A new fretting fatigue testing machine design, utilizing rotating bending principle approach

Transcription

1 Int J Adv Manuf Technol (2014) 70: DOI /s ORIGINAL ARTICLE A new fretting fatigue testing machine design, utilizing rotating bending principle approach E. Zalnezhad & Ahmed A. D. Sarhan & P. Jahanshahi Received: 5 February 2012 /Accepted: 24 October 2013 /Published online: 17 November 2013 # Springer-Verlag London 2013 Abstract Fretting fatigue is a phenomenon which occurs when two parts are contacted to each other and one of those parts or both are subjected to cyclic load. Fretting decreases fatigue life of materials drastically by half or even more. Therefore, investigation of fretting fatigue life of materials is an important subject. Fretting fatigue tests are usually performed using universal hydraulic testing devices. In this work, a rotating bending apparatus for fretting fatigue test is introduced in which the cyclic load is provided by an adjustable eccentric load. The apparatus called RBFF machine which is the abbreviation of rotating bending fretting fatigue. The eccentric load is measured by load cell. The coefficient of friction and fretting load are measured by foil strain gauges using a Wheatstone bridge configuration. The performance of the machine is verified with doing a comparison between fatigue lives of a number of AL7075-T6 alloy samples on a Shimadzu rotating bending fatigue testing machine and RBFF. The results shows very close assent between the operations of the two testing rigs. The main privileges of RBFF are its simplicity with respect to universal devices, cheapness and, coefficient of friction (between pads and specimen) evaluation during the test. The RBFF also has the capability of being used for any other soft and hard metals. It can be advanced further for high and low temperature. Keywords Fretting fatigue. Al7075-T6 alloy. Rotating bending fatigue E. Zalnezhad (*): A. A. D. Sarhan Department of Mechanical Engineering, Faculty of Engineering, University of Malaya, Kuala Lumpur 50603, Malaysia erfan_zalnezhad@yahoo.com P. Jahanshahi Photonics Research Group, Department of Electrical Engineering, Faculty of Engineering, University of Malaya, Kuala Lumpur 50603, Malaysia 1 Introduction Fretting is the surface damage that occurs when contacting surfaces between mating bodies experience an oscillatory motion of small amplitude. When fretting occurs under fluctuating loading condition, the process is termed fretting fatigue. Fretting fatigue increases the tensile and shear stresses at the contact surface producing surface defects which can act as stress concentration sites [1]. Fretting fatigue is a common occurrence in multitude of components or structures involving riveted joints, bolted joints, wire ropes, and dovetail joints of gas turbine engine [2 4]. Fretting decrease fatigue life of materials drastically. When fretting is applied on the surface of materials, the stress concentration at fretting region is increased and micro crack appears and starts to grow. Cracks may assert in some cases, while in others, they may propagate, eventually causing failure. Regions such as in the dovetail joint in a turbine engine, where the blade and disk are in contact and undergo vibratory stresses, are especially susceptible to fretting fatigue and have been subject of studies for a number of years [5, 6]. Structural members at macro- or micro-size scales by design or default involve contact between similar or dissimilar materials. Normally, the contacting surfaces are pressed together by a normal load, P and when one of the contacting bodies is subjected to an externally applied cyclic load, σ a asmall amount of relative displacement takes place at the contact interface and induces frictional force, f f at the contact zone. The combination of normal load, frictional force, and cyclic axial load develops a localized stress field at the contact interface, which leads to early nucleation of micro-cracks. The process of crack initiation at the contact interface with fretting due to the cyclic axial loading is termed as fretting fatigue. The fretting has been observed in riveted and flanged joints, steam and gas turbines, wire ropes, biomedical implants, micromechanical devices, ball bearings, and springs [7, 8]. Arora et al. [5, 9] designed a fretting fatigue test rig in such a

2 2212 Int J Adv Manuf Technol (2014) 70: Fig. 1 Schematic of rotating bending fretting fatigue test machine from different views. a 3Dview of RBFF. b Front and side views of RBFF 7 A pair of springs to apply the load Fretting ring Electrical rotating connector 8 Chuck (a) 3Dview of RBFF Compressive Load cell for measuring the bending load Pulley for applying the bending load Adjustable bearing for applying the load Compressive load cell (b) Front and side views of RBFF way that the initiated small cracks at the contact edge could be directly observed and the crack propagation could be recorded using the video microscope. The fretting fatigue test jig is mounted on an 810 MTS material test system of 100 kn capacity. On the loading jig, two frictional force load cells are mounted. Normal force to the specimen pad interface is applied through the normal force load cell by turning the loading bolt [10, 11] studied the crack growth in a shrink-fit assembly subjected to rotating bending. They observed a significant reduction in the fatigue crack propagation life, up to 50 %, as compared to an equivalent pure bending case. Rajasekaran and Nowell [12, 13] developed a biaxial fatigue experiment capable of simulating the loading experienced by a dovetail blade root in an aircraft gas turbine. The device can simulate the effects of centrifugal loading, disk expansion force, and blade vibration [14]. Ebara and Fujimura [15 17] used a fretting fatigue specimen with bolt-tightened shoe on both sides of the plate. The contact load was gained from the strain measured by a strain gauge fixed at the bolt center. A fretting test apparatus has been designed at Purdue University in USA to study fretting fatigue at elevated temperature. The fretting fatigue specimen is clamped between two contact pads and then loaded with a remote cyclic load that leads to eventual specimen cracking and failure. The main motivation behind the development of this device was to determine conditions that cause frettinginduced cracks to form in advanced turbine engine materials Digital microscope Cutoff switch Fig. 2 Rotating bending fretting fatigue test machine Load cell displayer RPM displayer

3 Int J Adv Manuf Technol (2014) 70: such as the rash of high cycle fatigue engine failures that led to the grounding of many aircrafts. Fretting fatigue may reduce the endurance limit of a component by half or more, in comparison to the normal fatigue conditions. Unfortunately, there is not a universal machine available in the market for fretting fatigue testing. Generally, the universal hydraulic fatigue testing devices are used for this sort of testing. However, these universal machines require accessories for producing contact pressure needed for fretting fatigue. Additionally, the frequency of these types of machines is low and, consequently, the fatigue tests are time consuming [7, 10]. In this work, a new design of rotating bending fretting fatigue test apparatus is introduced. The design and construction are much simpler and more economical than the fatigue test machines described above. A rotatable fretting ring accompanies with friction pads and strain gauges are designed to evaluate the friction coefficient between friction pads and specimens during fatigue test. Furthermore, the performance Fig. 3 Schematic of parallel specimen with two points loading accompanies with moment, stress graphs at slender region, and analysis of stress in different situation. a Schematic of specimen-two points loading. b Distribution of stress in the narrow part of specimen. c Stress concentration under friction pads without applying loads (F). d Combination of stress distribution and concentration on specimen Pulley for applying the bending load (a) Schematic of specimen-two points loading (b) Distribution of stress in the narrow part of specimen Friction pads (c) Stress concentration under friction pads without applying loads (F) Friction pads (d) Combination of stress distribution and concentration on specimen

4 2214 Int J Adv Manuf Technol (2014) 70: Strain gauge Fretting load adjusting Fretting fatigue specimen Strain gauges for friction pads Electrical rotating connector Fretting pads (strain gauge) Fretting ring (a) Data Acquisition Strain gauges for fretting ring Fig. 6 The schematic of the strain gauges connections with electrical rotating connector accompany with Wheatstone bridge circuits (b) Fig. 4 a Schematic and b real images of the fretting fatigue test rig of the new design (rotating bending fretting fatigue; RBFF) was examined by comparative investigation on fatigue life of Al7075-T6 alloy using Shimadzu fatigue test machine. 2 Description of RBFF machine principles The present apparatus relates to a fretting fatigue machine using rotating bending method are designed by the authors in this work. A general view and a schematic of the device (called in this paper as RBFF) are depicted in Figs. 1 and 2, respectively. The device consists of a base (1), a motor (2) 8mm Fig. 5 The fretting pads mounted to the base for generating rotational movement means, a pair of motor couplings (3) connecting a motor shaft (4) to the motor, a plurality of first bearings (5) mounted to the base for holding the motor shaft, a chuck (6) connected to the motor shaft for gripping a fretting fatigue specimen, a frame including a plurality of rods and holders, a second bearing coupled to a bearing holder for supporting the test specimen, a pulley including a lead screw (7), a load cell (8) connected to Force (N) Force (N) Strain (µm) (a) Strain (µm) (b) Fig. 7 Calibration curves for (a)frettingringand(b) friction pads

5 Int J Adv Manuf Technol (2014) 70: the lead screw for measuring the force applied from the lead screw to the test specimen, the instrumentation which consists of load cell for measuring the axial forces imposed on the specimens, the inverter for varying the frequency, and fatigue cycles counter. Also, there is an adjustable support between chuck and last bearing (it is placed at the end left side of machine). The support has two springs to balance the bending load. Another compressive load cell is located under the ball bearing of support, to adjusting the loads (making the loads equal) on the fretting fatigue specimen. The basis functionality of this machine is applying equal opposite loads at different places on the specimen to get high stress in the narrow part of it, so with adding the friction pads which is located at the middle of the fretting parts, maximum stresses are placed under the friction pads. The threedimensional (3D) model of fretting fatigue specimen, pads, and bearing were modeled according to the ISO-1143:2010 (E) in ABAQUS (ABAQUS Inc., Providence, RI, USA). The specimen was meshed by 3D quadratic tetrahedral element. Mesh sizes decreased in order to carry out the convergence test and mesh independency to achieve the higher accuracy of results in ABAQUS. The fretting specimen was considered as linear elastic material with young modulus of 72 GPa. The Poisson's ratio was considered constant and equal to 0.3. The bearing and pads were considered as rigid material. Surface to surface contact with the small sliding was considered between the pads and specimen, and bearings and specimen. The friction coefficient of 0.3 was assigned to the interfaces of pads and specimen. However, the interaction in the specimen and bearings was frictionless. The specimen movement was restricted in axial direction. Meanwhile, bearings and pads movement was limited in axial and perpendicular directions. The loads magnitude which applied on the bearings and pads were 150 and 100 Mpa, respectively. Figure 3a, b, c, d show the moment and stress graphs accompany with finite element analysis using ABAQUS software for the unplaced and placed friction pads on the specimen. Figure 3d also shows that the maximum stress is under friction pad when specimen is under bending loads. Combination of stress concentration and stress distribution can provide earlier fatigue failure. 2.1 Variable load and eccentric system The bending load is applied to the specimen by a pulley, which has the capability of being variable in eccentric mechanism maximum 2 mm, shown in Fig. 3a. The mechanism consists mainly of a pulley, rod, and bearing. This amount of displacement is sufficient for most of fretting fatigue tests and in particular if the deformations are supposed to remain within the elastic range. 2.2 Contact loading system The pressure required to produce fretting is transmitted to the contact area, via a calibrated proving ring, shown in Fig. 4,tothe contact area through the loading pads. Each pad (see Fig. 5)has two bases (contacting surfaces) through which load are exerted on the specimen. The pad span which is defined as the distance between the centers of the fretting pad feet was chosen to be 7 mm. The pad's base has a width of 2 mm. The pads were made of stainless steel 410 with σult=700 MPa and σy=420 MPa. The pads are revolved accompanied with the specimen, as this may affect the sliding oscillations between specimen and pads. The normal load is induced by an adjusting screw and is measured using a Wheatstone bridge circuit. Four strain gauges were bonded to the proving ring and wired to create a Wheatstone bridge circuit to measure the elastic strains induced by loading the ring through the load adjusting screw. Figure 6 shows the schematic of the strain gauges connections with electrical rotating connector accompanied with Wheatstone bridge circuits. The calibration curve of load versus strain was obtained by 10 N standard weights for the ring (see Fig. 7a). The normal load can be measured from the curve shown in Fig. 7a. The friction force, created by normal force and sliding movement between the specimen and pads, was calibrated before testing. Four strain gauges, two in direction of friction and two at right angle to this direction, were bonded to the pads to measure the elastic strain induced by friction force. This type of strain gauge configuration would cancel out the effect of temperature rise due to the fatigue loading in contact area. The pads were clamped to the specimen by the same Chuck Bearings Coupling d b F Fig. 8 Schematic of the main shaft accompany with the position of bending force and shaft's supports

6 2216 Int J Adv Manuf Technol (2014) 70: Table 1 Chemical composition of AL 7075-T6 (wt%) Al Cu Si Mg Cr Zn Mn normal loads as was to be used in real experiments. The specimen was then pulled by constant load using INSTRON tensile testing machine. The strain induced by this load was measured by Wheatstone Circuit Bridge attached to the pads. This procedure was repeated for different loads and the calibration curve of friction force versus strain was obtained. The calibration curve is depicted in Fig. 7b. The coefficient of friction (μ) can be determined by dividing the maximum measured friction force at a fretting pad foot by the contact load at that foot. 2.3 Instrumentation There are two load cells in RBFF machine for applying the equal loads on the specimen. The bending load is measured by a SMSB type 50 kg load cell (see Fig. 1b). The variation of load monitored by a digital displayer and can be transferred to a computer for further processing. The second load cell, which is placed bellow the adjustable bearing (support), is used to measure the load which is needed to convey the maximum stress to the middle of the narrow part of the specimen. This load should be equal to bending load for avoiding any shear force in specimen. The frequency of the device can vary using an inverter up to a maximum of 50 Hz. The machine is equipped with a cutoff switch which acts when specimen break down to avoid unnecessary revolution. There is a stationary microscope to see and capture an image from the cracks and propagation of cracks under the friction pads on the specimens when the RBFF machine is turned off. An electrical slip ring is used to transfer data from rotating strain gauges to stationary outlets. A data acquisition is utilized to measure coefficient of friction (data come from strain gauges) with a computer. 3 Design of RBFF machine's parts 3.1 Design of the main shaft The most important issues in designing of a shaft are stress and deflection. With choosing proper material and calculation of the shaft's diameter, this problem can be solved. A schematic of the main shaft, the supports, and the bending force accompanying their locations are depicted in Fig. 8. In order to minimize deflections, steel is the logical choice for a shaft material because of its high modulus of elasticity [15]. In this research work, the material of main shaft is selected SAE1020 steel, which has σ ut =65 kpsi, S y =38 kpsi. The mean shaft's deflection was obtained by Eq. (1), where y is the deflection of the shaft, E is the modulus of elasticity, and I is the momentum of inertia. y ¼ Fbþ ð c þ d 3EI Þ3 ð1þ The shaft's diameters were obtained by Eq. (2), where d is diameter of the shaft, N f is the safety factor, k f is the fatigue stress-concentration factor, M a is the alternating bending moment, S f is the fatigue strength, k fsm is the fatigue factor for mean torsional stress, T m is the transmitted torque which was calculated by Eq. (3), and S y yield stress [18]. d ¼ T m ¼ P ω ( " 32N f M 2 # )1 a T 2 3 m k f þ k fsm π S f S ð2þ y ð3þ Where P is the power of motor which is 1 hp and ω is angular velocity which is 2,800 rpm. Equations (4) and(5) are state fatigue strength and endurance limit of the material. Where Cs are correction factors; in this case, the amount are equal to one, except C surface, which is considered 0.84 for surface of steel [18]. Table 2 The plain fatigue and fretting fatigue lives of Al7075- T6 alloy using RBFF and Shimadzu fatigue test machines Stress (MPa) Shimadzu RBFF Plain fatigue error % Plain fatigue Fretting fatigue

7 Int J Adv Manuf Technol (2014) 70: Bending stress (MPa) Plain Fatigue E E E+07 No. of cycle to failure Fretting Fatigue Fig. 9 S/N curve of fatigue and fretting fatigue lives of Al7075-T6 alloy using RBFF machine Number of cycle % 8.22% 5.31% 4.54% SHIMADZU 3.45% 7.82% RBFF 5.43% Bending stress (MPa) Fig. 11 Comparison between fretting fatigue lives of Al7075-T6 alloy using Shimadzu and RBFF machines accompany with error percentage S f ¼ S e ¼ C load C size C surface C temp C reliable S e 0 ð4þ holder are 9,500 and 22,000 rpm, respectively, which are higher than the operating speed of 2,800 rpm. S e 0 ¼ 0:5S ut ð5þ 3.2 Design of the bearings There are three bearings in RBFF machine, two bearings for supporting the main shaft and one bearing to support the specimen. The aim of selecting the appropriate bearings is to find the bearings fatigue live at main shaft and specimen holder. The maximum moment is applied on point B, as it canbeseeninfig.8; it means that the support at point B suffers more force than other support at point A. From the calculation of the proper diameter for the shaft from previous section, the inner diameter of bearings is known. Using Eq. 6, the fatigue lives of bearings are found [18]. L ¼ C 3 ð6þ R Whereas L is fatigue life, C is dynamic load rating, and R is applied load on bearing A. The fatigue lives of bearings are around 9E10 for bearing A, 3E11 for bearing B, and 5.5E9 for bearing at point D, which support the specimens, respectively. Bearings limiting speed for shaft supports and specimen Coefficient of friction MPa Number of cycles Fig. 10 Friction coefficient versus fretting fatigue life 4 Experimental details to investigate the RBFF machine performance Aluminum alloy 7075-T6 was used in this investigation. The material's composition obtained using EDX apparatus is illustrated in Table 1. From a number of tensile tests, the yield stress and ultimate strength of Al7075-T6 were obtained as follows: σ y =520 MPa and σ ut =590 MPa, respectively. The round shape specimens used in this work were prepared in accordance with ISO standard. Fretting fatigue pads were fabricated from stainless steel 410 plate. The specimens were gripped and loaded rotationally in a rotating bending fretting fatigue test apparatus. When a fatigue specimen is subjected to cyclic stresses, fretting between the contact pads and specimen is generated. Plain and fretting fatigue testing were carried out at room temperature in a two-point loading rotating bending machine (R = 1) under constant stress amplitude at a rotational speed of 3,000 rpm. The relationship between the stress amplitude and the number of cycles to failure for the all the condition analyzed is defined by Eq. 1 [19, 20]. S ¼ AN b f ð7þ Where S is stress amplitude, A is fatigue strength coefficient, b is fatigue strength exponent, and N f is number of cycles to failure. 5 Fretting fatigue test results In order to examine the performance of the fretting fatigue testing apparatus, RBFF, a number of specimens were tested in this investigation. The experiments were conducted for stress ratio of R = 1, 50 Hz (for RBFF) at a constant contact

8 2218 Int J Adv Manuf Technol (2014) 70: Fig. 12 Cross-section view of fracture specimens under fretting fatigue a at bending stress 200 MPa and 1.5E10 6 cycles and b at bending stress 300 MPa and 1.2E10 5 cycles (a) Fretting zone Tensile zone (b) Fretting zone Tensile zone Fretting zone Fretting zone 1mm 1mm force of 1,200 N, and working stress amplitudes of MPa. A number of tests were also performed using rotating bending fatigue (Shimadzu, Japan) under the same conditions as stated above for comparison purposes. The plain fatigue and fretting fatigue results are given in Table 2.As the results indicate, the fatigue lives obtained from the Shimadzu testing machine and RBFF are very close. Figure 9 shows S/N curve of plain fatigue and fretting fatigue lives of Al7075-T6 alloy using RBFF machine. As it can be seen from the figure, fretting decreases the fatigue life of Al7075-T6 alloy drastically. The friction coefficient is calculated from the relation μ =F/P, inwhichp is the normal force applied by the adjusting bolts to the pads and is kept constant throughout the experiments (P =1,200 N). F is also measured from the experiments using the calibration curve shown in Fig. 7b. A typical curve of friction coefficient versus fretting fatigue life is shown in Fig. 10. Figure 11 presents comparison between fatigue lives of Al7075-T6 alloy using Shimadzu and RBFF machines accompanied with error percentage. As it can be seen from the figure, the results are in close agreement and the maximum and minimum errors obtained are 8.22 and 2.78 %, respectively. Fracture surface of tested specimens were examined using optical microscopy. A typical fractured specimen and crosssectional views are illustrated in Fig. 12. The figures clearly indicates that the fracture surface consists of two quite distinct regions; a fatigue zone created by crack propagation and a tensile region which gives rise to fracture of specimen when it is sufficiently weakened by the crack zone development. Crack-propagation growth rate is very small, on the order of 10 8 to 10 4 in per cycle, but this adds up over a large number of cycles. If the failed surface is viewed at high magnification, the striations due to each stress cycle can be seen as in Fig. 12, which shows the fracture specimens at 40 magnification along with a representation of the stress-cycle pattern that failed it. Figure 12a, b present the cross-section view of fracture specimens under fretting fatigue at bending stress of 200 MPa and 1.5E10 6 cycles and at bending stress of 300 MPa and 1.2E10 5 cycles, respectively. The occasional large-amplitude stress cycles show up as larger striations than the more frequent small amplitude once, indicating that higher stress amplitudes cause larger crack growth per cycle [18, 21]. 6 Conclusions A newly fretting fatigue testing machine design under rotating bending approach is presented by the authors in this work. The machine is capable of exerting adjustable bending and contact loads. The frequency of the apparatus can be varied up to 50 Hz. The performance of the machine is validated by fatigue tests which carried out using the Shimadzu fatigue testing machine. The result shows a close agreement between RBFF and Shimadzu fatigue testing machine. The device has a simple variable eccentric system for fluctuating loading. The system is pure mechanical with no need for hydraulic systems which are sophisticated, expensive, and slow. The machine is cheap, light, small, and manageable. The device is well instrumented for varying and measuring axial loads, fatigue cycles, and test frequency. Furthermore, RBFF machine has the capability to measure the friction coefficient between specimen and fretting pads, while the specimen is rotating by using an electrical rotating connector. Acknowledgments The authors acknowledge the financial support under the research grant with no. UM. TNC2/IPPP/UPGP/261/15 (BK ) from the University of Malaya, Malaysia. References 1. Lietch LC, Lee H, Mall S (2005) Fretting fatigue behavior of Ti-6Al- 4V under seawater environment. Mater Sci Eng A 403: Lee H, Mall S, Sathish S (2005) Investigation into effect of re-shotpeening on fretting fatigue behavior of Ti-6Al-4V. Mater Sci Eng A Suresh S (1998) Fatigue of materials. Cambridge University Press, Cambridge 4. Lee H, Jin O, Mall S (2003) Fatigue fract. Eng Mater Stract 26:

9 Int J Adv Manuf Technol (2014) 70: Huston A, Nicholas T, John R (2005) Fretting fatigue crack analysis in Ti-6Al-4V. Int J Fatigue 27: Mall S, Jain VK, Nanjoshi SA, Lykins CD (2003) In: Kinyon SE, Hoeppner DW, Mutoh Y (eds) Fretting fatigue: advances in basic understanding and application, ASTM STP ASTM International, West Conshohocken, pp Arora PR, Jacob MSD, Salti MS, Ahmed EM, Saleem M, Edi P (2007) Experimental evaluation of fretting fatigue test apparatus. Int J Fatigue 29: Hoeppner DW, Chandrasekaran V, Elliot CB (eds) Fretting fatigue: current technologies and practices, ASTM STP Hutson AL, Neslen C, Nicholas T (2003) Characterization of fretting fatigue crack initiation processes in Ti-6Al-4V. Tribol Int 36: Majzoobi GH, Hojjati R, Nematian M, Zalnejad E, Ahmadkhani AR, Hanifepoor E (2010) A new device for fretting fatigue testing. Trans Indian Inst Metals 63: Jin O, Mall S (2004) Shear force effect on fretting fatigue behavior of Ti-6Al-4V. Metall Mater Trans 35A: Zalnezhad E, Sarhan AAD, Hamdi M (2013) Investigating the fretting fatigue life of thin film titanium nitride coated aerospace Al7075-T6 alloy. Mater Sci Eng A 559: doi: /j.msea Naidu NKR, Raman SGS (2005) Effect of contact pressure on fretting fatigue behavior of AL-Mg-Si alloy AA6061. Int J Fatigue 27: Sarhan AAD, Zalnezhad E, Hamdi M (2013) The influence of higher surface hardness on fretting fatigue life of hard anodized aerospace AL7075-T6 Alloy. Int J Mater Sci Eng A 560: doi: / j.msea Ebara R, Fujimura M (2006) Fretting fatigue behavior of Ti 6Al 4V alloy under plane bending stress and contact stress. Tribol Int 39: Araujo JA, Nowell D (2002) The effect of rapidly varying contact stress fields on fretting fatigue. Int J Fatigue 24: Jin O, Mall S (2004) Effect of slip on fretting behavior: experiments and analysis. Wear 256: (2010) Machine design, an integrated approach, R. L. Norton, worecester politechic institute, Worcester, Massachusetts 19. Lin SK, Lee YL, Lu MW (2001) Evaluation of the staircase and the accelerated test methods for fatigue limit distributions. Int J Fatigue 23: Mugadu A, Hills DA, Nowell D (2002) Modifications to a frettingfatigue testing apparatus based upon an analysis of contact stresses at complete and nearly complete contacts. Wear 252: ASM Handbook (1997) Fatigue failures, Desk ed. ASM International