Engineering Failure Analysis 19 (2012) 7786

Contents lists available at SciVerse ScienceDirect

Engineering Failure Analysis

Spalling investigation of connecting rod

a r t i c l e

i n f o

a b s t r a c t
The customers of the vehicles reported high noise and vibration in the engine at an early stage of service lifetime. An analysis of the various components of the internal combustion engine was carried out. Subsurface cracks and pit marks were seen in the crank pin, roller bearings and big end surfaces of the connecting rod. It was found that high wear at the interface of these components was the main culprit. A laboratory test set-up was developed to correlate and reproduce the eld failures. The loads and boundary conditions obtained from the experiments were used in the nite element model of the connecting rod assembly. Results shows high interfacial pressure and stresses near the junction of web and ange of the connecting rod. The modied design of the connecting rod shows signicant reduction in the extreme pressure in FEM resulting in the signicant enhancement of durability life in laboratory test. A discussion of the spalling problem has been provided leading to the connection of pick pressure and spalling phenomena. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 28 April 2011 Received in revised form 7 September 2011 Accepted 12 September 2011 Available online 18 September 2011 Keywords: Connecting rod Spalling Wear Finite element model Contact pressure

1. Introduction and failure background One source of energy in automobile industry is internal combustion engine. IC engine converts chemical energy into mechanical energy in form of reciprocating motion of piston [1]. Crankshaft and connecting rod convert reciprocating motion into rotary motion. The connecting rod experiences various forces of piston acceleration and deceleration from the stroking motion, loads generated by friction and the load by the cylinder pressure during the combustion stroke [2,3]. Under normal operating condition, the design of the connecting rod for innite number of cycles is based on the these forces such that designed connecting rod does not exceed the desired strength during the life span. However, when the customer exceeds the rev limit of the engine, it adds to a larger boost levels, increasing the cylinder pressure and thus causing higher friction between the sliding members. Catastrophic engine failure results if the sign of failure modes are not investigated early [47]. In this paper we report the failure of the connecting rod and the pin due to spalling observed under such condition of motorcycle engine. Crankshaft assembly is shown in Fig. 1. The big end of the connecting rod is connected to the crankshaft assembly using a crank pin. To reduce the friction, roller bearing is used between the connecting rod and the pin. The roller bearings have non-conformal contact surfaces (line or point contact called hertzian contacts), which support the moving parts of the crankshaft assembly with rolling action at the interface [8]. An oil lm between the interfaces prevents metal-to-metal contact. The customers reported failure of these components after an average driving distance of 15,000 km against the desired life of 50,000 km. Failed sample of connecting rod, pin and the roller bearing is shown in Fig. 2. Scratches and rough marks on connecting rod, crank-pin and crankpin bearing can be seen with naked eyes. Such kind of failure is generally referred as fatigue spalling. Fatigue spalling initiates in the depth of material by subsurface stress eld close to the
Corresponding author. Address: TVS Motor Company Ltd., R&D (Design analysis group), P. Box No. 4, Harita, Hosur, Tamilnadu 635 109, India. Tel.: +91 4344 276780x3502, +914344270502 (Direct); fax: +91 4344 276649. E-mail addresses: [email protected], [email protected] (O.P. Singh).
1350-6307/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2011.09.007

78

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

Piston

Connecting rod Roller bearing Crank pin Crankshaft

Fig. 1. Computer aided design of the crankshaft assembly showing various components.

contacting surface [9]. It then grows to the surface in the form of micro-spall or crack [10]. When these localized inelastic subsurface deformations such a cracks interact, the material break up, disrupt the oil lm in the region of highest load and lowest oil lm thickness [8]. The interesting and bafing part of the failure is that high wear is seen near the I-beam section of the connecting rod (Fig. 2c). Noise levels radiated from the engine was reported to be high when the failure occurred. In this paper, the factors, which conduce to this fatigue spalling of the connecting rod, has been investigated in this paper. 2. Initial actions and laboratory test development 2.1. Visual inspection A close inspection of the failed components reveals the following modes of failure (see Fig. 3):  Surface defects like burr, surface mark, pit mark, dents have observed on crankpin. Fig. 3a shows damage marks on crankpin surface which acts as a stress concentration location [11,12].  Similarly burr and deep machine tool mark near oil hole (Fig. 3b). This can act as a stress raiser during rolling contact phenomenon [8]. These observation suggests that external contaminants like hard particles, wear debris, and akes that might have entered into the sliding zone. The hard particles abrasive contaminants cause indentation or cutting wear damage to the softer material. The average depth of the cavities measured at the contact surfaces was about 25 lm. It is nearly impossible to observe these kinds of failures during the engine running condition. Hence, different probable causes of failures are identied. The probable causes could be many; few important factors may be due to:      Hardness of connecting rod and crank pin. Connecting rod to crank pin bearing clearance is more then specied limit. High stresses at the critical location. Design of the connecting rod. Scarcity of the lubricating oil at sliding interfaces.

Before we investigate these failures further, a laboratory test method was developed and the failures observed in the eld are reproduced rst. 2.2. Laboratory test development Durability testing under real environment is time consuming. To reduce testing time connecting rod spalling failure has simulated on an engine dynamometer. The controlled testing method was established based on the four-quadrant matrix

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

79

Fig. 2. Failed pictures of (a) roller bearing spalling (a) crankpin spalling, and (b) connecting rod spalling.

method (e.g.) [13]. This method basically correlates the MTTF (mean time to failure) in eld-testing with the MTTF in lab testing. All the failure modes of the system components are identied and failure modes are correlated with the components. In lab test, vehicle was run on dynamo and followed standard running cycle (as shown in Fig. 4) with specied engine operating condition (like load on vehicle, and engine oil quantity). The time period of vehicle acceleration and deceleration constitute a cycle. Note that the deceleration time is smaller compared to the acceleration time. The engine speed at point A and B are 20 km/h and 60 km/h respectively. The time duration for AC is 18 s and for CD is 6 s. Engine operating condition was decided based on real durability test condition. Noise radiated from the engine was monitored. The normal engine (at the start of the test) radiated noise in the range of 7075 dB. When the noise levels was increased to 90 dB, the engine was stopped, parts were dismantled and the condition of connecting rod and pin was checked. At such high noise level, high vibration of the engine was also noticed. The same failure (see Fig. 2) was reproduced at about 20,000 cycles in the lab test. The same process was repeated for other engine tests.

2.3. Process improvement Surface defects as seen in Fig. 3 will act as a stress raiser during service of crank pin and connecting rod, and thus wear will initiate at this defect and shorten the life of the component. Machining process, cleaning process and material handling

80

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

(a)

(b)
Deep machine tool mark

Burr

Fig. 3. (a) Surface damage on the crankpin and (b) Surface defect near oil hole.

Fig. 4. Engine running cycle developed for the laboratory test.

process was improved to avoid surface defect on crankshaft and crankpin. However, no signicant improvement in the life the component was observed.

2.4. Material improvement Case depth hardening of the material was measured and improved as shown in Fig. 5. Near the surface, the hardness was below 800 HV and it was consistently low throughout the material depth of 1.3 mm with steep variation across the depth. The supplier process of heat treatment and material check controls were improved. The case depth was increased throughout the thickness with less steep variation.

2.5. Connecting rod to crank pin clearance In failed sample clearance between connecting rod and roller bearing was not as per the specication. The measured clearance was 15 lm. With high clearance the crankpin can oat and increase the rate of wear [8]. The machining process was improved to match design specication accurately. The new design clearance specication was controlled to 12 lm.

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

81

950

Hardness (HV 0.3 kg)

850 750 650 550 450 350

After improvement Before imrovement

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4

Distance from surface (mm)

Table 1 Contact parameters in the contact FEM. Contact type Contact element Target element Contact detection point Contact algorithm Initial penetration Effect of initial penetration Contact stiffness Penetration tolerance Pinball region Surface to surface (rigid exible contact) CONTA 174 TARGE 170 Gauss point Augmented lagrangian 10 lm Include offset only (exclude initial geometrical penetration or gap) with ramped effect 2.7792e6 N/mm 0.15 mm 6.045

(a)

(b)

Pressure force Inertia force

Crankpin

Force (N)

Crank Angle (degree)

Fig. 6. (a) Finite element model of the connecting rod and (b) Variation of pressure and inertia forces during engine operation.

82

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

2.6. Crankpin and roller bearing diameter Hertzian contact theory explains that contact pressure developed between two cylindrical rolling surface is inversely proportional to contact length and diameter of rolling surfaces. High pressure at the interface can result in the wear of the materials [8]. Eq. (1) shows the relations between rolling cylinder parameter and contact pressure develop between the surfaces. Maximum pressure between two rolling cylinder surfaces is given by [8],

Pmax
where,

2F

pal

 1 4Pk1 k2 R1 R2 2 l R1 R2

(a)

Contact pressure (MPa)

VonMises stress

(b)

Fig. 7. Results from FEM of the original connecting rod showing (a) pressure contours in the contact patch region and (b) VonMises stresses.

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

83

ki

1 m2 i pEi

subscript 1 and 2 denotes rst and second cylinder respectively, F is the normal load; R is cylinder diameter; E is elastic modulus; m is poissons ratio, l is contact length between two cylinder. Note that increasing the parameter a reduces the maximum pressure and hence, the stresses between the two mating cylinders. To reduce stresses, crankpin and roller bearing diameter was increased from 25 mm to 26 mm and 3 mm to 4 mm respectively while maintaining the same design clearance. As per the hertzian contact theory this increase in roller diameter would increase contact diameter by 14%, which in turn would reduce pressure at interfaces by 14%. After all initial checks and improvements crankshaft assembly was tested in engine endurance test. Similar connecting rod and crankpin-spalling failure was observed at MTTF of 30,000 cycles. Though a 50% improvement in life was achieved, it is still signicantly less than the desired life of 50,000 km. Hence; further investigations were conducted to improve the design of the connecting rod. In the next section we present the nite element model (FEM). The main purpose of the FEM study is to understand the effect of connecting rod design prole on contact behavior under engine operating condition.

3. Non-linear nite element analysis 3.1. FEA model A simplied non-linear contact model was developed to investigate the contact pattern between connecting rod and crankpin using Ansys code [14]. Rigid to exible surface-to-surface contact mechanics was used. Since the crankpin hardness is much more than connecting rod, the connecting rod as considered as exible surface and crankpin as rigid surface. The intersection face between connecting rod big end inner surface and crankpin outer surface is modeled as contact patch. FEA contact patch consist of contact and target surfaces; connecting rod big end inner surface was considered as contact (element CONTA 174) and crankpin outer surface as target (element TARGE 170). Table 1 shows values of other parameters of the model. Connecting rod was meshed with second order solid 92 elements. Fig. 6a shows the nite element model with loads and boundary conditions. Crankpin is modeled as rigid body. Small end side of the connecting rod is constrained in lateral and transverse direction to avoid rigid body motion. Force on the connecting rod is obtained from the cylinder pressure (measured experimentally) and inertial forces of the connecting rod. Fig. 6b shows the variation of cylinder pressure and inertial forces. Maximum force experienced by the connecting rod was applied as shown. Effect of thermal distortion has not been considered in the model. 3.2. FEA results Fig. 7a show contact pressure distribution between connecting rod and crankpin of the existing design. The pressure distribution is localized and it is seen near the I-section of the connecting rod. The maximum contact pressure is 487.51 MPa. Von-misses tress distribution shows similar distribution with maximum value of 271.44 MPa (Fig. 7b). It is interesting to note that high magnitude of pressure and stresses are observed near the beginning of the ange and end of the web of the I-section of the connecting rod. This is the location where high rate of spalling was observed (see Fig. 3c). An interpretation of results suggests that (a) at maximum engine load condition, the load is affectively share by only two central roller

Fig. 8. Four different designs of connecting rod. Notice how the nal web design is different from the previous three design.

84

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

(a)

Contact pressure (MPa)

VonMises stress

(b)

Fig. 9. Results from FEM of the nal design of the connecting rod showing (a) pressure contours in contact patch region and (b) contours of VonMises stresses.

bearings resulting in high-localized pressure and (b) The high-localized contact pressure would reduce the lubricating oil lm thickness and eventually lead to asperity interactions resulting in lower fatigue life. The high and localized pressure and stresses at the interface may not induce catastrophic failure of the crankshaft assembly. However, this would act as a feeding source for accelerated wear of the interfaces and hence the consequent reduction of useful life. Any further design modications of the connecting rod should reduce the high interfacial contact

Table 2 Magnitude of maximum pressure and stress at the interface between the connecting rod and crankpin. Parameter Max. contact pressure Max. VonMises stress Original design (MPa) 487.51 271.44 New design (MPa) 323.40 229.08 Change (%) 33.70 15.60

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

85

Fig. 10. Elastic strain contours in four designs investigated.

pressure and stresses. Various design modications were investigated using FEA and they are shown in Fig. 8. It is to be noted how the design of web and ange has changed progressively from the original design 1 to nal design 4. Such prole is expected to reduce the effective stiffness of the connecting rod at the roller bearing and connecting rod interface. The web area has increased when compared to the original design. The same FE model was used in the analysis. Fig. 9a shows the pressure contours in the contact patch region. Following points are noted: (a) the contact patch area has increased compared to the original design. Hence, the maximum pressure has reduced to 323.40 MPa, a reduction of 34% from the original design (see Table 2), (b) It can be seen from the contours that at a given instant of time, at least four roller bearing would share the maximum load. The magnitude of maximum VonMises stress (Fig. 9b) also reduced to 192 MPa, a reduction of 16% is achieved. Fig. 10 shows the non-dimensional displacements i.e. elastic strain in all the four designs (see Fig. 8) at the location where failure was observed. Elastic strains in design 2, 3 and 4 are respectively 3.4, 4.9 and 5.7 times the elastic strain in design 1. Higher deformation in design 4 is due to the reduction in stiffness that manifested in signicant reduction in interfacial pressure. It is to be noted further that how the concentrated deformation in design 1 has spread to a much larger areas in design 4. The nal design of the connecting rod passed the durability life target in the lab test. 4. Discussion The failure under investigation is a case of surface fatigue resulted from the high concentrated interfacial pressure and stresses. The roller bearing operates with EHD (elasto-hydrodynamic lubrication regime [15]. The hertzian or point contacts between the roller bearing and connecting rod produce extreme pressures and therefore elastically deform the surfaces locally to provide small elliptical contact areas. The repetitive and cyclic formation of these elastically deformed contacts eventually lead to the surface fatigue (see Fig. 2). The tendency of the roller element bearings to oat on the hydrodynamic lm would reduce under excessive interfacial pressure, as viscous drag will no longer be able to draw sufcient amount of oil into the load zone. Heat generated due to viscous dissipation at the interface would further aggravate the problem [2]. The bearing surface may reach the temperature at which the material locally melts. It is to be noted heat ux generated by viscous shear in the bearing is proportional to the pressure eld around the bearing surface. Hence rise in temperature can be approximated by [8],

dT

P _ qC Q

_ is the ll ratio, q is density and C is the heat capacity. In the two designs where, dT is the rise in temperature, P is pressure, Q of connecting rod discussed above (see Fig. 8), and assuming the denominator remains the same in Eq. (2), the decrease in

86

S. Khare et al. / Engineering Failure Analysis 19 (2012) 7786

bearing surface temperature in nal design would be about 33.70% (see Table 2) lower than the original design. This is substantial in the light of the fact that 50% reduction in oil life for every 10 C increases in temperature [16]. If the bearing surfaces continues to operate under high load and speed for prolonged period, lubrication oxidation may breakdown due to elevated temperature [17]. This will set the stage for very ne particles to be liberated from the rolling element and nally making way into the load zone. Further surface craters will be created that would act as stress concentration sites (see Fig. 3). The distinct region of pressure spike seen in the original design of the connecting rod (Fig. 7) would lead into the boundary regime of lubrication (metal-to-metal contact) triggering the spalling related phenomenon as described earlier. Though the initial actions taken as described earlier helped in improving the component life, it was the change in design of the connecting rod, which improved the desired life signicantly. This corroborates the fact that durability life of the components experiencing alternating forces is improved substantially by proper design [18]. The change in the prole of the connecting rod (Fig. 8b) has facilitated in reducing the extreme interfacial pressure (Fig. 9). The participation of more effective load sharing surfaces (Fig. 10) would result in reduced elliptical contact areas formation and hence, improving the lubrication. In the likely scenario of the engine being abused or revved for a signicant span of time, such design would provide enhanced the life the components especially when spalling related phenomenon is dominant. Four samples of this design of the connecting rod was tested in the lab under same operating conditions and passed the required durability life. No abnormal noise and vibration was observed even after 50,000 km of continuous run of the engine. 5. Conclusions The spalling of connecting rod, crank pin and roller bearing is attributed to the high-localized interfacial pressure that developed due to the design of the web and ange of the connecting rod. Load sharing interface area was less with only two roller bearings effectively participating at a given instant. The high pressure resulted in accelerated surface fatigue spalling leading to the early failure of the components. The modied design of the web provided signicant reduction in pressure and stresses in the contact patch region, hence improving the durability considerably. References
[1] Edward F. Obert, internal combustion engines: analysis and practice. 2nd ed. Pennsylvania; 1964. [2] Kim Byung-Jik, Kim Kyung-Woong. Thermo-elastohydrodynamic analysis of connecting rod bearing in internal combustion engine. J Tribol 2001;123(3). [3] Rabb Roger. Fatigue failure of a connecting rod. Eng Fail Anal 1996;3(1):1328. [4] Wang ZW, Luo YY, Zhou LJ, Xiao RF, Peng GJ. Computation of dynamic stresses in piston rods caused by unsteady hydraulic loads. Eng Fail Anal 2008;15(12):2837. [5] 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(5):15428. [6] Silva FS. Fatigue on engine pistons a compendium of case studies. Eng Fail Anal 2006;13(3):48092. [7] Atxaga G, Irisarri AM. Failure analysis of the end of a shaft of an engine. Eng Fail Anal 2010;17(4):71421. [8] Rahnejat H. Tribology and dynamics of engine and powertrain: fundamentals, applications and future trends. Cambridge, UK: Woodhead Publishing limited; 2010. [9] Gohar R, Rahnejat H. Fundamentals of tribology. London: Imperial College Press; 2008. [10] Kaneta K, Murakami y, Okazaki T. Growth mechanism of subsurface crack due to hertzian contact. J Tribol Trans ASME 1986;108. [11] Ktari A, Haddar N, Ayedi HF. Fatigue fracture expertise of train engine crankshafts. Eng Fail Anal 2011;18(3):108593. [12] Gagg Colin R, Lewis Peter R. Wear as a product failure mechanism overview and case studies. Eng Fail Anal 2007;14(8):161840. [13] Dazzi Giovanni, Savio Stefano, Firpo Pierluigi. Estimate of components reliability and maintenance strategies impact on trains delay. In: Proceedings 21st European conference on modelling and simulation; 2007. [14] ANSYS Version 10.0. ANSYS, Inc. Release 10.0 documentation. [15] Keer LM, Bryaut MD. A pitting model for rolling contact fatigue. J Lubricat Technol Trans ASME 1983;105. [16] Bartz WJ. Long-life and life-time lubrication possibilities and limitations. Lubricat Eng 1993;49:518. [17] Hoag Kevin L. Vehicular engine design. New York: Springer Wien; 2005. [18] Subbiah S, Singh OP, Mahan SK, Jeyaraj AP. Effect of mufer mounting bracket designs on durability. Eng Fail Anal 2011;18(3):1094107.