them, could serve as the basis for FIDAP simulations which would allow the optimization of stent
and drug loading parameters tailored to each patients specific needs. s
references:
N. Bulman-Fleming, R. Mongrain, and O.F. Bertrand, 3D Numerical Simulations of Stent Based Local Drug Delivery Using Realistic Post Implantation Geometries, Endocoronary Biomechanics and Restenosis Symposium, Paris, 2003. R. Mongrain, N. Bulman-Fleming, J.C. Tardif, S. Plante, and O.F. Bertrand, Numerical Simulations of Local Pharmacokinetics of a Drug Delivered from an Eluting Stent, in Advanced Materials for Biomedical Applications, edited by D. Mantovani, Canadian Institute of Mining, Metallurgy and Petroleum, ISBN 1-894475-25-9, 399, pp.213-223, 2002.
S6
drug delivery
Needle-Free
Drug Injections
By Y. Liu, and M.A.F. Kendall, Department of Engineering Science, University of Oxford, Oxford, UK helium micro cylinder powdered drug cassette
time (s)
novel, needle-less, powdered drug delivery system is currently being developed. It makes use of biolistics technology, where tiny particles are injected through cell walls by a high-powered gun. The technology provides a unique capability; it effectively delivers vaccines in micro-particle form through the skin, into an epidermal layer where Langerhans cells reside. By targeting these cells, an optimal immune response can be elicited, potentially decreasing both the risk of disease and cost of protection. In the gun, high-pressure helium gas is stored in a micro-cylinder, and the powdered vaccine is stored in a cassette. The gun accelerates the vaccine particles into human skin with a transient supersonic jet. To do this, the particles must be delivered with a narrow and controllable impact velocity range, and a wide, yet uniform spatial distribution. Ideally, the bulk of the particle cloud should be accelerated in the quasisteady supersonic flow (QSSF) regime. Proper choice of system geometry, gas species, and operating conditions can ensure that this condition is met. Practical constraints limit the device length and the duration of QSSF, when the particles are to be entrained. To better understand the delivery mechanism and biological interaction, the effects of these important parameters need to be identified and understood. FLUENT software offers comprehensive capabilities to model the biolistics system, from the transient gas and particle dynamics to interactions with the skin target. The species transport equations together with the standard k- turbulence model are used to solve for the multi-species gas phase flow. The coupled explicit solver is used to capture the main features of the unsteady motion of the shock wave process. An overall second order accuracy is satisfied both spatially and temporally. The particle trajectory equations, in conjunction with a drag correlation and inter-phase heat exchange, are advanced in time with the gas flow simulation. The drag correlations proposed by Igra & Takayama (1993), which consider unsteady effects and cover a wide range of Reynolds numbers (200 to 101,000), are implemented through user-defined functions (UDFs). Through the modeling efforts, the FLUENT simulations have allowed for the evaluation of key parameters, the visualization of different designs, and the gathering of new insights into the biolistics system. Simulations of the whole prototype biolistics system, as well as of the key components, have shown an excellent agreement with the static pressure measurements, Pitot probe survey, and images made using Doppler global velocimetry (DGV) and particle image velocimetry (PIV).1, 2 A space-time diagram that shows different gas flow regimes and particle cloud trajectories can be used to illustrate the performance of the prototype. The diagram demonstrates that the particles are accelerated to the nozzle exit, avoiding the starting process1 and the reflected expansion wave, and thereby remaining in the QSSF regime, as desired. s
contour shock tube
actuation pin
spacer
silencer A schematic of a prototype biolistics system, configured for clinical use1
Instantaneous contour plot of the gas velocity and particle trajectories for a silenced configuration, taken 120s after diaphragm rupture2
300 250 200 reflected wave 150 100 50 0 unsteady expansion
area change
nozzle
secondary shock
ry to jec tra
references:
1 M.A.F. Kendall, The Delivery of Particulate Vaccines and Drugs to Human Skin with a Practical, Hand-held Shock Tube-based System, Shock Waves Journal, 12(1), pp.2230, 2002. Y. Liu, M.A.F. Kendall, N.K. Truong, and B.J. Bellhouse, Numerical and Experimental Analysis of a High Speed Needle-free Powdered Vaccines Delivery Device, AIAA-20022807, Proc. 20th AIAA Applied Aerodynamics Conference, St. Louis, MO, USA, 2002.
ud clo le tic r pa
contact surface primary shock -60 -40 -20 0 20 40 axial position (mm) 60 80 100
The key gas flow regimes and particle cloud trajectories of the prototype biolistics system are shown together in the calculated space-time (x-t) diagram2
S7
healthcare industry
pharmaceuticals
healthcare industry
Improving Reaction Selectivity

by
Clever Mixing
n the pharmaceutical industry, many chemical reactions leading to desirable intermediate and end-products are accompanied by side reactions producing undesired by-products. By-products decrease reaction yield and complicate product separation. The simplest system that reveals this behavior is one of two parallel reactions with two reagents, B and C, present in a homogeneous mixture and competing for a common reagent, A. Competing reactions result in the formation of the desired product, R, and the undesired product, S:
By Jerzy Baldyga and Lukasz Makowski, Dept. of Chemical and Process Engineering, Warsaw University of Technology, Warsaw, Poland
Backmixing effect chemical reaction takes place through back diffusion into the feed pipe
1 A + B R k2 A + C S
Contours of concentration of the secondary product S in the stirred tank reactor
Visualization of reaction zone in the semibatch stirred tank reactor
A competitive-consecutive reaction scheme that behaves in a similar manner1 involves only reagents A and B, but delivers by-products through a reaction of A with R. An understanding of the process parameters that govern the competitive-parallel scheme therefore can be applied to the competitive-consecutive scheme as well. To improve selectivity, the competition between reactions can be enhanced. For example the addition of a homogeneous catalyst can increase the rate of the first, desired reaction, resulting in a more selective synthesis of R. However, there is a limit for such a procedure. When the first reaction becomes very fast, its rate becomes controlled by mixing, rather than the reaction kinetics. Competition is then between the mixing-controlled first reaction and the slower second reaction, which is often also affected by mixing. The problem of reactive mixing arises and is further complicated when the flow is turbulent and the Reynolds averaging procedure is applied. To model this process, it is useful to check which of the sequence of mixing processes can directly or indirectly affect the course of the chemical reactions. Comparison of the characteristic times for mixing and reaction can provide this information. For very long reaction times (minutes to hours) only the process time is relevant (feed time in the case of semibatch stirred tank reactors or mean residence time in continuous stirred-tank reactors (CSTRs)). When the reaction time is between seconds and minutes, consideration of the flow
pattern (macromixing) and turbulent diffusion (mesomixing) is required, but concentration fluctuations of reacting species can be neglected. For reaction times that are significantly smaller than 1 second, the reaction rate becomes micromixing-dependent and the effects of concentration fluctuations should be included. A general modeling approach has been developed and linked to FLUENT 6 through user-defined functions (UDFs). It incorporates all three regimes, and can be applied when the first reaction is instantaneous and the second one is fast. Involving macromixing, mesomixing, micromixing, and reaction kinetics2,3, it can be applied to semibatch or continuous operation. To model micromixing, a nonequilibrium multiple-time-scale mixing model is applied. The model includes mixing in the inertial-convective, viscous-convective and viscous-diffusive subranges of the spectrum. This means that the effects of the molecular diffusivity and viscosity on the rate of turbulent mixing are included. To express the averaged reaction rate in terms of other dependent variables, a conditional moment closure based on a linear interpolation of local instantaneous reactant concentration values is applied. The reactant concentrations are expressed by means of the mixture fraction, whereas the distribution of the mixture fraction is approximated using a beta distribution function. The model has been applied2,3 to a semibatch stirred tank reactor and a CSTR, predicting well all the trends observed in experiments, including the effects of residence and feed times, impeller speed, feed concentrations, feed pipe position, and backmixing. The results illustrate that by using this approach, it is possible to design the way in which the reagents are contacted and mixed to obtain the goals specific for the process, that is, to improve selectivity. s
references:
1 2 3 J. Baldyga and J.R. Bourne, Turbulent Mixing and Chemical Reactions, Chichester, Wiley, 1999. J. Baldyga, M. Henczka, L. Makowski, Chem. Eng. Res. Des., 79, Part A, pp.895-900, 2001. J. Baldyga and L. Makowski, Chem. Eng. Technol., 27(3), pp.225-230, 2004.
S8
medical equipment
Virtual Prototyping System

for
Blood Pump Design

tions allow accurate prediction of near-wall regions without a significant increase in the mesh size. For unsteady calculations, the frozen rotor solution, computed from the multiple reference frames (MRF) model, is used as an initial condition. Postprocessing of the CFD results includes a quantitative assessment of the pump design performance through predictions of pressure rise in the pump, rotor torque, forces, and stresses. Vector plots provide detailed pictures of the flow patterns, and contour plots show the locations of regions where the maximum potential blood damage would most likely occur. Journal files are used during postprocessing to prepare comprehensive reports on the computational results. If desired, special procedures are written using Scheme files, which can be read by FLUENT. These procedures allow for complex cyclical operations during solver execution and postprocessing that expand the capabilities of the code. Only after the optimal virtual design has been identified is the actual prototype built. The highly efficient virtual prototyping system that has been developed at ABI has significantly reduced design time and development costs. s
By Valentin Izraelev and Andrei Khodak, ABI, Hopkins, MN
The blood pump housing, showing the inlet at left and outlet at right
Cross-section of the pump in the outlet plane, showing velocity vectors
ince the late 1990s, ABI (Advanced Bionics, Inc.) has used Fluent software on a regular basis in the development of proprietary bearing- and seal-free rotary pump technology. Blood pumps for medical applications are a major part of ABIs business. The traditional theory of centrifugal pumps, used for over 100 years, does not apply directly to small blood pumps with flow rates between 0.3 1.5 gal/min and differential pressures of 1.0 6.0 psi. To complicate the matter, small blood pumps use variable speed drives, and have unique optimization criteria. In addition to traditional pump characteristics, such as flow rate, pressure, and efficiency, blood-handling characteristics such as hemolysis and thrombosis are important for blood pumps. Hemolysis is the destruction of red blood cells and is correlated to shear stress and residence time. Thrombosis is clot formation, and is correlated to wall shear stress. Thus, operation of an optimum blood pump should not lead to too much shear stress (to have low hemolysis), and should simultaneously maintain a certain level of wall shear stress (to prevent thrombosis). ABI has developed an expertise in analyzing blood pump performance, and has used this capability as a virtual prototyping design tool with a number of variables representing the pump dimensions. A number of actual blood pump tests
have shown that the computational and experimental results are very closely correlated. To perform an accurate analysis, the whole pump, and often a substantial portion of the outlet region, needs to be meshed. GAMBIT is used to build 3D grids with sliding, often non-conformal interfaces around the rotor region. This kind of flexibility allows accurate meshing for the blades, which have complex geometry. A typical mesh size for the calculations is about 300,000 cells. Journal files for GAMBIT provide an opportunity to make small changes in the geometry and automatically rebuild the computational grid. This means that the mesh generation time for each design change is significantly reduced. This parametric approach allows fast virtual prototyping of many variants from the predetermined matrix of geometric parameters. The parametric approach is also applied in FLUENT, where journal files are used during the solution stage and for postprocessing. The performance of each variant is tested at different operating conditions. Journal files allow automatic preparation of the case and data file for each regime from a prescribed set of rotation speed and flow rate boundary conditions. For solutions in the turbulent regime, the standard k- turbulence model is used. FLUENTs enhanced wall-func-
Contours of the wall shear stress on the rotor wall of a Tesla-type pump; half of the pump rotor is shown
S9
healthcare industry
respiration
healthcare industry
FSI Produces a
By Antoine Dozolme and Thierry Marchal, Fluent Benelux The surface mesh, and contours of deformation on the inner trachea wall during inspiration
Breath of Fresh Air

Below, deformation of the outer trachea wall with velocity vectors at the bronchi exits, with the shadow of the inner trachea wall shown
he airflow through the trachea and bronchi can be impacted by disease or injury. It can also be used to transport drugs from nasal sprays or inhalers to the blood stream through the airway walls. Recently, the fluid-structure interaction (FSI) model in FIDAP was used to study the airflow and induced wall deformation during a normal breathing cycle. When FSI is used, the stresses on, and subsequent deformation of the airways are calculated as a consequence of the airflow. Two approaches were investigated. For the simpler case of a 3D branched duct, the shell element option was used to model the trachea wall. This option treats the wall as a membrane of negligible thickness with an associated bending resistance, so that small deformations are possible. A more rigorous but more computationally expensive approach was applied to a patient-specific geometry. For this case, the trachea wall was modeled as a deformable entity of finite thickness. While the simplified approach yields quick qualitative results that are satisfactory for a preliminary study, the second approach is more appropriate for patient-specific geometries, when stresses and deformation of the trachea wall are the major focus. PreSTO, the new template capability in FIDAP, was used to setup the sim-
ulations, using a realistic geometry obtained from computerized tomography (CT) scans of a patient. To prescribe the inlet velocity boundary condition, breathing was modeled as a sinusoidal function of time. A breathing cycle of 5 seconds, appropriate for normal adult activities, was considered. Modifications to the breathing rate and/or amplitude were studied to investigate the differences between periods of exertion or sleeping. Only radial deformations of the airway walls, resulting from normal forces, were allowed to occur. Following such deformations, the elastic remeshing technique was applied. The mixing length model was used for turbulence. The results showed that deformations of the trachea were small, occurring at the junction of the trachea and bronchial pipes, as expected. The computed deformations are directly related to the pressure applied by the moving air. The airflow was found to separate into equal proportions between the right and left lung during inspiration, with no direct air exchange between the lungs during expiration. s
The authors wish to thank Dr. Thomas Frauenfelder, University Hospital of Zurich, for providing the CT data.
Breathing Easy with the Dynamic Mesh Model

By Craig Skinner and Rob Woolhouse, Fluent Europe Ltd.
he periodic breathing through the trachea and bronchial tubes of an idealized adult has been modeled using FLUENT. A sinusoidal expansion and contraction of the walls was prescribed using the dynamic mesh model, and an accompanying sinusoidal velocity profile was assigned to the inlet. Pressure boundaries were used for the flow outlets.
The results have been used to examine the trajectories of fine particles, representative of drugs delivered by an inhaler. Simulations such as this are opening the door to many other associated studies, such as investigations of inhaler designs for the purpose of optimizing the uniformity of drug particle dispersion in the lungs. s
Particle traces show the dispersion of drug particles during inhalation
S10
automotive
In-Cylinder
Power
By Xiao Hu and Lisa Mesaros, Fluent Inc.
s emissions regulations tighten and fuel economy requirements heighten, improving engine performance is one of the highest priorities of major automotive manufacturers. The overall vehicle performance is highly driven by the efficiency and combustion by-products of the internal combustion (IC) engine. Better control over the turbulent flow structures, air-fuel mixing rates, and ignition patterns can greatly impact the performance. However, the simulation of IC engines remains one of the most challenging applications for CFD modeling. Moving valves and pistons require a mechanism to handle a changing mesh, and this must be coupled with models of complex sprays and combustion processes, all within a highly turbulent flow environment. FLUENTs general purpose dynamic mesh tool is capable of simulating a wide range of moving boundary applications, including IC engines. Its numerous spray and combustion models have been validated through years of rigorous testing. The compatibility of these three models dynamic mesh, spray, and combustion makes FLUENT uniquely qualified to simulate a variety of situations inside internal combustion engines. For a dynamic mesh application, the domain is decomposed into different zones. This allows for different mesh motions and the use of different re-meshing algorithms in different regions within a single simulation. The unstructured smoothing and re-meshing approaches are popular choices for modeling the upper combustion chamber, since they greatly facilitate the process of tracking changes to the complex mesh in the vicinity of the valves. If only traditional structured approaches were available, it would be difficult to generate topologies that could accommodate the full range of valve motion in this region. Typically, such structured moving mesh approaches require special preprocessing tools and involve significant manual work. These tools and procedures are not required for the dynamic mesh model in FLUENT, where only the initial mesh and description of the bound-
ary movement are required. FLUENT also includes tools for treating arbitrarily complicated piston shapes. Engines such as these pose tremendous challenges for structured re-meshing approaches, but are easily handled with the unstructured tools in FLUENT. Several new models, specific to in-cylinder applications, have been implemented in the upcoming release of FLUENT 6.2. For premixed engines, two critical models have been incorporated. The first is a spark model, which is used to simulate the electric spark required to initiate the combustion process. The second is an autoignition model, which is used to predict knocking. Other new features include an ignition delay model based on the work of Hardenburg and Hase.1 This model has been developed specifically to simulate direct injection diesel engines. There is also a new wall film model, which is necessary for capturing the fuel formation along walls in direct-injection and port-fuel-injected gasoline engines, and some small-bore or cold-start diesel engines. Simulations have been ongoing at Fluent for some time to test the new IC engine capabilities. One study was carried out to validate several of the models using a single cylinder version of the Caterpillar 3400 series heavyduty diesel engine for which experimental data has been published.2 The objective of the study was to validate the newly implemented ignition model in conjunction with the dynamic mesh capability for six different load and speed conditions (modes) from a federal transient test procedure. This procedure, from the United States Environmental Protection Agency, evaluates exhaust pollution from diesel engines at different load points. A vehicle is run on a chassis dynamometer and measurements for fuel economy and emissions are made. Speed and load points are specified as a function of time and chosen to represent the typical output of the engine. Europe and Japan have their own procedures. For the CFD analysis, the diesel spray was described using the Lagrangian discrete phase model (DPM). The nozzle was treated as a solid cone atomizer, and the fuel breakup
A symmetric four-valve engine; local smoothing and remeshing are used in the upper part of the combustion chamber, and dynamic layering is used in the lower part, adjacent to the piston, and in the region above the valves to allow better resolution of the valve seat gap
Grid of an engine with arbitrarily complex shape6
23
automotive
Spray and wall film model applied to a port-fuel-injected engine showing the spray (left) and the wall film height (right)
Combustion flame development after the spark ignition of the DaimlerChrysler engine
Courtesy of DaimlerChrysler
was governed by the wave breakup model.3 Ignition delay was also accounted for using the Hardenburg and Hase model. After evaporation of the spray, the eddy-dissipation model in FLUENT was used to simulate mixing controlled combustion. Results for the mass-averaged cylinder pressure as a function of time were found to be in very good agreement with the experimentally obtained data for each of the six modes. Ignition delay was well represented and the peak pressure was predicted accurately for both magnitude and phasing. The models in FLUENT can also be applied to spark ignition (SI) engines. A study was carried out on a port-fuel-injected, SI gasoline engine from DaimlerChrysler using FLUENTs Zimont premixed turbulence combustion model.4 The Zimont model solves a progress variable equation to predict the turbulent flame speed and rate of energy release on a spatially resolved basis. To use this model, the fuel and air are assumed to be perfectly premixed, an acceptable assumption for many SI engines. Although no species information is available for pollutant calculations, the model runs quickly on a per cell basis so that the flow field can be well resolved without requiring excessive computational time. In the simulation, the electric spark model was used to initiate the combustion. The mesh count varied from 280,000 elements at top dead center to 724,000 elements at bottom dead center. The results were used to illustrate the combustion flame development after ignition. Good agreement between the CFD predictions and experimentally derived values for global burned gas mass fraction as a function of crank angle were obtained. With this tool, the impact of different spark timing on the engine power output can be studied to determine conditions that improve fuel economy. Engineers at DaimlerChrysler made further use of their CFD results for this case.
Convective heat transfer coefficients from the gas side were computed in FLUENT and used as thermal boundary conditions for a steadystate stress analysis calculation in the engine block and head. A steady stress analysis requires less computational time than a transient one, yet still accounts for non-uniform thermal loading. Thus spatially-resolved but time-averaged heat transfer coefficients were exported from FLUENT for import into a structural finite element analysis (FEA) code. Since only thermal information was needed for this purpose, the relatively inexpensive Zimont combustion model, capable of predicting heat release, was sufficient. As expected, the time-averaged heat transfer coefficient on the surface of the cylinder head was found to be highly non-uniform, especially at the exhaust valve seat region where the gas speed is high. The exhaust valve seat area has a heat transfer coefficient more than a factor of ten higher than other parts of the head, which correlates with a significant increase in thermal stress in that region. The piston, on the other hand, has a relatively uniform heat transfer coefficient distribution because of the uniformity of the velocity in that region. The study showed the importance of using CFD to obtain more accurate heat transfer coefficients for thermal stress analysis. It also showed that solutions of this type can be obtained with reasonable computational expense. Other complex cases have been studied to test other IC application areas. For example, the wall film model has been used in the simulation of a port-fuel-injected gasoline engine. For this analysis, the spray was described using the discrete phase model. The nozzle was treated as a solid cone atomizer, and the ORourke and Amsden5 model was used for the film on the intake valve. The study was done to investigate whether or not the fuel vapor is uniform when the spark goes off. In another case, the volume of fluid (VOF) model was combined
24
automotive
with the dynamic mesh model to predict engine cooling. In this simulation, a moving piston was modeled as a conducting solid with an upper temperature of 1300 K, and an oil jet at a temperature of 400 K was injected onto the piston from below. The high speed engine had a piston speed that was greater than the speed of the jet. The jet was found to cool the piston, as evidenced by surface temperature contours as a function of time. Validation and testing continues on these and other models in FLUENT 6.2 that are targeted at IC engines. Other advanced combustion capabilities are also being planned for future development, including unsteady flamelet approaches and multi-component vaporization. When combined with the flexibility of the dynamic mesh model, these options will allow for the most comprehensive suite of internal combustion modeling tools available in commercial software today. s
Contours of heat transfer coefficient on the cylinder head (left) and piston (right) for the DaimlerChrysler engine; this data was imported into a steady-state stress analysis code for analysis of heat transfer in the engine block and head
1.1 1.0 0.9 0.8 static pressure (pascal, e+06) 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.010
references:
1 H.O. Hardenburg and F.W. Hase, An Empirical Formula for Computing the Pressure Rise Delay of a Fuel from its Cetane Number and from the Relevant Parameters of Direct-Injection Diesel Engines, SAE 790493, 1979. D. Montgomery and R.D. Reitz, Six Mode Cycle Evaluation of the Effect of EGR and Multiple Injections on Particulate and NO Emission from a DI Diesel Engine, SAE 960316, 1996. R.D. Reitz, Modeling Atomization Processes in High-Pressure Vaporizing Sprays, Atomization and Spray Technology 3, pp.309-337, 1987. V.L. Zimont, To Computations of Turbulent Combustion of Partially Premixed Gases, Chemical Physics of Combustion and Explosion Processes; Combustion of Multi-phase and Gas Systems, Chernogolovka, OIKhF, pp.77-80 (Russian), 1977. P.J. ORourke and A.A. Amsden, A Particle Numerical Model for Wall Film Dynamics in Portinjected Engines, SAE paper 961961, pp.143-156, 1996. E.S. Suh and C.J. Rutland, Numerical Study of Fuel/Air Mixture Preparation in a GDI Engine, SAE 1999-01-3657, 1999.
FLUENT Experiment
0.013
0.016
0.019
0.022
0.025 0.028 time (m)
0.031
0.034
0.037
0.040
0.043
Mass averaged cylinder pressure for the six mode CAT engine simulation that includes internal combustion
1.0 mass fraction burnt 0.8 0.6 0.4 0.2 0.0 680
FLUENT Experiment
700
720 740 crank angle
760
Mass fraction of burned gas vs. crank angle for the DaimlerChrysler case shows very good agreement between FLUENT predictions and measurements
The VOF model is used with the dynamic mesh model to simulate piston cooling; surface temperatures are shown on the piston and oil surface at two times during the simulation
25
automotive
barrel
barrel
front mold
rear mold
barrel parison Mold and parison for an automotive water tank

Courtesy of MANN+HUMMEL GMBH
barrel
The Shape of Things to Come with
Automotive Gas & Water Tanks

By Thierry Marchal and Cathy Gomez, Fluent Benelux
flash flash front view rear view
hen designing the different components of a car, the gas and water (windshield washer fluid) tanks are often the last ones to be considered, because their shapes can be adjusted easily to fit the available space. The constraint is usually rather simple: maximize the inner volume of the tank by using as much of the remaining space as possible. This is why gas and water tanks often have complex, sometimes unusual shapes. For gas tanks, safety regulations require that the thickness of the walls be larger overall than a minimum value in order to avoid any failure in the event of an accident. If the gas tank walls are too thick, however, it needlessly increases the weight of the part and uses an excessive quantity of polymer, which impacts the cost. To balance these requirements, several companies are now using POLYFLOW software to simulate and optimize the blow molding process for gas and water tank manufacturing. The blow molding process consists of two stages: the extrusion of a cylindrical tube of polymer, also called the parison, and the blowing process, where the material is forced to conform to a mold. During the extrusion of the parison, gravity and die head motion allow the designer to adjust the parison thickness profile, which is normally non-uniform. Once extruded, the parison is positioned in between the two halves of the mold. These molds have the complex imprint of the desired tank. The top and bottom of the parison are squeezed by knives to seal the tube so that air can be blown into it. The air inflates the parison, just like a balloon, while the mold gradually closes. When the two halves of the mold are completely closed, an increase in pressure ensures that the blowing parison takes the exact shape of the mold, which is the desired gas and water tank shape. Using POLYFLOW, the free surface of the deforming parison is tracked as it is inflated. The material is allowed to stretch and thin until contact between the deforming free surface and moving mold is detected. The result of the numerical simulation illustrates not only the final shape of the part, which is similar to the mold, but more importantly, the thickness map of the blown parison. Often, the final thickness of a blow-molded product shows regions where the blown parison is too thin (dark blue) to be acceptable for safety reasons, or too thick (yellow and red) to be economically attractive. To compensate for a non-optimized blown parison, POLYFLOW has tools to suggest a different profile thickness for the extruded parison that will lead to a more uniform final thickness map for the blown product, independent of the complexity of the shape. Other results of the simulation include an extensions map, which indicates how much the material has stretched to reach the final shape, the permeability of the finished product, the weight of the flashes (waste material at the edges), as well as the inner volume. s
Thickness distribution of the EVOH layer for an automotive water tank; this low permeability material is one of several layers used in the parison
Courtesy of MANN+HUMMEL GMBH
26
automotive
Windshield Wipers at
High Speed
By Kimiyoshi Takada, Fuji Techno-Service; Toru Komoriya, Yukio Furugori, and Kazuhiro Kaminaga, Fuji Heavy Industries; Gunma, Japan
SUBARU LEGACY
s SUBARU invests in its IMPREZA for the World Rally Championship (WRC), the know-how and technology developed are fed back into their commercial line of cars, such as the SUBARU LEGACY. In the WRC, competitions take place under a variety of weather, road, and driving conditions, and ensuring visibility is one of the primary requirements for safe and fast driving. One issue at the forefront of wiper design is the tendency of wipers to lift off the surface of the windshield, or float, especially at high driving speeds. When this occurs, the wipers are not as effective in keeping the windshield clear. The problem has become more pronounced in recent years, with the reductions in aerodynamic drag on automobiles as a whole. In the same manner that a rear-mounted spoiler increases the down force on a performance car, a wiper-spoiler can also be used to increase the down force on a windshield wiper. By optimizing the spoiler design, the threshold speed at which floating begins can be increased. At SUBARU, FLUENT was recently used to calculate the airflow around a stationary wiper arm, with and without a spoiler, on a car moving at a range of speeds. The entire car, as well as the road, was included in the solution domain. The mesh of approximately 4.6 million cells was refined in the vicinity of the wiper. A uniform inlet velocity boundary condition was specified to simulate the moving car. The realizable k- model was used to calculate the separation vortices generated behind the wiper. Laboratory measurements were made to validate the CFD models. Experimental tests demonstrated that for all speeds tested (from 100 to 180 km/h), the spoiler greatly increased the down force of the wiper on the wind-
shield. For this range of driving conditions, FLUENT predictions of both down force and drag force were found to be in excellent agreement with experimental data for a wiper equipped with a spoiler. Both forces were found to increase in proportion to the square of the car speed, in accordance with the experimental results, and the agreement between the CFD calculations and experiment improved as the speed of the car increased. The detailed flow field surrounding the wiper was not measured experimentally, but the validation of the down and drag forces led designers to have confidence in the CFD predictions of the flow field for subsequent design changes that were proposed. CFD predictions of pressure in the vicinity of a wiper with a spoiler showed a larger pressure on top of the wiper than below it, even at the highest speed considered. The results showed that where air hits the top of the spoiler, a region of high pressure develops. Vortices generated behind the spoiler cause a corresponding drop in pressure in that region. The interaction of the wiper blade and spoiler was assumed to exist, but was not investigated. SUBARU engineers have also been using FLUENT to study windshield washer designs. Using the discrete phase model (DPM), the trajectories of water droplets were followed for a variety of car speeds and nozzle designs. The locations where each of the trajectories splashed onto the windshield in the experiments were compared to the CFD predictions. For two nozzle designs, the agreement was very good for driving speeds ranging from 100 to 180 km/h. The advances made using CFD for windshield wiper and washer design will be realized in the newest line of SUBARU LEGACY automobiles. s
6 5 force (N) 4 3 2 1 0 100 down force FLUENT experiment 160 120 140 vehicle speed (km/h) 180 drag force FLUENT experiment
Comparison of down force and drag force predicted by FLUENT and measured results
Pressure distribution on the wiper arm
Velocity vectors on a slice through the wiper arm
27
automotive
to oil, and air to water. Through these heat exchangers, the major part of energy produced by the engines internal combustion is released to the environment. The auxiliary weight and size of these components are considered drawbacks in terms of performance and efficiency loss. Thus, there is a need to design heat exchangers that guarantee maximal heat rejection with reduced overall dimensions. Usually, the designs are based on the experience acquired from previous projects and from experimental data obtained from costly prototyping. Approximate formulae and correlations for heat exchanger design can also be used, but they do not help designers achieve a proper balance between cost and performance. For the coupled analysis, a FLUENT simulation was first performed for the entire motorcycle, starting from the CAD surfaces of the vehicle provided by Ducati. The external flow analysis provided velocity distributions on the surfaces of the two heat exchangers. In the FLUENT simulations, the heat exchangers were modeled by a porous media in order to simulate the pressure drop across the components and the deviation experienced by the air when crossing them. The surface velocities predicted by the CFD analysis were then used as boundary conditions for the distributed parameters calculation for the heat exchangers. This approach uses a formulation that is based on the discretization of a system with n elements or entities, each of which represents a part of the system analyzed. In each of these entities, an integral form of balance equations is solved. For the heat exchangers, the distributed parameters calculation provides a detailed temperature distribution throughout the heat exchanger. The CFD results illustrated some interesting features of the flow in the vicinity of the heat exchangers. The velocity distributions on the heat exchanger surfaces clearly show the influence of the front fork and front wheel. The upstream components also give rise to a non-uniform turbulence intensity on the surfaces of the heat exchangers, which may influence their operating efficiency. The distributed parameter predictions for outlet oil temperature are in good agreement with experimental data. The success of the coupled calculations suggests that the numerical multiscale method is a valuable tool for quantifying the effects of surface velocity distribution on heat exchanger performance. It is cost-effective, too, since the analysis time is considerably less than that of a full CFD analysis of the motorcycle and heat exchanger detail for each proposed design change. s
Pathlines around the vehicle
Cool Ducati
Heat Exchangers
By A. Cantagalli and R. Rossi, Laboratorio di Termofluidodinamica Computazionale Seconda Facolt di Ingegneria di Forl, Forl, Italy; S. Di Piazza and A. Ferraresi, Ducati Motor Holding S.p.A., Bologna, Italy
Air-water (top) and air-oil (bottom) heat exchangers
he Computational Thermo-FluidDynamics Laboratory of the Forl School of Engineering and Ducati Motor Holding have developed an innovative approach to the analysis of motorcycle heat exchanger performance. Called the numerical multiscale method, the approach makes use of different orders of spatial discretization to analyze two heat exchangers embedded in the complex geometry of a motorcycle. The approach, which has been carried out on the cooling system of the Ducati 999 motorcycle, couples FLUENT with an in-house distributed parameters code. Heat exchangers represent the main component of a motorcycles cooling system. Two types are commonly used: air
The full vehicle model solved in FLUENT
Contours of velocity magnitude on the heat exchanger surfaces
28
marine
Modeling
Marine
By Shin Hyung Rhee and Shitalkumar Joshi, Fluent Inc.
Propellers
The static pressure distribution on the back (left) and face (right) sides of the propeller for an advance ratio of 1.1, showing the expected features, including the low pressure region near the tip on the face side, where the tip vortex originates
ompared to fans and turbines, a marine propeller has a very complex geometry because it has variable section profiles, chord lengths, and pitch angles. In addition, to meet the heavy loads required for todays high-speed vessels, marine propellers are usually operated at high rotation speeds, and this results in even higher skewness in the propeller design. These factors make propeller flow one of the most challenging problems in CFD and to date, many researchers and designers have relied on experimental techniques and simplified numerical methods for analysis. Recent modeling efforts have shown, however, that CFD can provide valuable insights into the flow field generated by a propeller, including the forces and moments due to the rotating blades. The inviscid potential flow methods used previously were successfully applied to many propellers and are still popular among field designers. Yet this approach often requires cumbersome procedures and a considerable amount of preliminary knowledge. During the past decade, CFD methods for solving the RANS equations have been increasingly applied to various marine propeller geometries. While these studies have shown great advancement in the technology, some issues still need to be addressed for more practicable procedures. These include mesh generation strategies and turbulence model selection. Regarding the mesh challenges, the strong twisting of the blade central plane, a stagnation point on the hub close to the blades, and limited space for grid generation behind the ship have combined to make structured meshes impractical. Unstructured mesh generation offers many benefits, but the recent introduction of hybrid meshing has the most promise. Hybrid meshes allow for fine resolution of the boundary layers, and concentrated detail in the tip vortex and wake regions.
Vradial
To properly capture the highly swirling flow and tip vortex, the turbulence model must be chosen appropriately. Many previous studies have under-predicted the size and strength of the tip vortex, and the cause is usually attributed to an over-estimation of the isotropic turbulent viscosity. Recent studies have shown that both the k- and algebraic Reynolds stress models perform far better than previously used models in predicting the mean flow features of tip vortices. In a recent study, a five-bladed, highly skewed propeller of typical modern design, the P5168, was simulated using the moving (rotating) reference frame model and the k- turbulence model in FLUENT. The simulations followed open water measurements1 made at a wide range of advance ratio, J (the ratio of freestream speed to tip speed). Thrust and torque coefficients were found to be within 8% and 11% of measured values, respectively. Circumferentially averaged axial, tangential, and radial velocity components were found to reproduce the experimental trends well. In particular, the maximum axial velocity in the mid-span area, where the blade has its highest pitch angle, was captured. The tangential component was found to increase with increasing radius, as expected, and negative radial velocity, which indicates flow contraction due to the propeller action, was also in good agreement with data. Finally, the increasing magnitude of all velocity components with increasing load and decreasing J was also correctly captured. s
1.5
J = 0.98 J = 1.10 J = 1.27 J = 1.51
1.2 Vaxial 0.9 0.6 0.25
0.50 r/R
0.75
1.00
3.5 3.0 Vtangential 2.5 2.0 1.5 1.0 0.5 0.25
J = 0.98 J = 1.10 J = 1.27 J = 1.51
0.50 r/R
0.75
1.00
.05
.00
J = 1.51
-.05
J = 1.27 J = 1.10 J = 0.98
reference:
1 C. Chesnakas and S. Jessup, Proc. 22nd Symposium on Naval Hydrodynamics, Washington DC, 1998.
-.10
FLUENT Chesnakas and Jessup 0.50 r/R 0.75 1.00
-.15 0.25
Circumferentially averaged velocity components normalized by tip speed, as functions of normalized radial coordinate (r/R) on an axial plane
29
chemical
Realistic Distillation Modeling

By Valmor de Almeida, Oak Ridge National Laboratory, Oak Ridge, TN; Rafiqul Khan and David Schowalter, Fluent Inc. Pressure contours on the surface of structured packing elements in a distillation column
Pathlines, colored by pressure, illustrate the flow through the structured packing elements
uch of the energy required for processing chemicals in the industrialized world is used in distillation columns. Because the multiphase physics and intricate geometry involved in these operations are so complicated, their design remains semiempirical. CFD modeling offers an opportunity to improve the performance of distillation columns, translating into energy savings, better product purity, and reduced environmental impact. In fact, the US Department of Energy (DOE) estimates that optimization could save trillions of BTUs per year. The Office of Industrial Technology at DOE has supported efforts undertaken by a team of engineers from ORNL, Fluent, and the Separations Research Program (SRP) at the University of Texas, Austin to modernize the modeling practices in distillation columns and similar chemical processing units. Other partners on the team include Dow Chemical, John Zink Co., Praxair, Sulzer Chemtech, and 3D Imaging and Development Inc. The project has focused on distillation columns that employ a structured packing technique, where packings are fabricated from thin, corrugated metal sheets, and arranged parallel to one another. The approach combines detailed geometry
modeling, first-principles simulation, and validation experiments to deliver a realistic simulation tool. A collection of macros, GraSPI (Graphical Structured Packing Interface), has been developed and integrated into GAMBIT and FLUENT. GraSPI coordinates the generation of geometries and meshes for typical commercial structured packing elements and flow simulation. Simulations to date have involved gas flows through different packing geometries, and comparisons of predicted pressure drop with experiment have identified the minimum mesh size required for a certain level of solution accuracy. GraSPI is currently being extended to handle other packing geometries, such as perforated sheets and cross-stacked cylindrical elements. FLUENT simulations of gas-liquid flows are also being carried out using the VOF model. The ultimate goal is to be able to simulate gas-liquid countercurrent flows through as many packing elements as necessary to produce a realistic model of SRPs pilot packed column. Achievement of this goal will show that accurate, large-scale, first-principles simulation of distillation in structured packed columns is finally at hand. s
Size Distribution Modeling in FLUENT

By Kumar Dhanasekharan and Jay Sanyal, Fluent Inc.
n many industrial applications involving multiphase flow, the size distribution of particles, bubbles, or droplets can evolve in conjunction with transport and chemical reaction. The evolutionary processes can be a combination of different phenomena, such as nucleation, growth, dispersion, dissolution, aggregation, and breakage. In CFD models of these systems, a balance equation is required to describe the changes in the particle population. Population balance can be applied to a wide range of applications such as crystallizers, bubble columns, fluidized bed reactors, sprays, soot formation, oilwater separators, and aerosols. There are several approaches to solving the population balance equations, and each is more suited to one application area than another. A discrete method divides the particle population into a finite number of size intervals or bins, and keeps track of particle transfer among the bins. A moment method solves for moments of the population balance equation, providing aver-
age and total properties of the distribution. In the classical moments approach, no assumptions are made about the size distribution and the equations are formulated in a closed form involving only functions of the moments themselves. However, this exact closure requirement poses a serious limitation as aggregation and breakage phenomena cannot be written as functions of moments. The quadrature method of moments overcomes this limitation with an approximate closure using Gaussian quadrature. Population balance is a salient component of modeling transport phenomena and chemical reactions. Fluent has undertaken significant development efforts in this area during the last three years through a Department of Energy Office of Industrial Technologies (DOE-OIT) program. All three of the size distribution modeling approaches have been implemented, and these are currently being offered in the form of application-specific consulting services for interested clients. s
4 particle number density (x 1017 particles/m3-m)
10 20 30 40 particle diameter (microns)
50
The predicted particle size distribution of KCl in a batch crystallizer using the discrete method; the solubility curve of KCl is modeled as a linear function of temperature, and crystal nucleation and growth are modeled using a typical supersaturation power law
30
chemical
Predicting Residence Time Distribution with FLUENT

By Terry Ring, Bin Wang, and Byung Sang-Choi, University of Utah, Salt Lake City, UT, and Kumar Dhanasekharan, Fluent Inc.
n chemical reaction engineering, the concept of residence time distribution (RTD) is fundamental to reactor design. RTD is the exit age distribution of fluid molecules leaving a reactor. The residence time is the total time spent by the fluid molecules within the reactor. The RTD for process equipment is typically measured using stimulus-response tracer experiments to detect design flaws such as bypass, channeling, and dead zones, in addition to characterizing a reactors mean residence time and standard deviation. This measurement has been done, historically, when a complete velocity distribution map for the fluid in the vessel is not available. Because CFD is capable of predicting the complete velocity distribution in a vessel, it provides an alternative, indeed, simpler means of determining the RTD. As part of a Department of Energy Chemical Industry of the Future proj-
ect for modeling crystallization, Fluent is working with the University of Utah to predict RTD using FLUENT, with the results being compared to experimental measurements. On the experimental side, a 1.4 liter stirred tank with complex multiple feed tube geometry has been used to measure the residence time distribution as a function of flow rate. The vessel is driven by a Rushton impeller and is either baffled or not baffled for the different cases studied. At both ideal and non-ideal conditions, experiments and simulations of the RTD have been performed. For the CFD work, there are multiple approaches for predicting residence time distribution. In one approach, the tracer fluid is represented by a large number of discrete particles and Lagrangian particle tracking analysis is done with the discrete phase model (DPM). A histogram of time at the outlet is the residence time
distribution. One drawback of this method, however, is that a large number of particles are required to ensure proper statistics. Alternatively, the tracer fluid can be treated as a continuum by solving a transport equation for the tracer species. It is the latter continuum approach that has been used for the crystallization project. A single-species flow field was first obtained using the k- turbulence model. A passive tracer was then introduced with a step change in its concentration in the feed. The tracer was modeled using a user-defined scalar transport equation. The surface-area averaged tracer concentration was monitored as a function of time at the outlet. The results provided the exit age distribution of the tracer in the reactor, and therefore, represent the RTD. The FLUENT predictions were found to be in excellent agreement with the experiment over a broad range of flow rates and impeller RPM values, for both
baffled and non-baffled tanks. Since residence time distributions are routinely, and rather easily measured for process equipment, the RTD predictions allow a rather painless way to validate a complex flow simulation. In addition, with the aid of FLUENT simulations, process scale-up can be facilitated without expensive work on an intermediate scale in a pilot plant. The question of whether to construct a pilot plant or not depends on whether engineers are in control of all the major variables for the process. Using FLUENT, a small-scale reactor can be modeled to accurately predict its residence time distribution. Various potential largescale reactor designs can also be modeled in the same manner, and the residence times predicted to verify that upon scale-up, the large-scale reactor controls the major variables for the process in the same way that the small-scale reactor does. s
FLUENT perfectly mixed CSTR conductivity data pH data
E(t) x10-4
1000
2000 time (s)
3000
4000
5000
Fluid flow simulation of a 1.4 liter baffled stirred tank operating with a Rushton impeller
Comparison of experimental and calculated residence time distribution function, E(t), for a CSTR; operating conditions are 200 rpm, 40 ml/min
31
chemical
Taking the
in Outer
Pressure Off
Space
A
few seconds before the lift-off of European Space Launchers, cryogenic tanks containing propellants are pressurized in order to have the necessary pressure conditions for the engine. If the Synchronized Sequence (a list of procedures to be performed before launch) is stopped (i.e. if the launch is aborted), the tanks must be returned to atmospheric pressure in order to control the temperatures and perform a new Synchronized Sequence later on. For future space launchers, this depressurization process will also occur during the long ballistic phase in microgravity conditions. In normal gravity, the pressure relief valve, positioned on the top of the tank, releases only gas to the environment, acting to lower the temperature of the liquid and ensuring the structural integrity of the vessel. In a microgravity environment, however, the liquid and gas phases will not separate by virtue of the gravitational body force, so it will be more difficult to deal with pressure relief by a release of gas alone. Since experimental measurements are not feasible in this regime, numerical methods are the most promising option for understanding the process. Experimental measurements of a normal gravity depressurization process performed at Air Liquide/DTA have been compared to simulations using FLUENT. Validation of this work has allowed engineers to put confidence in other depressurization simulations performed under microgravity conditions. Experiments were performed using liquid nitrogen in a glass cryostat, 800 cm in height. Pressure and temperature probes were positioned inside the vessel. Several rates of depressurization were examined for their impact on the thermal stratification and subsequent boiling patterns. For one rate studied, the level of liquid was found to increase slightly following depressurization, with bubbles observed to form, coalesce, and rise to the top of the liquid after 1.4 seconds, distorting the surface. CFD simulations of this case were performed using the VOF model in FLUENT. Both surface tension and contact angle were specified, and source terms were added to simulate mass transfer during boiling1. Predictions of volume fraction showed that the results are qualitatively the same as those observed in the experiments. Both an increase in the liquid level and the onset of boiling were predicted. The CFD solution predicted more of an increase in the liquid level than that observed in the experiments, however. This is because in the 2D axisymmetric model, bubbles correspond to torii, so they take up more space than the spherical bubbles in the actual experiment, thereby forcing the liquid to rise more as a result of their presence. After validation of this depressurization case, engineers have used CFD to explore depressurization in microgravity conditions. The same rate of depressurization has been studied, with g = 0.001 m/s3, and the results show a very different scenario. Bubbles form more rapidly, and the free surface is destabilized more easily. Rather than rise and coalesce, as in the normal gravity case, bubbles expand and entrain liquid throughout the entire domain. In one of the most significant findings of the study, the gas bubbles entrain droplets of liquid, and a mixture of gas and liquid, rather than pure gas, escapes out the pressure release valve at the top of the vessel. s
By Jerome Lacapere, Air Liquide DTA, Sassenage, France; and Bruno Vieille, CNES, Evry Cedex, France
t=0 t=0.5s t=1s t=1.4s The onset of boiling during depressurization of the cryostat in the experiment; the liquid level increases slightly before boiling begins
t=0
t=0.5s
t=1s
t=1.1s
t=1.4s
Volume fraction of liquid nitrogen (red) and nitrogen gas (blue) during depressurization in normal gravity conditions
reference:
t=0 t=1s t=1.3s t=1.6s t=2.7s
1
Volume fraction of liquid nitrogen (red) and nitrogen gas (blue) during depressurization in microgravity conditions
S.W.J. Welch and J. Wilson, A Volume of Fluid-Based Method for Fluid Flows with Phase Change, J. Comput. Phys., 160, pp.662-682, 2000.
32
power generation
ATEX for Explosive Atmospheres

By Mark Keating, Fluent Europe Ltd.
as turbine installations commonly involve housing the gas generator unit inside a confined acoustic enclosure, designed to attenuate noise levels and provide ventilation for controlled convective heat rejection and gas fuel leak dilution. Throughout the European Union, the potential for fuel leakage means that such enclosures must conform to the newly introduced ATEX Directives legislation (from the French Atmospheres Explosibles) as described in HSE (Health and Safety Executive) Guidance Note PM84. In the detection of gas leaks it is important to balance the ventilation system so that leaks can be diluted and removed from the enclosure, preventing build-up to explosive levels. Dilution should not be so great, however, that the true leak magnitude is masked at the detection point. The philosophy is that leaks that cannot be detected are diluted and those that cannot be diluted are detected. PM84 declares limits on the size of the 50%LEL (lower explosive limit) gas cloud before detection as less than 0.1% of the net enclosure volume, where 100%LEL represents the fuel/air mixture just rich enough to be ignitable. Additionally, PM84 requires that the sensor alarm threshold be set at ideally less than 5%LEL and no more than 10%LEL. FLUENT was used in a recent case study to investigate the conformance to the PM84 guidelines at one particular power plant site in the UK. A representative model of the inside of the enclosure, including the gas generator, intake, diffuser, pipework, and enclosure walls, was built using GAMBIT. Roof extraction fans were used to establish a negative gauge pressure inside the enclosure, and side wall inlet vents were used for air supply. A tetrahedral mesh of approximately 3.1 million cells was used, with cells concentrated around the fuel manifolds. The level of heat rejected from the engine casing was calculated from a heat balance across the engine, and applied to the casing surface as the driving thermal boundary condition. In conjunction with site measurements of internal pressure and temperature, FLUENT was used to establish confidence in a baseline operating case prior to leak introduction. For the safety analysis, a range of leak source flow rates, directions, and locations were modeled in an attempt to create the largest leak cloud possible. The leak flow rates used were based on the detector alarm threshold levels in the 3-10%LEL range. The leaks were introduced one at a time, as point sources in the worst possible locations regions with relatively high ventilation air residence time and low local air velocity. A two-species mixture model (methane/air) was used to examine diffusion of the gases and a user-defined function (UDF) was used to calculate the sizes of the gas clouds. The sizes of the steady state 50%LEL gas clouds for the worst leak considered were predicted at the 3, 5 and 10%LEL detection levels. These were determined to be 0.083% and 1.61% of the net enclosure volume for the first and last cases, respectively. The sizes that these gas clouds would reach before being detected depend on the exit detection levels. In this instance, it was decided to install detection capable of reaching the 3%LEL level rather than use a less sensitive limit coupled with localized dilution or ventilation system modifications. Detection at the 3%LEL level ensured compliance with the PM84 guidelines in that the cloud size was <0.1% of the net enclosure volume. s
Rendering of the gas turbine model created in GAMBIT, showing ventilation inlet (blue) and exit (red) ports; access doors (orange) are superimposed as a visual guide
An iso-surface of the 50%LEL cloud at (above) 3%LEL detection (corresponding to 0.083% of net enclosure volume), and (below) 10%LEL detection (corresponding to 1.61% of net enclosure volume); the turbine casing and enclosure walls have been removed for visualization
33
turbomachinery
Mapping the
Eckardt Centrifugal
By Michael R. Ruith and Franklyn J. Kelecy, Fluent Inc. diffuser hub impeller spinner
The impeller design for the Eckardt Rotor O, showing static pressure contours on the surfaces of the compressor for a mass flow rate of 6 kg/s and a rotor speed of 14,000 rpm
Surface mesh on spinner, hub, blade and diffuser for a single passage with the shroud removed; the meshed volume is outlined by the blue lines
FD has been used extensively over the past three decades to predict the performance of compressors for both single and multiple blade row configurations. Today, with the power of modern computers, steadystate solutions are carried out on a routine basis, and can be considered as part of the design process. As CFD algorithms and software have continued to be developed and refined, it remains essential that validation studies be conducted in order to ensure that the results are both sufficiently accurate and can be obtained in a robust and predictable manner. One of the most widely-used validation cases for compressors is the study carried out by Eckardt1,2,3 for a centrifugal compressor |impeller, also known simply as the Eckardt rotor. Eckardt published a series of papers in the 1970s discussing experiments conducted on two 20blade centrifugal compressor rotors, known as Rotor O and Rotor A. The resulting impeller performance maps as well as the investigation of local effects have been used extensively over the years to test the accuracy and robustness of CFD codes for turbomachinery. FLUENT 6 has a long history of application in a variety of industries, including turbomachinery, and thus has been validated for a wide range of geometries and flow conditions. However, validation studies for turbomachinery have mostly been limited to the calculation of a single operating point or a speed line. With the advent of robust and accurate solution techniques in FLUENT 6.2, it was decided that a more extensive validation study should be conducted than had been attempted in the past. Specifically, it was desired to carry out a series of calculations for multiple speed lines, where each speed line was analyzed from choke to near stall. The Eckardt centrifugal compressor1 was deemed to be well suited for this study given the availability of data for a wide range of conditions with which the numerical predictions could be compared. The compressor consists of a centrifugal impeller with 20 radially ending blades, together with a spinner attached to the hub. The inner and outer intake diameters are 90 and 280mm, respectively, while the impeller tip diameter is 400mm. The blade camber lines have ellipsoidal shapes in cylindrical sections.
The vaneless diffuser has a constant flow area to radius ratio of two. Experimental tests revealed that the tangential compressor outlet produced a severe distortion of the flow field, with a strong asymmetry of the flow field within the diffuser1. In an attempt to avoid these effects, Eckardt placed an additional throttle ring near the outlet of the diffuser. This ring substantially constrained the maximum mass flow, causing the compressor to choke early. However, the distortion of the flow field in the remaining portion of the compressor map was reduced to an insignificant level. This observation is important because in the present study, the throttle ring was not modeled, so the choke line defined in Eckardts compressor map is not applicable. The geometry and mesh were generated using GAMBIT. Assuming 18-degree rotationally periodic conditions, only a single blade passage was modeled. The flow volume extends from the inflow position 200mm upstream of the leading edge of the blade, through the blade passage, and finally into the vaneless diffuser section to the outlet placed at a radial position of 350mm. Due to the sharp angle at the axis, triangular elements were chosen for the surface mesh on the spinner close to the center. These were connected to hexahedral elements in the remainder of the spinner region through a conformal interface. In order to adequately resolve the leading edge of the blade as well as the flow channel, hexahedral elements and a non-conformal interface were used for the remainder of the domain. This choice also permitted a lower total cell count by reducing the resolution upstream of the rotor. The final hybrid mesh consisted of approximately 500,000 elements. Since the stationary shroud and diffuser surfaces were surfaces of revolution, a single rotating frame was employed for the entire flow domain, thus permitting the use of steady-state modeling procedures. The fluid was assumed to be air, modeled as an ideal gas. The turbulent flow was modeled with the realizable k- model, using a non-equilibrium, (pressure gradient sensitized) wall treatment. The implicit, coupled, density-based solver was used for the flow calculations, since high subsonic Mach numbers were expected in the blade passage (especially close to the trailing edge).
34
turbomachinery
Compressor
As the main goal of the present investigation was to reproduce the entire compressor map of the Eckardt rotor from choke to surge, the specification of inlet and outlet boundary conditions required some additional attention. For compressors, the total pressure, total temperature, mass flow, and flow angles are typically known for the inlet of the machine, whereas the outlet static pressure is unknown and a desired result of the simulation. However, a well-posed inlet and outlet boundary condition set for compressible flows requires knowing the total conditions and flow angles at the inlet and a static pressure at the outlet. To resolve this problem, two approaches were adopted to cover the range of compressor operating conditions. For the higher flowrates (up to choke), a conventional pressure inlet was used in conjunction with a pressure outlet augmented with the mass flow outlet option. The mass flow outlet adjusts the exit pressure such that a target mass flow rate (prescribed by the user) is obtained at convergence. At lower mass flow rates (near the approximate surge point), a mass flow inlet (with a mass flux profile prescribed using a user-defined function) paired with a conventional pressure outlet was used. This approach permitted stable solutions to be obtained at much lower flowrates than was possible with the traditional pressure inlet. The solutions were initialized using the unstructured full approximation scheme, or FAS, initialization algorithm which has been recently implemented in FLUENT 6.2. This procedure makes use of a series of coarse grid levels (based on multigrid groupings) upon which flow solutions are obtained, starting from the coarsest mesh, interpolating the solution to the next finest mesh, and so on until the finest level is reached. Calculations initialized in this manner can employ much higher Courant numbers than usual, thereby reducing the number of (fine grid) iterations required to reach convergence. Moreover, as the solutions on the coarse meshes are inexpensive, the amount of CPU time required to perform FAS is relatively small. In the present study, the FAS initialization capability provided a speed increase of about a factor of five for most cases. Three sets of speed line calculations were performed to examine the flow through the compressor at rotational speeds of 12,000, 14,000, and 16,000 rpm. The flow rate was varied from choke to surge for each speed. Both the total pressure ratio across the impeller and the isentropic efficiency were used to assess the results. A comparison of the present computations with the compressor map reported by Eckardt1 show that the CFD results are in good agreement with the Eckardt compressor map. It should be noted that since the Eckardt experiments were performed using a throttle ring in the diffuser, the choke point for the numerical simulations extends beyond the experimental choke point indicated in the plot. These results demonstrate that modern CFD codes can be employed to study complex flows in turbomachinery over a wide range of operating conditions. Moreover, the efficiency and robustness of these kinds of calculations can be significantly enhanced through techniques like FAS initialization and the mass flow outlet boundary condition. s
2.75 corrected shaft speed (rpm) 2.50 16000
ch thr oke w ott ith le r ing
ge
total pressure ratio
2.00
14000
1.75
sur
2.25
12000
choke without throttle ring FLUENT Experiment
1.50
1.25
1.00
4 5 6 7 corrected mass flow rate (kg/s)
Comparison of the experimentally (blue) and numerically (red) obtained compressor map, showing good agreement between the plotted surge and choke lines with the throttle ring; the difference is greater in the choke region without the throttle ring
references:
1 D. Eckardt, Instantaneous Measurements in the Jet-Wake Discharge Flow of a Centrifugal Compressor Impeller. J. Eng. Power, 337, 1975. D. Eckardt, Detailed Flow Investigations Within a High-Speed Centrifugal Compressor Impeller. Trans. ASME, September, 1976. D. Eckardt, Flow Field Analysis of Radial and Backswept Centrifugal Compressor Impellers Part I: Flow Measurements using a Laser Velocimeter. In Performance Prediction of Centrifugal Pumps and Compressors, edited by S. Gopalakrishnan, P. Cooper, C. Grennan, and J. Switzer, ASME, 1980.
Vectors of relative velocity, colored by the local Mach number for choke condition at n = 16,000 rpm; the perspective view, showing the corner between blade suction side (left) and shroud (right) illustrates strong flow separation, in line with experimental results
35
academic news
Archaeological reconstruction of a typical furnace (insert shows the front wall face on)
Courtesy Dr. Gillian Juleff
CFD in the 1
he manufacture of iron and steel requires high temperatures and precise control of chemical conditions. Todays sophisticated blast furnaces are the heirs to a long period of technological development and innovation stretching back into prehistory. One ancient technique for smelting iron is the bloomery process, in which iron ore and charcoal are heated to temperatures in excess of 1100C in order to initiate a reduction process that results in a mixture of lumps of iron metal and lumps of slag being produced. For a long time, archaeometallurgists believed that the high temperatures necessary for this process could only be generated by means of bellows-driven furnaces. Archaeological excavations carried out in the 1990s in Sri Lanka by Dr. Gillian Juleff1 (also from the University of Exeter) have uncovered an extensive, yet previously unknown, iron-producing industry dating to the 1st Millennium AD. The industry displayed a high degree of organization and produced high quality iron and steel, which was traded through-
st
Millennium AD
A typical furnace was modeled in FLUENT, with the bed being represented as a semi-porous region at a fixed temperature (1200C or 1500C). The furnace walls were assigned the material properties of clay brick, and sandstone was used as the material for the hillside. Full 3D calculations were performed on a mesh of 1.1 million cells, taking into account heat transfer and buoyancy effects. The results2 provided great insight into the flow patterns through and around the furnace, and proved conclusively that these wind-powered furnaces easily provide sufficient air flow through the bed to create the high temperatures necessary to smelt iron. In fact, these results suggest that the output from the furnaces could have been even higher quality carbon steel. s
By Gavin Tabor, School of Engineering, Computer Science and Mathematics, University of Exeter, Exeter, Devon, UK
out the region, reaching as far as the Islamic world. Confounding expectations, no evidence of bellows was found. On the contrary, it was suggested that the furnaces may have been shaped so that the strong winds in the area acted to draw air through the furnace. Having suggested this mechanism, it was important to demonstrate its validity, both through experimentation (building a replica furnace) and through detailed modeling. FLUENT was an ideal choice to model the air flows around and through the furnace and the heat transfer from the furnace bed to the surrounding environment. All the furnaces excavated at the Sri Lanka site (around 80 in all) consist of long trenches dug into the crests of hills, lined with clay, and each with a clay front wall punctuated by air vents. As the wind blows up the hillside, it is diverted over the top of the furnace, creating a high pressure region in front of the furnace wall and a low pressure region above the furnace bed. This pressure differential draws air into the bed through the vents in the front wall.
references:
1 2 G. Juleff, An ancient wind-powered iron-smelting technology in Sri Lanka, Nature, 379, p.60, 1996. G. Tabor, D. Molinari and G. Juleff, Computational simulation of air flows through a Sri Lankan wind driven furnace, submitted to J. Arch. Sci, 2003.
Pathlines colored by temperature
Air flow through the furnace as computed by FLUENT
36
academic news
Southampton Partnership
Optimizes
By Andy Keane and Neil Bressloff, School of Engineering Sciences, University of Southampton, Southampton, UK
Collaboration
outhampton Universitys School of Engineering Sciences is home to the Computational Engineering and Design (CED) group, which uses FLUENT extensively in a number of academic research projects. The group has an international reputation in the field of multi-disciplinary design optimization, and hosts the University Technology Partnership for design supported by RollsRoyce and BAE Systems. The industrial partners are eager for the CED group to use an industry standard code like FLUENT for optimization of aerodynamic performance and geometry. One important project undertaken by the group was the development of an automated design environment to exploit a heterogeneous mixture of local and remote computing facilities. Template files are used to drive CAD-based parametric geometry definition (e.g. CATIA, ICAD, Pro/ENGINEER, IDEAS), mesh generation (GAMBIT) and CFD solutions (FLUENT). In a typical application for the aerospace industry, for example, the system has been used to optimize the shape and orientation of the flap-track fairing on a commercial aircraft wing, which houses the mechanism used to deploy the trailing edge flaps. Because the group has strong links with industry, it concentrates on methods that combine flexibility with robustness and efficiency. The design search and optimization strategies under investigation by the group attempt to deal with the computationally intensive nature of CFD in a number of ways:
Poor (left) and improved (right) designs for a flap track fairing
Application of enhanced response surface methods Use of partially converged CFD solutions to build accurate response surface models in a fraction of the time required for full convergence Data fusion to combine low and high fidelity CFD data CFD mesh morphing to progress rapidly towards optimum geometry
Illustration of a response surface for the scarf angle optimization of a nacelle (top) with pressure coefficient contours (bottom)
The work in these areas continues, but the methods under development show considerable potential for improving the speed with which design optimization based on CFD can progress. In recognition of its expertise, the group has been awarded funds to host one of the UK regional e-Science centers, which is tasked with the promotion and development of a distributed computing infrastructure at both local and national levels. One of the supported projects is called GEODISE (Global Engineering Optimization and Design Search for Engineering). The project activities include the use of FLUENT to optimize the scarf angle of the inlet face of an aero-engine nacelle, making the compromise between ground noise levels and aerodynamic performance. The optimization strategy combines a design of experiment (DOE) study with response surface modeling. The experience gained in using FLUENT in aircraft optimization projects is now being extended to a broader range of applications. Racing car components, curved diffusers, river bank erosion, and the influence of bio-geometry on arterial disease all form the subjects of current research projects within the group. Southampton University demonstrates once again the power of FLUENT as a tool for the academic researcher. s
more.info@
Southampton Computational Engineering and Design group: www.ses.soton.ac.uk/projects/Comp_Eng_Des/ comp_eng_des.html The Southampton Regional e-Science Centre: www.e-science.soton.ac.uk The Geodise Project: www.geodise.org
references:
W. Song, A.J. Keane, and S.J. Cox, CFD-Based Shape Optimisation with Grid-Enabled Design Search Toolkits, Proc. UK e-Science All Hands Meeting, ISBN 1-904425-11-9, ed. S.J. Cox, Nottingham, pp.619-627, 2003. A.I.J. Forrester, N.W. Bressloff, and A.J. Keane. Response Surface Model Evolution, 16th AIAA Computational Fluid Dynamics Conference, Orlando, Florida, pp.23-26, June 2003. A.J.Keane, Wing Optimization Using Design of Experiment, Response Surface, and Data Fusion Methods, J. Aircraft, 40(4), pp.741-750, 2003.
37
academic news
Cavitation and Hydraulic Flip

By H. Vahedi Tafreshi and B. Pourdeyhimi, Nonwovens Cooperative Research Center, North Carolina State University, Raleigh, NC
n unsteady simulation of a sharp-edged orifice (r/d = 0.01, where r is the inlet radius of curvature and d the inlet diameter) conducting a waterjet under high pressure (150 bar) has been performed to examine the onset and evolution of cavitation clouds inside a nozzle. Using FLUENT, the simulation showed that if the pressure is high enough and the inlet edge sufficiently sharp, the cavitation cloud grows and reaches the nozzle outlet. As the cloud reaches the outlet, the downstream ambient air finds a way to flow into the nozzle, resulting in a so-called hydraulic flip. Once the hydraulic flip condition occurs, cavitation immediately stops because
Contours of stream function at 26 s show the presence of cavitation
the cloud region becomes filled with air, separating the waterjet from the nozzle wall. This effect keeps the waterjet surface from cavitation- and friction-induced instabilities. Constricted waterjets, enveloped by air inside a nozzle, stay intact for a significantly greater distance than non-constricted jets1. These waterjets have diverse applications, including nonwoven fabric manufacturing via a process called hydroentangling. This process is used for mechanically bonding a web of loose fibers to form uniform entangled sheets of fibers1. The impact of the waterjets with the fibers displaces and rotates them with respect to their neighbors. During these relative displacements, some of the fibers twist and entangle around others and inter-lock with them through fiber-to-fiber friction. The cavitation model in FLUENT 6 was used for a 17,000-cell axisymmetric simulation of flow through a sharp-edged nozzle. The cavitation model tracks two interpenetrating fluids (liquid and vapor) using a volume fraction equation and a single momentum equation. Bubbles form when the local pressure becomes less than the vaporization pressure, and these bubbles may grow and form cavities. Pressure inlet and pressure outlet boundary conditions were used along with the RNG k- turbulence model. The two-layer zonal method was used for the wall treatment, with y+ values close to unity in the cells adjacent to the solid surface. For the cavitation simulation, the bubble number density (BND) should be known in advance. Acquiring such information is difficult because it requires sophisticated
experimental facilities. Instead, a minimum BND that results in the occurrence of a hydraulic flip was determined and used in the simulation. To determine the value, a series of simulations were run with different BND values. Starting from a low value, the BND was gradually increased until the hydraulic flip occurred. The final value (6x109 bubbles/m3) was in agreement with a range of values (108 to 1012 bubbles/m3) reported in the literature2. (Any value greater than 6x109 bubbles/m3 would cause the hydraulic flip to occur as well.) To simplify the problem, the pressure outlet was set up with a vapor (rather than air) volume fraction of unity. This means that any backflow through the outlet enters the calculation domain in the form of vapor rather than air. This approximation is valid since the densities of vapor and air are similar when compared to that of liquid water. The CFD simulations were very successful in predicting the hydraulic flip. The discharge coefficient obtained from the simulation, Cd = 0.63, defined as the ratio of the actual flow rate from a nozzle to that calculated by inviscid one-dimensional theory (the Bernoulli equation), was found to be in excellent agreement with experimental data. s
references:
1 2 H. Vahedi Tafreshi and B. Pourdeyhimi, Experiments in Fluids, 35(4), pp.364-371, 2003. R.A. Bunnell, S.D. Heister, C. Yen, and S.H. Collicott, Atomization and Sprays, 9, pp.445-465, 1999. H. Vahedi Tafreshi and B. Pourdeyhimi, Textile Research Journal, 74(4), 2004.
Contours of mixture density inside the nozzle after 10, 30, 50, and 60 s of operation
38
academic news
CFD in the Politecnico di Milano Classroom

By Emanuela Colombo, Politecnico di Milano, Milan, Italy
n the 1980s, students involved in fluid dynamics research were encouraged to use in-house, rather than commercial CFD software. By the mid-1990s, following the spread of CFD into different industrial sectors, many technical universities began to realize that future engineers would have more job opportunities if they also learned how to properly use commercial CFD codes. Such training would teach the students how to apply the scientific method to CFD analyses of industrial problems. At Politecnico di Milano in 2001, Prof. Fabio Inzoli proposed a CFD course for PhD students
with specialization in areas related to fluid mechanics (energy, mechanical, bioengineering, hydraulics, aerospace, and nuclear). Over the past three years, the 70-hour course has grown to accommodate 32 students each year, utilizing a staff of eight professors with diverse scientific backgrounds. The students appreciate the multidisciplinary approach and the practical hands-on sessions. Since 2001, Gerard De Neuville, a Corporate VP of Fluent, has been an invited speaker at the opening lesson, and he has been joined by Giorgio Buccilli from Fluent Italia. In 2004, Dr. Emanuela Colombo, who worked at Fluent Italia until 2001,
has been assigned to teach the CFD module in the Wind Engineering Master with theoretical and practical sessions supported by the new wind tunnel facility. A further collaboration between Fluent and the academic world was the publication earlier this year of the second edition of the first Italian book on CFD, Fondamenti di Termofluidodinamica Computazionale, with the first chapter, a historical perspective on CFD, authored by Dr. De Neuville and Prof. Colombo. It is hoped that efforts such as these will work to bridge the gap between commercial CFD development and academic research and education in the future. s
King Fahd University of Petroleum & Minerals Adopts Fluent Software

By Latfi Karim, Fluid Codes Ltd. (Fluent Distributor for the Middle East), Greenford, Middlesex, UK
he Middle East is an emerging region in terms of its CFD usage. Many industries that are native to the region have, in other parts of the world, invested heavily in CFD analysis. Accordingly, Fluid Codes, launched in 1996, has successfully promoted CFD within the academic sector. Today, leading academic institutions are making use of the technology in their classrooms and research labs, so that their students, once out in the working world, can continue to advance the technology of the regions businesses. For over forty years, King Fahd University of Petroleum & Minerals (KFUPM) in Saudi Arabia has served as the center of academic engineering activities in the area, providing the entire Gulf region with qualified engineering manpower for many of the engineering sectors, including petrochemical, water treatment, and HVAC. Several departments at the university are now using Fluent software products exclusively. In the chemical and mechanical engineering departments, for example, FLUENT, MixSim, and POLYFLOW are being used for a variety of activities. Applications under the direction of Associate Professor Habib Zughbi in the chemical engineering department include jet mixing, pipeline mixing, reacting flow, and gas-solid multiphase flow. It is hoped that through cooperation between Fluent Europe Ltd., Fluid Codes Ltd., and the University, the use of CFD will continue to grow throughout the Middle East region, particularly at some of the worlds leading petrochemical companies in countries such as Saudi Arabia, Qatar, and the United Arab Emirates. s
Simulation of flow and chemical reactions in a Claus Converter using the finite rate model; surfaces of SO2 mass fraction are shown
Temperature contours on the surface and mid-plane at three times, showing the progress of mixing in a liquid jet agitated tank
39
product news
CFD Without Borders

Using
FloWizard
Guests are invited to the hosts session
By Andr Bakker and Scott Gilmore, Fluent Inc.
ow that CFD is making it into the mainstream of engineering design, many organizations have a need for tools that go beyond the standalone software approach that has characterized the CFD industry for so many years. An efficient data exchange system is crucial to the success of teams that include CAD engineers, designers, analysts, and researchers. Not only does the CFD software have to be able to exchange files with other CAD, meshing, and CAE analysis packages, it should also enable communication between the different team members. To meet this important need, FloWizard will be the first CFD software on the market that includes collaboration as a standard feature. Collaboration allows two or more people on a network to share in a FloWizard session. Thus, people in different locations can work together to review a CFD model at the same time. Collaboration is useful for design review meetings, and other team cooperation and support functions. To enable a collaborative session, a host user, or owner, starts FloWizard and enables collaboration from the FloWizard menu. He can then invite guest users to his session. FloWizard generates a unique session ID and password, and these are e-mailed to the invited guests. When the invited guests start FloWizard and supply the session ID and password, they are connected to the hosts FloWizard session. For collaboration, there are two main approaches, both of which can be used with FloWizard. The first one is referred
Guests connect to the hosts session
to as WYSIWIS (What You See Is What I See). Whatever is shown on the hosts desktop is broadcast to all guests. The host can also cede some control of his computer to others. The WYSIWIS approach is quite generic, and application independent. However, because broadcasting is data intensive, it is not suited for applications that involve large models, complex 3-D graphics, moving objects, or animations. Rather, it is most commonly used for simple applications such as online slide presentations or reviewing meeting notes, for example. Third party software is required for this type of collaboration. The other approach is to create a state level mode of collaboration between users. This approach uses tools that are built-in to the standard release of FloWizard. With this approach, whenever the state of the model is changed, an updated set of data is broadcast to all guests. All local interactivity (rotate, zoom, pan) with the model happens on the users machines independently. This approach offers a much higher level of interactive response for models with large data sets, while retaining the benefits of synchronization. It is, therefore, much more useful for data intensive workflow processes, such as CFD model reviews. With FloWizard, not only is CFD modeling itself much easier, but the whole design process is streamlined from start to finish. The new collaboration feature, in particular, will help the successes and efficiencies of product design teams. s
40
product news
Taking
Meshing
Level
GAMBIT
In GAMBIT 2.2, planned for a summer 2004 release, the focus continues to be on CAD import, geometry conditioning, and automatic meshing. Several new features have been added to help streamline CAD import. The IGES, CATIA V4, STEP, and Parasolid CAD translators can now be upgraded for 2.2 and subsequent releases of GAMBIT. A new smoothing algorithm, together with enhanced healing, will provide high quality surfaces after import. Holes in the model can easily be patched with the muchenhanced cleanup hole tool, and duplicate faces are automatically removed or connected, providing a clean geometry, ready for volume creation. Multivolume stitching is now fully automatic and the fluid domain can now be obtained easily using GAMBITs powerful Boolean operators. For disconnected assemblies, GAMBITs improved face connect capability seamlessly hooks parts together, making them ready for the fully automatic conditioning tools. These tools make for quick removal of unwanted features such as short edges, sliver faces, and sharp angles. The art of providing an optimal mesh density variation is left to the size functions in GAMBIT 2.2. With speed improvements ranging from fifty to several hundred percent, and a number of new functionalities (meshed source and 2D curvature, for example), GAMBIT is a state-ofthe-art meshing tool. In addition, the automatic unstructured quad paver, as well as the boundary layer capability have seen several enhancements related to functionality, quality, and stability. s
to a New
By Erling Eklund, GAMBIT Product Manager
This GAMBIT mesh illustrates improved quad paving using the new 2D curvature size function and HexCore meshing using the surface mesh size function
TGrid
TGrid 3.6 was released in March with a new HexCore meshing capability, previously only available in GAMBIT. Passing over forty industrial test cases using the HexCore technology, with cell counts reaching 7 million, TGrid 3.6 continues to be the best-in-class volume mesher for the automotive industry. Mesh density distribution can be controlled both locally and globally, and the Cartesian mesh is automatically adjusted to prism layer growth for minimal cell count at maintained quality levels. The combination of overall cellcount reduction and increased cell quality results in better CFD solutions and increased productivity. s
The new HexCore meshing capability in TGrid is illustrated for a cabin interior
41
product news
Whats New in
By Nicole Diana, FLUENT Product Manager
FLUENT 6.2
produce more realistic results in certain circumstances. The granular model is now available with the mixture model, providing an economical alternative to the full Eulerian model when appropriate. Several improvements have been made to the VOF model. A new variable time-stepping scheme provides improved robustness and efficiency for cases where the velocity changes significantly with time, as with tank sloshing and filling. An interface-capture scheme provides more efficient convergence for free surface problems in which the final steady state solution is more important than intermediate solutions, as is common in certain ship hydrodynamics applications. An open channel boundary condition and support for inviscid flows further improves FLUENTs capabilities for marine applications. A new wall film model has been added to FLUENTs suite of spray models. It is particularly important for in-cylinder combustion simulations. New tracking schemes, automated tracking scheme selection, and error-controlled time-step adaption result in more accurate and efficient particle tracking with the discrete phase model.
he fall release of FLUENT 6.2 promises many exciting new features and capabilities. It focuses on areas such as dynamic mesh, multiphase flows, acoustics, reacting flows, rotating equipment, and heat transfer and phase change. Application areas include automotive in-cylinder flows, positive displacement pumps, store separation and missile launch, bubble column reactors, fluidized bed reactors, flow-induced noise prediction, turbomachinery, fuel cells, underhood flows, and automotive climate control.
Velocity vectors are shown for a 2D rotary engine simulation, solved using the dynamic mesh model
dynamic mesh
While the dynamic mesh model has served as the foundation for automotive in-cylinder applications, it has been widely used for other industrial applications as well, ranging from chemical process to biomedical. Dynamic hanging node adaption, new in FLUENT 6.2, allows for more accurate solutions with lower overall cell counts. The remeshing and layering capabilities have been expanded, and include the use of symmetry boundary conditions with local re-meshing, and a more flexible 2.5-D remeshing technique that is particularly useful for simulating certain types of pumps. A new builtin six degrees of freedom solver is also available for applications with unconstrained motion, including store separation, missile launch, and tank sloshing. Parallel scalability has also been improved.
multiphase flows
A major new improvement for multiphase systems is the ability to include species transport and homogeneous reactions with all multiphase models, including the Eulerian and VOF models. For heterogeneous reactions, the reaction rates must be specified in a user-defined function (UDF); however, the stoichiometry is easily specified in the graphical user interface (GUI). The Reynolds stress turbulence model can now be used in conjunction with the Eulerian model for the prediction of highly swirling multiphase mixtures. Several enhancements have been added to the granular model to
Fischer-Tropsch synthesis in a bubble column, used to produce liquid hydrocarbons from carbon monoxide and hydrogen, is modeled using the Eulerian multiphase model with chemical reactions; the liquid volume fraction (left) and bubble velocity vector field, colored by the dissolution rate of H2 (right) are shown
42
product news
A wet steam model has been implemented in FLUENT 6.2. Two-phase flow in steam turbine applications is one area that will benefit. Cases with shell conduction no longer require encapsulation in parallel; this significantly improves calculation efficiency. Several enhancements have also been made to both the cavitation and heat exchanger models.
solver improvements
Several enhancements to the numerics in FLUENT 6.2 result in improved solver accuracy and efficiency. For the segregated solver, an improved RhieChow interpolation technique has been implemented that can significantly improve accuracy on tetrahedral meshes. In addition, a new, locally 3rd order discretization scheme has been implemented for both the segregated and coupled solvers.
Prominent noise sources on a sedan, computed using the acoustics capabilities in FLUENT
acoustics
FLUENT 6.2 provides an unparalleled suite of acoustics capabilities. The Ffowcs-Williams Hawkings (FW-H) model that was introduced in FLUENT 6.1 has been extended to handle rotating surfaces, so that fan noise can be modeled. When used in conjunction with rotating reference frames, fan noise predictions can be obtained based on a steady-state simulation, providing an efficient and powerful tool. The FW-H model is now also available with the coupled solver. A new set of broadband source models allows acoustic sources based on the results of steady-state simulations to be estimated. These models are practical tools for evaluating design modifications.
time-dependent flows
FLUENT 6.2 contains some significant improvements in non-iterative transient solution methods. A non-iterative PISO solution scheme is now available, and an implicit fractional step method has been implemented for incompressible and mildly compressible flows. Both methods result in significant speed-ups compared with the iterative transient solution method. Neither, however, can be used with dynamic mesh, sliding mesh, reacting, or Eulerian multiphase flows. Some improvements have also been made to the coupled solver that result in better accuracy along with a speed-up in the time required per iteration. Additional modifications were made for unsteady cases that result in improved accuracy and convergence for the coupled solver when implicit time-stepping is used. In particular, the coupled implicit solver now captures pressure wave propagation with accuracy similar to the coupled explicit solver. Further, the number of iterations required for these problems is significantly reduced.
reacting flows
FLUENT 6.2 includes models for diesel autoignition and gasoline spark ignition. (See the article on page 23.) A new stiff chemistry algorithm for the segregated solver is available for laminar flows. When combined with ISAT, laminar finite rate chemistry problems can be solved more quickly and efficiently. Several enhancements have also been added to the NOx model. Surface chemistry mechanisms can be imported in the Surface CHEMKIN format. Another useful new feature is the ability to create custom material databases. This allows material properties and reaction mechanisms that have been specified in one case to be used again in a different case. PrePDF has been completely integrated into FLUENT 6.2. In addition to improved usability, the integration has also resulted in reduced computation times.
turbulence
In the area of turbulence modeling, there is an effort underway to improve the large eddy simulation (LES) capability. FLUENT 6.2 includes several new sub-grid scale viscosity models, including two dynamic models that improve the quality of LES predictions. To help accelerate convergence, an LES simulation can be launched from a steady-state RANS solution using turbulence synthesization. Two new methods for specifying the stochastic velocity boundary condition at inlets, more rigorous than previous offerings, are now available. The usability of the detached eddy simulation (DES) model has also been enhanced.
rotating equipment
A new sliding mesh model is available in FLUENT 6.2 that addresses the robustness issues associated with earlier versions of the model. The improvements also apply to non-conformal interfaces in general. Of particular interest to turbomachinery users is automated full multigrid initialization. This feature allows an initial solution field to be computed so that a higher Courant number can be used with the coupled implicit solver, resulting in faster overall convergence. Non-reflecting boundary conditions are now supported for unsteady flows.
and more
FLUENT 6.2 also includes add-on modules for proton-exchange membrane (PEM) fuel cells and solid oxide fuel cells (SOFC). The modules will be included on the FLUENT 6.2 CD, but a special license will be required to activate them. There are several other enhancements in FLUENT 6.2, including substantial improvements to the speed of reading cases files into the parallel version of FLUENT, new particle animation capabilities, the ability to display multiple monitors in a single graphics window, and more. s
heat transfer and phase change

A new solar load model has been implemented in FLUENT 6.2. The solar load model allows the radiant energy transport between the sun and a wall surface to be solved, based on the intensity and position of the sun. This model is important for automotive and building climate control applications. For the discrete ordinates (DO) model, a second order discretization option is now available for obtaining more accurate results, particularly for cases with high optical thickness. A UDF can be used to customize the absorption coefficient for each band with the gray-banded DO model. Several improvements have also been made to the surface-to-surface radiation model, including the ability to specify the number of faces per surface cluster for different boundaries, allowing fine clustering to be used only where it is needed. The result is a more accurate solution without the memory penalty associated with using fine clustering everywhere.
Solids distribution in an unbaffled stirred tank containing four LIGHTNIN A310 impellers, solved using the Eulerian granular multiphase and Reynolds stress models
43
support corner
The Fluent
By Kiran Kumar, Director of Support, Fluent Inc.
User Services Center

he Fluent User Services Center (USC, www.fluentusers.com) is a convenient gateway to a host of product and support resources that can help make the most out of your Fluent software simulations. It provides 24/7 access to the latest software releases, and all the information needed to ensure success when solving engineering problems. In this article, some tips on finding and using the many resources on the USC are presented. To access the site, you can register online at www.fluent.com/ support/fuscreg.htm.
Maximize the Return on your

T
product downloads and information

The latest releases of Fluent software products can be downloaded from the site as long as your current license contract entitles you to software upgrades. You can access the software download center either on the home page (under Services) or from each of the product pages under the Product Information section. An installation guide as well as answers to frequently asked installation questions are available from the software download center. To learn what is new about a particular software release, review the Release Information on the appropriate product page. Release notes, feature specifications, and lists of fixed and known defects will help bring you up to speed on the new version. Details on how to use the new and existing features are provided in the Users Guide, Tutorial Guide, and other manuals that are available under Documentation on the product pages. Additional application- and feature-specific tutorials and examples are also available to help you apply the new or enhanced models to your engineering problems. In addition to the lists of fixed and known defects, a searchable database of defects is available on each product page under Defects/Workarounds. Here, you can search for defect reports that may be related to a problem you have encountered. The defect reports will also tell you the software version in which the defect has been fixed or will be fixed and any known workarounds.
The USC home page
training resources
Training is essential for keeping up with, and learning to effectively use the ever-expanding capabilities of the software. The USC provides a wide variety of training resources under Training@LearningCFD.com on the home page. Whether you are a new user or an experienced one, you will find many useful training materials here. These materials include presentations or lectures, hands-on tutorials, as well as archived live presentations that contain slide shows with embedded audio of the speakers voice. These materials have been organized by product and course level to help you find the most appropriate materials. Simply select the product, course level, and type of material from the drop-down lists on the CFD Learning Center page to access the available options.
The CFD Learning Center home page
44
support corner
Fluent Software Investment

support resources
From time to time, you may have questions or run into problems that need to be resolved. In addition to the Documentation and Defects/Workarounds resources described earlier, the USC has other support resources that can be very valuable in helping you obtain timely answers to your questions. Frequently Asked Questions This link, present on the home page, will take you to a collection of answers to frequently asked installation, usage (how to), and troubleshooting questions. Online Technical Support This portal enables you to a) access our knowledge base of solutions on various topics and b) request support from your local Fluent support team and monitor the status of your support requests. The portal communicates in real time with a support database that contains solutions to product issues and answers to various how to questions contributed by our worldwide technical staff. Use the Search Solutions page on this site to access the wealth of technical information available. When you request support by creating a Case, it is automatically forwarded to your local Fluent support team. You will then be notified via e-mail when your case is updated. You can further communicate with the support staff by adding notes to your Case, and you can close it if it is resolved to your satisfaction. Detailed instructions on using this portal are available in a Quick Reference Guide on the home page of the portal. Discussion Forums these are user community forums where you can discuss topics of mutual interest, and exchange information with other Fluent software users around the world. Be sure to take advantage of the wealth of experience contained in this extensive user community.
The Online Technical Support portal home page
more services and resources

While solving your CFD problem, you may also need access to technical papers in your area of interest, published by experts in the CFD and general fluid mechanics community. A new Online CFD Literature Database (CFDLit) is available on the USC under the Resources section of the home page. CFDLit is a unique collection of technical literature relevant to computational fluid dynamics, created by Fluent engineers. It consists of papers published in academic and industry trade journals, conference proceedings, and book chapters that are useful in the development and support of Fluent products. The database is offered for reference and research as a service to Fluent customers. You may search and view the authors, titles, and abstracts of any of the papers in the database and request copies. More services such as the Fluent Remote Simulation Facility (RSF) and eLicense (Realtime On-Line Licensing System, or ROLLS) are available are available under the Services section of the home page. The RSF provides you with surge compute capacity by enabling you to run FLUENT calculations on a large computer cluster located at Fluent via a secure internet connection. The ROLLS service is a new online overdraft licensing system that allows existing Fluent software customers to acquire surge license capacity in the form of short term extra licenses to cover peak demands. s
The CFDLit home page
45
around Fluent
Fluent Worldwide
Corporate headquarters
Fluent Inc. 10 Cavendish Court Lebanon, NH 03766, USA Tel: 603 643 2600 Fax: 603 643 3967 Email: info@fluent.com
Fluent India Moves to Pune Infotech Park

By Sajitha Jayaprakash, Fluent India
USA regional offices

Ann Arbor, MI 48104 Tel: 734 213 6821 Evanston, IL 60201 Tel: 847 491 0200 Santa Clara, CA 95051 Tel: 408 522 8726 Morgantown, WV 26505 Tel: 304 598 3770
European regional offices

Fluent Benelux Wavre, Belgium Tel: 32 1045 2861 Email: info@fluent.be
luent initiated its operations in India in 1994 with a team of five engineers. Today, Fluent India is the fastest growing business unit of Fluent worldwide. With a staff of over 110, it is the largest CFD organization in India and Southeast Asia. To cater to its growth, Fluent India has moved into a new facility at the Pune Infotech Park. Present at the inauguration on February 4, 2004 were Mr. Richard Rothman (US Trade Commissioner), Mr. Shyamal Ghosh (Administrator, Universal Service Obligation Fund, Department of Telecommunication), Dr. Ferit Boysan (COO, Fluent), Mr. Peter Christie (CFO, Fluent) and Mr. Brian Drew (VP, Development, Fluent).
The world-class facility is spread over 130,000 sq.ft, with a building area of 40,000 sq.ft., designed to accommodate 160 people. Adequate space has been allocated for future expansion. The new infrastructure consists of massive computational power and a high-speed internet link of 2.5 Mbps. This one-of-a-kind facility even boasts a beautiful landscaped butterfly garden surrounding the building. Fluent Indias activities include CFD software development, applications consulting, technical support, product quality assurance, documentation, sales, and marketing. In addition to its support of Fluent offices worldwide, the company serves Indian and Southeast Asian markets directly. s
Fluent Deutschland GmbH Darmstadt, Germany Tel: 49 6151 36440 Email: info@fluent.de Fluent Europe Ltd. Sheffield, England Tel: 44 114 281 8888 Email: info@fluent.co.uk Fluent France SA Montigny le Bretonneux, France Tel: 33 1 3060 9897 Email: info@fluent.fr Fluent Italia Milano, Italy Tel: 39 02 8901 3378 Email: info@fluent.it Fluent Sweden AB Goteborg, Sweden Tel: 46 31 771 8780 Email: info@fluent.se
Upcoming UGM Dates

ATES Korea: November 10-12 Fluent Asia Pacific: November 18-19 China: to be announced US Fluent UGM: June 8-10 FLUENT, FIDAP, POLYFLOW, Automotive Fluent India: Planned for Q4, 2005
Asian regional offices

Fluent Asia Pacific Co., Ltd. Tokyo, Japan Tel: 81 3 5324 7301 Email: info@fluent.co.jp Osaka, Japan Tel: 81 6 6445 5690 Fluent India Pvt. Ltd. Pune, India Tel: 91 20 4142500 Email: info@fluent.co.in
Distributors
Anova Ltd. Turkey ATES Korea Beijing Hi-key Technology Corporation Ltd. China & Hong Kong Cavendish Instruments de Mexico, S.A. de C.V. (CIM) Mexico, Venezuela, Argentina, Chile, Colombia FEM++ Israel (POLYFLOW only) Fluid Codes Ltd. UK (serving Middle East) INNOTECH Ltd. Hungary J-ROM Ltd. Israel LEAP Australia Pty., Ltd. Australia & New Zealand Process Flow Finland & Baltics Regional Technologies Corp. Ukraine Simcon International (Pvt.) Ltd. Pakistan SimTec Ltd. Southeastern Europe SMARTtech Services & Systems, Ltd. Brazil SymKom Poland Taiwan Auto-Design Company (TADC) Taiwan Techsoft Engineering s.r.o Czech Republic & Slovak Republic TENSOR SRL Romania Thermal Technologies/QFINSOFT South Africa
Fluent Sweden: September 20-21 Fluent Europe: September 23-24 Fluent Germany: September 29-30 Fluent Benelux: October 5-6 Fluent France: October 19 Fluent Italy: October, to be announced Fluent in Spain: October, to be announced
46

Fluent Spring 2004

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Fluent Spring 2004

Hochgeladen von

Copyright:

Verfügbare Formate

VOL XIII ISSUE 1 SPRING 2004

Healthcare Industry Supplement Inside!

Fluent News is published by

Liz Marshall fluentnews@fluent.com

healthcare industry supplement

Oilflow lines over the body colored by dynamic pressure

Fluent NEWS spring 2004

Speedo Goes for Gold with CFD

The CFD modeling we have done has

Fluent NEWS spring 2004

Speedo Fastskin FSII swimsuit in action

Fluent NEWS spring 2004

Alive & Kicking

Fluent NEWS spring 2004

By Keith Hanna, Fluent News

Pulse of the Olympics

Fluent NEWS spring 2004

Fluent NEWS spring 2004

The Heat is On But Not at Fluents

Fluent NEWS spring 2004

Plan view of temperature contours two feet above the floor

Fluent NEWS spring 2004

Shooting the Breeze

A schematic of the proposed addition

By Erik Olsen, Andrew Gilles, and Jessica Wojtas, OWP/P, Chicago, IL

Fluent NEWS spring 2004

Feeling the Breeze on Arabian Nights

A typical wind tower house in Dubai

Fluent NEWS spring 2004

Snow Drift Modeling

Fluent NEWS spring 2004

Fluent NEWS spring 2004

Nuclear Power Plant

Map and computational domain of the overall river model

Temperature contours on the water surface

0 depth (ft) 5 10 15 20 70 75 80 temperature (F) 85 90 measured FLUENT prediction

Temperature comparison at one of the downstream monitoring stations

Fluent NEWS spring 2004

The model of buildings prepared for FLUENT

100 normalized mean concentration wind tunnel FLUENT prediction 10

R9 R11 R13 R15 U1 measurement points

Fluent NEWS spring 2004

Fluent NEWS spring 2004

Pulp Fiction to Pulp Fact in Paper Recycling

Computer model of an 8-ft pulper with a new low-power rotor

An 81.6-inch-diameter Vortech rotor installed in a 10D Hydrapulper

Fluent NEWS spring 2004

Modeling Clothing and Fabrics with FLUENT

Circular pattern of chemical agent droplets on a fabric swatch in a test cell

Fluent NEWS spring 2004

Notebook Computer Batteries Meet the

discharge power 10.0 W

1500 time (sec)

Fluent NEWS spring 2004

100 mass fraction of air left in FOUP (%)

plenum inlet purge 600 time (sec) 900

Comparison of different FOUP nitrogen purge times

Fluent NEWS spring 2004

high mobility ions

Fluent NEWS spring 2004

Fluent NEWS spring 2004

Heart Valve Dynamics

Fluent NEWS spring 2004

FSI Through Partnerships