- Email: [email protected]
Effect of shot blasting on plain fatigue and fretting fatigue behaviour of Al–Mg–Si alloy AA6061
International Journal of Fatigue 27 (2005) 323–331 www.elsevier.com/locate/ijfatigue
Technical Note
Effect of shot blasting on plain fatigue and fretting fatigue behaviour of Al–Mg–Si alloy AA6061 N.K. Ramakrishna Naidu, S.Ganesh Sundara Raman* Department of Metallurgical & Materials Engineering, Indian Institute of Technology Madras, Chennai 600 036, India Received 24 September 2003; received in revised form 17 June 2004; accepted 26 July 2004
Abstract The present work describes the effect of shot blasting on plain fatigue (without fretting) and fretting fatigue behaviour of Al–Mg–Si alloy AA6061. Shot blasting significantly increased the plain fatigue life by a factor 2.8 and fretting fatigue life by a factor 2.4 at a maximum cyclic stress (smax) of 169 MPa, but at higher stress levels shot blasting slightly reduced both the plain fatigue and the fretting fatigue lives. At a maximum cyclic stress of 169 MPa, the stabilized value of coefficient of friction in the shot blasted condition was lower than that in the T6 condition (0.55 against 0.60), but at a maximum cyclic stress of 265 MPa, it was around 0.82 in the T6 condition and 0.84 in the shot blasted condition. q 2004 Elsevier Ltd. All rights reserved. Keywords: Fatigue; Fretting fatigue; Shot blasting; Aluminium alloy; Coefficient of friction
1. Introduction Fretting is a small amplitude oscillatory relative sliding motion which may occur between two contacting surfaces, which are nominally at rest with respect to each other, subjected to vibrations or cyclic loads. Fretting fatigue is the combined action of fretting and fatigue. Fretting greatly accelerates the fatigue crack initiation process. The joint action of fretting and fatigue may produce strength reduction factors of 2–5 or even greater [1]. There are many surface modification methods used to mitigate fretting damage [2–5]. However, the factors responsible for the improvements in fretting fatigue resistance are quite different. Five main factors [6] to enhance fretting fatigue resistance can be summarized as: (a) inducing compressive residual stresses; (b) decreasing the coefficient of friction; (c) increasing the hardness (to prevent adhesion and to increase abrasive wear resistance); (d) increasing the surface roughness; (e) altering the surface chemistry. * Corresponding author. Tel.: C91 44 2257 8608; fax: C91 44 2257 0509. E-mail address: [email protected] (S.G.S. Raman). 0142-1123/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijfatigue.2004.07.007
Inducing compressive residual stresses in the surfaces of materials, is one of the most widely used surface engineering techniques in improving the fatigue life. Shot peening is one of the best methods of inducing the compressive residual stresses. The residual surface compressive stresses reduce the possibility of a propagating fatigue crack by reducing the peak applied tensile stress. The localized plastic flow at the surface, resulting in the shot peening process, causes work hardening of the surface, general roughening of the surface along with the generation of the compressive residual stresses [7]. All of these factors can be expected to affect both the plain fatigue (without fretting) and fretting fatigue properties of a material. The deliberate introduction of residual compressive surface stresses by shot peening is beneficial in fretting fatigue as well as in plain fatigue. However, in fretting fatigue, there are further two advantages. Waterhouse [8] has reported that coefficient of friction measured during fretting fatigue tests was lower in shot peened condition than that in unpeened condition and also the rough surface resulting from shot peening was beneficial in improving fretting fatigue life. In case of a rough surface, the real surface contact areas are broken down into small discrete areas and hence the volume of material affected by fretting
324
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
action in any one contact is much reduced leading to a reduced possibility of initiating a crack [8]. Improvements in fretting fatigue strength due to shot peening have been reported in different materials, e.g. steels, Al, Mg and Ti alloys. Mutoh et al. [9] have reported that shot peening improved the fretting fatigue strength of a steam turbine steel by a factor of 1.8 at room temperature and 773 K. They have observed that the shot peening was effective even for long-term practical use at elevated temperatures. Kondoh et al. [10] have studied the effect of a wide peening and cleaning (WPC) treatment on the fretting fatigue strength of AA6061-T6 material. In the WPC treatment, steel balls were shot on the test material for 30 s followed by the shot using ceramic balls for 30 s. The WPC treatment enhanced the value of fretting fatigue strength from 50 to 70 MPa. Lindley and Nix [11] have reported that in AA2014 aluminium alloy, the most effective single palliative treatment to combat fretting fatigue was glass bead peening, which introduced surface compressive residual stresses and inhibited the growth of small fretting cracks. Shot blasting is one of the most frequently used treatments for getting rough surface prior to plasma spraying to ensure a strong mechanical bond between coating and substrate. While shot peening is intentionally employed to improve fatigue strength/life of materials, shot blasting is used only for preparing the surface to be coated. In the present work, an attempt has been made to study the effect of shot blasting on the plain fatigue (without fretting) and fretting fatigue behaviour of Al–Mg–Si alloy AA6061.
2. Experimental details 2.1. Material, test specimen design and shot blasting The chemical composition of the material AA6061 used in the present investigation is given in Table 1 (taken from Ref. [12]). 8.3 mm thick specimens of 65 mm gauge length and 10 mm width were used for both plain fatigue and fretting fatigue tests (Fig. 1). The alloy was solution treated at 540 8C for 1 h and water quenched to room temperature and subsequently aged at 160 8C for 12 h. This heat treatment will be referred to as T6 condition. The room temperature mechanical properties of the material in the T6 condition are given in Table 2 (taken from Ref. [12]). Shot blasting, using spherical balls of aluminium oxide abrasives of size 120 mm, was done on some specimens
(in T6 condition) on the thickness side, where fretting was induced during fretting fatigue testing. The shot blasting details are: 100 mm stand off distance, 0.4 MPa average blasting pressure, 100 s blasting time, 100% coverage (coverage is the proportion of surface area blasted) and nozzle kept normal to the surface blasted. It may be noted that the normal blasting pressure used for getting a rough surface before plasma spraying is 0.245 MPa. However, in order to obtain higher residual compressive stresses, the blasting pressure was increased to 0.4 MPa in the present study. The blasting time of 100 s was chosen based on many shot blasting trials with different blasting duration. The magnitude of the residual compressive stresses induced due to shot blasting was determined as 110 MPa by the X-ray diffraction technique employing Cu Ka radiation. Higher values of residual stresses would have resulted if glass beads or steel shots were used. However, it was decided to study only the influence of shot blasting in the present investigation. The depth of the effect of shot blasting was not determined. 2.2. Fretting fatigue test assembly An experimental facility, with a ring-type load cell and bridge-type fretting pads, which can simulate the fretting fatigue conditions was designed and fabricated. Fretting pads were made of the same test material AA6061 alloy in the T6 condition. Fig. 2 shows schematic diagrams of the fretting fatigue test assembly and fretting pad assembly. Fig. 3 shows the photograph of the test assembly used in the present study. Proving ring (load cell), which was used to apply contact load in fretting fatigue tests, had strain gauges pasted on opposite sides of the ring (two on the inner surface and two on the outer surface). For calibration, the ring was loaded (through the axis of the load adjusting screw) in a Schenck servo-hydraulic testing machine and the elastic strains induced in the ring were recorded by means of a strain amplifier and data acquisition system (HBM-Spider8600 Hz and catman Express 4.0 software). Then a calibration curve of load vs. ring strain was plotted and the slope was used to determine the contact (normal) load. Bridge-type fretting pads, shown in Fig. 2(b), with a pad span of 30 mm were clamped to the specimen by means of the proving ring as shown in Fig. 3. Frictional force between the fretting pad and the specimen was measured by bonding strain gauges to the underside of the fretting pads, with the strain gauge grid centered between the pad feet. Split specimen technique [13] was employed to calibrate the fretting pads. 2.3. Test procedure
Table 1 Chemical composition of the material used (wt%) [12] Mg
Si
Fe
Mn
Cr
Cu
Zn
Ti
Al
0.63
0.55
0.21
0.30
0.25
0.31
0.25
0.15
Rest
Plain fatigue and fretting fatigue tests were conducted in laboratory air (approximately 60–70% relative humidity) at room temperature on the Schenck servo-hydraulic machine
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
325
Fig. 1. Specimen design (not to scale; all dimensions are in mm).
with a stress ratio of 0.1 at different cyclic stress amplitudes (see Table 3). For plain fatigue tests, the cycle frequency was 30 Hz and for fretting fatigue tests it was 10 Hz. A constant contact pressure of 100 MPa was used in the fretting fatigue tests. During fretting fatigue testing, the friction force values were recorded, and the contact load applied by the proving ring was monitored and maintained constant throughout the test with the help of the data acquisition system. The fretting scar regions in the tested specimens were observed at low magnification and photographs were taken. The roughness profiles across the fretting scar were obtained using a perthometer. A scanning electron microscope was employed to observe the fretting scars and the fracture surfaces of the specimens.
3. Results and discussion Fig. 4 shows the results corresponding to plain fatigue (PF) and fretting fatigue (FF) tests. As the tests were conducted with a stress ratio of 0.1, mean stress (smean) values were different for different tests (see Table 3). King and Lindley [14] have shown that the mean stress has greater effect on fretting fatigue strength than on plain fatigue strength. They have reported that raising the mean stress from 0 to 300 MPa reduced the plain fatigue strength of a low alloy rotor steel by 27% whereas in fretting fatigue, the reduction was 60%. In the present study, as the mean stress used was in the range 93–146 MPa, it was assumed that the effect of mean stress would not have been very significant. The fretting fatigue lives (Nf fretting fatigue) were shorter compared to plain fatigue lives (Nf plain fatigue). The fatigue life was reduced due to fretting to a greater extent at low cyclic stress levels. The life reduction factor (defined as Nf plain fatigue/Nf fretting fatigue) was 4.0 for the specimens in the T6 condition and 4.6 for the specimens in the shot blasted condition at a maximum cyclic stress (smax) of 169 MPa. Since the fretting plays a severe role in crack Table 2 Mechanical properties of the material used (in T6 condition) [12] Yield strength (MPa)
Ultimate tensile strength (MPa)
Elongation (%)
Hardness HV 5
281
348
16
115
Fig. 2. (a) Fretting fatigue test assembly. (b) Fretting pad assembly.
326
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
Fig. 3. Photograph of the test assembly used in the present study.
initiation period and as crack initiation occupies a major portion of the total life, the influence of fretting on the fatigue life may be expected to be of a greater extent at lower stresses. The beneficial effect of shot blasting in enhancing the life was clearly evident in both plain fatigue and fretting fatigue tests at low stress levels. The life improvement factor due to shot blasting was 2.8 in case of plain fatigue loading and it was 2.4 in case of fretting fatigue loading corresponding to a maximum cyclic stress of 169 MPa. As the cyclic stress increased, the beneficial effect decreased. At the highest maximum cyclic stress of 265 MPa, the effect of shot blasting was detrimental under both plain fatigue and fretting fatigue loading conditions. However, the reduction in life was not significant. The influence of shot blasting on plain fatigue life may be explained as follows. The shot blasting may be expected to give rise to similar effects on the surface as that of shot peening—localized plastic flow causing work hardening, general roughening and generation of compressive residual stresses. All these effects are expected to influence the fatigue life. At low cyclic stresses, the favourable effects of compressive residual stresses and work hardening may be expected to dominate over the detrimental effect of surface roughening, and hence resulting in life enhancement. As the cyclic stress increases, the residual stresses will relax quickly and the extent of life enhancement will decrease. At a particular stress level, the effect of residual stress and work hardening would balance the effect of surface roughening. Table 3 Stress values employed in the present study (stress ratio, 0.1) Cyclic stress amplitude, sa (MPa)
Mean stress, smean (MPa)
Maximum cyclic stress, smax (MPa)
76.05 86.85 97.65 108.45 119.25
92.95 106.15 119.35 132.55 145.75
169 193 217 241 265
Fig. 4. Effect of shot blasting on the plain fatigue (PF) and fretting fatigue (FF) lives.
Hence, shot blasting would not have any effect on the fatigue life at all. However, above this stress level, the adverse effect of roughening and/or surface cracking would dominate leading to reduction in life [15]. The effect of shot blasting on fretting fatigue life may be explained based on the values of coefficient of friction (Zfrictional force/normal load) in addition to the effect of residual stresses, work hardening and surface roughening (see Figs. 5 and 6). Fig. 5 shows the variation of coefficient of friction (m) at the fretting interface with number of fretting cycles at different cyclic stress levels in both T6 and shot blasted conditions. The friction force is considered to be created by either direct interlocking of surface asperities, or by trapping oxide debris in between the surface asperities. The coefficient of friction increased rapidly during the early stages of fretting fatigue life (below 3000 cycles) and then it remained almost constant for the remaining life period. Rayaprolu and Cook [13] have given an explanation for this type of behaviour. Under constant amplitude loading conditions, an initial ‘bedding-in’ phase, associated with gross slip between the contacting bodies is present. The bedding-in period consists of a gradual build up in the magnitude of frictional forces and the reduction in the degree of macroslip. During the macroslip conditions, larger contacting asperities are worn out leading to an increase in the area of asperity contact, which promotes microslip conditions. Towards the end of the bedding-in process, under microslip conditions, a point is reached where the entire frictional load causes only elastic deformation of contacting asperities. From this point a relatively constant value of frictional force occurs throughout the remainder of the fretting fatigue cycles. Fig. 6 shows the variation of stabilized m values with the maximum cyclic stress. A cross over in the plots for
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
327
stresses. Similar remarks have been made by Lindley and Nix [11] for the case of shot peening. Fig. 7(a)–(d) shows the appearance of fretting scars on tested specimens in the T6 condition and the shot blasted condition corresponding to two different cyclic stresses. It can be observed that the fretting damage width at high stress level is more than that at low stress level. It is very clear that the extent of fretting damage induced in the shot blasted condition is less than that in the T6 condition. This effect may be due to increased surface roughness, increased hardness due to work hardening of the surface and nonadhesion of the surface. The roughness profiles taken across fretting scar regions of the specimens are shown in Fig. 8. The middle portion of the profiles shown within dotted lines correspond to the fretting scar region and the end portions correspond to unfretted region. It is evident that the fretting damage increased the surface roughness of the specimens in the T6 condition. The roughness value (Ra) for a specimen tested at a higher stress was more than that for a sample tested at a lower stress. In case of the shot blasted specimens, the roughness value for a specimen tested at a higher stress was lower than that for a sample tested at a lower stress. This may be due to high rates of wear and large slip ranges at higher stresses. It may be noted that the shot blasted specimens had rough surfaces prior to fretting fatigue deformation. Due to the deformation, the surface ondulations were reduced and the roughness value decreased. As the cyclic stress increased, wear rate and relative slip increased. So the extent of reduction in the surface roughness value from the original value was more for samples tested at higher cyclic stresses. Fig. 9(a)–(d) shows scanning electron micrographs of the scars. A fatigue crack initiated from the fretting damage region may be seen. The roughness introduced due to shot blasting may be seen in Fig. 9(c) and (d). Fig. 5. Variation of coefficient of friction with number of cycles. (a) T6 condition; (b) shot blasted condition.
the material in T6 and shot blasted conditions may be seen similar to that was observed in Fig. 4. At a maximum cyclic stress of 169 MPa, the stabilized value of m in the shot blasted condition is 0.55 compared to 0.60 in the T6 condition. As the maximum cyclic stress increases, the difference between the values of m in both the conditions decreases. At a maximum cyclic stress of 265 MPa, the value of m is 0.82 in the T6 condition against 0.84 in the shot blasted condition. Though the difference between the values of coefficient of friction for both the conditions is small, it is assumed that in addition to the three effects discussed earlier, the difference in m values may also be responsible for the difference in the fretting fatigue lives of specimens in the T6 and the shot blasted conditions. Large relative slip ranges and high rates of wear at high cyclic stresses would have reduced the benefits of shot blasting by removing the near surface layer containing the compressive residual
Fig. 6. Variation of stabilized value of coefficient of friction with maximum cyclic stress (smax).
328
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
Fig. 7. Appearance of fretting scars in specimens tested at different maximum cyclic stress levels (smax). (a) T6 condition, 265 MPa; (b) T6 condition, 169 MPa; (c) shot blasted condition, 265 MPa; (d) shot blasted condition, 169 MPa.
Fig. 8. Roughness profiles taken across fretting scar in specimens tested at different maximum cyclic stress levels (smax).
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
329
Fig. 9. Fatigue crack (shown by an arrow) initiated in fretting damage region of specimens tested at different maximum cyclic stress levels (smax). (a) T6 condition, 265 MPa; (b) T6 condition, 169 MPa; (c) shot blasted condition, 265 MPa; (d) shot blasted condition, 169 MPa.
The crack initiation in the specimens both in the T6 condition as well as the shot blasted condition was at the trailing edge of the fretting scar region. However, in the shot blasted specimens, the initial direction of the crack followed
a much shallower angle to the surface. As had been pointed out by Leadbeater et al. [15], it is a direct consequence of the residual compressive stresses existing in the surface. This had the effect of moving the principal tensile stress in
Fig. 10. Fracture surface of a shot blasted specimen tested under plain fatigue with a maximum cyclic stress (smax) of 217 MPa. (a) Near crack initiation region; (b) striations in stage II propagation region; (c) final overload region.
330
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
Fig. 11. Fracture surface of a specimen in the T6 condition tested under fretting fatigue with a maximum cyclic stress (smax) of 169 MPa. (a) Multi crack initiation sites; (b) transition from stage I to stage II propagation region.
the surface layers to a direction more nearly normal to the surface. A shallow angle of crack propagation resulted since a fatigue crack will always propagate in a direction perpendicular to the principal tensile stress. However, once the crack had moved out of the zone influenced by the fretting stresses, the direction of cracking was always at 908 to the principal tensile stress. Fig. 10(a)–(c) shows the appearance of fracture surfaces of a shot blasted specimen subjected to plain fatigue. No distinct features were observed near the crack initiation zone (Fig. 10(a)). Secondary cracking perpendicular to the main crack plane was observed. Striations were clearly observed in the stage II propagation region (Fig. 10(b)). Holes seen in the figure are the prior inclusion sites. Fig. 10(c) shows the final overload fracture region containing dimples. The fracture surfaces of samples in the T6 condition and the shot blasted condition tested under fretting fatigue loading showed similar features. Multi crack initiation sites were seen (see Fig. 11(a)). The transition from stage I to stage II of fatigue fracture was characterized by a change of orientation of the main fracture plane in each grain from one or two shear planes to many parallel plateaus separated by longitudinal ridges (see Fig. 11(b)).
4. Conclusions Based on the results obtained in the present study on the effect of shot blasting on the plain fatigue and fretting fatigue behaviour of AA6061 material, the following conclusions were drawn. (a) Fatigue life reduction due to fretting was high at low cyclic stress levels. At a maximum cyclic stress (smax) of 169 MPa, the life reduction factor was 4.0 in the T6 condition and 4.6 in the shot blasted condition. Towards higher cyclic stress levels, the extent of reduction in life decreased. (b) Shot blasting significantly increased the plain fatigue life by a factor 2.8 and fretting fatigue life by a factor
2.4 at a maximum cyclic stress of 169 MPa. However, towards higher stress levels, the extent of increase in fatigue lives decreased and at a maximum cyclic stress of 265 MPa both plain fatigue and fretting fatigue lives slightly reduced in the shot blasted condition. (c) The coefficient of friction (m) at the fretting interface increased rapidly during the early stages of fretting fatigue life (!3000 cycles) and then it remained almost constant. (d) At a maximum cyclic stress of 169 MPa, the stabilized value of m was less in the shot blasted condition than that in the T6 condition (0.55 against 0.60). As stress levels increased, the difference between the values of m in both the conditions decreased. At a maximum cyclic stress of 265 MPa, the value of m was less in the T6 condition than that in the shot blasted condition (0.82 against 0.84). (e) Shot blasting can also be used as one of the cheapest and best methods to improve the fretting fatigue life at low service loads.
Acknowledgements This work was supported by the Department of Science and Technology, Government of India under SERC Fast Track Scheme 2001–2002 (project number SR/FTP/ET183/2001).
References [1] Waterhouse RB. Fretting fatigue. Int Mater Rev 1992;(37):77–97. [2] Beard J. Palliatives for fretting fatigue. In: Waterhouse RB, Lindley TC, editors. Fretting fatigue, ESIS 18. London: Mechanical Engineering Publications; 1994. p. 419–36. [3] Gordelier SC, Chivers TC. A literature review of palliatives for fretting fatigue. Wear 1979;56:177–90. [4] Chivers TC, Gordelier SC. Fretting fatigue palliatives: some comparative experiments. Wear 1984;96:153–75. [5] Gabel MBK, Bathke JJ. Coatings for fretting prevention. Wear 1978; 46:81–96.
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331 [6] Fu Y, Wei J, Batchelor AW. Some considerations on the mitigation of fretting damage by the application of surface modification technologies. J Mater Proc Technol 2000;99:231–45. [7] Li CX, Sun Y, Bell T. Shot peening of plasma nitrided steel for fretting fatigue strength enhancement. Mater Sci Technol 2000;(16): 1067–72. [8] Waterhouse RB. Effect of material and surface condition on fretting fatigue. In: Waterhouse RB, Lindley TC, editors. Fretting fatigue, ESIS 18. London: Mechanical Engineering Publications; 1994. p. 339–49. [9] Mutoh Y, Satoh T, Tsunoda E. Improving fretting fatigue strength at elevated temperatures by shot peening in steam turbine steel. In: Attia MH, Waterhouse RB, editors. Standardization of fretting fatigue test methods and equipment, ASTM STP 1159. Philadelphia: American Society for Testing and Materials; 1992. p. 199–209. [10] Kondoh K, Mutoh Y, Mabuchi M, Kobayashi T. Fretting fatigue behavior of surface-treated aluminium alloys. In: Wu X-R, Wang ZG, editors. Fatigue’99: Proceedings of the international congress, 8–12 June, vol. 2. Beijing: Higher Education Press; 1999. p. 1347–51.
331
[11] Lindley TC, Nix KJ. Fretting fatigue in the power generation industry: experiments, analysis and integrity assessment. Fatigue’99: Proceedings of the international congress, 8–12 June 1999. p. 153–169. [12] Karunanidhi S. Fatigue and short crack growth behaviour of an age hardenable Al–Mg–Si alloy. MS Thesis. IIT Madras; 1999. [13] Rayaprolu DB, Cook R. A critical review of fretting fatigue investigations at the royal aerospace establishment. In: Attia MH, Waterhouse RB, editors. Standardization of fretting fatigue tests and equipment, ASTM STP 1159. Philadelphia: American Society for Testing and Materials; 1992. p. 129–52. [14] King RN, Lindley TC. Fretting fatigue in a 3.5%NiCrMoV rotor steel. Advances in fracture research, vol. 2. Oxford: Pergamon Press; 1981. p. 631–40. [15] Leadbeater G, Noble B, Waterhouse RB. The fatigue of aluminium alloys produced by fretting on a shot peened surface. In: Valluri SR, editor. Advances in fracture research. Proceedings of the sixth international conference on fracture, vol. 3. Oxford: Pergamon Press; 1984. p. 2125–32.
Technical Note
Effect of shot blasting on plain fatigue and fretting fatigue behaviour of Al–Mg–Si alloy AA6061 N.K. Ramakrishna Naidu, S.Ganesh Sundara Raman* Department of Metallurgical & Materials Engineering, Indian Institute of Technology Madras, Chennai 600 036, India Received 24 September 2003; received in revised form 17 June 2004; accepted 26 July 2004
Abstract The present work describes the effect of shot blasting on plain fatigue (without fretting) and fretting fatigue behaviour of Al–Mg–Si alloy AA6061. Shot blasting significantly increased the plain fatigue life by a factor 2.8 and fretting fatigue life by a factor 2.4 at a maximum cyclic stress (smax) of 169 MPa, but at higher stress levels shot blasting slightly reduced both the plain fatigue and the fretting fatigue lives. At a maximum cyclic stress of 169 MPa, the stabilized value of coefficient of friction in the shot blasted condition was lower than that in the T6 condition (0.55 against 0.60), but at a maximum cyclic stress of 265 MPa, it was around 0.82 in the T6 condition and 0.84 in the shot blasted condition. q 2004 Elsevier Ltd. All rights reserved. Keywords: Fatigue; Fretting fatigue; Shot blasting; Aluminium alloy; Coefficient of friction
1. Introduction Fretting is a small amplitude oscillatory relative sliding motion which may occur between two contacting surfaces, which are nominally at rest with respect to each other, subjected to vibrations or cyclic loads. Fretting fatigue is the combined action of fretting and fatigue. Fretting greatly accelerates the fatigue crack initiation process. The joint action of fretting and fatigue may produce strength reduction factors of 2–5 or even greater [1]. There are many surface modification methods used to mitigate fretting damage [2–5]. However, the factors responsible for the improvements in fretting fatigue resistance are quite different. Five main factors [6] to enhance fretting fatigue resistance can be summarized as: (a) inducing compressive residual stresses; (b) decreasing the coefficient of friction; (c) increasing the hardness (to prevent adhesion and to increase abrasive wear resistance); (d) increasing the surface roughness; (e) altering the surface chemistry. * Corresponding author. Tel.: C91 44 2257 8608; fax: C91 44 2257 0509. E-mail address: [email protected] (S.G.S. Raman). 0142-1123/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijfatigue.2004.07.007
Inducing compressive residual stresses in the surfaces of materials, is one of the most widely used surface engineering techniques in improving the fatigue life. Shot peening is one of the best methods of inducing the compressive residual stresses. The residual surface compressive stresses reduce the possibility of a propagating fatigue crack by reducing the peak applied tensile stress. The localized plastic flow at the surface, resulting in the shot peening process, causes work hardening of the surface, general roughening of the surface along with the generation of the compressive residual stresses [7]. All of these factors can be expected to affect both the plain fatigue (without fretting) and fretting fatigue properties of a material. The deliberate introduction of residual compressive surface stresses by shot peening is beneficial in fretting fatigue as well as in plain fatigue. However, in fretting fatigue, there are further two advantages. Waterhouse [8] has reported that coefficient of friction measured during fretting fatigue tests was lower in shot peened condition than that in unpeened condition and also the rough surface resulting from shot peening was beneficial in improving fretting fatigue life. In case of a rough surface, the real surface contact areas are broken down into small discrete areas and hence the volume of material affected by fretting
324
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
action in any one contact is much reduced leading to a reduced possibility of initiating a crack [8]. Improvements in fretting fatigue strength due to shot peening have been reported in different materials, e.g. steels, Al, Mg and Ti alloys. Mutoh et al. [9] have reported that shot peening improved the fretting fatigue strength of a steam turbine steel by a factor of 1.8 at room temperature and 773 K. They have observed that the shot peening was effective even for long-term practical use at elevated temperatures. Kondoh et al. [10] have studied the effect of a wide peening and cleaning (WPC) treatment on the fretting fatigue strength of AA6061-T6 material. In the WPC treatment, steel balls were shot on the test material for 30 s followed by the shot using ceramic balls for 30 s. The WPC treatment enhanced the value of fretting fatigue strength from 50 to 70 MPa. Lindley and Nix [11] have reported that in AA2014 aluminium alloy, the most effective single palliative treatment to combat fretting fatigue was glass bead peening, which introduced surface compressive residual stresses and inhibited the growth of small fretting cracks. Shot blasting is one of the most frequently used treatments for getting rough surface prior to plasma spraying to ensure a strong mechanical bond between coating and substrate. While shot peening is intentionally employed to improve fatigue strength/life of materials, shot blasting is used only for preparing the surface to be coated. In the present work, an attempt has been made to study the effect of shot blasting on the plain fatigue (without fretting) and fretting fatigue behaviour of Al–Mg–Si alloy AA6061.
2. Experimental details 2.1. Material, test specimen design and shot blasting The chemical composition of the material AA6061 used in the present investigation is given in Table 1 (taken from Ref. [12]). 8.3 mm thick specimens of 65 mm gauge length and 10 mm width were used for both plain fatigue and fretting fatigue tests (Fig. 1). The alloy was solution treated at 540 8C for 1 h and water quenched to room temperature and subsequently aged at 160 8C for 12 h. This heat treatment will be referred to as T6 condition. The room temperature mechanical properties of the material in the T6 condition are given in Table 2 (taken from Ref. [12]). Shot blasting, using spherical balls of aluminium oxide abrasives of size 120 mm, was done on some specimens
(in T6 condition) on the thickness side, where fretting was induced during fretting fatigue testing. The shot blasting details are: 100 mm stand off distance, 0.4 MPa average blasting pressure, 100 s blasting time, 100% coverage (coverage is the proportion of surface area blasted) and nozzle kept normal to the surface blasted. It may be noted that the normal blasting pressure used for getting a rough surface before plasma spraying is 0.245 MPa. However, in order to obtain higher residual compressive stresses, the blasting pressure was increased to 0.4 MPa in the present study. The blasting time of 100 s was chosen based on many shot blasting trials with different blasting duration. The magnitude of the residual compressive stresses induced due to shot blasting was determined as 110 MPa by the X-ray diffraction technique employing Cu Ka radiation. Higher values of residual stresses would have resulted if glass beads or steel shots were used. However, it was decided to study only the influence of shot blasting in the present investigation. The depth of the effect of shot blasting was not determined. 2.2. Fretting fatigue test assembly An experimental facility, with a ring-type load cell and bridge-type fretting pads, which can simulate the fretting fatigue conditions was designed and fabricated. Fretting pads were made of the same test material AA6061 alloy in the T6 condition. Fig. 2 shows schematic diagrams of the fretting fatigue test assembly and fretting pad assembly. Fig. 3 shows the photograph of the test assembly used in the present study. Proving ring (load cell), which was used to apply contact load in fretting fatigue tests, had strain gauges pasted on opposite sides of the ring (two on the inner surface and two on the outer surface). For calibration, the ring was loaded (through the axis of the load adjusting screw) in a Schenck servo-hydraulic testing machine and the elastic strains induced in the ring were recorded by means of a strain amplifier and data acquisition system (HBM-Spider8600 Hz and catman Express 4.0 software). Then a calibration curve of load vs. ring strain was plotted and the slope was used to determine the contact (normal) load. Bridge-type fretting pads, shown in Fig. 2(b), with a pad span of 30 mm were clamped to the specimen by means of the proving ring as shown in Fig. 3. Frictional force between the fretting pad and the specimen was measured by bonding strain gauges to the underside of the fretting pads, with the strain gauge grid centered between the pad feet. Split specimen technique [13] was employed to calibrate the fretting pads. 2.3. Test procedure
Table 1 Chemical composition of the material used (wt%) [12] Mg
Si
Fe
Mn
Cr
Cu
Zn
Ti
Al
0.63
0.55
0.21
0.30
0.25
0.31
0.25
0.15
Rest
Plain fatigue and fretting fatigue tests were conducted in laboratory air (approximately 60–70% relative humidity) at room temperature on the Schenck servo-hydraulic machine
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
325
Fig. 1. Specimen design (not to scale; all dimensions are in mm).
with a stress ratio of 0.1 at different cyclic stress amplitudes (see Table 3). For plain fatigue tests, the cycle frequency was 30 Hz and for fretting fatigue tests it was 10 Hz. A constant contact pressure of 100 MPa was used in the fretting fatigue tests. During fretting fatigue testing, the friction force values were recorded, and the contact load applied by the proving ring was monitored and maintained constant throughout the test with the help of the data acquisition system. The fretting scar regions in the tested specimens were observed at low magnification and photographs were taken. The roughness profiles across the fretting scar were obtained using a perthometer. A scanning electron microscope was employed to observe the fretting scars and the fracture surfaces of the specimens.
3. Results and discussion Fig. 4 shows the results corresponding to plain fatigue (PF) and fretting fatigue (FF) tests. As the tests were conducted with a stress ratio of 0.1, mean stress (smean) values were different for different tests (see Table 3). King and Lindley [14] have shown that the mean stress has greater effect on fretting fatigue strength than on plain fatigue strength. They have reported that raising the mean stress from 0 to 300 MPa reduced the plain fatigue strength of a low alloy rotor steel by 27% whereas in fretting fatigue, the reduction was 60%. In the present study, as the mean stress used was in the range 93–146 MPa, it was assumed that the effect of mean stress would not have been very significant. The fretting fatigue lives (Nf fretting fatigue) were shorter compared to plain fatigue lives (Nf plain fatigue). The fatigue life was reduced due to fretting to a greater extent at low cyclic stress levels. The life reduction factor (defined as Nf plain fatigue/Nf fretting fatigue) was 4.0 for the specimens in the T6 condition and 4.6 for the specimens in the shot blasted condition at a maximum cyclic stress (smax) of 169 MPa. Since the fretting plays a severe role in crack Table 2 Mechanical properties of the material used (in T6 condition) [12] Yield strength (MPa)
Ultimate tensile strength (MPa)
Elongation (%)
Hardness HV 5
281
348
16
115
Fig. 2. (a) Fretting fatigue test assembly. (b) Fretting pad assembly.
326
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
Fig. 3. Photograph of the test assembly used in the present study.
initiation period and as crack initiation occupies a major portion of the total life, the influence of fretting on the fatigue life may be expected to be of a greater extent at lower stresses. The beneficial effect of shot blasting in enhancing the life was clearly evident in both plain fatigue and fretting fatigue tests at low stress levels. The life improvement factor due to shot blasting was 2.8 in case of plain fatigue loading and it was 2.4 in case of fretting fatigue loading corresponding to a maximum cyclic stress of 169 MPa. As the cyclic stress increased, the beneficial effect decreased. At the highest maximum cyclic stress of 265 MPa, the effect of shot blasting was detrimental under both plain fatigue and fretting fatigue loading conditions. However, the reduction in life was not significant. The influence of shot blasting on plain fatigue life may be explained as follows. The shot blasting may be expected to give rise to similar effects on the surface as that of shot peening—localized plastic flow causing work hardening, general roughening and generation of compressive residual stresses. All these effects are expected to influence the fatigue life. At low cyclic stresses, the favourable effects of compressive residual stresses and work hardening may be expected to dominate over the detrimental effect of surface roughening, and hence resulting in life enhancement. As the cyclic stress increases, the residual stresses will relax quickly and the extent of life enhancement will decrease. At a particular stress level, the effect of residual stress and work hardening would balance the effect of surface roughening. Table 3 Stress values employed in the present study (stress ratio, 0.1) Cyclic stress amplitude, sa (MPa)
Mean stress, smean (MPa)
Maximum cyclic stress, smax (MPa)
76.05 86.85 97.65 108.45 119.25
92.95 106.15 119.35 132.55 145.75
169 193 217 241 265
Fig. 4. Effect of shot blasting on the plain fatigue (PF) and fretting fatigue (FF) lives.
Hence, shot blasting would not have any effect on the fatigue life at all. However, above this stress level, the adverse effect of roughening and/or surface cracking would dominate leading to reduction in life [15]. The effect of shot blasting on fretting fatigue life may be explained based on the values of coefficient of friction (Zfrictional force/normal load) in addition to the effect of residual stresses, work hardening and surface roughening (see Figs. 5 and 6). Fig. 5 shows the variation of coefficient of friction (m) at the fretting interface with number of fretting cycles at different cyclic stress levels in both T6 and shot blasted conditions. The friction force is considered to be created by either direct interlocking of surface asperities, or by trapping oxide debris in between the surface asperities. The coefficient of friction increased rapidly during the early stages of fretting fatigue life (below 3000 cycles) and then it remained almost constant for the remaining life period. Rayaprolu and Cook [13] have given an explanation for this type of behaviour. Under constant amplitude loading conditions, an initial ‘bedding-in’ phase, associated with gross slip between the contacting bodies is present. The bedding-in period consists of a gradual build up in the magnitude of frictional forces and the reduction in the degree of macroslip. During the macroslip conditions, larger contacting asperities are worn out leading to an increase in the area of asperity contact, which promotes microslip conditions. Towards the end of the bedding-in process, under microslip conditions, a point is reached where the entire frictional load causes only elastic deformation of contacting asperities. From this point a relatively constant value of frictional force occurs throughout the remainder of the fretting fatigue cycles. Fig. 6 shows the variation of stabilized m values with the maximum cyclic stress. A cross over in the plots for
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
327
stresses. Similar remarks have been made by Lindley and Nix [11] for the case of shot peening. Fig. 7(a)–(d) shows the appearance of fretting scars on tested specimens in the T6 condition and the shot blasted condition corresponding to two different cyclic stresses. It can be observed that the fretting damage width at high stress level is more than that at low stress level. It is very clear that the extent of fretting damage induced in the shot blasted condition is less than that in the T6 condition. This effect may be due to increased surface roughness, increased hardness due to work hardening of the surface and nonadhesion of the surface. The roughness profiles taken across fretting scar regions of the specimens are shown in Fig. 8. The middle portion of the profiles shown within dotted lines correspond to the fretting scar region and the end portions correspond to unfretted region. It is evident that the fretting damage increased the surface roughness of the specimens in the T6 condition. The roughness value (Ra) for a specimen tested at a higher stress was more than that for a sample tested at a lower stress. In case of the shot blasted specimens, the roughness value for a specimen tested at a higher stress was lower than that for a sample tested at a lower stress. This may be due to high rates of wear and large slip ranges at higher stresses. It may be noted that the shot blasted specimens had rough surfaces prior to fretting fatigue deformation. Due to the deformation, the surface ondulations were reduced and the roughness value decreased. As the cyclic stress increased, wear rate and relative slip increased. So the extent of reduction in the surface roughness value from the original value was more for samples tested at higher cyclic stresses. Fig. 9(a)–(d) shows scanning electron micrographs of the scars. A fatigue crack initiated from the fretting damage region may be seen. The roughness introduced due to shot blasting may be seen in Fig. 9(c) and (d). Fig. 5. Variation of coefficient of friction with number of cycles. (a) T6 condition; (b) shot blasted condition.
the material in T6 and shot blasted conditions may be seen similar to that was observed in Fig. 4. At a maximum cyclic stress of 169 MPa, the stabilized value of m in the shot blasted condition is 0.55 compared to 0.60 in the T6 condition. As the maximum cyclic stress increases, the difference between the values of m in both the conditions decreases. At a maximum cyclic stress of 265 MPa, the value of m is 0.82 in the T6 condition against 0.84 in the shot blasted condition. Though the difference between the values of coefficient of friction for both the conditions is small, it is assumed that in addition to the three effects discussed earlier, the difference in m values may also be responsible for the difference in the fretting fatigue lives of specimens in the T6 and the shot blasted conditions. Large relative slip ranges and high rates of wear at high cyclic stresses would have reduced the benefits of shot blasting by removing the near surface layer containing the compressive residual
Fig. 6. Variation of stabilized value of coefficient of friction with maximum cyclic stress (smax).
328
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
Fig. 7. Appearance of fretting scars in specimens tested at different maximum cyclic stress levels (smax). (a) T6 condition, 265 MPa; (b) T6 condition, 169 MPa; (c) shot blasted condition, 265 MPa; (d) shot blasted condition, 169 MPa.
Fig. 8. Roughness profiles taken across fretting scar in specimens tested at different maximum cyclic stress levels (smax).
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
329
Fig. 9. Fatigue crack (shown by an arrow) initiated in fretting damage region of specimens tested at different maximum cyclic stress levels (smax). (a) T6 condition, 265 MPa; (b) T6 condition, 169 MPa; (c) shot blasted condition, 265 MPa; (d) shot blasted condition, 169 MPa.
The crack initiation in the specimens both in the T6 condition as well as the shot blasted condition was at the trailing edge of the fretting scar region. However, in the shot blasted specimens, the initial direction of the crack followed
a much shallower angle to the surface. As had been pointed out by Leadbeater et al. [15], it is a direct consequence of the residual compressive stresses existing in the surface. This had the effect of moving the principal tensile stress in
Fig. 10. Fracture surface of a shot blasted specimen tested under plain fatigue with a maximum cyclic stress (smax) of 217 MPa. (a) Near crack initiation region; (b) striations in stage II propagation region; (c) final overload region.
330
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331
Fig. 11. Fracture surface of a specimen in the T6 condition tested under fretting fatigue with a maximum cyclic stress (smax) of 169 MPa. (a) Multi crack initiation sites; (b) transition from stage I to stage II propagation region.
the surface layers to a direction more nearly normal to the surface. A shallow angle of crack propagation resulted since a fatigue crack will always propagate in a direction perpendicular to the principal tensile stress. However, once the crack had moved out of the zone influenced by the fretting stresses, the direction of cracking was always at 908 to the principal tensile stress. Fig. 10(a)–(c) shows the appearance of fracture surfaces of a shot blasted specimen subjected to plain fatigue. No distinct features were observed near the crack initiation zone (Fig. 10(a)). Secondary cracking perpendicular to the main crack plane was observed. Striations were clearly observed in the stage II propagation region (Fig. 10(b)). Holes seen in the figure are the prior inclusion sites. Fig. 10(c) shows the final overload fracture region containing dimples. The fracture surfaces of samples in the T6 condition and the shot blasted condition tested under fretting fatigue loading showed similar features. Multi crack initiation sites were seen (see Fig. 11(a)). The transition from stage I to stage II of fatigue fracture was characterized by a change of orientation of the main fracture plane in each grain from one or two shear planes to many parallel plateaus separated by longitudinal ridges (see Fig. 11(b)).
4. Conclusions Based on the results obtained in the present study on the effect of shot blasting on the plain fatigue and fretting fatigue behaviour of AA6061 material, the following conclusions were drawn. (a) Fatigue life reduction due to fretting was high at low cyclic stress levels. At a maximum cyclic stress (smax) of 169 MPa, the life reduction factor was 4.0 in the T6 condition and 4.6 in the shot blasted condition. Towards higher cyclic stress levels, the extent of reduction in life decreased. (b) Shot blasting significantly increased the plain fatigue life by a factor 2.8 and fretting fatigue life by a factor
2.4 at a maximum cyclic stress of 169 MPa. However, towards higher stress levels, the extent of increase in fatigue lives decreased and at a maximum cyclic stress of 265 MPa both plain fatigue and fretting fatigue lives slightly reduced in the shot blasted condition. (c) The coefficient of friction (m) at the fretting interface increased rapidly during the early stages of fretting fatigue life (!3000 cycles) and then it remained almost constant. (d) At a maximum cyclic stress of 169 MPa, the stabilized value of m was less in the shot blasted condition than that in the T6 condition (0.55 against 0.60). As stress levels increased, the difference between the values of m in both the conditions decreased. At a maximum cyclic stress of 265 MPa, the value of m was less in the T6 condition than that in the shot blasted condition (0.82 against 0.84). (e) Shot blasting can also be used as one of the cheapest and best methods to improve the fretting fatigue life at low service loads.
Acknowledgements This work was supported by the Department of Science and Technology, Government of India under SERC Fast Track Scheme 2001–2002 (project number SR/FTP/ET183/2001).
References [1] Waterhouse RB. Fretting fatigue. Int Mater Rev 1992;(37):77–97. [2] Beard J. Palliatives for fretting fatigue. In: Waterhouse RB, Lindley TC, editors. Fretting fatigue, ESIS 18. London: Mechanical Engineering Publications; 1994. p. 419–36. [3] Gordelier SC, Chivers TC. A literature review of palliatives for fretting fatigue. Wear 1979;56:177–90. [4] Chivers TC, Gordelier SC. Fretting fatigue palliatives: some comparative experiments. Wear 1984;96:153–75. [5] Gabel MBK, Bathke JJ. Coatings for fretting prevention. Wear 1978; 46:81–96.
N.K.R. Naidu, S.G.S. Raman / International Journal of Fatigue 27 (2005) 323–331 [6] Fu Y, Wei J, Batchelor AW. Some considerations on the mitigation of fretting damage by the application of surface modification technologies. J Mater Proc Technol 2000;99:231–45. [7] Li CX, Sun Y, Bell T. Shot peening of plasma nitrided steel for fretting fatigue strength enhancement. Mater Sci Technol 2000;(16): 1067–72. [8] Waterhouse RB. Effect of material and surface condition on fretting fatigue. In: Waterhouse RB, Lindley TC, editors. Fretting fatigue, ESIS 18. London: Mechanical Engineering Publications; 1994. p. 339–49. [9] Mutoh Y, Satoh T, Tsunoda E. Improving fretting fatigue strength at elevated temperatures by shot peening in steam turbine steel. In: Attia MH, Waterhouse RB, editors. Standardization of fretting fatigue test methods and equipment, ASTM STP 1159. Philadelphia: American Society for Testing and Materials; 1992. p. 199–209. [10] Kondoh K, Mutoh Y, Mabuchi M, Kobayashi T. Fretting fatigue behavior of surface-treated aluminium alloys. In: Wu X-R, Wang ZG, editors. Fatigue’99: Proceedings of the international congress, 8–12 June, vol. 2. Beijing: Higher Education Press; 1999. p. 1347–51.
331
[11] Lindley TC, Nix KJ. Fretting fatigue in the power generation industry: experiments, analysis and integrity assessment. Fatigue’99: Proceedings of the international congress, 8–12 June 1999. p. 153–169. [12] Karunanidhi S. Fatigue and short crack growth behaviour of an age hardenable Al–Mg–Si alloy. MS Thesis. IIT Madras; 1999. [13] Rayaprolu DB, Cook R. A critical review of fretting fatigue investigations at the royal aerospace establishment. In: Attia MH, Waterhouse RB, editors. Standardization of fretting fatigue tests and equipment, ASTM STP 1159. Philadelphia: American Society for Testing and Materials; 1992. p. 129–52. [14] King RN, Lindley TC. Fretting fatigue in a 3.5%NiCrMoV rotor steel. Advances in fracture research, vol. 2. Oxford: Pergamon Press; 1981. p. 631–40. [15] Leadbeater G, Noble B, Waterhouse RB. The fatigue of aluminium alloys produced by fretting on a shot peened surface. In: Valluri SR, editor. Advances in fracture research. Proceedings of the sixth international conference on fracture, vol. 3. Oxford: Pergamon Press; 1984. p. 2125–32.