Engineering Failure Analysis 32 (2013) 115

Contents lists available at SciVerse ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Failure analysis of an aircraft engine cylinder head

Branimir Krstic a,, Bosko Rasuo b, Dragan Trifkovic a, Igor Radisavljevic c, Zoran Rajic c,
Mirko Dinulovic b
a
University of Defence in Belgrade, Military Academy, Generala Pavla Jurisica Sturma 33, 11000 Belgrade, Serbia
b
University of Belgrade, Faculty of Mechanical Engineering, Kraljice Marije 16, 11000 Belgrade, Serbia
c
Military Technical Institute, Ratka Resanovica 1, 11000 Belgrade, Serbia

a r t i c l e i n f o a b s t r a c t

Article history: The piston engine of the training aircraft malfunctioned during the ight due to the crack-
Received 3 March 2013 ing of its cylinder head (CH), which is manufactured from an aluminum casting alloy. Based
Accepted 6 March 2013 on the fractographic examination of the mating fracture surfaces, the characteristic ratchet
Available online 15 March 2013
and beach marks were observed indicating the occurrence of fatigue failure. The crack was
initiated from multiple origins located on the inner ange llet on the exhaust side of the
Keywords: CH. The metallography examination has shown that the fatigue was promoted from pre-
Aircraft
existing material defect due to an elevated presence of shrinkage pores at the crack initi-
Cylinder head
Failure analysis
ation zone and was most likely associated with the manufacturing process of casting. The
Fatigue failure nite element (FE) method, utilized to determine the stress state of the CH subjected to gas
Finite element analysis pressure, also conrmed that the crack origin was located at the most stress area.
2013 Elsevier Ltd. All rights reserved.

1. Introduction

Apart from pilot errors, the majority of aviation accidents have been caused by failures of aircraft structure or failures of
some technical systems. In recent years, an increasing attention has been given to the investigation of different aircraft tech-
nical systems contribution to the aircraft losses and their potential for failure that lead to repeating incidents and accidents.
One of the most important technical systems in aircraft is an engine, nevertheless if the aircraft is propelled by a jet or a
piston engine. The extremely changeable operating conditions for components in aircraft engine together with their age
and maintenance procedure can inuence failures.
Bhaumik et al. [1] have investigated the reason of crankshaft fracture in the transport aircraft piston engine. They have
concluded that the shaft had been subjected to high contact pressure and cyclic loading leading to pitting and spalling at the
region of the journals where the fatigue cracks appeared. Also, fatigue was found as a major cause of failure of many aircraft
piston engine components like connecting rods [2]. In this concrete case, it has been reported that enormous friction
between the bearing and the crank pin had contributed to fatigue crack initiation.
In addition to fatigue, some material discontinuities caused by a casting process can attribute to failure appearance in
engine parts. The case study of cracking in a gray-iron cylinder head has revealed that high operational stresses of the engine
acting on a material microporosity can produce cracking in the area adjacent to the vent plug [3]. There are a few of reasons
for porosity in casting material among them inadequate pouring temperature is predominant.
The training aircraft Utva-75 (Fig. 1), with two pilots on board, was on a training ight when pilots reported an engine
problem. Shortly after takeoff, pilots heard a very loud bang and the engine began to run rough and vibrate. The engine

Corresponding author. Address: University of Defence in Belgrade, Military Academy, Belgrade, Department of Military Aviation, Generala Pavla Jurisica
Sturma 33, 11000 Belgrade, Serbia. Tel.: +381 11 3603 365; fax: +381 11 3005 190.
E-mail address: branimir.krstic@va.mod.gov.rs (B. Krstic).

1350-6307/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.engfailanal.2013.03.004
2 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115

Fig. 1. Utva-75 training aircraft.

working parameters degraded and pilots decided to return at the airport. After successful forced landing at the airport, the
engine was removed from the aircraft and signicant damage to the engine was observed. However, no damage to the air-
craft structure has been found. An examination of the aircraft engine revealed a crack in cylinder head (CH), between the fth
and the sixth cooling ns (counting from the barrel). The damaged cylinder was sent to the Military Technical Institute,
Structural and Technical Materials Laboratory to determine the cause of the failure.
Similar two cases happened in Canada. First one, in June 2009 the airplane Glastar struck trees and the engine was
found with one cylinder head separated from the barrel. The engine was an Aero Sport Power which is identical to the
Lycoming O-360-A2A model. It has been reported by Transportation Safety Board of Canada Operational Services Branch
that the separation of the cylinder head occurred due to fatigue cracks developed at the thread. Stress concentration
formed by the sharp corners of the barrel threads led to initiation of metal fatigue [4]. The next year, in May 2010, Cirrus
Design SR20 aircraft crashed onto the roof of a building due to the same problem, i.e. cylinder head separation from the
cylinder head barrel. The aircraft engine was Teledyne Continental Motors model IO-360-ES. Engineering examination sug-
gested that the cylinder head separated as a result of the instantaneous overstress extension of a pre-existing high-cycle
fatigue crack. Combination of corrosion damage and a sharp corner on the aluminum casting thread contributed to the fa-
tigue initiation [5].
Damaged cylinder from Utva-75 had been installed on the engine in the number one position (Fig. 2). By the time of the
occurrence ight, it had accumulated 1560 ight hours since new and 360 h since overhaul. The recommended time between
overhauls for this type of engine is 2000 h. The last periodic (50 h) compression test was conducted at 1545 h of total ight
time or 15 h before accident. The result of differential pressure test is presented in Table 1.
The Utva-75 training aircraft is propelled by four cylinder, direct drive, horizontally opposed, air cooled engine. The cyl-
inders are of conventional air cooled construction with the two major parts, head and barrel, screwed and shrunk together.
The heads are made from an aluminum casting alloy with a fully machined combustion chamber. The cylinder barrels are
machined from chrome nickel molybdenum steel forging with deep integral cooling ns. The interior of the barrels are
ground and honed to a specied nish. The engine specications [6] are presented in Table 2.

Fig. 2. Utva-75 engine.

 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115 3

Table 1
Compression test.

Cylinder
Cylinder no. 1 Cylinder no. 2 Cylinder no. 3 Cylinder no. 4
Compression reading (psi) 64 72 74 72

Table 2
Engine specications.

Parameter Value
Rated horsepower 180 hp
Rated speed 2700 rpm
Bore 0.130 m
Stroke length 0.111 m
Volumetric displacement 5.893 L
Number of cylinders 4
Compression ratio 8.5:1
Firing order 1-3-2-4
Spark occurs 25 before top dead center
Propeller drive ratio 1:1
Propeller drive rotation (viewed from rear) Clockwise
Weight 136.5 kg

2. Failure analysis procedure

This section presents the failure analysis procedure employed for determining the root cause of the failure of aircraft en-
gine CH.
In order to obtain information about general state of the failed cylinder assembly (CH and cylinder barrel), it was sub-
jected to the external visual inspection by unaided human eye in the as-received condition. The next step was a detailed
visual examination of the fracture surfaces. Since the upper part of the CH was not completely separated from the lower part,
the CH was mechanically opened under laboratory conditions to perform this task. Once this visual examination of fracture
surfaces by means of unaided human eye had been nished, the fracture surface attached to the upper part of the CH was cut
off into ring shape to allow introducing it under stereomicroscope. The macroscopic analysis of the fracture surface was
undertaken using Leica M205A stereomicroscope at low magnication (650). This macroscopic examination may provide
information about failure mechanism and the most interesting areas for further analysis. The obtained ring with fracture sur-
face was subsequently cleaned ultrasonically in a citrus-based liquid degreaser followed by cleaning in acetone and ethanol
to eliminate residues produced by the combustion. The central portion of the fatigue fracture surface was cut off from the
ring and observed using scanning electron microscopy on a JEOL JSM-6610LV low vacuum, high-performance scanning elec-
tron microscope (SEM). The metallographic specimen was taken perpendicular to the fracture surface, at the location which
approximately corresponds to the main fatigue crack origin. The metallographic preparations included wet mechanical
grinding, using silicon carbide papers down to 2500 grit. This was followed by polishing using 7/5 lm diamond paste and
nal polishing with 3/2 lm diamond paste. The sample was examined in the as-polished condition under LeitzMetalloplan
light microscope. The objectives of microstructural analysis are to determine the microstructures of the body of CH and to
reveal the presence of any imperfections that might have led to failure. The chemical composition of the CH material was
determined by X-ray uorescence spectroscopy (XRF) chemical analysis method. The obtained results were compared with
standard specication for aluminum casting alloy specied for production of air cooled cylinder heads. The Brinell hardness
test of the material was conducted using a 2.5 mm diameter hardened steel ball with a 612.9 N load applied for a period of
30 s (HBS 2.5/62.5/3000 ). The hardness measurements were taken from the ring plane, opposite to fracture surface, and the
measurement points were radially distributed around the ring circumference. The identication of stress concentration areas
of the CH was done by applying linear nite element (FE) method. For this purpose, 3D solid model, identical to the physical
model of CH, was developed using CATIA V5R22. The nominal load of the engine was applied.

3. Results

3.1. Visual inspection

The cylinder assembly was photographed as-received and shown in Fig. 3a. The external inspection of damaged cylinder
assembly in the as-received condition revealed a clearly visible crack in the CH, between the fth and the sixth cooling ns
(Fig. 3b) with approximately 0.5 mm width. Actually, after detailed visual observation of CH, it was concluded that revealed
clearly visible crack was just a minor part of the crack which was extended about 80% of the CH circumference. Also, it should
4 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115

Fig. 3. (a) Cylinder assembly photographed as-received and (b) close-up view showing crack in the CH between the fth and the sixth cooling ns (region
marked in (a)).

be indicated that the crack propagation plane runs through the position of the top plane of cylinder barrel (Fig. 4). On the
other hand, any damage to the cylinder barrel or irregularities was not noticed.

3.2. Macrofractography

In order to determine the root cause of the crack in the CH, the failure analysis was continued with the visual examination
of the fracture surfaces. For this purpose, the crack was mechanically opened under laboratory conditions and the obtained
parts, Part I and Part II, with associated fracture surfaces are shown in Fig. 5. In particular, Part I represents the upper part of
CH fractured and separated from the cylinder assembly and Part II is the remaining part of CH which stayed screwed and
shrunk together with cylinder barrel, respectively. Visual examination of the mating fracture surfaces was performed with
unaided human eye and with Leica M205A stereomicroscope at low magnication (650).
The fracture surfaces revealed the presence of ve characteristic morphologies (Fig. 6), marked as A, B, C, D and E. The
sector A, located on the exhaust, deep nned, side of the CH, showed a smooth and at surface as well as clearly visible ratch-
et marks, which represent steps between fatigue crack regions propagated from different origins on slightly different planes
(the larger of them are marked with a white arrowheads in Fig. 7). This sector corresponds to the area over which the crack
propagated slowly and has a distinct arc-shaped boundary with sector B. The segment of fracture surface which contains
sector A was cut off from the ring and examined under stereomicroscope. This examination conrmed that the fatigue crack
initiated from different origins located at the inner edge of CH (Fig. 8). The macroscopic fracture surface of sector B (Fig. 9) is
at approximately 45 to the CH axis, which suggests ductile fracture under mostly tensile load. The narrow rim of overstress
fracture is an integral part of this area and extends around the entire outer boundary of the slant fracture. Also, visual obser-
vation revealed dull and rough texture of the slant fracture surface. However, in this stage of fracture life, crack depth
reached the thickness of the CH wall. Feature characteristic of fatigue fracture, beach marks, can be clearly distinguished

Fig. 4. (a) Schematic drawing of cylinder, piston and valve components (adapted from [7]) and (b) close-up view showing crack position in the CH (region
marked in (a)).
 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115 5

Fig. 5. Macrograph of the Part I and Part II of CH obtained after crack was mechanically opened.

Fig. 6. Macrograph showing characteristic morphologies on fracture surface. Part II.

Fig. 7. View of the sector A of the fracture surface depicting ratchet marks (arrowheads). Part II.

Fig. 8. Stereomacrograph showing the sector A of fracture surface with ratchet marks (arrowheads). Part I.
6 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115

Fig. 9. Closer slightly different low angle view of the sector B showing the place where crack reached the thickness of the CH wall (red arrowhead). Part II.

in sector C and are indicated in Fig. 10. The fracture surface of unstable fracture (sector D) exhibited a rough, dull with coarse
grainy morphology, where the overload was the main fracture mechanism. This nal overstress fracture of the CH wall
occurred when the total area of crack had reached a critical size. Finally, the sector E is area of mechanically forced open
fracture characterized by the bright appearance resulting from the reective qualities of cleaved crystals. This sector is rep-
resenting about 20% of the CH cross section. Also, the visual inspection disclosed a varying degree of exposure the fracture
surface to the combustion gases. Consequently, areas of apparent fatigue and overstress fracture (sectors A, B, C and D) were
signicantly contaminated by these gases.

3.3. Microfractography

In order to obtain detailed information about mechanism leading to failure, the fracture surface attached to Part I was
observed by SEM. However, although the fracture surface was thoroughly cleaned before introducing it under SEM, the
examination revealed signicant amounts of deposits on the entire fracture surface (Fig. 11), except the area of mechanically
forced open fracture. Deposits, visible on SEM images, were analyzed by energy dispersive spectroscopy (EDS). They con-
tained high concentrations of lead and bromine (Fig. 12), which is typical for aviation fuel combustion products. Aviation

Fig. 10. Photographs showing fatigue beach marks in sector C. Part II.
 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115 7

Fig. 11. Morphology of the central portion of the fatigue fracture surface covered by deposits.

Fig. 12. EDS spectrum from deposit on the fatigue fracture surface (region marked in Fig. 11).

gasoline normally contains lead, while bromine is an additive which scavenges lead by forming volatile lead bromide [8].
Due to the inability to remove the deposits from the most interesting sectors (sectors A, B and C) of the fracture surface, using
nondestructive methods, it was impossible to obtain signicant information.

Fig. 13. Fatigue fracture surface shows evidence of ratchet marks (black arrowheads).
8 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115

Fig. 14. Transition from fatigue to ductile overstress fracture.

On the other hand, regardless of the presence of deposits on fracture surface, close to the crack origin zone, SEM obser-
vations revealed the ne ratchet marks along the entire fracture initiation region on the CH inner surface (Fig. 13). These
observations also conrmed the existence of numerous fatigue fracture origin sites, which suggest that it was not a single
particular defect that caused fatigue crack nucleation. In addition, transition from fatigue to ductile overstress fracture on
the boundary between sectors A and B is shown in Fig. 14.

3.4. Metallography

A longitudinal metallographic section oriented perpendicularly to the fracture surface was prepared at the location which
approximately corresponds to the main fatigue crack origin (indicated by 1yellow dotted line in Fig. 15). The microstructure
revealed the typical structure of an aluminumcopper casting alloy with intermetallic phases precipitates located at the grain
boundaries [9] (Fig. 16). Metallographic examination discovered a presence of casting defects in the CH material. A large amount
of microporosity, in form of grouped shrinkage pores surrounded by primary dendrites and eutectic phase [10], was found near
to the CH surface (Fig. 17). Some of these microporosities reached approximately 400 lm in length, as it shown in Fig. 18. Inter-
granular cracks, formed by linking up of other shrinkage pores in material, were also observed (Fig. 19).

3.5. Chemical composition

The chemical analysis of specimens sampled from the undamaged part of the CH was carried out using X-ray uorescence
spectroscopy (XRF) chemical analysis method, and its result is presented in Table 3. The result shows that the chemical com-
position matches the standard specication for aluminum casting alloy 242.0 (according to American National Standard
Institute designation system). It should be noted that the aluminum casting alloy 242.0 is specied for production of air
cooled cylinder heads [11].

3.6. Hardness measurements

Hardness measurements were carried out on the cross section of the CH (ring plane opposite to the fracture surface) in
accordance with the Brinell standard method. The measurement points were radially distributed around the CH wall circum-
ference at the equal angular spacing of 45 (blue points in Fig. 20). The obtained hardness data leaded to a value of 111HB
(measurement point 6) that is signicantly higher than those recorded in other measurement points. Since the location of
the measurement point 6 corresponds to the central part of the most interesting area of the fracture surface, additional mea-
surements were performed at points in the vicinity of the measurement point 6 (red points in Fig. 20). The nal hardness
data that were obtained from the testing appear in Table 4.

3.7. Finite element analysis

To identify the stress state of the CH, linear FE method was applied on FE model. The CH FE model was developed using
CATIA V5R22 Generative Structural Analysis Workbench. Continuum of the previously modeled 3D geometry of the CH
(Fig. 21) was meshed by a 4-nodes linear tetrahedral solid elements generating FE model of 698,737 nodes and 3,044,753

1
For interpretation of color in Figs. 9, 15, 20 and 26 the reader is referred to the web version of this article.
 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115 9

Fig. 15. View of the fracture surface with location where metallographic sample was taken. Part I.

Fig. 16. Optical micrograph of the CH microstructure.

Fig. 17. Microporosity of CHs material in a form of grouped shrinkage pores.

elements, Fig. 22. Mechanical properties of the material, applied to the 3D geometry of the CH, were taken from [12] for
242.0 aluminum casting alloy.
The inner surface of the CH was affected by gas pressure which was represented by the mean effective pressure. Based on
classic models for internal combustion engine and using available information of the engine shown in Table 1, the mean
10 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115

Fig. 18. Typical shrinkage pore (black) surrounded by primary dendrites (white) and eutectic phase (gray needles in a white matrix).

Fig. 19. Metallographic section showing the intergranular crack due to the linking of the shrinkage pores.

Table 3
Cylinder head chemical composition (wt.%).

Element
Cu Mg Mn Si Fe Ni Ti Cr Zn Al
Cylinder head 4.28 1.65 0.03 0.22 0.24 1.90 0.22 0.20 Bal.
Standard specication 3.54.5 1.21.8 60.35 60.7 61.0 1.72.3 60.25 60.25 60.35 Bal.

effective pressure (998 kPa) was calculated. This pressure was applied to all entities of the inner surface of the CH geometry,
which are in contact with combustion gases. The obtained value of mean effective pressure is within the range of typical
values of the mean effective pressure for naturally aspirated spark ignition engines [13]. The appropriate contact denitions
have been assigned to the contact zones of the members of the cylinder assembly [14]. In accordance with this, the nodes of
the lowest cross-section of the CH were xed by restraining all degrees of freedom. The displacements were free in the rest of
the CH.
The stress eld of the CH obtained according to the Von Mises criterion and visualization of Von Mises stress eld patterns
is shown in Fig. 23. Finite element analysis conrmed that the origins of the crack in the CH are located at the highest stress
area. The stress maximum value appears at the llet entity on the inner surface of exhaust side of the CH, Fig. 24. The node
where stress reaches the maximum value corresponds to the main fatigue crack origin, Figs. 25 and 26.

4. Analysis of the results

Fractography features revealed fatigue as the main cause of the failure of the CH. Tracing the fatigue evidences in form of
ratchet marks, the crack origin region was determined. The fatigue originated from multiple origins located on the inner
 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115 11

Fig. 20. Distribution of the hardness measurement points around the CHs wall circumference.

Table 4
Hardness measurements.

Measurement point
1 2 3 4 5 50 500 5000 6 60 600 6000 7 8
Hardness (HB) 71 75 76 76 86 93 100 102 111 102 96 92 87 71

Fig. 21. 3D model of the CH.

Fig. 22. FE model of the CH.

12 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115

Fig. 23. FE model showing the highest stress area of the CH.

Fig. 24. FE model showing the nodes with minimum and maximum stress values.
 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115 13

Fig. 25. Stress distribution on the longitudinal section of the FE model through the node with maximum stress value.

Fig. 26. Stress distribution on the cross section of the FE model through the node with maximum stress value (red arrowhead).

ange llet of the CH, which suggests that it was not a single particular defect that caused fatigue crack nucleation. Also, the
FE analysis conrmed that the fatigue crack origins were in correspondence of the most stressed area of the CH. Merging
initially occurred cracks into the single one, the crack front prole grew across the CH wall with a semi-elliptical surface
shape. The fatigue crack propagated slowly in radial direction up to about 30% of the wall thickness of the CH. Since the fa-
tigue crack covered substantial portion of the local CH wall thickness, the load capacity of the remaining local cross section of
the CH wall has been signicantly reduced. Further, the reduce load capacity combined with the local material porosity con-
tributed to the crack advance up to the outer surface of the CH by the overload mechanism. It was not found the evidence of
abnormal combustion that could boost overload fracture as well. When the crack had reached the outer surface of the CH, the
gases from the combustion chamber started to leak outside through the crack. Therefore, the pressure loads were reduced
and the process of overloaded fracture was stopped. It is most likely that the engine continued work in that condition for a
while. The evidence for this statement is a presence of large amount of combustion residue that had been deposited in the
surrounding area of external edge of the crack and on the matting fracture surface. While the gas was leaking it was acted on
matting fracture surfaces in both axial directions accelerating subsequent fatigue appearance. Then nal overstress crack of
the remaining load-bearing cross-section of the CH wall occurred when the total area of two fatigue cracks, on left and right
side, reached the critical size. The nal overstress crack did not totally separated CH into two parts due to sharp decrease of
gas pressure caused by huge leaking. They remained linked by non-fractured 20% of cross-section of CH, which was subse-
quently mechanically opened to examine the fracture surfaces. Metallographic examination discovered a presence of casting
defects in CHs material exactly at the area where cracks initiation began. The most of the observed subsurface shrinkage
pores were in the range 100300 lm. The largest of these reached about 400 lm in length. Intergranular cracks in material
in fracture initiation area were formed by linking up of shrinkage pores. The results of the investigation concerning the fail-
ures of aeronautical and engineering components due to the presence of casting defects are widely presented in scientic
papers [1517]. Such defects, determining the fracture initiation site, compromise the mechanical properties as well as
the fatigue of the material [18]. They have a detrimental effect on fatigue life by shortening not only the fatigue crack prop-
agation, but also initiation period. The decrease in fatigue life is directly correlated with the increase of defect size [1924].
Also, the fatigue life of cast aluminum alloys containing defects can be one or two orders of magnitude lower than in
14 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115

defect-free cast components [2530]. The complex geometry of the CH body, with distinct change in the section thickness,
combined with the casting process parameters (inadequate pouring temperature, insufcient number of pouring gates, inad-
equate gate cross section, excessive pouring temperature, or inadequate feeding) also represent other sources of failure. The
signicant increase of the hardness of the CH material in cracks initiation region is frequently correlated with toughness
loosening and therefore might assist the fast crack propagation [3133]. Chemical composition and microstructure were
found in conformity with those expected from the selected aluminum casting alloy. It is worth mentioning that the others
series of such cylinder assemblies also had a history of the CH separation due to fatigue cracking and was covered by the
manufacturers bulletin [34] and the Federal Aviation Administration Airworthiness Directive [35]. For this purpose, early
detection of the crack in the CH has a crucial importance to prevent loss of engine power and possible engine failure caused
by separation of a CH, which could result in loss of control of aircraft. The manufacturers bulletin precisely denes inspec-
tion procedures to determining the existence of crack in the CH. For damaged cylinder (cylinder no. 1), the result of the last
conducted periodic compression test (Table 1) was not in agreement with the minimum value, required by the aforemen-
tioned bulletin. During compression test the cylinder pressure gauge readings (64 psi) was below the minimum allowable
value of 70 psi. In this case, the soap bubble leak check, as a reliable verication method, should be accomplished, but inves-
tigators had no evidence that it committed. All above leads to the conclusion that is most likely that the crack already had
reached the thickness of the CH, at the time when the last periodical inspection was carried out. Since the aircraft returned in
service after this inspection, the crack continued to grow and it was eventually led to a complete failure of cylinder assembly.

5. Conclusions

The failure investigation evidenced a fatigue failure of the CH, which was promoted by a pre-existing material defect and
identied as an elevated presence of shrinkage pores in the most stressed region on the inner ange llet of the exhaust side
of the CH. The defects were probably caused during the manufactured stage of the casting.

Acknowledgements

The authors would like to acknowledge Military Technical Institute (Ministry of Defence of the Republic of Serbia) for
experimental program. This research has been supported by the University of Defence in Belgrade within the project titled
Contribution to the development of the forensic engineering in the domain of the failure analysis of cylindrical structures.

References

[1] Bhaumik SK, Rangaraju R, Venkataswamy MA, Bhaskaran TA, Parameswara MA. Fatigue fracture of crankshaft of an aircraft engine. Eng Fail Anal
2002;9:25563.
[2] Bhaumik SK, Sujata M, Venkataswamy MA. Fatigue failure of aircraft components. Eng Fail Anal 2008;15:67594.
[3] Snyder JJ. Failure related to casting. In: Failure analysis and prevention, vol. 11. ASM handbook, ASM International; 2002. p. 10355.
[4] Guertsman V, Dionne S, Crosby T. Engineering report number LP085/2009. Transportation Safety Board of Canada; 2009.
[5] Guertsman V, Dionne S, Crosby T. Engineering report number LP077/2010. Transportation Safety Board of Canada; 2010.
[6] Operators manual Textron Lycoming O-360, HO-360, IO-360, AIO-360, HIO-360 & TIO-360 series aircraft engines, Part No. 60297-12, Rev. No. 60297-
12-6, Textron Lycoming; 2000.
[7] Overhaul manual-lycoming direct drive aircraft engines, Part No. 60294-7, Rev. No. 60294-7-10, Textron Lycoming; 2007.
[8] ASTM Standard D910-11. Standard specication for aviation gasolines. West Conshohocken (PA): ASTM International; 2011. DOI:10.1520/D910-11,
www.astm.org.
[9] Zolotorevsky VS, Belov NA, Glazoff MV. Casting aluminum alloys. 1st ed. Elsevier; 2007.
[10] Anson JP, Gruzleski JE. The quantitative discrimination between shrinkage and gas microporosity in cast aluminum alloys using spatial data analysis.
Mater Charact 1999;43:31955.
[11] Rasuo B. Aircraft production technology. Belgrade: Faculty of Mechanical Engineering; 1995 [in Serbian].
[12] Kearney AL. Properties of cast aluminum alloys. In: Properties and selection: nonferrous alloys and special-purpose materials, vol. 02. ASM handbook,
4th ed. ASM International; 1995.
[13] Heywood JB. Internal combustion engine fundamentals. 1st ed. New York: McGraw-Hill; 1988.
[14] Trifkovic D et al. Failure analysis of the combat jet aircraft rudder shaft. Eng Fail Anal 2011;18:19982007.
[15] Riad WT, Hussain BS, Shalaby HM. Cracking of aluminum cast pistons of fuel gas reciprocating compressors. Eng Fail Anal 2010;17:4406.
[16] Handbook of case histories in failure analysis, vol. 1. ASM International; 1994.
[17] Handbook of case histories in failure analysis, vol. 2. ASM International; 1994.
[18] Bagnoli F, Bernabei M, Ciliberto A. Failure analysis of an aircraft auxiliary power unit air intake door. Eng Fail Anal 2011;18:28494.
[19] Wang QG, Apelian D, Lados DA. Fatigue behavior of A356T6 aluminum cast alloys. Part I. effect of casting defects. J Light Metals 2001;1:7384.
[20] Couper MJ, Nesson AE, Grifths JR. Casting defects and the fatigue behavior of an aluminum casting alloy. Fatigue Fract Eng Mater Struct
1990;13:21327.
[21] Petrasinovic D, Rasuo B, Petrasinovic N. Extended nite element method (XFEM) applied to aircraft duralumin spar fatigue life estimation. Tech Gaz
2012;19:55762.
[22] Grbovic A, Rasuo B. FEM based fatigue crack growth predictions for spar of light aircraft under variable amplitude loading. Eng Fail Anal 2012;26:564.
[23] Rasuo B. Experimental techniques for evaluation of fatigue characteristics of laminated constructions from composite materials: full-scale testing of
the helicopter rotor blades. J Testing Eval ASTM 2011;39:23742.
[24] Grbovic A, Rasuo B, Vidanovic N, Peric M. Simulation of crack propagation in titanium mini dental implants (MDI). FME Trans 2011;39:16570.
[25] Yi JZ, Gao YX, Lee PD, Flower HM, Lindley TC. Scatter in fatigue life due to effects of porosity in cast A356T6 aluminumsilicon alloys. Metall Mater
Trans A 2003;33:187990.
[26] Yi JZ, Gao YX, Lee PD, Lindley TC. Microstructure-based fatigue life prediction for cast A356T6 aluminumsilicon alloys. Metall Mater Trans B
2006;37:30111.
[27] Gao YX, Yi JZ, Lee PD, Lindley TC. The effect of porosity on fatigue life of cast aluminumsilicon alloys. Fatigue Fract Eng Mater Struct 2004;27:55970.
 B. Krstic et al. / Engineering Failure Analysis 32 (2013) 115 15

[28] Kuwazuru O, Murata Y, Hangai Y, Utsunomiya T, Kitahara S, Yoshikawa N. X-ray CT inspection for porosities and its effect on fatigue of die cast
aluminum alloy. J Solid Mech Mater Eng 2008;2:122031.
[29] Skallerud B, Iveland T, Harkegard G. Fatigue life assessment of aluminum alloys with casting defects. Eng Fract Mech 1993;44:85774.
[30] Maksimovic S. Fatigue life analysis of aircraft structural components. Sci Tech Rev 2005;1:1522.
[31] Mateo A, Heredero F, Fargas G. Failure investigation of a centrifuge duplex stainless steel basket. Eng Fail Anal 2011;18:216578.
[32] Murakami Y. Metal fatigue: effects of small defects and non-metallic inclusions. Oxford (England): Elsevier; 2002.
[33] McEvily AJ. Metal failures: mechanisms, analysis, prevention. 1st ed. New York: John Wiley & Sons, Inc.; 2002.
[34] ECi Mandatory Service Bulletin, No. 08-1, Rev. 3 of 19 August 2008.
[35] Airworthiness Directive (AD) 2008-19-05, Federal Aviation Administration (FAA) of the USA.