Influence of different combined severe shot peening and laser surface melting treatments on the fatigue performance of 20CrMnTi steel gear

Materials Science & Engineering A 658 (2016) 77–85

Contents lists available at ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Influence of different combined severe shot peening and laser surface melting treatments on the fatigue performance of 20CrMnTi steel gear You Lv n, Liqun Lei, Lina Sun School of Mechanical Engineering, Jilin Agricultural Science and Technology University, Jilin 132101, PR China

art ic l e i nf o

a b s t r a c t

Article history: Received 7 December 2015 Received in revised form 16 January 2016 Accepted 18 January 2016 Available online 19 January 2016

In this paper, the effect of severe shot peening combined with laser melting (LSMSSP for short) on the fatigue resistance of 20CrMiTi steel gears is investigated in comparison with the effect of traditional shot peening on the fatigue resistance of the laser surface melted (LSMTSP for short) 20CrMiTi steel gear. The surface characteristics of the gear have been analyzed by a scanning electron microscope (SEM) and an X-ray diffractometer (XRD). The Forschungsstelle für Zahnräder und Getriebebau (FZG) back-to-back spur gear test rig was used for fatigue experiments. Experimental results showed that the residual stresses, full width at half maximum (FWHM), microhardness and retained austenite of the LSMSSP gears and LSMTSP gear were entirely different. Although the LSMSSP gears had higher surface roughness than the LSMTSP gear, the LSMSSP gears still had better fatigue resistance than the LSMTSP gear and laser surface melted gear. The nanocrystallized surface layer on the gear tooth flank created by severe shot peening might be a very important factor for improving the fatigue property of the LSMSSP gears. & 2016 Elsevier B.V. All rights reserved.

Keywords: Gear Severe shot peening Laser surface melting Residual stresses

1. Introduction A gear drive is the most important form of mechanical transmission [1]. It has many advantages, such as high transmission efficiency, compact structure, stable transmission ratio, high reliability, long durability, and so on [2]. Gear drives are widely used in many fields. Gears have become the most important part in the machinery industry. With the continuous improvements of the industrial level, increasing demands have also been raised for the strength of gears. In the transmission process, gears are subjected to a variety of forces, including bending stress, contact stress and impact force, so gears are peculiarly prone to multiple forms of failure, encompassing tooth flank pitting, spalling, wear, scoring and tooth bending fatigue [3,4]. Pitting is one of the most typical characteristics of gear contact fatigue. The main factors that affect the pitting damage are shown in Fig. 1. Pitting occurrence as less as possible will play an important role in improving the fatigue strength of gears. For a long time, traditional shot peening (TSP for short) has been one of the most important means to improve the fatigue strength of gears in the industrial production because of its high flexibility, low production costs and little environmental pollution. In recent years, there have been many new types of peening, such as laser shock peening [5,6], ultrasonic shot peening [7,8], fine n

Corresponding author. E-mail address: [email protected] (Y. Lv).

http://dx.doi.org/10.1016/j.msea.2016.01.050 0921-5093/& 2016 Elsevier B.V. All rights reserved.

particle peening [9,10], water cavitation peening [11] and re-peening [12]. Other than those peening treatment, there is another peening treatment called “severe shot peening” (SSP for short). Comparing with TSP, SSP uses higher Almen intensities in the shot peening process. After SSP, a nano-crystalline layer has been created on the surface of the metallic material, and thus the SSPed sample has superior mechanical properties. So far, SSP has been using air blast equipment with higher Almen intensities in order to get higher kinetic energy as shown in Fig. 2, accumulating a huge mass of plastic deformation, and decreasing the grain size of the material to the level of nanometer. Severe shot peening can also be combined with other surface treatment. S.M. Hassani-Gangaraj et al. [13,14] investigated the effect of combining nitriding and SSP on the fatigue behavior and micro-structure of ESKY-LOS6959 low-alloy steel. The experimental results show that the combination of SSP and nitriding is more effective in improving the fatigue properties of ESKYLOS6959 low-alloy steel, compared with the single treatment. Authors [15] investigated the effect of shot peening on the fatigue resistance of laser surface melted (LSM for short) 20CrMnTi steel gear. Previous experimental results showed that the fatigue properties of the LSMþ SP gears were much better than that of the laser surface melted gear. This paper studied the effect of SSP combined with laser surface melting on the fatigue resistance of 20CrMnTi steel gear, in comparison with the effect of TSP on the fatigue resistance of the laser surface melted 20CrMnTi steel gear. These research results will have important guiding significance to the future industrial production. The microhardness, residual

78

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

Table 2 Laser parameters for laser surface-melted 20CrMnTi steel gear.

Fig. 1. The major effect factors of pitting damage.

Samples no. Energy density (MJ/m2) Pulse duration (ms)

Scanning speed (mm/s)

Frequency (Hz)

LSM-1 LSM-2 LSM-3

2

8

19 25 31

20

LSM tooth consisted of the melted zone (MZ), the heat-affected zone (HAZ) and the substrate. The microstructure of the MZ of the LSM teeth processed under different conditions was shown in Fig. 3(b)–(d). In the MZ, the maximum temperature exceeded Ac3 owing to the direct radiation of laser beam, and the components are uniformly distributed. Carbon and alloying elements fully diffused at high temperature to form a homogeneous austenite structure. In the subsequent process of rapid cooling, the carbonpoor needle-like martensite and a small amount of retained austenite were formed along the direction of maximum temperature gradient as shown in Fig. 3(b)–(d). With the increasing of laser energy density, the needle-like martensitic structure of the LSM gear was much more refined, especially compared with the as-cast one. The XRD pattern of the LSM-3 gear was shown in Fig. 8. All of the LSM gears were composed of phases: ferrite (α), martensite, carbide and retained austenite (γ). 2.2. Shot peening

Fig. 2. A diagram of severe shot peening of gear.

stress, surface roughness, and retained austenite of LSMSSP gears would be studied under various initial surface states.

2. Experiment details 2.1. Gear material and LSM treatment The chemical composition of 20CrMnTi steel is listed in Table 1. The laser surface melted treatment was carried out using a 4.5 kW Nd:YAG pulsed laser with a wavelength of 1064 nm, laser energy densities of 19 MJ/m2, 25 MJ/m2 and 31 MJ/m2, a laser beam diameter of 1 mm, a frequency of 8 Hz, a laser scanning speed of 2 mm s  1, a duration time of 20 ms and argon flowing at 20 l min  1 used as shielding gas. The laser-melted processing was executed by a computer numerical control program. The overlapping ratio of the successive melted tracks was 50%. The designations of the LSM gears were presented in Table 2. The as-cast 20CrMnTi steel was composed of a coarse pearliticferritic matrix, as shown in Fig. 3(a). The LSM layer in the gear tooth was free of cracks and pores. From the top of the surface, the Table 1 Chemical composition of 20CrMnTi steel. Element

Composition (Wt%)

C Si Cr Mn Ti S P Cu Fe

0.17–0.23 1.00–1.30 0.17–0.37 0.80–1.10 0.04–0.10% r0.035 r0.035 r0.025 Balance

The shot peening treatment was completed by a pneumatic numerical control air blast shot peening machine. The angle of shot peening was set through a four-axis robotic arm. The direction of shot peening was manipulated by a four-axis automatic control system to ensure that the best result of shot peening would be obtained for gears. The LSM gears were treated by shot peening with different Almen intensities. Table 3 shows the shot peening parameters for gears in detail. The shot diameter was 0.58 mm. The peening intensities for the LSMSSP gears increased from A34-36 to C6-8 Almen intensities. With an increase in the Almen intensity, the plastic deformation rate of the LSMSSP gears also increased. For comparison, the LSM gear was treated by traditional shot peening parameters, called “LSMTSP gear”. When the LSM-1 gear was treated by SSP-1 treatment, the gear was called “LSM-1-SSP-1 gear”. 2.3. Surface roughness measurement The roughness of the gears was measured by Veeco Wyko NT1100 optical profiler 5 times each, and the surface roughness parameters could be obtained. 2.4. Residual stress measurement The residual stresses were measured by XRD on the (211) interference line of the LSM gear tooth flank using Cr Kα radiation. The measured interference peaks were evaluated by the sin2ψ method and the angle of ψ varied every 10° in the range of  70° to þ70°. The depth profiles of the residual stresses were determined by iterative electrolytic removal of thin surface layers and subsequent X-ray measurements. 2.5. Microhardness measurement Microhardness along the depth of the cross-section of gear teeth was measured by a Vickers hardness tester by applying 500 g load with the dwell time of 15 s. Microhardness was the average

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

79

Fig. 3. SEM image of (a) the cross-sectional appearance of the LSM tooth; the microstructure of the MZ of (b) LSM-1 gear, (c) LSM-2 gear and (d) LSM-3 gear. Table 3 Shot peening parameters for the LSM 20CrMnTi steel gear.

Table 4 Geometrical parameters of test gears.

Samples no. Shot type and diameter (mm)

Almen intensity (0.0001 in.)

Coverage (%) Shot speed (m/s)

LSMSSP-1 LSMSSP-2 LSMSSP-3 LSMTSP

A34-36 C4-6 C6-8 A16-18

1000

90

100

30

S230(steel, ؼ 0.58)

microhardness value of 5 points at the same depth of teeth.

Parameters

Symbol

Units

Module Number of teeth Pressure angle Helix angle Face width Normal contact ratio Complementary contact ratio Overall contact ratio Center distance Tooth flank roughness

m Z1:Z2 α β b εα εβ εγ

(mm) (/) (deg) (deg) (mm) (/) (/) (/) (mm)

R

a

4.5 16:24 20 0 14 1.43 0 1.43 91.5 0.3

2.6. Fatigue tests of gears The fatigue tests were run by using the FZG machine, as shown in Fig. 4 [16]. The geometric parameters of test gears are presented in Table 4. The properties of oil samples are illustrated in the reference [15]. The gears in the tests were placed in an oil bath for lubrication at fixed 90 °C. The rotation speed of pinion was 2000 rpm. The stress levels of the experiment were 1841 MPa, 1691 MPa, 1539 MPa, 1232 MPa, 1080 MPa and 929 MPa, respectively. Each type of gear had 5 experimental points. The test machine was equipped with a vibration warning device and an automatic stopping design in order to end the experiment.

Considering the dispersion of the contact fatigue test results of gears, the Weibull distribution probability function has been used to calculate the fatigue life of gears. When the fatigue life was distributed according to Weibull, the failure distribution function was shown in Eq. (1). t m

F (t ) = 1 − e−( η )

(1) 1

η was the characteristic life. η = t m t0 was the size parameter. m was the shape parameter. It is difficult to estimate m and η parameters directly, so Eq. (1) could be formulated into Eq. (2). ⎞ ⎛ 1 ln ln ⎜ ⎟ = m ln t − ln t0 ⎝ 1 − F (t ) ⎠

(2)

Therefore, the independent variable was transformed into ln t in view of fatigue life for calculation.

3. Results and discussions 3.1. Residual stresses

Fig. 4. Schematic diagram of the FZG machine [16].

Okan Unal et al. [17] found that the microstrain evaluations by using X-ray diffraction helped to determine the influence of the

80

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

Fig. 5. The residual stress distributions of gears.

slightly. With an increase in Almen intensity, compressive residual stresses on the surface of LSMSSP gears were sharply increasing. Compressive residual stress on the surface of LSM-3-SSP gear was slightly less than that of LSM-1-SSP gear because of counteraction of initial residual tensile stress and higher surface roughness of LSM-3-SSP gear. However, residual tensile stresses on the surface of LSM-1-SSP-3 (-412 MPa) and LSM-3-SSP-3 (  400 MPa) were very close, which indicated that compressive residual stresses on the surface of LSMSSP gears had reached saturation. When Almen intensity reached a certain threshold value, compressive residual stress on the surface of LSMSSP gear could not continue to increase. 3.2. Full width at half maximum (FWHM)

Fig. 6. The FWHM distributions of gears.

deformation process such as residual stresses and FWHM, after the severe plastic deformation methods. Fig. 5a shows that the residual stress distributions of the LSM-3-SSP gears and the LSM-3TSP gear are very different in a direction following the depth. The maximum compressive residual stresses of the LSM-3-SSP gears were  535 MPa (at a depth of 100 μm),  560 MPa (at a depth of 160 μm), and  592 MPa (at a depth of 200 μm), respectively. It can be seen that the peak and maximum depth of compressive residual stress of the LSM-3-SSP gears gradually increased with an increase in the Almen intensity. The most obvious reason for this phenomenon was that the increase in the Almen intensity produced more kinetic energy, and induced a greater degree of plastic deformation. The changes of residual compressive stress values and depths of the LSM-3-SSP gears were resulted from a large amount of plastic deformation. The average compressive residual stress of the LSM-3-SSP gears (  469 MPa) was 33.6% more than that of the LSM-3-TSP gear (  351 MPa). The average depth of the compressive residual stress of the LSM-3-SSP gears (1030 μm) was 186% more than that of the LSM-3-TSP gear (360 μm). The initial tensile stress state of the LSM-3 gear became the compressive stress state, after the TSP and SSP treatment. Our previous study [15] showed that the initial tensile stress state of the LSM gear would offset part of the compressive stresses which was generated by shot peening. Thus, the general level of residual stress of the LSMþ SP gears was reduced. In this research, the compressive residual stresses of the LSM-3-SSP gears must have been reduced by the initial tensile stress state of the LSM gears. Fig. 5b shows the surface residual stress of the LSM gears and the LSMSSP gears. With an increase in laser energy density, residual tensile stresses on the surface of LSM gears increased

With the help of an X-350A X-ray diffraction tester and electrochemical stripping technologies, the diffraction peak width at half the maximum intensity (FWHM) distributions of the LSM-3SSP and LSM-3-TSP gears in the depth-direction were obtained and presented in Fig. 6. The maximum values of the FWHM were located on the surfaces of the LSM-3-SSP and LSM-3-TSP gears, and gradually decreased with an increase in depth, finally reaching a constant value. The maximum FWHM values of the LSM-3-SSP gears were 6.85°, 6.93°, and 6.99°, respectively. The average FWHM values of the LSM-3-SSP gears (6.47°) were 4.4% more than that of the LSM-3-TSP gear (6.2°). The maximum FWHM values of the LSM-3-SSP gears increased with an increment of the Almen intensity owing to the increase of kinetic energy generated by SSP. The FWHM depth can be regarded as the depth of the work hardening layer. As thus, the depths of the work hardening layer of the LSM-3-SSP gears were 400 μm, 400 μm, and 450 μm, respectively. The average depth of the work hardening layer of the LSM3-SSP gears (416.7 μm) was 108.3% more than that of the LSM-3TSP gear (200 μm). Bagherifard et al. [18] found that the SSP treatment to be more effective in work-hardening with respect to CSP specimen and the effect of SSP treatment could be observed up to the depth of 1 mm. With growing Almen intensities, the accumulated plastic strain of the LSM-3-SSP gears increased, the results led to increasing depths of the work hardening layer of the LSM-3-SSP gears. The plastic strain accumulation increased the FWHM of the LSM-3-SSP gears, which was related to grain distortion, dislocation density, grain size, and the so called “type II micro residual stresses” [19] of the LSMSSP gear tooth flanks. This suggests that the LSM-3-SSP gear surface layer has more dramatic microdistortion of the crystalline lattice, higher dislocation density and finer grain, comparing with the LSM-3-TSP gear, and thus a conclusion is drawn that SSP is stronger than TSP in respect to the work hardening effect.

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

81

Fig. 7. The microhardness distribution curves of gears.

3.3. Microhardness Fig. 7a shows the microhardness distributions of the LSM-3-SSP gears and the LSM-3-TSP gear along the depth-direction. The maximum microhardness of the LSM-3-SSP gears and the LSM-3TSP gear appeared on the tooth surface, gradually decreasing with increasing depth and finally reaching a constant value, which was caused by the reduction of plastic deformation generated by shot peening. The decreasing trend of the LSM-3-SSP gear microhardness was relatively gentle, because the presence of a high degree of plastic deformation could prevent the sharp changes in the microhardness. The maximum microhardness of the LSM-3-SSP gears was 731.2 HV0.5, 745.5 HV0.5, and 767.5 HV0.5, respectively. The average microhardness of the LSM-3-SSP gears (665.7 HV0.5) was 4.8% more than that of the LSM-3-TSP gear (635.4 HV0.5). With an increment in the Almen intensity and higher degrees of plastic deformation, the maximum microhardness of the LSM-3-SSP gears increased due to grain refinement and higher dislocation density generated by SSP. The same experimental results were found by Okan Unal et al. [20]. Hassani-Gangaraj et al. [21] got the following surface nanocrystallized mechanism of SSP. Under the repeated impacts of the high-energy shot, the deeper position of the material was affected by the action of the external force. When the external force reached the critical stress required for the dislocation movement, the slip of dislocations was generated, and a netlike structure was formed. On the same slip plane, a lot of dislocations were hindered by grain boundaries, stacked at the grain boundaries and formed groups of dislocation pile-ups. With an increase in the number of groups of dislocation pile-ups, there were more and more forces acting on a part of dislocations, and stress concentration was generated. When stress concentrations reached a certain degree, grain boundaries were destroyed, deformation twins occurred and twins of single-series were produced. When the stress continued to increase, the twins of singleseries would go over into the twins of multi-series, and then coarse grains were separated into small pieces. With a drop in the distance from sample surface, the stress and microstrain became bigger and bigger, grain size continually decreased, and twin boundaries repeatedly increased. Due to the high energy of grain boundaries and atomic activities, it’s easier for the new phase to nucleate preferentially at a grain boundary. At the same time, twins of multi-series accumulated very high deformation storage energy to overcome the resistance of martensitic transformation. Therefore, during grain refinement, strain induced martensite was formed. The stress and microstrain continued to increase near the sample surface, and finally the equiaxed and randomly oriented martensitic nanocrystalline was generated under the combined action of the large strain, high strain rate and multi-direction loading. Research by Okan Unal [22] found that a nanocrystalline layer was produced on the surface of AISI 1017 low carbon steel

through SSP treatment. Nanocrystal was produced under the influence of plastic deformation. The plastic deformation rate increased with an increase in the Almen intensity, which led to more fine grains. However, grain size had a limit, and the extent of grain refinement would not change, when the plastic deformation of the SSPed sample reached the saturation point. Furthermore, the questions of whether nanocrystalline existed on the surface layer of the LSMSSP gear tooth flank, and whether nanocrystalline size reached the limit would be investigated in the future. The influenced depths of the LSM-3-SSP gears were 400 μm, 425 μm, and 450 μm, respectively. Katarína Miková et al. [23] found that the effect of both TSP and SSP in terms of HV was not notable till a depth of about 0.5 mm. The influenced depths of the LSM-3-SSP gears increased with an increment in the Almen intensity. The average influenced depth of the LSM-3-SSP gears (425 μm) was 70% more than that of the LSM-3-TSP gear (250 μm). Figs.6a and 7a show that the influenced depth of microhardness of the LSM-3-SSP gear is essentially in agreement with the maximum depth of the FWHM (i.e. the depth of work hardening layer). Fig. 7b shows surface hardness of LSM gears and LSMSSP gears. With an increase in laser energy density, comparing with surface hardness of as-received gear (270 HV0.5), that of LSM gears increased by 103.8%, 116.9% and 124.4%, respectively, because of grain refinement, solid solution strengthening and dislocation strengthening. With an increase in Almen intensity, surface hardness of LSMSSP gear gradually increased. However, surface hardness of LSM-1-SSP-3 gear (762.7 HV0.5) and LSM-3-SSP-3 gear (767.5 HV0.5) was almost the same, suggesting that plastic deformation of LSMSSP gears was limited. When Almen intensity reached the critical point, surface hardness of LSMSSP gear had a peak value. 3.4. Retained austenite Fig. 8 shows the XRD spectrums of the LSM-3-SSP and LSM-3TSP gear tooth flanks. It can be seen that the γ-Chromium Iron Niokel (111) peak existed on the LSM-3 gear tooth flank. After TSP treatment, the γ-Chromium Iron Nickel (111) peak remained on the LSM-3-TSP gear tooth flanks. With an increase in the Almen intensity, the γ-Chromium Iron Nickel (111) peaks of the LSM-3SSP-1 and LSM-3-SSP-2 gears gradually faded away, and the γChromium Iron Nickel characteristic peak completely disappeared on the LSM-3-SSP-3 gear tooth flank due to the martensitic transformation of retained austenite, which also caused an increase in microhardness of the LSM-3-SSP gears. This demonstrates that TSP can reduce the content of retained austenite in the LSM gear surface, but cannot eliminate it thoroughly like SSP does. Other than that, a part of the characteristic peaks of the LSM-3-SSP

82

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

Fig. 8. The XRD spectrums of gears.

gears have broadened with an increment in the Almen intensity, signifying that the microstructure of the LSM-3-SSP gear surface layer is refined. 3.5. Surface roughness Table 5 shows the surface roughness of the LSM gears, the LSMSSP gears and the LSPTSP gears. Surface roughness of LSM gear increased with an increase in laser energy density. When the laser energy density reached a certain value, gasification of the local material occurred in the molten zone, and the liquid metal radially flowed in the molten pool from the center to the edge. After the melting and resolidification in the LSM gear surface layer, the uneven surface morphology was formed in the LSM teeth flank. With an increase in laser energy density, the radial flow of the liquid material in the molten pool of the LSM gear became more intense during laser irradiation, and the gasification and melting extent of the material increased. As a result, surface roughness of LSM gear was enlarged. The surface parameters Ra, Rq, Rz and Rt represent the arithmetic mean, maximum height of the profile, micro-roughness of ten-point height, and root mean square based on the definitions of ISO 4287 [24]. All of the surface roughness of the LSMSSP gears was higher than that of the LSMTSP gear. Sara Bagherifard et al. [25] found that the arithmetic-mean value (Ra) increased from 0.57 μm for NP specimen to 3.53 μm for CSPed specimen and eventually noticeably increased to 8.58 μm for SSPed specimen. With an increase in the Table 5 Surface roughness parameters of test gears. Treatment

LT (mm)

LM(mm)

Ra(μm)

Rq (μm)

Rz(μm)

Rt (μm)

LSM-1 LSM-1-TSP LSM-1-SSP-1 LSM-1-SSP-2 LSM-1-SSP-3 LSM-2 LSM-2-TSP LSM-2-SSP-1 LSM-2-SSP-2 LSM-2-SSP-3 LSM-3 LSM-3-TSP LSM-3-SSP-1 LSM-3-SSP-2 LSM-3-SSP-3

0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80

4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00

2.17 1.46 3.19 3.45 4.17 2.34 1.75 2.93 3.38 3.78 2.61 2.09 2.51 2.81 2.86

2.74 1.79 3.95 4.21 5.35 2.89 2.15 3.49 5.17 4.85 3.21 2.59 3.04 3.48 3.72

11.73 7.22 14.51 16.15 21.02 12.46 9.03 13.71 21.36 19.79 13.65 10.78 11.31 13.62 14.59

14.92 9.11 19.59 25.48 29.03 22.58 11.82 18.93 29.85 29.94 17.49 13.49 15.35 19.24 21.91

Almen intensity, impactive energy of SSP increased, and the surface roughness of the LSMSSP gears were improved. Our previous studies [15] had found that the surface roughness of the LSMþ SP gears was gradually reduced with an increment in air pressure. However, when the air pressure exceeded a certain critical value, the surface roughness of the LSM þSP gears began to increase. In this study, the effect of SSP on surface roughness of LSMSSP gears was complex. Comparing with the Ra of the LSM-1 gear (2.17 μm), the arithmetic-mean values (Ra) of LSM-1-SSP gears were 3.19 μm, 3.45 μm and 4.17 μm, respectively. With an increase in Almen intensity, surface roughness of LSM-1-SSP gear increased by 47%, 58.9% and 92.1%, respectively. The reason for this phenomenon was that the grain size of LSM-1 gear was the largest, and its resistance to surface plastic deformation was the weakest. Consequently, the surface plastic deformation degree of LSM-1SSP gear increased, and the magnitude of surface roughness increase had also enhanced. The arithmetic-mean values (Ra) of LSM-3-SSP gears were 2.51 μm, 2.81 μm and 2.86 μm, respectively. Comparing with the Ra of the LSM-3 gear (2.61 μm), the surface roughness of LSM-3-SSP gears firstly decreased by 3.8%, and then increased by 7.7% and 9.6% with increasing Almen intensity. When Almen intensity was low, the impact of shot peening on LSM-3 gear was equivalent to surface polishing, so surface roughness of LSM-3-SSP-1 gear was reduced at first. With continuously increasing Almen intensity, the surface plastic deformation degree of LSM-3-SSP gears had raised, and surface roughness increased gradually since the grain size of LSM-3 gear was the smallest, and its resistance to surface plastic deformation was the most powerful. The surface roughness of LSM-3-SSP gears was less than that of LSM-1-SSP gears. It can be seen that surface roughness of LSMSSP gear was resulted from the comprehensive effects of multiple factors, including initial surface roughness, grain size, surface hardness and residual stress, etc. Bagherifard et al. [18] and Katarína Miková et al. [23] also found that the surface roughness increased with an increment in the Almen intensity. Their experimental results were basically in consonance with the conclusion of this research. Okan Unal et al. [22] found that the increase of the coverage could effectively reduce the surface roughness of the SSPed sample for a constant Almen intensity. Whether or not the coverage has the same effect on the surface roughness of the LSMSSP gears would be investigated in the future. 3.6. Fatigue properties 1

The linear relationship between ln ln ( 1 − F (t ) ) and ln t was presented in Fig. 9. m was the slope of a straight line. According to Eq. (2), the average life and the characteristic life of the Weibull distribution can be obtained. Experimental data and calculated results were presented in Table 6 and plotted in Fig. 9. It is indicated by the Weibull distribution curves that the farther distance between the curve and the vertical axis, the longer the fatigue life of gears. There were multiple relationships between these values and the fatigue lives of gears in Table 6. Firstly, “ the average life (h) “ was the average of fatigue lives of gears. The bigger the fatigue life was, the higher the average life (h) was. Secondly, “ slope m” was the slope of the Weibull distribution curve of fatigue life of gear. The greater the slope was, the more concentrated the fatigue life of gears was. In this way, a higher “m” was also considered a better repeatability of gear tests. Thirdly, “characteristic life (h) “ was increased with the increase of fatigue life of gear. The average fatigue lives of LSM-3-TSP, LSM-3-SSP-2 and LSM-3-SSP-3 gears exceeded the average fatigue life of LSM-3 gear by 6.23 times, 19.67 times and 32.43 times, respectively. It is suggested that the fatigue resistance of the LSMSSP gear is

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

Fig. 9. The Weibull curves of different test gears.

Fig.10. The σ − ln t curve of the LSM-3-SSP-3 gear.

Table 6 The fatigue lives of different test gears. (Stress level σ = 1232MPa ). Sample number

1 2 3 4 5 Weibull statistics

The failure time of different gears (h)

slope m average life (h) characteristic life (h)

LSM-3 LSM-3TSP

LSM-3SSP-2

LSM-3SSP-3

1.2 1.5 1.5 2 2 5.36 1.645 1.789

15 24 42 42 44 3.18 34 37.9

40 45 55 65 70 3.52 55 60.7

4 6 9 16 20 1.61 11.9 13.2

Table 7 The fatigue lives of the LSM-3-SSP-3 gear under different stress levels. Stress level (MPa) Sample number

1841 1691 1539 1232 1080

1 2 3 4 5 Weibull average life (h)

4 6 7.5 8.5 9 7.15

8 12.5 18 23.5 27 18.6

15 24 42 42 44 22.9

42 45 48 48 50 47

929

36 120 60 120 84 120 120 (unfailure) 120 120 (unfailure) 120 53 120

83

(unfailure) (unfailure) (unfailure) (unfailure) (unfailure)

improved with an increment in the Almen intensity. The dispersion degree of the fatigue life of the LSM-3-TSP gear was the largest. The fatigue performance of the LSM-3-SSP-3 was the best. Next, under six kinds of load conditions, the experiment of the LSM-3-SSP-3 gear was carried out further. The experimental data of the LSM-3-SSP-3 gear were statistically analyzed on the basis of Weibull, and the results were shown in Table 7. Fig. 10 shows the σ − ln t curve for the contact fatigue characteristic of the LSM-3SSP-3 gear. It can be seen that σ and ln t are linear. The relationship was σ = A − B ln t . In accordance with the principle of least squares, it is obtained that A = 224.5 and B = 16.6. The fatigue limit of the LSM-3-SSP-3 gear was 929 MPa. When the gear load was less than the stress level of 929 MPa, the LSM-3-SSP-3 gear would not appear fatigue failure. Fig. 11 showed the surface morphologies of gear tooth flanks. The largest and deepest pits appeared on the LSM-3 tooth flank. The number, depth and size of pits on the LSM-3-TSP tooth flank

were much higher than that on the LSM-3-SSP tooth flank. The number of pits on the LSM-3-SSP-2 tooth flank was similar to that on the LSM-3-SSP-3 gear, but the depth and size of pits on the LSM-3-SSP-3 tooth flank was smaller that on the LSM-3-SSP-2 tooth flank. Bagherifard et al. [26] found that the considerable fatigue strength improvement of the SSP specimens increased by 246%, compared with the surface NC specimens. There were four factors affecting the fatigue resistance of the LSMSSP gears, which are grain refinement, residual stress field, retained austenite and surface roughness. 3.6.1. Grain refinement An ultra-fine/nanostructured surface layer had a beneficial effect on the fatigue properties of the LSMSSP gears. The smaller the grain size was, the more the grain boundaries were. A grain boundary is a disordered region of atomic arrangement in the crystal. When the fatigue crack attempted to propagate through the grain boundary, the grain boundary resistance to the propagation of the fatigue crack was very high because of different grain orientations on both sides of the grain boundaries. When the fatigue crack has already passed the grain boundaries and continued to propagate, still because of different grain orientations on both sides of the grain boundaries, the fatigue crack had to change the initial direction of propagation constantly as a result of extension paths of fatigue crack propagation and an increase in the required energy of fatigue crack propagation. In addition, according to the Hall–Petch relationship, grain refinement can improve the strength and toughness of materials, which helps to improve the fatigue crack propagation resistance of the LSMSSP gears. 3.6.2. Residual stress field As can be seen in Fig. 5, the values and depths of compressive residual stresses of the LSMSSP gears were higher than those of the LSMTSP gear. It is generally known that fatigue cracks are usually initiated at the gear surface without the presence of compressive residual stresses. The existence of the residual compressive stress field of the LSMSSP gears could push the initiation position of the fatigue cracks from the surface to the subsurface of the gears. With an increase in the alternating stress, the fatigue crack initiation position will gradually shift to the LSMSSP gear surface. In this process, the fatigue performance of the LSMSSP gears was improved. The influences of SSP on the fatigue performance of the LSMSSP gears were analyzed as follows, from the perspectives of the fatigue crack initiation and propagation in the residual stress field of the LSMSSP gears.

84

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

Fig.11. Surface morphologies of (a) the LSM-3 gear, (b) LSM-3-TSP gear, (c) LSM-3-SSP-2 gear and (d) LSM-3-SSP-3 gear.

3.6.2.1. Effect of residual stress fields on fatigue crack initiation. Fatigue crack initiation in a residual stress field includes two cases, which are surface crack initiation and subsurface crack initiation. (a) Surface crack initiation. The surface roughness of the LSM gear was very high and stress concentration existed on the LSM gear surface. When experimental loading was applied, the fatigue crack source was often located on the LSM gear surface. Thus, the LSM gears had a shorter fatigue life than that of other gears. After SSP and TSP treatment, the surface roughness of the LSMSSP gears and the LSMTSP gear decreased; the location of the surface crack was reduced accordingly. At this moment, the higher compressive residual stress of the LSMSSP gears contained the fatigue crack initiation at the LSMSSP gear surface, and forced fatigue cracks to initiate in the internal weak area of the LSMSSP gears (i.e. a tensile residual stress field). Therefore, the fatigue strength of the LSMSSP gears has been improved by compressive residual stresses. (b) Subsurface crack initiation. The local loading stress of fatigue crack initiation was reduced, and there was no effect of surface damage and lubricant on the fatigue crack initiation of the LSMSSP gears in the subsurface. Meanwhile, the dislocation slip is restrained. Therefore, the local fatigue resistance of the LSMSSP gears was improved. 3.6.2.2. Effect of residual stress fields on the fatigue crack propagation. Zabeen et al. [27] found that a compressive residual stress field can not only greatly reduce the fatigue crack growth rate da/ dN, but also turn fatigue cracks into non-propagation cracks, and improve the fatigue crack closure force. Thus the fatigue strength of the material was improved by the compressive residual stress. When the fatigue cracks began to propagate, the fatigue crack growth rate da/dN and stress intensity factor △K of the LSMSSP gears decreased due to the presence of the compressive residual stress field in the LSMSSP gears. When the stress intensity factor △K was lower than the threshold value of the fatigue crack propagation, the fatigue crack propagation of the LSMSSP gears would even stop propagating. Libor Trško et al. [28] measured the residual stresses of the SSP specimens, and the residual stresses partially relaxed during the test. During the LSMSSP gear running, the compressive residual stress field of the LSMSSP gears

inevitably had a certain degree of relaxation, which made da/dN and △K of the LSMSSP gears increased again. Through mutually bridging between the fatigue cracks, finally pitting was formed on the LSMSSP gear tooth flanks. 3.6.3. Retained austenite Retained austenite is a kind of phase with lower hardness and a smaller specific volume. The retained austenite that existed in the LSMSSP gear surface layer would decrease the plastic deformation resistance and effective compressive residual stress of the LSMSSP gears. Fig. 7 shows that the strain induced martensitic transformation of retained austenite has taken place in the SSP process. The hardening layer was generated on the LSMSSP gear tooth flanks. Besides, SSP was better than TSP with reference to the elimination effect of retained austenite. A decline in retained austenite content had a very positive influence on the contact fatigue property of the LSMSSP gears. 3.6.4. Surface roughness Higher surface roughness had a bad effect on the fatigue resistance of the LSMSSP gears. Stress concentration occurred easily at the peak-valley position of the LSMSSP gear tooth flanks. Under the action of the alternating stress, the stress concentration region preferentially became the fatigue crack source, the fatigue crack initiated and propagated. However, it is interesting that the properties of the LSMSSP gears improved with an increase in the surface roughness. It suggests that the fatigue cracks of the LSMSSP gears did not mainly initiate on the surface but in the subsurface layer. Therefore, higher surface roughness did not affect the fatigue performance of the LSMSSP gears obviously. Moreover, the harmful effects of surface roughness can also be offset by the improvements of other aspects of the LSMSSP gears after SSP treatment.

4. Conclusions This paper investigated the effect of severe shot peening and traditional shot peening on the fatigue resistance of laser surface

Y. Lv et al. / Materials Science & Engineering A 658 (2016) 77–85

melted 20CrMnTi steel gears. According to analyses, this study came to the following conclusions: (1) The average compressive residual stress of the LSM-3-SSP gears (  469 MPa) was 33.6% more than that of the LSM-3-TSP gear ( 351 MPa). The average depth of compressive residual stress of the LSM-3-SSP gears (1030 μm) was 186% more than that of the LSM-3-TSP gear (360 μm). Comparing with TSP, SSP had higher energy kinetic and induced a greater degree of plastic deformation, thereby resulting in higher values and depths of compressive residual stresses of the LSMSSP gears. With increasing Almen intensity, residual compressive stress of LSMSSP gears gradually increased, but saturated values of surface residual stresses of LSMSSP gears were almost the same. (2) The average values of the FWHM of the LSM-3-SSP gears (6.47°) were 4.4% more than that of the LSM-3-TSP gear (6.2°). The average depth of the work hardening layer of the LSM-3SSP gears (416.7 μm) was 108.3% more than that of the LSM-3TSP gear (200 μm). The results showed that SSP was stronger than TSP in respect to the work hardening effect. (3) The average microhardness of the LSM-3-SSP gears (665.7 HV0.5) was 4.8% more than that of the LSM-3-TSP gear (635.4 HV0.5). The average influenced depth of the LSM-3-SSP gears (425 μm) was 70% more than that of the LSM-3-TSP gear (250 μm). An increment in microhardness of the LSMSSP gears was caused by grain refinement and higher dislocation densities. With an increase in Almen intensity, surface hardness of LSMSSP gear gradually increased, but the peak values of surface hardness of all LSMSSP gears were very similar. (4) All of the LSMSSP gears had greater surface roughness than the LSMTSP gear. With an increase in the Almen intensities, impactive energy of SSP increased, and the surface roughness of the LSMSSP gears were improved. Under various Almen intensities, the change trends of surface roughness of LSMSSP gears were different due to the distinctions in initial surface roughness of LSM gears. (5) All of the fatigue lives of the LSMSSP gears were longer than that of the LSMTSP gear. Fatigue lives of the LSMSSP gears increased with an increase in the Almen intensity. Factors that determined the fatigue lives of the LSMSSP gears were grain refinement, residual stress fields, retained austenite and surface roughness.

Acknowledgment The authors would like to acknowledge the Project Doctor Scientific Research Startup Fund (No. 2012.301) and the Project Seed Fund (No. 2015.247) of Jilin Agricultural Science and

85

Technology University, the Scientific Research Project of Jilin Provincial Education Department (JiJiaoKeHeZi. 2015.No.378), the Project Youth Scientific Research Fund (No. 20150520107JH) of Jilin Province Science and Technology Development Program for financial support.

References [1] Carlos M.C.G. Fernandes, Pedro M.T. Marques, Ramiro C. Martins, Tribol. Int. 88 (2015) 309–316. [2] Carlos M.C.G. Fernandes, A. Hernandez Battez, R. González, Tribol. Int. 90 (2015) 306–314. [3] Carlo Dengo, Giovanni Meneghetti, Manuele Dabalà, Int. J. Fatigue 80 (2015) 145–161. [4] Vincent Savaria, Hossein Monajati, Florent Bridier, J. Mater. Process. Technol. 220 (2015) 113–123. [5] C. Correa, D. Peral, J.A. Porro, Opt. Laser. Technol. 73 (2015) 179–187. [6] C. Correa, L. Ruiz de Lara, M. Díaz, Int. J. Fatigue 79 (2015) 1–9. [7] Thomas Rousseau, Thierry Hoc, Philippe Gilles, J. Mater. Process. Technol. 225 (2015) 413–420. [8] J. Marteau, M. Bigerelle, P.-E. Mazeran, Tribol. Int. 82 (2015) 319–329. [9] Yutaka Kameyama, Hitoshi Ohmori, Hiroshi Kasuga, CIRP Ann. – Manuf. Technol. 64 (2015) 549–552. [10] Yutaka Kameyama, Kotaro Nishimura, Hideaki Sato, Tribol. Int. 78 (2014) 115–124. [11] Seong-Jong Kim, Koang-Yong Hyun, Seok-Ki Jang, Curr. Appl. Phys. 12 (2012) S24–S30. [12] Anoop Vasu, Koorosh Gobal, Ramana V. Grandhi, Int. J. Fatigue 70 (2015) 395–405. [13] S.M. Hassani-Gangaraj, A. Moridi, M. Guagliano, Int. J. Fatigue 62 (2014) 67–76. [14] S.M. Hassani-Gangaraj, A. Moridi, M. Guagliano, Mater. Des. 55 (2014) 492–498. [15] You Lv, Liqun Lei, Lina Sun, Mater. Sci. Eng. A – Struct. 629 (2015) 8–15. [16] You Lv, Mater. Sci. Eng. A – Struct. 564 (2013) 1–7. [17] Okan Unal, Remzi Varol, Appl. Surf. Sci. 351 (2015) 289–295. [18] S. Bagherifard, I. Fernandez-Pariente, R. Ghelichi, Int. J. Fatigue 65 (2014) 64–70. [19] S. Bagherifard, R. Ghelichi, M. Guagliano, Surf. Coat. Technol. 204 (2010) 4081–4090. [20] Okan Unal, A. Cahit Karaoglanli, Remzi Varol, Vacuum 110 (2014) 202–206. [21] S.M. Hassani-Gangaraj, K.S. Cho, H.-J.L. Voigt, Acta Mater. 97 (2015) 105–115. [22] Okan Unal, Remzi Varol, Appl. Surf. Sci. 290 (2014) 40–47. [23] Katarína Miková, Sara Bagherifard, Otakar Bokuvka, Int. J. Fatigue 55 (2013) 33–42. [24] ISO 4278, Geometrical Product Specifications (GPS) – Surface Texture: Profile Method – Terms, Definitions and Surface Texture Parameters, 1st ed., 1997. [25] Sara Bagherifard, Mario Guagliano, Eng. Fract. Mech. 81 (2012) 56–68. [26] S. Bagherifard, I. Fernandez-Pariente, R. Ghelichi, Mater. Des. 45 (2013) 497–503. [27] S. Zabeen, M. Preuss, P.J. Withers, Acta Mater. 83 (2015) 216–226. [28] Libor Trško, Otakar Bokůvka, František Nový, Mater. Des. 57 (2014) 103–113.