Effects of multi-pass turning on surface properties of AISI 52100 bearing steel

Hard turning is extensively used in the machining of bearings. The turning process has a significant influence on the properties of machined surface. In this paper, multi-pass turning experiments were machined on AISI 52100 bearing steel, and corresponding simulation model was established. The effects of multiple pass turning on microhardness and residual stress were investigated. The results demonstrate that the machined surface hardness of single, double, and triple pass turning is 30.0%, 25.2%, and 24.5% higher than the initial surface, respectively, at a turning depth of 0.1mm and a speed of 100m/min. Double pass turning significantly reduces the residual stress of machined surface. With the change of turning speed, the residual stress after double pass turning is 88 MPa lower than single pass on average, while the difference of residual stress is relatively minor after double pass and triple pass turning. At a cutting depth of 0.05 mm, the residual stresses after double pass turning and triple pass turning are 152 MPa lower than those after one-pass turning. As the turning depth increases, the influence of the previous pass turning gradually decreases in terms of residual stress.


Introduction
Bearing steels are widely used in nearly all mechanical equipment for its outstanding fatigue performance and excellent wear resistance [1,2]. However, bearing failure is a common fault in rotating systems [3]. Both residual stress [4] and hardness [5] have an essential impact on the service life of bearings. Hard turning is a popular machining method in bearing manufacturing. Nevertheless, turning inevitably affects the residual stress [6], hardness [7], and other properties of machined surface. Therefore, investigations on residual stress and microhardness of turned surface are prerequisite for a better performance of bearings.
In recent years, many researchers have focused on the residual stress generated by turning bearing steel. Cutting speed [8], cutting depth [9], and feed rate [10] had a critical impact on residual stress. Ekkard et al. [11] concluded that in turning thin walls of AISI 52100 steel, feed rate played the most important effects on turning surface residual stress, then speed, wall thickness, and lastly depth of cut. Turning tool and workpiece hardness were also factors affecting residual stress [12]. Greater hone radius tools resulted in deeper negative penetration [13], and the residual stresses had the similar tendency with the tool-to-workpiece contact length [14]. Moreover, Carusoa et al. [15] pointed out that the higher material hardness leaded to greater compressive residual stresses and deeper beneficial depths.
Machining conditions and processing methods also have an essential influence on the performance of bearings. Bertolini et al. [16] and Jamil et al. [17] studied on liquid nitrogen and dry ice assisted turning of AISI 52100, respectively. The experimental results demonstrated residual stresses were improved. Caruso et al. [18] indicated that cryogenic cooling conditions leaded to less compressive residual stress on surface and limited the white layer thickness. Tong et al. [19] stated that ultrasonic assisted turning reduces residual stress, which was beneficial to the fatigue life. Aiay et al. [20] compared the surface integrity of AISI 52100 bearing steel during hard turning in different near-dry environments. The results shown that residual stress, surface roughness, and white layer performed better in Hybrid Nanofluid Minimum Quantity Cooling and Lubrication condition. Singh [21] noted that compared with dry turning, the surface roughness was improved by 8~15% in the condition of using solid lubricant. In terms of material removal rate and tool life, AISI 52100 bearing steel hard turning with high-pressure coolant supply had more advantages [22].
In addition to the single pass turning mentioned above, multi-pass machining is also one of the most important methods in processing and manufacturing. Iqbal et al. [23] showed that double pass turning had advantages over single pass machining in the aspects of cutting energy, surface roughness, and turning forces. Continuous cutting resulted in thicker white layer [24], and the higher residual austenite content at smaller feed rates was that the compressive stresses on the turning surface restricted martensitic phase transformation. However, many researchers studied multi-pass turning focused on production cost and environmental protection. Lu et al. [25] developed a multi-objective optimization model for multi-pass turning, which considered machining quality and energy consumption. Phengky et al. [26] established a multi-pass turning process parameter optimization model with multiple objectives of carbon emission, productivity, and energy consumption.
Multi-pass turning is an excellent alternative solution for large turning depth machining, and maybe a feasible method to improve the integrity of the machined surface. The previous pass turning, which changes the current surface and subsurface properties, has an essential influence on microhardness and residual stress of the workpiece after the next pass turning. Few researches studied the evolution of surface properties in multiple pass turnings process. Therefore, the current study investigates the effects of multi-pass cutting on surface hardness and residual stress at different turning speeds and turning depths.

Experimental procedures
The mental material used in present study was AISI52100, and the initial surface microhardness was about 233 HV. The cylindrical sample is with a diameter of 40 mm and a length of 400 mm. The chemical composition (in wt.%) of the material is listed in Table 1. Firstly, grooving was performed on the blank around the circumference, spaced at 8 mm intervals. The depth and width of the grooves were 3 mm and 2 mm, correspondingly. Then, the experiments were conducted on HTC2050 CNC machine as shown in Fig. 1. Considering the effects of tool wear on microhardness and residual stress, a new carbide turning edge (WNMG080408-BM2) was adopted for each pass turning of each specimen in the experiments. The tool hardness was approximately 98.9 HRC, with a 11° rake angle and 0° relief angle, a cutting-edge radius of 0.02 mm and a nose radius of 0.8 mm.
In the experimental machining processes, the cutting speed and depth were varied and the feed was kept at 0.05 mm/rev. According to different turning parameters, 5 types of cutting process were set up, and each type was performed in single, double, and triple pass turning. The experimental processing parameters are given in Table 2. The depth of turning is 0.1mm×2, which means that double-pass turning are performed, and the depth of each pass turning is 0.1mm. The turning depth is 0.15mm×3, which means that triple-pass turning are performed, and the depth of each pass turning is 0.15mm. After the previous cutting, the workpiece was static for 30min and cooled down to room temperature. Then, the next pass turning process began.
The surface hardness has an essential impact on the wear resistance and fatigue strength of the component. Therefore, microhardness measurements were performed on the machined surface and cross-section. The experiments were conducted on a Vickers hardness tester (402MVD, China). During microhardness test procedure, every specimen was subjected to a load of 50g and remained for 10s. To ensure accurate surface microhardness, each specimen was tested four times, and the average of the tested values was considered the microhardness of the specimen.
The measurement of residual stresses on turned surfaces was carried out with an X-ray stress diffractometer (PROTO

Finite element modeling
Hard turning is a complicated process with complex factors such as high temperature, large deformation, and large strain rate. In the simulation of metal turning, reasonable configurations are selected to assure accurate results. Figure 2 shows the experimental turning model. When viewed from the XY plane, 3D turning could be regarded as 2D turning, and the result is shown in Fig. 3. Among the directions, the X-direction and Y-direction are cutting speed direction and cutting depth direction, respectively. The X-direction in 2D simulation model is the hoop direction in 3D experiment model. A workpiece with a height H of 2 mm and a length L of 5 mm was modeled. The workpiece material was AISI52100 bearing steel, and the constitutive model used in this paper is default. The rake angle and relief angle of the turning tool were 11° and 0.01° respectively, and the radius of cutting edge was 0.02mm. The initial temperature was 25°C and the default friction coefficient was 0.5. The meshes of finite element model were relatively dense meshes near the plastic deformation region and relatively sparse meshes in other regions, so that a better mesh division was obtained. The maximum and minimum element size of the workpiece were 0.1mm and 0.02mm, respectively. The depth of mesh refinement for residual stress analysis was 0.4mm. The turning length of turning simulation model was 3mm. The turning width was 0.05mm, which was equal to the of feed in experimental turnings.
After the previous pass turning, nine positions were selected evenly in the machined area and the residual stresses distributed along the depth were taken. Then, the nine-stress data were averaged for the corresponding depths and directions, and the averaged stress distribution was regarded as the initial stress for next pass turning. In the single pass turning simulation model, the initial stress of the workpiece was not set, and the measured surface microhardness was assumed as the initial condition. The simulated residual stress and the microhardness after single pass turning were taken as the initial conditions of the workpiece in the double pass turning simulation model. The simulated residual stress and measured microhardness after the double pass turning were considered the initial conditions for the triple pass turning simulation model.
The turning process parameters of simulation model were added as shown in Table 3. One-degree polynomial fits to the measured microhardness were performed according to turning speed and turning depth, and the fitting results were regarded as initial microhardness of the added simulation model.

Microhardness
The cross-sections of the turned samples were polished, and then, the samples were etched with a mixture of 4% nitric acid and 96% alcohol. The microstructures were observed using an optical microscope. A cross-sectional optical microscopy image of the triple pass turning of T5 is illustrated in Fig. 4. A dark layer approximately 20 μm exists between the turned surface and the bulk material. In the hard turning of bearing steel, the causes of white layer formation on the workpiece surface are related to excessive heat or mechanical loading [27]. The initial hardness of the material and turning parameters play a crucial role in the white layer formation. When the material microhardness or turning speed is low, the workpiece surface may not develop a white layer after hard turning.
In hard machining, the surface of the workpiece is at high temperatures because of intense extrusion and friction. Thermal effect weakens the surface hardness, while plastic deformation strengthens it [28]. The influences of turning speed and turning depth on surface hardness at multiple turning are shown in Fig. 5. Turning hardens the machined surface. At a turning depth of 0.1 mm and a speed of 100 m/min, the single, double, and triple pass turning result in the machined surface 30.0%, 25.2%, and 24.5% harder than initial surface microhardness respectively. As shown in Fig. 5(a), the microhardness after single, double, and triple pass turning increases by 4.5%, 1.5%, and 2.4%, as the turning speed adds from 75 to 125 m/min. Figure 5 presents a tendency that the surface microhardness decreases as the number of turning pass increases, at the same turning parameters. As shown in Fig. 5(a) and (b), the microhardness decreases by an average of 3.5% and 3.9% after triple pass turning compared to single pass turning, respectively. The difference of microhardness is not obvious after double and triple pass turning. The thermal softening effect may result in softer microhardness of latter pass turning than previous pass turning. Figure 6 shows the microhardness at different depths under multi-pass turning conditions. The microhardness decreases sharply within the 0~20 μm below the machined surface, while beyond 20 μm, the value of microhardness varies around 160 HV. As the number of turning passes increased, the microhardness of the subsurface layers does not differ significantly. Multi-pass turning have essentially no effect on the depth of the hardened layer. Figure 7 displays the temperatures and chip morphologies for different turning parameters, at the moment of cutting length of 0.4 mm in multi-pass turning conditions. Correspondingly compared Fig. 7(a)-(f), when turning speed is changed from 75 to 100 m/min, the maximum temperature increases from 413 to 432 °C in the single pass turning, from 430 to 456°C in the double pass turning, and from 426 to 455 °C in the triple pass turning. The higher turning speed increases the rate of plastic deformation, which has a stronger hardening effect on the material than weakening effect produced by cutting heat, leading to an increase in surface hardness. As shown in Fig. 5(b), the microhardness after one-pass, two-pass, and three-pass turning increased by 6.7%, 4.8%, and 5.0%, as the turning depth changed from 0.075 to 0.15 mm. Correspondingly compared Fig. 7(d)-(f) and Fig. 7(g)-(i), the maximum temperature of the workpiece surface increases with deeper turning depth. The maximum surface temperatures rise by 14 °C, 8 °C, and 6 °C for single, double, and triple pass turning, respectively. Plastic deformation dominates the hardening of the workpiece surface as the turning depth increases.
Harder materials make the turning temperature higher. As illustrated in Fig. 7, the temperatures of double and triple pass turning are greater than single pass turning. The thermal softening effect results in softer surface microhardness of latter pass turning than previous pass turning.
As shown in Fig. 7(a)-(c), the curl degree of chip formation is described by chip radius r. The smaller the chip radius is, the greater the crimp degree is, and vice versa. The comparison of the swarf bending radius produced by multi-pass turning is illustrated in Fig. 7. In the simulation model S2, compared with single pass turning, the chip bending radius is reduced by 14.2% and 15.1%, respectively, after two-pass and three-pass turning. Fig. 7(d)-(i) display that multi-pass turning promotes chip curling. In the simulation models S3 and S10, bulges appear on the side of chip curl during multi-pass turning, which is caused by plastic flow at high temperatures. In multi-pass turning, the thermal softening effect reduces the chip curl radius. Compared Fig. 7(a)-(f), the increase in turning speed results in a reduction in chip radius. Deeper turning depth leads to less chip radius as displayed in Fig. 7(d)-(i). The more curled the chip, the easier it is to break the chip, which reduce the entanglement of the chip on the workpiece and effectively remove the turning heat. Therefore, the decrease in the chip radius is advantageous for the turned surface quality. , f triple pass turning of S3, g single pass turning of S10, h double pass turning of S10, i triple pass turning of S10 1 3

Residual stress
The distributions of simulated turning residual stresses are presented in Fig. 8. The maximum tensile stresses occur approximately 0.025 mm, and compressive residual stresses appear about 0.1 mm below the surface. With the same Y-coordinate, the residual stresses at different X-coordinate are slightly different. The longer the turning length is, the more heat accumulates, which leads to more residual tensile stress. In Fig. 8(b), (c), (e) and (f), more compressive residual stresses are located at the beginning of turning, at 0.1 mm below the surface. More tensile residual stresses appear at the ending part of turning as illustrated in Fig. 8(g)-(h). As illustrated in Fig. 8(d), the residual stress distribution is relatively uniform, with a shape of butterfly. For the comparison of Fig. 8(a)-(f), the difference in surface residual stresses is less when turning speed is changed from 50 to 200m/min. In contrast to Fig. 8(g)-(l), the surface residual stress moves toward the tensile range while the turning depth varies from 0.05 to 0.2mm. At the same turning parameters, the residual stress changes minor after the double and triple pass turning, while they are more compressive than after one-pass turning. Multi-pass turning promotes greater compressive residual stress.
Nine residual stress values are evenly sampled on the surface of the simulated turning model, and the average of the nine values is considered the surface residual stress. The residual stress of each turning pass simulation was fitted by Fig. 8 Residual stress contours of multi-pass turning at different turning parameters: a single pass turning of S1, b double pass turning of S1, c triple pass turning of S1, d single pass turning of S7, e double pass turning of S7, f triple pass turning of S7, g single pass turning of S8, h double pass turning of S8, i triple pass turning of S8, j single pass turning of S13, k double pass turning of S13, l triple pass turning of S13 1 3 quadratic polynomial according to turning speeds and turning depths. A comparison of the experimental and simulated residual stresses is illustrated in Fig. 9. The maximal deviations of experimentally measured and simulated residual stress occur in the 0.075 mm×2 group, with corresponding values of −69.5 MPa and −59.5 MPa. The simulated residual stresses and the experimental measurements are consistent in trend for different turning parameters. One possible explanatory factor for the difference between measurement and simulation of residual stresses is the error related to the measured technique. In summary, the residual stresses of the finite element turning simulation model established in this paper match well with the experimental residual stresses. Accordingly, the simulation model can be used to investigate the effects of multi-pass turning on surface residual stress of AISI 52100 bearing steel. Figure 10(a) illustrates the effects of multi-pass turning on hoop residual stress at different turning speeds. The peak tensile stresses are 125 MPa, 12 MPa, and 51 MPa for single, double, and triple-pass turning, correspondingly, while turning speed exceeds 150 m/min, the residual stress gradually decreases. Figure 10(b) demonstrates the effects of multipass turning, with different turning depths, on hoop residual stress. The residual stress increases from 52 to 152 MPa in single-pass turning, from −100 to 164 MPa in the doublepass turning, and from −100 to 163 MPa in the triple-pass turning, as the turning depth changes from 0.05 to 0.2 mm.
The combined actions of mechanical effect, thermal effect, and phase change cause the residual stress to be altered [29]. Mechanical effect leads to greater cutting force and thus to higher compressive residual stress. Thermal effect causes tensile residual stress, and high temperature may result in phase transformation, which induces compressive residual stress. Ramesh et al. [30] stated that plastic deformation and residual stress could seriously affect the phase transition temperature. Phase transformations began  Fig. 11(a), at a turning speed of 150m/min, the average temperatures during single pass, double pass, and triple pass turning were 547°C, 573°C, and 570°C, respectively. When the speed is more than 150m/min, the high temperature may cause the phase transformation. As presented in Fig. 12(a), the change of the hoop turning force is not noticeable with the turning speed increasing. Therefore, the turning force results in small residual stress at low speeds. While the speed changes from 50 to 150m/min, the thermal effect increases the residual stress on turning surface. When the speed exceeds 150m/min, the excessive temperature causes phase change, which reduces residual stress. Figure 11(b) displays that as the turning depth increases from 0.05 to 0.2 mm, the average temperature changes from 446 to 533°C for single pass turning, from 472 to 568°C for the double pass turning and from 469 to 564°C for the triple pass turning. The turning force grows with the turning depth increasing, as shown in Fig. 12(b). At larger turning depths, the phase transition and mechanical effect are not sufficient to dominate the change in residual stress, which is mainly driven by the tensile stresses produced by thermal effects. Therefore, the turning heat is the primary reason for the increase of residual stress as the deeper turning depth. Figure 10(a) also demonstrates that, at the same turning speed conditions, the hoop residual stress is the highest after single pass turning, and the residual stress is the lowest after double pass turning. The residual stress changes after double pass and triple pass turning are not obvious, and the residual stress after double pass turning is 88 MPa lower than that after single pass turning on average. This is due to greater turning force produces higher compressive residual stress at the same turning speed. In Fig. 10(b), at low turning depths, the larger turning force makes the smaller residual stress after double pass and triple pass turning than that after single pass turning. With the increase of turning depth, the influence of the previous turning on the latter turning is gradually weakened. At a turning depth of 0.2 mm, the combined effects of thermal effect and turning force make the residual stresses tend to be the same after multiple pass turning.

Conclusions
The effects of multi-pass turning on the microhardness and the hoop residual stress of AISI52100 bearing steel are studied by experiments and simulations. The residual stress of multiple pass turning, compared single pass turning, is more preferable. The following is a summary of the main findings: • Turning hardens the surface of the workpiece. At such conditions with a speed of 100 m/min and a depth of 0.15 mm, the surface microhardness of the workpiece after turning is the highest (315.1 HV) in the one-pass turning, which is 35.2% higher than the initial hardness of the surface. • The surface microhardness is harder with the increase of turning speed and turning depth. At the same turning parameters, the surface microhardness is lower after double pass turning than single pass turning, while that is not significant changed after triple pass turning. • In the cutting speed range of 50~150m/min, the increase in speed causes greater residual stresses. After the onepass, two-pass, and three-pass turning, the peak residual stresses are 125 MPa, 12 MPa, and 51 MPa respectively. When turning speed exceeds 150m/min, the residual stress decreases gradually. At the same turning parameters, the residual stress after two-pass turning is lower than one-pass, while the residual stresses change relatively minor after three-pass turning. • As the turning depth changes from 0.05 to 0.2mm, the hoop residual stresses on turned surface increase. At a turning depth of 0.05 mm, the residual stresses move toward compression range after double and triple pass turning, compared to single pass turning. At relatively large turning depths, the increase in turning pass has almost no effect on the residual stress.

Fig. 12
Influence of turning parameters on turning force: a turning force at h=0.1mm, b turning force at V=100m/min