1 Introduction

Rolling bearings are a common mechanical element used in many manufactured products, such as automobiles, electrical appliances, and aircraft engines [1,2,3,4,5,6,7]. Bearing outer rings are an important component in rolling bearings and are usually manufactured by forging because forging can eliminate some defects caused by the raw materials. The main form of bearing failure is fatigue spalling [8, 9], which is mainly related to flow-line outcrop. Under high temperature, high pressure, and alternating stress conditions, the bearing outer ring experiences pitting and stress corrosion due to flow-line outcrop, which eventually causes the groove to undergo fatigue spalling. During the forging process, second phases in metallic materials are fragmented and distributed as a chain or strip along the elongation direction. These chains or strips are macroscopic “flow lines” and can cause anisotropy in the mechanical properties. The flow lines are preferably parallel to the working surface, but the ones of concern are perpendicular to the working surface. Therefore, if the flow-line distribution is reasonable, the product life can be greatly extended [10]. Chen et al. [11] found that relatively reasonable flow lines help the bolt resist tensile stress and shear stress. Zhang et al. [12] studied the effect of flow lines on the mechanical properties forged by isothermal precision. However, there are few reports on the flow lines of bearing rings; thus, it is necessary to study the metal flow lines of bearing rings. For bearings, the expected flow lines are distributed along the contour of the bearing outer ring, which greatly improves the contact fatigue resistance and service live of the bearing ring.

The excellent properties of M50 steel are related to its alloying elements and carbides [13], which result in secondary hardening as well as poor hot-working plasticity for M50 steel. Thus, M50 bearing steel typically exhibits decreased deformability only in the relatively narrow processing range for hot working. Moreover, the carbides in the M50 steel tend to aggregate in a band-like distribution, resulting in an extremely uneven chemical composition, greatly dividing the matrix and decreasing the impact toughness. The coarse carbides simultaneously cause stress concentrations in the matrix during service and form butterfly wings (BWs); a BW near the groove position converges, eventually causing the groove to peel off [14,15,16]. The current research on M50 steel mainly focuses on surface modification and improving the surface properties of the bearing ring to reduce the possibility of groove spalling [17,18,19]. There are few studies on the distribution of bearing ring carbides after the forging process.

Traditional defect analysis is usually based on experiments, but some defects are difficult to clearly explain due to metal flow [12], especially for the multistep processing of bearing rings. FEM simulation is a convenient method for forging defect analysis, and it makes the formation of defects more intuitive [11]. Therefore, the formation mechanism and evolution law of forging defects can be accurately analyzed, and effective measures can be taken to avoid forging defects. The forming quality of bearing rings is considerably affected by the parameters of the hot-working process. Therefore, the influence of different forging process parameters on the forming properties must be studied. In this research, the formation mechanism and evolution law of flow lines during the bearing ring forging process were investigated using FEM simulations and experiments, and control methods for optimizing the flow lines are proposed. The microstructure and mechanical properties of the bearing rings forged by the optimized process were also studied.

2 Methodology for FEM simulations and experiments

2.1 Experimental materials

M50 bearing steel supplied by AECC Harbin Bearing Co., Ltd. was used as the experimental material. The specific chemical composition is shown in Table 1. The M50 bearing steel was melted by vacuum induction melting and vacuum arc remelting, followed by a forging process. The M50 steel used in the experiment was in a spheroidized annealed state. The raw billet for process one and process two consisted of ϕ80 × 121 mm and ϕ75 × 113 mm bar materials, respectively. The upsetting ratios for process one and process two were 2.28 and 2.11, and the rolling ratios were 2.49 and 1.93, respectively.

Table 1 Chemical composition (in wt%) of M50 steel
Full size table

2.2 Forging process

Figure 1(a) shows a schematic of process one for manufacturing bearing outer rings. The deformation of the billet can be divided into three steps. In the first step, the billet was upset to reduce the thickness under the impact of the upper board. The upsetting process was conducted with an air hammer. Then, it was necessary to operate the punching blind hole in the middle position of the billet. The punching recess was placed at a position 1/3 of the distance from the bottom of the billet. Finally, the billet was placed on a D51 ring rolling mill to perform a ring rolling test, and a core roller with a protrusion was used to form the bearing outer ring and its groove. To obtain a bearing outer ring with a reasonable distribution of flow lines, process two was proposed. The process concept is shown in Fig. 1(b), which involved placing the punching recess at the bottom surface. The billet was put into the furnace at 830 °C and heated to 1120 °C while inside the furnace to ensure uniform heating of the billet. Then, the upsetting and punching processes were carried out, followed by three repeated expanding processes to grow the billet to the required size. To ensure that there were no defects in the forging process, the final forging temperature needed to be higher than 950 °C. After returning the expanding billet to the furnace for a period of time, the rolling process was carried out. Finally, the bearing outer ring was subjected to a spheroidizing annealing treatment after the ring rolling process. As a hypereutectoid steel, the spheroidizing annealing temperature of M50 steel is from 30 to 50 °C on AC1, and it was placed near the AC1 temperature for an extended period of time. The large carbides were spheroidized to avoid stress concentrations caused by large sharp carbides during service. After the spheroidizing annealing process, the billet was slowly cooled to 550 °C and then cooled in the air.

Fig. 1
figure 1

Bearing outer ring forging processes. a Process one with straight punching pin and b process two with V-shaped punching pin

Full size image

2.3 Simulation parameters

DEFORM 3D-11.0 FEM software was used to simulate flow-line formation with rigid-plastic material models. The mold was defined as rigid, and the workpiece was set as plastic. The forming process for the bearing outer ring was simulated to analyze the forming mechanism and the evolution law of flow lines. The simulation parameters were as follows:

(1) The top die speed was 5 mm/s.

(2) The selected shear friction coefficient was 0.4 [20].

(3) The temperature of the workpiece was 1120 °C, and those of the molds were 150 °C.

(4) The material used in the simulation was M50 steel, and the material of the mold was AISI H13 (tool steel).

(5) The drive roller speed and feed rate were 7.85 mm/s and 6 mm/s, respectively.

2.4 Microstructure observation

Figs. 2(a) and (b) show that a sample with a thickness of 10 mm was cut along the longitudinal section of the bearing outer ring, and the reasonable flow-line distribution is shown in Fig. 2(c). The surface of the sample was smoothed by sandpaper, followed by mechanical polishing, and finally, the flow-line distribution was revealed by using a specific etchant. In this experiment, the polished billet was placed in a 50% HCl and 50% H2O etchant for 30 min, and the temperature was between 60 and 70 °C. An Olympus stereomicroscope was used to observe the flow-line distribution.

Fig. 2
figure 2

Bearing outer rings. a Outer bearing forging, b sampling position, and c expected flow lines

Full size image

Optical microscopy (OM) samples were prepared by grinding with 200, 400, 600, and 800 grit SiC papers, polishing with a diamond paste, and finally etching in a supersaturated picric acid aqueous solution at 60 °C. Metallographic observation was performed using an Olympus metallographic microscope, and the average grain sizes were measured using a linear intercept method.

The microstructure and morphology before and after forging were observed by scanning electron microscopy (SEM) on an FEI Quanta 200F instrument and field-emission transmission electron microscopy (TEM) on a Talos F200x instrument. The specimens for TEM were manually ground to 50–80 μm using waterproof abrasive papers. The specimens were consequently prepared by twin-jet electropolishing containing a 10% perchloric acid alcohol solution.

2.5 Mechanical properties

The tensile tests were conducted with an Instron-5569R electronic tension machine at ambient temperature, and the tensile rate was 0.54 mm/min. Tensile samples with a gage dimension of 3 × 1.2 × 1 mm were cut from the regions shown in Fig. 2(b) with the tensile axis parallel to the axial direction (AD) and perpendicular to the radial direction (RD). To ensure the reliability of the results, the tensile strength was the average of five repeated tensile tests.

3 Results and discussion

3.1 Microstructure of original annealed M50 steel

The optical and SEM micrographs of the original annealed M50 steel are shown in Fig. 3. The grains are equiaxed, and the average grain size is approximately 22 μm, as shown in Figs. 3(a) and (b). Figures 3(c) and (d) show that the carbides of the original annealed M50 steel are distributed along the AD, and agglomeration occurs between the carbides. The energy spectrum analysis of the bulk carbides reveals that the main elements in this carbide are carbon, molybdenum, and vanadium, which is consistent with an M2C-type carbide reported in reference [21]. In general, The M2C carbides are rod-shaped and very irregular. The length of a single M2C carbide is approximately 20–30 μm, while the length of an agglomerated carbide is greater than 50 μm or even greater than 200 μm. A large number of nanometer-sized granular carbides are uniformly dispersed in the matrix. These carbides are not calibrated due to their small size. Therefore, TEM is required to continue the analysis.

Fig. 3
figure 3

Microstructure of the original M50 steel. a OM micrograph of AD, b OM micrograph of AD at high magnification, c SEM micrograph of AD, and d SEM micrograph of AD at high magnification

Full size image

Figure 4 shows the TEM and high-angle angular dark-field (HAADF) micrographs of the original annealed M50 steel. It can be seen from Figs. 4(a)–(h) that large irregular carbide particles are enriched in molybdenum, vanadium, and carbon; thus, it can be confirmed that the carbide is an MC-type carbide [22, 23]. These granular carbides are enriched in chromium, molybdenum, and carbon; thus, it can be confirmed that the carbide is M23C6-type [21, 23].

Fig. 4
figure 4

TEM micrograph of original annealed M50 steel. a HAADF image. Element distributions of b Fe, c C, d Cr, e Mo, and f V

Full size image

3.2 FEM analysis of flow-line evolution during the whole forging process

Due to the transferability of the flow-line distribution, it is necessary to analyze the influence of the upsetting and punching processes on the flow-line distribution. Therefore, FEM was used to simulate the evolution of the flow lines during the upsetting and punching processes. The initial flow lines are distributed along the AD before upsetting, and the flow-line distribution after upsetting is presented in Figs. 5(a) and (b). Due to the frictional force, the distribution of the flow lines will change after upsetting, and the flow-line distribution at the middle sides are curved. After upsetting, the billet is subjected to a punching process. At this time, the center of the billet is thinned by the upper and lower molds, and the flow lines before the final punching are shown in Fig. 5(c). The flow lines at the position where the billet is in contact with the upper and lower molds are disordered, but a distribution of flow lines away from the inner hole of the billet does not occur. The flow-line distribution after punching is shown in Fig. 5(d). After punching, the flow lines are disordered, and there are vortex defects at the inner diameter of the billet. At the position where the billet is in contact with the upper and lower molds for punching, the turbulence of the flow lines is significantly stronger than that in other positions, which indicates that the punching process has an important influence on the distribution of the billet flow lines.

Fig. 5
figure 5

Evolution of flow lines during process one. a Initial flow lines, b after upsetting, c after punching, and d after punching the recess

Full size image

Preforming is the key process for obtaining reasonable flow lines, and the size of the billet is consistent with the design dimensions. To provide a high-quality ring billet for ring rolling, it is necessary to accurately design the deformation modes and deformation dimensions of each pass during the forging process. The billet is forged by swage forging, and the upper die is formed by a splitter taper. The punching recess is placed on the bottom surface to eliminate various defects caused by the punching recess surface in the working area of the groove, and the defects generated by the bottom punching recess can be completely removed by subsequent processing. The flow lines can be shunted by the splitter taper to avoid the cut-off phenomenon of flow lines during punching. After the billet is released from the mold, the surface is cut with a small punch. Next, the expanding process is carried out three times. Each pass has to control the position of the billet to ensure the position of the V-shaped hole. The through-hole punch is used to eliminate the taper of the inner hole into a straight hole, and process two is depicted in Fig. 1(b). Figure 6 shows the FEM simulation of the billet flow lines after the upsetting and punching processes. The flow-line distribution is reasonable, and there is no flow-line disturbance or vortex. Therefore, process two proved to be reasonable and feasible.

Fig. 6
figure 6

Distribution of flow lines during process two. a Initial flow lines, b after upsetting, c after punching, d after punching the recess, and e after expansion

Full size image

3.3 Stress and strain analysis of two forging processes

During plastic deformation, the metal mainly flows in the direction in which the maximum principal stress increases, and the metal flow has an important influence on the evolution of the flow lines. Therefore, stress and strain analysis is performed for each process. The upset deformation zone can be divided into three zones according to the degree of deformation. Among the three zones, the deformation degree of Zone I is the smallest, the deformation degree of Zone II is the largest, and the deformation degree of Zone III is moderate (Fig. 7(a)). The friction between the tool and the billet and the rapid temperature drop are the main causes of inhomogeneous deformation during the hot upsetting process. Therefore, the degree of metal deformation in Zone I is small, and the stress in Zone I is shown in Fig. 7(a).

Fig. 7
figure 7

Stress and strain analysis of both process. a Zone I stress analysis, b stress and strain analysis in Zone C, (c) effective strain distribution in process one, d effective strain distribution in process two, e stress and strain analysis in Zone B, and f stress and strain analysis in process two

Full size image

After the upsetting process, the punching process is performed. The punching process is mainly divided into a direct stress in Zone A, an intermediate stress in Zone B, and a difficult deformation in Zone C, as shown in Fig. 7(c). The general deformation tendency is as follows: the RD experiences compression deformation, the tangential direction experiences elongation deformation, and the AD depends on the value of D (billet outer diameter)/d (billet inner diameter). The metals in Zone A are similar to those that experienced upsetting under the circumference of the annular metals. The metal in Zone A is compressed, resulting in a decrease in the height, and the metal in this zone flows radially outward. Therefore, the stress state of Zone A is a three-dimensional compressive stress. The metal in Zone B is deformed by the radial outflow of the metal from Zone A. Since the D/d value of process one is 2.5, the radial stress |σr| is small and the tangential stress |σθ| is large; consequently, the stress order is σθ, σz, and σr, i.e., σ1, σ2, and σ3. At this time, the RD and AD experience a negative strain, and the tangential direction experiences a positive strain; thus, the height of the billet after punching is reduced (Fig. 7(e)). Because of the friction and temperature reduction factors, the metal in the contact portion of the punch (Zone C) is in a difficult deformation region. In addition to the plastic deformation of the adjacent metal, the radial force σ1 in Zone C must overcome the effect of the friction force. Therefore, it is difficult for plastic deformation to occur in Zone C, and the stress profile is as shown in Fig. 7(b). Since Zone C is less prone to plastic deformation, the flow lines in this region remain substantially unchanged and are still distributed along the AD, as shown by the arrow in Fig. 5(c). However, the plasticity and fluidity of the metal in Zone B is good, resulting in the flow lines curving downward along the AD. Therefore, opposite metal flow directions occur between Zones A and C, which leads to the occurrence of a vortex (Fig. 5(d)). Since D/d = 6.7 for process two, the outer sides σθ, σz, and σr are small, and the deformation is mainly concentrated on the inner side (Fig. 7(d)). At this time, the inner σr and σθ experience compressive stresses, and the stress order is σz, σθ, and σr, i.e., σ1, σ2, and σ3 (Fig. 7(f)). According to the stress-strain relationship, the AD is elongated and deformed; thus, the height of the billet after punching increases. The flow-line distribution of process two is reasonable because the deformation is mainly concentrated on the inner side of the billet, and the flow lines are still distributed according to the upsetting process.

3.4 Analysis of flow-line distribution in process one and process two

Using the bearing outer ring manufacturing process one, the punching recess is placed at a position 1/3 of the distance from the bottom of the billet, and the raised core roller is also placed at this position, as shown in Fig. 8(a). The flow-line distribution of the bearing outer ring is shown in Fig. 8(c). A vortex appears at the 1/3 position of the groove, as shown in the oval frame. The flow lines of this position do not distribute along its outline position but deviate from the groove position. The distribution of flow lines does not meet the actual requirements. Due to the disorder of the flow lines at the location of the punching recess and the strong movement of the metal with the raised core roller, a vortex appears.

Fig. 8
figure 8

Ring rolling. a Core roller setting at the position 1/3 of the distance from the top of the billet and b distribution of flow lines. c Core roller setting at the position 1/3 of the distance from the bottom of the billet and d distribution of flow lines

Full size image

Since the flow lines are still unreasonable when forging is performed by process one, the position of the core roller is changed. The convex position of the core roller is set at the position 1/3 of the distance from the top of the billet. At this time, the convex position of the core roller deviates from the position of the punching recess, as shown in the schematic in Fig. 8(b). The vortex condition of the bearing outer ring groove position is greatly reduced, but the vortex phenomenon still exists, as indicated in Fig. 8(d). This result shows that even if the position of the core roller avoids the position of the punching recess, the adverse effect of the punching recess on the flow-line distribution cannot be eliminated. The experimental results are consistent with the FEM simulation results.

Figure 9(a) shows the FEM simulation results of the flow-line distribution after forging with process two. As shown in Fig. 9(b), by optimizing the process parameters, the flow lines of the groove are ultimately densely distributed and exhibit an orderly distribution. The flow-line density at the bottom of the groove is relatively large, and uniform flow lines are gradually formed along the groove toward the edge. There is no vortex or disorder, and the flow-line distribution is reasonable, which is consistent with the FEM simulation results.

Fig. 9
figure 9

Distribution of flow lines in the bearing outer ring. a FEM simulation and b forging result

Full size image

3.5 Microstructure analysis of process one and process two

Figure 10 shows the microstructure micrographs of process one after ring rolling. The grains at three positions are mainly equiaxed, indicating that dynamic recrystallization occurs. The distribution of carbides is consistent with the direction of the flow lines. Figures. 10(a) and (d) reveal that grains are refined to 9.69 μm, and many M2C carbides are distributed along the groove. However, the phenomenon of M2C carbide agglomeration has been significantly improved. Figures. 10(b) and (e) show the grain size and carbide distribution at the middle position of the bearing outer ring. It can be seen that grains are refined to 9.83 μm, and the M2C carbide fragment is smaller than that at the groove position. Figures. 10(c) and (f) show the grain size and carbide distribution at the outside position. Grains are refined to 9.98 μm, and the phenomenon of M2C carbide agglomeration becomes more obvious.

Fig. 10
figure 10

Optical photographs of the bearing outer ring forged by process one at the a groove position, b middle position, and c outer position. SEM micrographs of the d groove position, e middle position, and f outer position

Full size image

Figure 11 presents the micrographs of the process two microstructure after ring rolling. The grain morphology and M2C carbide distribution at three positions are consistent with process one. The distribution of carbides is consistent with the direction of the flow lines. The grain sizes at three positions are refined to 9.4, 9.7, and 9.7 μm, respectively. Figure 11(d) reveals that the phenomenon of M2C carbide agglomeration at the groove position has been significantly improved. The length of the M2C carbide agglomerates is 30 μm, and a single carbide near the groove is broken up to a size of 10 μm, which is much lower than that of the original M50 steel. In Fig. 11(e) and (f), the agglomeration of carbides becomes apparent along with the location away from the groove position. The grain size at different positions does not vary substantially, indicating a homogeneous grain size after forging with process two followed by a spheroidizing annealing treatment.

Fig. 11
figure 11

Optical photographs of the bearing outer ring forged by process two at the a groove position, b middle position, and c outer position. SEM micrographs of the d groove position, e middle position, f outer position, and g a magnified view of the groove position. h Effective strain distribution of the bearing outer ring

Full size image

The deformation of the ring rolling process is inhomogeneous, and therefore, the deformed microstructure shows different characteristics throughout the thickness of the ring wall [24]. During the ring rolling process, a core roller applies a rolling force in the RD, and the driving roller exerts a reaction force on the ring to drive the ring to rotate. Therefore, the plastic deformation first starts closest to the groove position and continues toward the outside position. The groove position has the maximum plastic deformation, and the strain at the outer position is the smallest (Fig. 11(h)). Due to the deformation being mainly concentrated on the inner side of the billet, a large M2C carbide near the groove position is fragmented, and the M2C carbide agglomeration phenomenon decreases. After the bearing outer ring is forged, the grain is uniformly equiaxed due to the spheroidizing annealing treatment, but the M2C carbide agglomerates cannot be changed by the spheroidizing annealing treatment and can be fragmented only by forging. The spheroidized annealed microstructure contains pearlite, which is formed by the distribution of small granular carbides and a small amount of primary carbides in the ferrite matrix. The amounts of granular carbides are greater than those in the original M50 steel, and the distribution is more uniform, as shown in Fig. 11(g). In addition, the corners of the granular material are very rounded, which reduces the stress concentration caused by sharp carbides during the bearing outer ring service.

3.6 Mechanical anisotropy analysis

Figure 12 shows the tensile curves and mechanical properties of bearing outer ring samples forged by process two at ambient temperature, in which the ultimate tensile strength (UTS) and elongation (EL) are presented. The bearing outer ring exhibits obvious mechanical anisotropy. The UTS (899.3 MPa) and EL (24.3%) in the AD are significantly better than those in the RD. The AD is parallel to the direction of the flow lines, and the RD is perpendicular to the direction of the flow lines, which indicates that the mechanical properties in the lateral direction of the flow-line distribution are superior to those in the longitudinal direction. The reason for the mechanical anisotropy is that the cross-sectional area of the carbides in the lateral direction is small, whereas the cross-sectional area in the longitudinal section is large.

Fig. 12
figure 12

Tensile properties of the bearing outer ring in different directions. a Engineering stress-strain curves and b comparison of RD and AD

Full size image

4 Conclusions

In this paper, FEM simulation was used to assist the actual forging experiment to reveal the formation mechanism and evolution law of flow lines, and some control methods for optimizing the flow lines were proposed. The main results can be summarized as follows:

  1. (1)

    In process one, the flow lines exhibit opposite metal flow directions and vortexes due to the presence of difficult deformation zones.

  2. (2)

    A bearing outer ring with a reasonable flow-line distribution can be obtained by setting the punching recess at the bottom of the billet and utilizing the V-shaped punching pin.

  3. (3)

    Process two, followed by a spheroidizing annealing treatment, can produce homogeneous equiaxed grains and reduce the average grain size from 22 μm to approximately 9 μm. However, the M2C carbide agglomerates cannot be changed by a spheroidizing annealing treatment and can be fragmented only by forging.

  4. (4)

    The simulation results are consistent with the results of the physical experiments, and the process two bearing outer ring forging methods are verified.

  5. (5)

    The bearing outer ring exhibits evident mechanical anisotropy, and the mechanical properties in the AD are significantly better than those in the RD.