Study on optimization of surface processing technology of silicon nitride bearing ring

Based on the difficult machining characteristics of silicon nitride materials, this manuscript focuses on optimizing the precision machining process of silicon nitride bearing components and improving the machining efficiency and quality of silicon nitride bearing components. Firstly, the mechanism of crack formation and propagation in hard and brittle materials under the action of abrasive particles is discussed in this paper, and based on this, a kinematic model of single abrasive grain cutting on hard and brittle materials is established. The effect of workpiece linear velocity, grinding wheel linear velocity, grinding wheel oscillation velocity and feed velocity on inner surface roughness of Si3N4 ring was discussed through grinding test. The experimental results show that the surface roughness can reach about Ra0.20 ∼Ra0.33 μm through a large number of grinding experiments by adjusting the combination of process parameters μm. On this basis, use the optimized process parameters calculated by the surface quality prediction model constructed in this manuscript to conduct grinding test again, and the surface roughness value reaches about Ra0.19 ∼Ra0.23 μm. The purpose of optimizing the process parameters is realized. Finally, the surface roughness can be further reduced and maintained at Ra0.05 ∼Ra0.06 μm by further superfinishing the surface after the process optimization with an oilstone. The research work of this manuscript realized the optimization of precision machining process of silicon nitride bearing ring and the rapid optimization of machining process through the established prediction model, which improved the precision machining efficiency of ceramic bearing components. Through the study of this manuscript, a process route of controllable processing of inner surface quality of Si3N4 ring is formed, from preliminary selection of process parameters by optimization model to precision grinding and then to ultra-finishing of whetstone, which provides theoretical reference for efficient and precise manufacturing of practical bearing components.


Introduction
Silicon nitride (Si 3 N 4 ) ceramics have the performance advantages of high hardness, high temperature resistance, acid and alkali corrosion resistance, self-lubricating, good wear resistance, low coefficient of thermal expansion, etc. In recent years, ceramic precision parts such as ceramic bearings, ceramic spindles, ceramic bushings have been widely used in aerospace, national defense and military industry, CNC machine tools, energy and chemical industries and other industrial fields.. Taking Si 3 N 4 all-ceramic bearings as an example, ceramic bearings have obvious advantages compared with traditional metal bearings in ultra-high speed, ultra-high temperature, ultralow temperature and other extreme working conditions [1]. Since the first application of ceramic bearings in the 1970s, the design, manufacture and service of ceramic bearings have always been the common key technical issues in the field of ceramic materials, bearing design and manufacturing. In the process of ceramic bearing technology development due to the hard brittle characteristics of ceramic material itself has a very high, the traditional metal bearing processing technology is difficult to guarantee the surface quality of the ceramic bearing components, ceramic surface roughness cannot accurately control in the processing, surface micro cracks produced extremely easily, affect the rotation in the process of precision ceramic bearing service, lead to bearing producing vibration and noise in the process of operation, finally, the surface wear failure affects the service life of the bearing. Therefore, improving the surface machining quality of ceramic bearing assembly has become the key technology to improve the overall service performance and reliability of ceramic bearing [2].
At present, the research work on improving the surface machining quality of Si 3 N 4 ceramic materials mainly focuses on the grinding mechanism of ceramic materials, the micro morphology of ceramic surface and the core key technologies such as diamond grinding wheel dressing. Representative research work includes literature [3][4][5][6][7], which discusses the influence of different grinding processes on the surface roughness, surface morphology and sub-surface damage of Si 3 N 4 based on the brittle fracture characteristics of Si 3 N 4 materials in the traditional grinding process. At the same time, it is consistently proposed that the process from brittle removal to plastic removal in Si 3 N 4 grinding can obviously improve the grinding surface quality. The transformation of grinding mechanism is closely related to the grinding surface temperature. The grinding temperature range, the influence of grinding heat on crack propagation, the composition of surface burn and metamorphic layer can be deeply analyzed by establishing the grinding thermal characteristics model of Si 3 N 4 surface. Therefore, the research on the grinding mechanism of Si 3 N 4 materials mainly explores the influence on the surface quality from the perspectives of fracture characteristics of Si 3 N 4 materials, process parameter optimization and grinding environment influence. Literature [8][9][10][11] discussed the influence of different processing technologies on the removal and surface quality of Si 3 N 4 materials. For the removal and processing of hard and brittle materials such as Si 3 N 4 ceramics, in addition to traditional mechanical grinding, a variety of non-contact processing methods have also been applied to the processing of Si 3 N 4 materials, such as waterassisted ultrasonic and laser composite processing. Using laser local environment of high temperature promoted the hydrolysis of Si 3 N 4 to improve the machining efficiency, Si 3 N 4 materials on the conductive phase, for example, raise its conductivity for wedm processing, this method can obviously materials processing precision and efficiency, but non-contact surface easy to leave after processing method for processing a large number of pits, Surface roughness, surface residual stress, surface fatigue life and other parameters are greatly affected. Literature [12,13] has discussed the Si 3 N 4 chemical assistance processing technology, and compared the removal mechanism of Si 3 N 4 material and the influence of different abrasive types on the surface quality. However, so far, there is still a large space for improvement in the selection of reagents and processing accuracy of chemical assisted processing methods for ceramic materials with different components. In addition, a variety of precision machining processes for silicon nitride bearing components are included, as shown in figure 1. Up to now, among the many processing methods surrounding silicon nitride ceramic materials, the traditional mechanical grinding is still the most mature technology and the most widely used processing method [14][15][16][17].
In conclusion, the surface processing quality of Si 3 N 4 is the key factor to improve the overall service performance of Si 3 N 4 components. Currently, researches on the surface processing quality of Si 3 N 4 are only conducted comparative analysis from the perspectives of material removal mechanism, material removal model and material removal process. In the process of determining the optimal process parameters to improve the processing quality of Si 3 N 4 surface, a large number of repetitive validation processing tests are needed. At the same time, due to the inevitable error of the test results, the scientific research method of obtaining the process parameters to optimize the surface quality of Si 3 N 4 through the test needs to be further verified. Based on the previous research work, this manuscript proposes a prediction model of silicon nitride grinding surface quality on the basis of silicon nitride material grinding mechanism. The surface quality obtained by different processing parameters is compared and verified through prediction model calculation and processing test, and better processing parameters can be obtained through comparison. The research in this paper can reduce the costs related to the process research and processing test of ceramic bearing components to a certain extent. These costs basically come from the traditional process research methods. The traditional process research methods can only obtain the optimal process parameters by eliminating and comparing the test results through large-scale tests. Based on the model established in this paper, it can be gradually optimized and extended to the research and production of precision machining technology of other related components of ceramic bearings, providing a technical reference for achieving precision controllable machining of silicon nitride material surface quality.
2. Removal mechanism and Kinematics modeling of Si 3 N 4 ceramics in grinding 2.1. Removal mechanism analysis of Si 3 N 4 ceramics in grinding The removal mechanism of Si 3 N 4 ceramic materials mainly includes indentation fracture mechanical model and abrasive cutting model. This manuscript explains and discusses the processing characteristics of Si 3 N 4 from two aspects of crack propagation and processed surface breakage respectively. Figure 2(a) shows the indentation fracture mechanics model [18][19][20] constructed based on the indentation experiment of ceramic materials, which can reflect both radial and transverse cracks. According to this model, it can be found that the ceramic materials under the action of load produce plastic deformation zone and crack growth zone at the center of the contact zone. The cracks produced include radial cracks and transverse cracks. When the abrasive particles are pressed into the workpiece, A plastic deformation zone will be generated near the contact zone at first. With the gradual increase of the load and the load exceeds the critical value * P m for radial crack generation of the material, radial cracks will be generated and gradually spread just below the contact zone. When the load further increases to the critical value * P L for transverse crack generation, Transverse cracks will be generated from the plastic deformation zone to the periphery and gradually spread.
The mathematical relation of material properties for radial crack generation critical value * P m and transverse crack generation critical value * P L is as follows: In the formula, λ, η, γ, ζ is the constant related to the pressing of abrasive particles, and f (E/H) is the load attenuation function.
Based on the indentation fracture mechanical model, the abrasive cutting model as shown in figure 2(b) is constructed. When the abrasive particles are pressed into the surface of the workpiece, plastic deformation zone and cracks in different directions will be generated in the contact area. With the abrasive particles cutting on the surface of the workpiece, the cracks continue to expand. When the transverse crack expands to the surface of the material, the material spalling will occur, and then brittle breakage and surface defects will be formed on the processed surface. Therefore, theoretically, the number of broken pits on the workpiece surface and the actual abrasive particles can be used to calculate the breakage rate of the machined surface [21,22].
As shown in figure 2(b), C l is the horizontal length of the transverse crack, and h is the depth of the potential brittle fracture zone. When C l » h, the theoretical model of C l can be obtained according to the template principle as follows [21]: Where, ζ a , ζ b , A 0 is the coefficient dependent on the grinding wheel, ψ is the pressing cone Angle, ¢ F n is the normal grinding force of a single grain of the grinding wheel, and F 0 is the critical value of the transition from brittleness to plastic removal mechanism, since the material is removed in A brittle manner, ¢  F F , n 0 the above equation can be simplified as: Let the contact area between the grinding wheel and the workpiece be A, the effective number of grinding grains per unit area of the grinding wheel be Ns, A•Ns represents the number of broken pits on the surface of the processing zone, and then obtain the theoretical model of the breakage rate of the machining surface: Where, dg is the average diameter of abrasive particles, f is the actual ratio of abrasive particles involved in processing, take 0.5, b is the contact width. It is assumed that the grinding wheel is in an ideal state during grinding, that is, the breakage rate of the machining contact zone in the process is constant, so the formula can be derived as follows: The formula can be further derived: Where, ¢ F n is the normal grinding force of a single abrasive particle, which is related to the cutting depth of abrasive particle, grinding wheel linear velocity, workpiece linear velocity and other factors. The parameter that has the greatest impact on ¢ F n is the cutting depth of abrasive particle a p , which is much larger than other factors. Therefore, the model of surface breakage rate is optimized as follows: Where, ξ is the coefficient related to grinding wheel, x is the influence index of abrasive penetration depth. According to the above equation, π, ζ b , ψ, A 0 , N s , ξ can be found to be the correlation coefficient with the grinding wheel, which is represented by λ in equation (8). Kc, H and E are all the correlation coefficients of workpiece materials, which are represented by η in equation (8). Therefore, the model of surface breakage rate can be simplified as: It can be seen from equation (8) that the main factors affecting the surface quality of engineering ceramics include grinding wheel performance, ceramic material properties and grinding process.

Kinematic analysis of inner grinding of single grinding wheel
The proportion of abrasive particles exposed on the grinding wheel surface is not very high in inner grinding raceway, and it is the abrasive particles with cutting ability that can remove materials in grinding process. In order to make the grinding removal more uniform and improve the processing quality of the removed surface, it is necessary to carry out kinematic analysis on the grinding surface of the workpiece.
The motion of grinding wheel and bearing ring in internal circular grinding is shown in figure 3. OXYZ coordinate system is constructed with the workpiece center O, and oxyz coordinate system is constructed with the grinding wheel center o. Suppose that an abrasive particle at a fixed position on the grinding wheel is point P, and the motion track of the abrasive particle P in the xyz coordinate system is P n . The grinding of the ring with the grinding particle P every turn of the grinding wheel is denoted as point Jn, and the position of the grinding point J n at the end of the machining after it moves with the ring is denoted as G n .
According to the figure above, the trajectory of P n in the oxyz coordinate system is: Where, t n is the grinding wheel oscillation time (s), t total is the total grinding time (s), r G is the grinding wheel radius (mm), and b is the set oscillation amount (mm) of the machine tool, n G is the grinding wheel speed (r min) −1 . The conversion relation between n G and V G of grinding wheel linear velocity is as follows: In equation (9), z n is a continuous non-differentiable periodic function that changes with time: Where, L n represents the reference value of z n , and the value of L n is determined by t n and grinding wheel oscillation speed V z . The derivation formula of L n is as follows: In the actual processing, grinding wheel oscillates along the z axis, so V z in the above formula changes direction continuously with the processing, and its value is equal to the grinding wheel oscillation speed V o set by the machine tool, as follows: Where, n is the oscillation times, S is the oscillation length completed by the grinding wheel at t n , and its value is equal to: Put P n into the coordinate system OXYZ, and its trajectory equation is: Where, (e x ,e y ,e z ) is the position vector of o relative to O in the OXYZ coordinate system    Oo Since point P of the grinding wheel only contacts and grinds with the ring once per revolution, the grinding point J n is a set of discrete points, and its position in the OXYZ coordinate system during each grinding is:  Where, t q is the total time of particle P movement in each grinding action, and its value is: Where, q n is the grinding times, q total is the total grinding times. J n , the grinding point on the ring, rotates with the ring, and the position at the end of processing is denoted as G n , whose equation is: Where, r W is the radius of the workpiece inner circle (mm), t J is the movement time of J n with the ring after grinding (s), and n W is the speed of the ring (r min) −1 . The conversion relation between t J and n W and the workpiece linear velocity V W is as follows: The kinematic analysis of internal grinding process shows that the grinding times of single grain P are positively correlated with the grinding wheel linear velocity and the processing time. The factors affecting the distribution of grinding point G n in the inner circle include grinding wheel linear velocity V G , workpiece linear velocity V W , grinding wheel oscillation velocity V o and processing time. Assume that the width of the workpiece and the oscillation quantity b of the grinding wheel are both 15 mm. When V W is 1 m s −1 , V G is 50 m s −1 and V o is 400 mm min −1 , Stotal of the total oscillation length of the grinding wheel is 10b. The final calculation results are shown in figure 4(a). Where P n is the discretization track of the motion of abrasive P in the OXYZ coordinate system when the step size of t n is 0.00001s. The discretization result of the grinding point G n in the ring in the OXYZ coordinate system ( figure 4(b)) is expanded along the inner circle bus ( figure 4(c)). Set the search radius as 1 and use Jet color map matrix to obtain the visual calculation result of grinding density of inner circle ( figure 4(d))

Experimental process
The test uses internal and external grinding Machine for grinding (MK2710, Wuxi Machine Tools Co.,Ltd., China), the highest speed of the Machine tool spindle 36000 r min −1 , Si 3 N 4 ceramic ring as shown in figure 5, Si 4 N 3 bearing ring inner diameter 54 mm. The diamond grinding wheel used in the test was resin binder with particle size of 100#. The machining process of the Si 3 N 4 ring blank used in the experiment is divided into two parts: rough grinding and fine grinding. In the coarse grinding stage, the same parameters are used (the linear speed of the workpiece is 1 m s −1 , the linear speed of the grinding wheel is 50 m s −1 , the oscillating speed of the grinding wheel is 400 mm min −1 , and the feed speed is 10 μm min) −1 Roughly grind the workpiece, and the uniform depth of the specimen is 300 μm. After rough grinding, carry out fine grinding test machining according to the specific parameters shown in table 1, and deeply discuss the influence of four factors, namely workpiece linear speed, grinding wheel linear speed, grinding wheel swing speed and feed speed, on the internal grinding quality of Si 3 N 4 ring. According to the test group serial number in table 1, each group of tests has carried out three repeatable tests. Therefore, 72 Si 3 N 4 ring of the same specification are selected for this test, and the uniform grinding depth of fine grinding test pieces is 100 μm. Both rough and finish grinding process take uniform grinding depth as the processing target, so the actual processing time varies according to the selection of different process parameters in the finish grinding process.
Superfinishing grinding of whetstone is an important step in raceway machining of bearing ring. General superfinishing processing methods can be divided into cut type and through-through type. At present, most superfinishing methods used in bearing rings are cut type. In order to improve the surface quality, the inner ring surface of Si 3 N 4 bearing was further ground based on the basic principle of plunk-in super finishing raceway of bearing ring. In this paper, the workpiece surface after grinding test shown in table 1 will be superfinished to further improve the workpiece surface quality. The ceramic bond diamond whetstone with particle size of W2.5 and hardness of O grade was selected in the experiment. The superfinishing process is completed by the Bearing star111K inner and outer ring bearing raceway superfinishing machine (Thielenhaus, Germany). The superfinishing pressure is 0.8MPa, the workpiece tangent speed is 225 m min −1 , the swing frequency is 700 Hz, and the superfinishing time is 13 s. Figure 6 shows the schematic diagram and the surface topography of the oil stone for ultra-fine grinding of bearing inner ring surface.

Detection method
The Taylor contact roughness meter (surtronic25 Taylor roughness meter) was used to measure the surface roughness, and the sliding direction of the probe was perpendicular to the end face of the ring. Each ring was measured 5 times, and the final surface roughness value was based on the average value. As shown in figure 7, a diamond wire cutting machine was used to sample the area to be observed of the Si 3 N 4 ring, and scanning electron microscope (SEM, HITACHI S-4800) was used to observe the micro-surface topography of the inner circle of the ring after grinding.

Results and discussion
4.1. Influence of internal grinding parameters on raceway surface quality Figures 8(a)-(d) show the effects of process parameters such as workpiece linear velocity V W , grinding wheel linear velocity V G , grinding wheel oscillation velocity V o and grinding wheel feed velocity on surface roughness  Ra. It can be seen from figure 8(a) that when the grinding wheel linear speed of 50 m s −1 , grinding wheel oscillation speed of 400 mm min −1 and radial feed speed of 10 μm min −1 remain unchanged, and the workpiece linear speed increases from 0.4 m s −1 to 2.4 m s −1 , the average surface roughness of the workpiece ranges from Ra 0.25 μm to Ra 0.27 μm. It can be seen from the trend diagram that the workpiece linear velocity has little effect on the surface roughness value of internal grinding, but increasing the workpiece linear velocity is beneficial to reducing the roughness value. In Test 7 ∼Test 12, when the linear speed of the grinding wheel increases from 35 m s −1 to 60 m s −1 , the variation trend of the surface roughness value of the workpiece is shown in figure 8(b). Because the linear speed of the grinding wheel is far greater than the linear speed of the workpiece, changing the linear speed of the grinding wheel has a greater impact on the grinding quality. It can be seen from the results that when the linear speed of the grinding wheel is gradually increased to 50 m s − 1 ∼55 m s −1 , the surface roughness of the workpiece decreases to Ra 0.22 μm. However, when the linear speed of the grinding wheel exceeds the range of 50 m s − 1 ∼55 m s −1 , the surface roughness of the workpiece increases. According to Formula (21), as the linear speed of the grinding wheel increases, the maximum undeformed cutting thickness of the grinding wheel decreases. Therefore, the actual cutting thickness of a single abrasive particle is deformed, which ultimately leads to an increase in the proportion of plastic removal in the grinding process, thus improving the surface quality of the workpiece. However, with the further increase of the linear speed of the grinding wheel, the cutting fluid sprayed into the processing area is reduced, which leads to the poor heat dissipation capacity and chip removal capacity of the grinding wheel, reduces the cutting capacity of the grinding wheel, and finally increases the surface roughness value [22]. Where: Ns-effective abrasive grains of grinding wheel;   Figure 8(c) shows the trend of roughness value changing with the grinding wheel oscillation speed. It can be seen from the figure that the roughness value decreases first and then increases. When the grinding wheel oscillation speed is around 500 mm min −1 , the surface roughness is better. If the grinding wheel oscillation speed is too slow, the grinding wheel and the ring will contact for a long time in the grinding zone, which can not be fully cooled. Moreover, the total number of grinding wheel oscillation in the same processing time is less, leading to uneven grinding of the ring.
The grinding wheel linear speed of 50 m s −1 , workpiece linear speed of 1 m s −1 and grinding wheel oscillation speed of 400 mm min −1 are kept unchanged, and the radial feed speed is gradually increased from 3 μm min −1 to 18 μm min −1 . Figure 8(d) shows the influence trend of different feed speeds on roughness values. It can be seen that the surface roughness values gradually increase with the increase of grinding wheel feed speeds. This is consistent with the conclusion of the brittle fracture model. As the feed rate increases, the grinding depth of the grinding grain increases, and the furrow depth of the grinding grain pressed into the workpiece surface increases. At the same time, the proportion of brittle fracture on the surface increases, resulting in the increase of the surface roughness value. Combined with the above results of the influence of grinding parameters on surface quality, the single abrasive grinding kinematic model constructed based on the above mentioned can analyze the influence of different process parameters on abrasive machining trajectory and surface topography.
Since the radial feed of the grinding wheel is far less than the diameter of the inner circle, the trace feed has almost no influence on the track of the grinding point. Therefore, only the influences of the workpiece linear velocity V W , grinding wheel linear velocity V G and grinding wheel oscillation velocity V o on the grinding uniformity of the inner circle of the ring are discussed. The calculation results of experiments 1 to 18 are shown in figure 9.
According to the calculation results in figure 9, it can be found that under the conditions of 10, 11 and 16 sets of process parameters in table 1, by analyzing the influence rules of V W , V G and V o on the grinding uniformity of the inner circle of the ring, the grinding point set G n obtained is more uniform. Meanwhile, by comparing the grinding test results in figure 8 with the same process parameters, it can be found that, The surface roughness values corresponding to the three groups of test parameters are all in a relatively low position.Therefore, when setting grinding parameters, the influence of workpiece linear speed, grinding wheel linear speed, and grinding wheel oscillating speed on the ground density of the inner circle of the ferrule should be considered. During precision grinding, the calculation of single abrasive motion grinding proposed above can be used to optimize the precision grinding process parameters.
It is found that the brittle removal is related to the indentation fracture and crack propagation. When the sub-surface crack is subjected to the continuous action of load, it will deepen and lengthen gradually and form fracture, thus appearing brittle removal. The load on the machined surface will affect the depth of abrasive grain cutting into the workpiece. Increasing the grinding depth will increase the proportion of brittle fracture on the surface. By sampling and observing the surface after grinding, the machining traces of inner grinding can be clearly found. Figure 10 is the SEM figure of the inner surface of the ring after grinding with diamond grinding wheel using the process parameters of Test 7, which randomly selected the above non-optimal test results. It can be seen from the figure that the processed surface is broken greatly and the surface is obviously composed of gullies left by different grinding grains. The removal forms of different gullies are different and unevenly arranged.
In addition, the 10th group of test parameters in table 1 is selected to grind the workpiece. The test results are shown in figure 11. The surface quality of the workpiece obtained from the surface reorganization test process parameters is significantly improved, the roughness value is significantly reduced, and the workpiece surface roughness value was stable between ∼Ra0.19 μm and ∼Ra0.23 μm.
According to the grinding test results, it can be seen that the surface with smaller roughness can be obtained by increasing the grinding wheel linear speed, and the grooves left by abrasive cutting are shallower and more uniform, the surface brittle removal is obviously reduced, and the ductile plastic removal is increased. As the inner ring surface of high-precision bearings, it is very important to further improve the surface quality of the inner ring. Because there are many furrows and microscopic breakage on the surface after grinding with abrasive grains, the technological process of Si 3 N 4 ceramic bearing ring still needs to be improved to obtain the ideal raceway surface.

4.2.
Influence of ultra-finishing grinding process on raceway surface quality Superfinishing process is generally divided into two stages, namely coarse superfinishing stage and superfinishing stage. Coarse timeout requires increasing the whetstone oscillation frequency f G and reducing the ring speed n W to obtain larger cutting angle and enhance the cutting ability of the whetstone. Fining timeout requires increasing the ring speed n W and reducing the whetstone oscillation frequency f G , so as to obtain better superfining surface quality with smaller cutting angle.
In addition to the whetstone oscillation width, whetstone oscillation frequency, workpiece surface diameter and ring speed, the process parameters that affect the quality of ultra-finishing are also the time and pressure of ultra-finishing. The relationship between material removal volume and process parameters described by Preston equation can be expressed by the following equation [22][23][24][25].
In the formula, k is the test condition coefficient, v is the relative velocity between the workpiece surface and the whetstone abrasive particle, p is the whetstone pressure when superfinishing, and t is the processing time.
To sum up, it can be found that the final surface quality of bearing inner ring is not only related to the motion relationship between the whetstone and the ring during superfinishing, but also closely related to the superfinishing system. According to the study of Chang S H et al [26], it was found that in ultra-finishing, process parameters would affect the surface texture, and the surface roughness would gradually decrease to a stable state with the progress of ultra-finishing, but would not always decrease.
When the inner ring is overfinished, the maximum undeformed cutting thickness hm can be transformed into equation (21). Since the cutting depth is very small when the inner ring is overfinished, the maximum undeformed cutting thickness hm is also very small, and the critical depth of brittle plastic deformation in material cutting remains unchanged, the surface quality of whepstone after overfinished grinding is much better than that of grinding wheel. Figure 12 shows the comparison between the surface topography of the inner ring after grinding with grinding wheel and the surface topography of the inner ring after super polishing with whetstone. Figure 13 shows the surface quality test results of three different machining processes for machining the inner ring of silicon nitride bearing race. The three machining processes are respectively the seventh group of test process parameters in table 1, the tenth group of process parameters in table 1, and the process of superfinishing the workpiece surface based on the test results in table 1. It can be seen from the figure that the grinding process parameters of the grinding wheel screened by the parameter optimization model (the 10th group parameter in table 1, black line) can reduce the average roughness from about Ra0.28 μm before optimization to about Ra0.21 μm after optimization, and the surface roughness of the ultra-fine grinding of the oil stone further reduces to about Ra0.05 (red line) on the basis of the grinding surface of the grinding wheel. In addition, the roughness oscillation amplitude of the optimized grinding wheel is significantly reduced. By comparing the surface micromorphology obtained by grinding and the surface micro-morphology after superfinishing, it can be found that the superfinishing of oil stone is more to remove the high point on the micro-surface and keep the gully, which is more conducive to the storage of lubricating media for bearing raceway, so as to improve the life and quality of bearing.
The above content discusses the processing of Si 3 N 4 ring by combining grinding wheel with ultra-finishing of whetstone, so that better quality Si 3 N 4 ring surface can be obtained. This paper focuses on the effects of linear speed of grinding wheel, linear speed of workpiece, oscillating speed of grinding wheel and feed speed of grinding wheel on machining accuracy of Si 3 N 4 ring. However, the factors affecting the machining accuracy in the actual production process should also include many technical parameters, such as grinding mode, grinding wheel particle size, grinding wheel binder type, grinding particle shape, machine tool stiffness, clamping form of fixture, etc. Therefore, the subsequent research work can involve more analysis of the influence of process parameters.

Conclusion
(1) The test involves a variety of changes in processing parameters. The combined process has limited ability to control the surface grinding quality of silicon nitride ceramic materials. The surface roughness value is between ∼Ra0.20 and ∼Ra 0.33 after verified by the grinding experiments of multiple grinding wheels.
(2) Combined with the established calculation results of the grinding surface topography distribution of a single abrasive grain and referring to the corresponding process parameters, the workpiece is re ground, and the surface roughness value achieved can be stably controlled between ∼Ra 0.19 and ∼Ra 0.23.
(3) Ultra-precision machining of Si 3 N 4 inner ring surface by oilstone can further improve the machining quality of Si 3 N 4 ring surface, and the average surface roughness value can reach ∼Ra 0.05 -∼Ra0.06.
(4) Through this study, it can be clear that the best process of silicon nitride ceramic ferrule internal grinding is the combination of precision grinding with grinding wheel and superfinishing with oilstone. Through the established distribution map of grinding points on the single abrasive grinding surface, the best grinding process parameters are found, so as to improve the efficiency of grinding process selection and achieve controllable surface quality machining.