Engineering Failure Analysis 56 (2015) 142149

Contents lists available at ScienceDirect

Engineering Failure Analysis

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

Failure analysis and fatigue performance evaluation of a failed

connecting rod of reciprocating air compressor
M.N. Ilman , R.A. Barizy
Department of Mechanical and Industrial Engineering, Gadjah Mada University, Indonesia

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

Article history: A connecting rod of a reciprocating air compressor is subjected to complex dynamic loads
Received 1 September 2014 therefore it is of a critical machine element. Failure of this type of connecting rod was
Received in revised form 31 December 2014 reported to occur at the rounded llet of the big connecting rod end. The present investiga-
Accepted 11 March 2015
tion is aimed to identify the cause of failure and to evaluate fatigue performance of the
Available online 1 April 2015
failed connecting rod. Factors affecting failure including structural design, type of material
and dynamic loads were assessed using standard failure analysis method. This method
Keywords:
included analysis of chemical composition, microstructural examination using optical
Connecting rod
Rounded llet
microscopy, hardness and tensile tests, scanning electron microscopy (SEM) fractography
Inclusions and stress analysis. To evaluate fatigue performance, fatigue crack growth rate (FCGR) test
Fatigue was performed using a sinusoidal load with a constant load amplitude. Results of this
investigation suggest that the cause of failure was low cycle fatigue and the initial crack
location was consistent with high stress concentration, i.e. llet radius. From metallurgical
point of view, the connecting rod was made of cast steel, not forged steel, with a con-
siderable number of non metallic inclusions such as Al2O3, SiO2 and FeO. These inclusions
which were present near the surface of the rounded llet seemed to act as stress raiser and
they were responsible for crack initiation. In addition, the presence of inclusions could
increase fatigue crack growth rate, da/dN (in m/cycle) as indicated by a high value of
Paris constant (n), typically of 5.2141.
2015 Elsevier Ltd. All rights reserved.

1. Introduction

Reciprocating compressors nd their applications in gas pipelines, chemical plants, natural gas processing plants and oil
reneries. A reciprocating compressor is a device that converts power from prime movers such as an electric motor, a diesel
engine, a steam engine or a turbine, into kinetic energy by compressing and pressurizing air or gas then discharging it into
receiver or pressure system. In this mechanism, the compression is accomplished by transforming the rotary motion of a
crank shaft into reciprocating piston motion in a cylinder through a connecting rod. In such operating condition, the connect-
ing rod experiences complex dynamic loads during service and therefore, it is considered to be a critical component [1].
Materials used for connecting rods are required to have good mechanical properties, in particular rigidity, hardness, ten-
sile and fatigue strengths [2]. To date, the selected materials for connecting rods are malleable irons, nodular cast irons, high
carbon steels and microalloyed steels and they are manufactured by forging, casting, powder metallurgy or more recently,
fracture splitting technology. The use of forging for production of connecting rods has advantages, i.e. the connecting rod

Corresponding author. Tel./fax: +62 274 521673.

E-mail address: ilman_noer@ugm.ac.id (M.N. Ilman).

http://dx.doi.org/10.1016/j.engfailanal.2015.03.010
1350-6307/ 2015 Elsevier Ltd. All rights reserved.
 M.N. Ilman, R.A. Barizy / Engineering Failure Analysis 56 (2015) 142149 143

materials become compact and lightweight hence reducing inertia forces during service whereas cast connecting rods are
cheaper but their strength is relatively low so that their use is limited to small and medium size connecting rods.
Failure of connecting rods is mainly caused by fatigue and factors which contribute to the failure include improper mate-
rial selection, poor design or fabrication defects [3,4]. The majority of damages have been reported to take place at some
parts of connecting rod such as a small head of the connecting rod, crank pin, roller bearing and connecting rod bolt [5].
A number of efforts have been made to improve performance of connecting rods. Apart from experimental studies using frac-
ture analysis and microstructural examination, the use of a nite element model (FEM) could aid to improve the design of a
connecting rod [6,7].
Recently, premature failure of a connecting rod of a reciprocating air compressor has been reported to occur and based on
visual examination as shown in Fig. 1, the initial crack was found at the rounded llet of the big connecting rod end which
was connected to a crankshaft assembly. The cross section of the connecting rod was designed in the form of I-prole with a
drilled oil passage was located in the middle I-section. This oil passage was designed for lubrication system. As fracture
occurred, the connecting rod shank (or I-beam) was crashed for many times before crankshaft motion was stopped and
as a result, the rod shank near the small connecting rod end was broken. The present investigation aims to study the probable
cause of failure and to propose recommendations for improving the quality of the connecting rod.

2. Materials and experimental methods

The rst step of the failure analysis procedure is to inspect the location and appearance of fractured surface followed by a
complete analysis of the failure. Fig. 2 shows the locations of test specimens for fractography, chemical analysis, metallogra-
phy, hardness measurement, tensile test and fatigue test. The fractured surface of the connecting rod was maintained free
from scratch for fractography study using scanning electron microscopy (SEM). Microstructural examinations were carried
out using optical microscopy and SEM where samples were cut at a region near the fractured surface. The samples were pre-
pared according to standard metallographic technique including grinding, polishing and etching using nital (2% HNO3 + 98%
propanol). The hardness was measured using Vickers microhardness tester whereas tensile test specimens were taken par-
allel to the axis of I-beam.
The fatigue behaviour of the connecting rod was studied using fatigue crack growth rate (FCGR) test. Specimens for FCGR
test were machined in the form of compact tension specimen (CTS) according to ASTM E647 with the initial crack was par-
allel to the actual crack direction to simulate the fatigue crack growth behaviour.

3. Results and discussions

3.1. Chemical composition

The chemical composition of the connecting rod material is given in Table 1. The amount of C in steel material was around
0.462 with low percentages of alloying elements suggesting that the connecting rod was made from a medium carbon steel.
According to AISI standard, this steel is categorized as AISI 1045.

Fig. 1. The failed connecting rod: (a) location of initial crack, (b) magnied photograph of fractured surface outlined by a square in this gure (a), and (c) I-
section of the connecting rod.
144 M.N. Ilman, R.A. Barizy / Engineering Failure Analysis 56 (2015) 142149

Fig. 2. A schematic of the connecting rod with the locations of test specimens.

Table 1
Chemical composition of the connecting rod (%wt.).

Material C Mn Si P S Ni Cr Mo Cu
Connecting rod 0.462 0.761 0.411 0.002 0.007 0.019 0.046 0.003 0.0754
AISI 1045 0.430.5 0.60.9 0.100.35

3.2. Microstructural examination

Specimens for microstructural examinations were taken from the region near the fractured surface. Fig. 3a shows a pho-
tomicrograph of the connecting rod examined using optical microscopy. It can be seen that the microstructure is composed
of ferrite (light-etched) and pearlite (dark-etched). The presence of pearlite is conrmed from SEM photomicrograph (Fig. 3b)
which clearly shows parallel plates of ferrite with cementite (Fe3C) precipitated within the plates. Of note is that the dark-
etched area in the optical photomicrograph, i.e. pearlite appears to be bright in SEM photomicrograph and vice versa. Results
of microstructural examination suggest that the connecting rod material is cast medium carbon steel.

3.3. Mechanical properties

The connecting rod was subjected to tensile/compressive axial forces during operation so that the position of tensile test
specimens was made parallel to the axis of the connecting rod. Results of tensile and hardness tests are given in Table 2.
Consistent with the chemical composition analysis, the connecting rod material is closely matched to AISI 1045 steel.

Fig. 3. Microstructure of the connecting rod examined using: (a) optical microscopy and (b) SEM.
 M.N. Ilman, R.A. Barizy / Engineering Failure Analysis 56 (2015) 142149 145

Table 2
Results of tensile and hardness tests.

Material Tensile stresses Hardness (VHN)

YS (MPa) UTS (MPa) % Ductility
Connecting rod 449.32 648.03 15.08% 225.3
AISI 1045 300450 570700 1430 180220

3.4. Analysis of fractured surface

Results of visual examination on fracture surface of the connecting rod in Fig. 1, show that fatigue seems to be the main
cause of failure as indicated by the presence of initial crack, crack propagation and nal failure. As stated previously, the
crack was found to nucleate at the rounded llet of the big connecting rod end where stress concentration was relatively
high as shown in Fig. 1a. Further examination on the fractured surface shows that the fatigue area as seen in Fig. 1b and
c is relatively small, i.e. around 30% of the total fractured area suggesting that this type of fatigue is a low cycle fatigue
due to high stress level during operation.
SEM fractographs of the initial crack region outlined by a square in Fig. 1 are shown in Fig. 4. The presence of large inclu-
sions at subsurface region could be responsible for crack initiation. It seems that the inclusions can act as stress raisers lead-
ing to crack initiation. This argument is supported by the presence of cracked inclusions in Fig. 4b. A number of researchers
[810] has reported that as a crack nucleates at the interface between inclusion and matrix or crack initiates through crack-
ing of the inclusion, the stresses within inclusion are relieved leading to stress-free defect or pores. Subsequently, the pores
ahead of the crack tip coalescence hence promoting fatigue crack growth rate. Such mechanism produces small dimples
which represent coalesced voids as shown in Fig. 4a and b.
Microanalysis of inclusions in Fig. 4a and b shows that the inclusions are mainly composed of Al2O3, SiO2 and FeO. The
presence of a signicant number of inclusions is likely resulted from deoxidation of steel melt using aluminium or silicon

Fig. 4. SEM fractographs: (a) initial crack region, (b) cracked inclusions, (c) EDX-spectra of inclusion marked A in 4b, and (d) EDX-spectra of inclusion
marked B in 4b.
146 M.N. Ilman, R.A. Barizy / Engineering Failure Analysis 56 (2015) 142149

resulting in Al2O3 and SiO2 respectively. Another possible factor promoting inclusions is high percentage of oxygen present in
the steel melt since FeO can exist if the oxygen content is high [11].
Fig. 5 shows SEM fractograph of fatigue crack growth region of the failed connecting rod where crack propagates from left
to right. It seems that fatigue striations are not well developed and this is usually the case in steels such as AISI 1045. This is
because steels have limited slip systems and easy cross slip to accommodate the crack front and to facilitate continuity of the
crack front through adjacent grains. The most important nding in this investigation is that the averaged distance between
striations is relatively wide indicative of a low cycle fatigue.

3.5. Stress analysis in connecting rod

Forces acting on a connecting rod of a compressor are periodic and complex since a connecting rod is used to transform
rotary motion of crankshaft into piston reciprocating motion as shown in Fig. 6. A compressor operates at various speeds,
typically in the range of 10002000 rpm and hence it generates periodic disturbing forces, especially when the reciprocating
parts are not balanced.
Stress analysis of a connecting rod has been proposed by Timoshenko et al. [12]. For simplicity, analysis of the forces is
made based on assumptions that the connecting rod with the length of l can be considered as two masses, one at big con-
necting rod end (M1) and the second at the piston (M2). As the crankshafts with the radius of r rotates at angular velocity of
x, the angular displacement (h) at any given time t is h xt whereas the inertia force of the mass M1 is F c M1 x2 r. The
horizontal component of Fc is F1 which equals:

F 1 M1 x2 r cos xt 1
Referring to Eq. (1) and Fig. 6, it can be seen that tensile load starts to develop at the big end of connecting rod as the crank
angle (h) is equal to 90. On further motion, the maximum tensile load occurs at h of 180 and then the load decreases up to
zero on moving from 180 to 270. Subsequently, as the big connecting rod end leaves this position, the load becomes com-
pressive and the maximum compressive load is achieved at the position of 360 or 0.
From the geometry of Fig. 6, the displacement of piston designated as x due to the connecting rod with the length of l is
given by:
x l1  cos a r1  cos xt 2
and
r
sin a sin xt 3
l
The value of cos a can be expressed in term of sin xt using Eq. (3) as follow:
s
r2 2 r2 2
cos a 1  2 sin xt 1  2 sin xt 4
l 2l
Substituting Eq. (4) into Eq. (2) results in:
r2 2
x r1  cos xt sin xt 5
2l
The second differentiation of displacement x results in acceleration so that the inertia force of mass M2 is given by:

Fig. 5. Fatigue crack growth region of the failed connecting rod showing striations as indicated by arrows.
 M.N. Ilman, R.A. Barizy / Engineering Failure Analysis 56 (2015) 142149 147

Fig. 6. Forces acting on a connecting rod.

 r 
F 2 M2 x2 r cos xt sin 2xt 6
l
Combination of Eqs. (1) and (6) results in the resultant force acting on the connecting rod as follow:
r
Ft M1 M2 x2 r cos xt  M2 x2 r sin 2xt 7
l
Referring to Eq. (7), it is concluded that the connecting rod is subjected to time-varying axial forces (tensile and compres-
sive forces) during service as expressed by cosinusoidal and/or sinusoidal functions. Such complex forces acting at the ends
of a connecting rod may become critical, especially as the peak loads are achieved. These dynamic loads have a potency to
cause fatigue failure and in some cases, fatigue failure occurred at regions near connecting rod ends.
Apart from analytical approach above, the modelling of connecting rod was also performed using nite element analysis
(FEA) to conduct a linear elastic stress analysis. The model was meshed in ABAQUS 6.11 with the details are given in Table 3.
Static analysis was carried out for tensile condition where a tensile load, typically of 50 kN [13] was applied at the big end of
the connecting rod whereas the small one restrained. Constitutive properties for all elements including Young modulus and
Poissons ratio were assumed to be 200 GPa and 0.3 respectively. The Von Misses distribution obtained from FEA is shown in
Fig. 7 with the maximum stress is given in Table 3. It can be seen that the maximum stress is around 295.75 MPa and it
occurs at the llet region near the big connecting rod end consistent with the results of fractography study. Of note is that
despite, the maximum operating stress is less than the yield stress of the connecting rod material, i.e. 449.32 MPa, fatigue
failure can occur under dynamic loads due to high stress concentration and material defects.

3.6. Fatigue crack growth behaviour

Fatigue crack growth behaviour of the connecting rod given in loglog plots of da/dNDK is shown in Fig. 8. It can be seen
that the stage II of da/dNDK curve in the gure is linear and it is consistent with the Paris power law that is da/
dN = 1.1279E14(DK)5.2141. According to Boyer [14], microstructure strongly inuences n value and hence drastically change
fatigue crack growth rate (da/dN) whereas stress levels do not signicantly inuence fatigue crack growth rate when the
stress levels are below the yielding point. This means that the high n value of Paris power law of the failed connecting
rod could be linked to the microstructure.
The n values of Paris constant for most of metals are in the range of 24 [15] whereas the n value of the connecting rod
obtained in this study is around 5.2141 higher than those observed in most metals. The high n value of the connecting mate-
rial may be associated with as cast steel condition with a considerable number of inclusions. To prevent premature failure
due to fatigue, improvements need to be done in terms of design, material selection and manufacturing process.

4. Conclusions and recommendations

1. Results of chemical composition, microstructural analysis, hardness measurement and tensile test led to the conclusion
that the connecting rod material was cast medium carbon steel equivalent to AISI 1045.

Table 3
Net characteristics of a connecting rod with force simulated.

Load (kN) No. of elements No. of nodes Element type Von Misses stress (MPa)
50 18,096 72,384 Tetrahedral 295.75 (Max.)
148 M.N. Ilman, R.A. Barizy / Engineering Failure Analysis 56 (2015) 142149

Fig. 7. Results of nite element analysis showing stress spots at the llets of a big connecting rod end.

Fig. 8. Plot of da/dNDK.

2. The failure of connecting rod was caused by two factors. First, the presence of a low llet radius near a big connecting rod
end which had a high stress concentration from which crack nucleated under dynamic loads. Secondly, inclusions present
at subsurface of the connecting rod material seemed to act as stress raisers hence promoting crack initiation.
Subsequently, the high n value of Paris constant obtained in this investigation may have contributed to the shorter crack
growth.
3. To prevent future failures, material and design need to be improved including:
The use of forged steel rather than cast steel for connecting rod material is recommended. In addition, steel melt
should be treated using vacuum degassing process to obtain inclusion-free clean steel.
The connecting rod should be re-designed using a longer llet radius near the big connecting rod end to reduce stress
concentration.

Acknowledgements

This work was carried out under the continued support of Gadjah Mada University, Indonesia and therefore, it is pleasure
to acknowledge this support with thanks.
 M.N. Ilman, R.A. Barizy / Engineering Failure Analysis 56 (2015) 142149 149

References

[1] Shenoy PS, Fatemi A. Dynamic analysis of loads and stresses in connecting rods. Proc IMechE 2006;220:61524.
[2] Gu Z, Yang S, Ku S, Zhao Y, Dai X. Fracture splitting technology of automobile engine connecting rod. Int J Adv Manuf Technol 2005;25:8837.
[3] He BY, Shi GD, Sun JB, Chen SZ, Nie R. Crack analysis on the toothed mating surfaces of a diesel engine connecting rod. Eng Fail Anal 2013;34:44350.
[4] Xu XL, Yu ZW. Failure analysis of a diesel engine connecting rod. J Fail Anal Prevent 2007;7:31620.
[5] Khare S, Singh OP, Dora KB, Sasun C. Spalling investigation of connecting rod. Eng Fail Anal 2012;19:7786.
[6] Rabb R. Fatigue failure of a connecting rod. Eng Fail Anal 1996;3(1):1328.
[7] Griza S, Bertoni F, Zanon G, Reguly A, Strohaecker TR. Fatigue in engine connecting rod bolt due to forming laps. Eng Fail Anal 2009;16:15428.
[8] Murakami Y. Metal fatigue: effects of small defects and nonmetallic inclusions. 1st ed. Oxford: Elsevier Science Ltd.; 2002.
[9] Alfredsson B, Olsson E. Multi-axial fatigue initiation at inclusions and subsequent crack growth in a bainitic high strength roller bearing steel at
uniaxial experiments. Int J Fatigue 2012;41:1309.
[10] Spriestersbach D, Grad P, Kerscher E. Inuence of different non-metallic inclusion types on the crack initiation in high-strength steels in the VHCF
regime. Int J Fatigue 2014;64:11420.
[11] Campbell J. Castings. 1st ed. Oxford: Butterworth-Heinemann; 1991. p. 1991.
[12] Timoshenko S, Young DH, Weaver Jr W. Vibration problems in engineering. 4th ed. New York: John Wiley & Sons; 1974.
[13] Phad SA, Burande DH. Static and dynamic analysis of connecting rod of compressor. Int J Automobile Eng Res Dev 2013;3(3):2330.
[14] Boyer EB. Atlas of fatigue curves. Ohio: American Society for Metals; 1986.
[15] Broek D. Elementary Engineering Fracture Mechanics. Leyden, The Netherlands: Noordhoff International Publishing Co.; 1974.