A method for predicting material removal rate based on bonded abrasive grinding

By constructing a theoretical model of single abrasive cutting depth and a database of abrasive-workpiece contact pressure distribution, and combining nonlinear correction and weighted fusion methods, the problem of insufficient accuracy in predicting material removal rate during the grinding of optoelectronic functional materials with bonded abrasives in the existing technology is solved, and higher accuracy in predicting material removal rate is achieved. This method is applicable to ultra-precision machining of materials such as polycrystalline zinc sulfide, monocrystalline silicon, and sapphire.

CN123116733APending Publication Date: 2026-11-17KUNMING UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611610617.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-19
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing Preston macroscopic empirical models cannot accurately describe the nonlinear material removal behavior during the consolidation abrasive grinding process of optoelectronic functional materials. They lack a systematic description of microscopic factors such as abrasive morphology differences and the evolution of local contact states at the interface, and also lack a complete method for predicting material removal rates.

Method used

A theoretical model of single abrasive cutting depth is constructed, an abrasive-workpiece contact pressure distribution model is established, an abrasive-workpiece contact pressure distribution database is generated, a grinding pad geometry model is performed, a grinding motion process trajectory field model is established, and the abrasive geometry, spatial distribution, workpiece plastic deformation and resin matrix elastic deformation are comprehensively considered. The material removal rate is predicted by nonlinear correction and weighted fusion methods.

Benefits of technology

It improves the accuracy of material removal rate prediction, reduces the average relative error to 6.02%, and verifies that the experimental error is controlled within 11%. It is suitable for the consolidation abrasive grinding process of optoelectronic functional materials such as polycrystalline zinc sulfide, and can be extended to the ultra-precision machining of other brittle materials such as single-crystal silicon and sapphire.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116733A_ABST
    Figure CN123116733A_ABST
Patent Text Reader

Abstract

This invention provides a method for predicting material removal rate based on bonded abrasive grinding, relating to the field of ultra-precision machining technology. The method includes: constructing a theoretical model of single abrasive cutting depth and an abrasive-workpiece contact pressure distribution model; establishing an abrasive-workpiece contact pressure distribution database; geometric modeling of the grinding pad; trajectory field modeling of the grinding motion process; modeling of the material removal process; correcting the model parameters using material removal rate data measured in grinding experiments; and calculating the predicted material removal rate value during the grinding process using the corrected empirical material removal rate prediction model. This invention can dynamically simulate the evolution process of material removal on the grinding surface, achieving accurate prediction of the material removal rate. It can also be used to analyze the influence of process parameters on material removal rate and surface quality, thereby improving the prediction accuracy and process control level of the grinding process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of ultra-precision machining technology, specifically to a method for predicting material removal rate based on bonded abrasive grinding. Background Technology

[0002] Optoelectronic functional materials (such as polycrystalline zinc sulfide, monocrystalline silicon, and sapphire) have important applications in infrared optical windows, aerospace, and optoelectronic information, and the surface quality of their processing directly determines the performance of the components. Bonded abrasive grinding technology, due to its advantages of high material removal efficiency and good surface quality, has become an important means of ultra-precision machining of optoelectronic functional materials.

[0003] In the study of material removal mechanisms in grinding processes, the Preston equation is widely used as a classic macroscopic empirical model. This equation indicates that under stable processing conditions, the material removal rate is directly proportional to the normal contact pressure and relative sliding velocity on the workpiece. Due to its simple form and clear physical meaning, it has been used for a long time and continues to be developed and improved in the fields of grinding, polishing and ultra-precision machining.

[0004] However, this model has significant limitations: First, the Preston macroscopic empirical model assumes a linear relationship between material removal rate and contact pressure and relative velocity, simplifying the abrasive action and failing to consider microscopic factors such as abrasive morphology differences and the evolution of local contact states at the interface. Therefore, it is difficult to characterize the nonlinear material removal behavior in the actual grinding process. Second, the material removal model lacks a systematic description of the high randomness of abrasive exposure, the elastic yielding effect of the resin matrix, and the spatial distribution characteristics of abrasive particles. Existing methods separate the abrasive contact mechanical behavior from kinematic analysis, failing to achieve dynamic evolution simulation of material removal under multi-factor coupling. Furthermore, for the consolidation abrasive grinding process of optoelectronic functional materials, there is still a lack of a complete material removal rate prediction method that integrates theoretical modeling, numerical simulation, and experimental verification from the microscopic action of a single abrasive particle to the macroscopic material removal characterization. Summary of the Invention

[0005] To address the technical problems existing in the background art, this invention provides a method for predicting material removal rate based on bonded abrasive grinding. The specific steps are as follows:

[0006] S1. Construct a theoretical model of single abrasive cutting depth and establish a model of abrasive-workpiece contact pressure distribution;

[0007] S2. Based on the abrasive-workpiece contact pressure distribution model, establish an abrasive-workpiece contact pressure distribution database;

[0008] S3. Perform geometric modeling on the grinding pad to obtain the surface geometric features of the grinding pad and the spatial distribution information of the abrasive.

[0009] S4. Establish a trajectory field model of the grinding motion process based on the motion parameters of the grinding process;

[0010] S5. By integrating the single abrasive cutting depth, abrasive-workpiece contact pressure, grinding pad geometry, abrasive distribution characteristics, and motion trajectory field, a material removal process model is established.

[0011] S6. Correct the model parameters by measuring the material removal rate data using grinding experiments;

[0012] S7. Calculate the predicted value of material removal rate during the grinding process using the modified empirical material removal rate prediction model.

[0013] Furthermore, in step S1, the theoretical model for the single abrasive cutting depth is as follows:

[0014]

[0015] In the formula, δ(y) ave Let P0 be the desired depth of cut, R0 be the grinding pressure, D be the workpiece radius, D be the volumetric equivalent particle size of the abrasive, and E(A) be the contact projection area formed by a single abrasive particle on the surface of the grinding pad. H represents the volume fraction of abrasive in the grinding pad, γ represents the average contact ratio between the surface area of ​​the grinding pad protrusions and the surface area of ​​the workpiece, and H represents the volume fraction of abrasive in the grinding pad. M Where E represents the Mayer hardness of the workpiece material, E represents the elastic modulus of the resin matrix of the abrasive pad, and n p ≈1.2 represents the relative contact depth coefficient under fully plastic deformation conditions, used to characterize the relationship between indentation geometry and contact state, where θ is the angle between the conical abrasive and the grinding pad plane.

[0016] Furthermore, in step S1, the abrasive-workpiece contact pressure distribution model is established through ANSYS simulation, and the specific steps are as follows:

[0017] Select the abrasive shape and set the tilt angle of the abrasive, construct the contact boundary conditions and local stress distribution characteristics, and set the parameters for step-by-step loading solution control.

[0018] Furthermore, the steps for establishing the abrasive-workpiece contact pressure distribution database in step S2 are as follows:

[0019] MATLAB was used to read nodal pressure data under various working conditions, and the discrete nodal pressure data under different pressure conditions were uniformly normalized to reduce the number of nodes to be extracted for subsequent material removal pressure. The maximum radius of the contact area was calculated based on the effective contact nodes, and a linear interpolation algorithm was used to map the discrete nodal pressure onto a regular two-dimensional mesh. The pressure field interpolation is expressed as:

[0020]

[0021] in, This represents a two-dimensional interpolation operation based on a linear interpolation algorithm. Let be the pressure at a node under the k-th pressure distribution. Let x and y be the coordinates of node i under the k-th pressure distribution. The pressure of discrete points after two-dimensional interpolation is mapped to the x and y coordinates in a regular two-dimensional grid.

[0022] The null value regions generated by interpolation are assigned zeros while retaining the data within the contact circle domain to obtain a physically valid two-dimensional pressure matrix. A standardized single abrasive pressure field template is then constructed and stored in the abrasive-workpiece contact pressure distribution database.

[0023] Furthermore, the geometric modeling step of the abrasive pad described in step S3 includes:

[0024] The spiral groove structure of the grinding pad is generated based on the involute equation. A groove band with a finite width is constructed by normal offset, and a periodic spiral structure is formed by circumferential array.

[0025] Abrasive points are randomly generated within the effective area of ​​the grinding pad and a minimum spacing constraint is applied. Based on the structural characteristics of the protrusions and grooves of the grinding pad, abrasive points located in the groove area are eliminated, so that the abrasive is only distributed in the effective processing protrusion area. The abrasive is displayed in the form of scattered points in the grinding pad coordinate system to complete the initialization of the abrasive field. A trajectory storage structure is assigned to each abrasive to record its motion path during the simulation process.

[0026] Furthermore, the steps for establishing the trajectory field model of the grinding motion process described in step S4 are as follows:

[0027] Establish an absolute coordinate system with the center of the grinding pad as the origin and a workpiece coordinate system with the center of the workpiece as the origin. The absolute coordinate system {XOY} has the center of the grinding pad as the center, and the X and Y axes are fixed and do not move with the grinding pad. c Let the center of the carrier be e1, and the eccentricity between the center of the grinding pad and the center of the workpiece coordinate system be e1; w O w Y w}With workpiece center O w Centered on the circle, the X-axis direction is parallel to the vector. Perpendicular to the vector, the Y-axis direction is perpendicular to the vector. The grinding pads are collinear, with an eccentricity of e2 relative to the center of the carrier. The grinding pad rotates about its center with an angular velocity ω1, and the carrier rotates about its center with an angular velocity ω2. At time t, the initial position of a certain abrasive grain p on the grinding pad in the absolute coordinate system is... ;

[0028] The motion trajectory expression of any abrasive in the workpiece coordinate system is obtained as follows:

[0029]

[0030] Apply a motion expression of a single abrasive grain in the workpiece coordinate system to each abrasive grain, and discretize the time to obtain its instantaneous position at any time. When the abrasive grain passes through the workpiece surface area, statistically analyze its corresponding two-dimensional coordinates, and connect the coordinate points at consecutive times to form the abrasive grain motion trajectory, thereby obtaining the motion trajectory field of the abrasive grain actually participating in the processing on the workpiece surface.

[0031] Furthermore, the steps for establishing the material removal process model described in step S5 are as follows:

[0032] The workpiece surface is discretized into a two-dimensional regular mesh. A pressure template is randomly assigned to each abrasive grain from the abrasive-workpiece contact pressure distribution database. When the abrasive center enters the effective area of ​​the workpiece, the single-step material removal amount of a single abrasive grain on a single mesh node within the current time step Δt is calculated based on the Preston equation:

[0033]

[0034] In the formula, g is the Preston coefficient. This represents the maximum pressure value across all abrasive-workpiece contact pressure distributions. For the normalized first The pressure at a node under a certain pressure distribution. The relative velocity between the abrasive and the workpiece at the point of contact;

[0035] After sampling the material removal amount for all mesh nodes, it is mapped to the corresponding mesh cells to update the height field:

[0036]

[0037] in, for The amount of material removed after a certain time. Let n be the amount of material removed at time t. k The number of discrete points on the workpiece surface;

[0038] Based on the updated workpiece height matrix The calculable material removal rate is:

[0039]

[0040] Where MRR is the material removal rate. H represents the simulation time, H is the updated workpiece height matrix, and H0 is the initial workpiece height matrix. This represents the average height of material removed from the workpiece surface.

[0041] Furthermore, the model parameter correction step described in step S6 is as follows:

[0042] Orthogonal grinding experiments were designed and implemented to obtain experimental values ​​of material removal rate and surface roughness data under different combinations of process parameters. Based on the orthogonal experimental data, range analysis was used to determine the primary and secondary relationships of the influence of each process parameter on material removal rate and surface quality. Based on the material removal rate data measured by the orthogonal experiments and the corresponding simulation input parameters, the Preston coefficients were calibrated by linear fitting. Pressure and velocity indices were introduced into the traditional linear Preston model to construct a nonlinear modified material removal rate prediction model. With the goal of minimizing the average relative error, different combinations of pressure and velocity indices were traversed within a set range, and the combination of indices that minimizes the average relative error between the simulation prediction value and the experimental measurement value was selected to complete the model correction and obtain the corrected empirical material removal rate prediction model.

[0043] Furthermore, after completing the nonlinear correction, the model prediction results are post-processed and calibrated using a weighted fusion method. The specific steps are as follows:

[0044] A weighted function is constructed that integrates simulation predictions and experimental measurements. The contribution ratio of simulation values ​​and experimental values ​​to the final prediction result is adjusted by weighting coefficients. The optimal weighting coefficients are determined with the goal of minimizing the average relative error under all experimental conditions. The optimal weighting coefficients are substituted into the weighted function to calculate the final corrected material removal rate prediction model.

[0045] Furthermore, the calculation steps for the predicted material removal rate in the grinding process described in step S7 are as follows:

[0046] Using a single-factor variable control method, while keeping other process parameters constant, the grinding pressure, grinding pad speed, and workpiece disk speed were changed respectively. Numerical simulation was performed using a modified empirical material removal rate prediction model to calculate the material removal rate and workpiece surface morphology evolution results under different parameter levels.

[0047] Based on the simulation results, the mapping relationships between grinding pressure and material removal rate, and between grinding speed and material removal rate were established respectively, and the influence trends of each process parameter on the micro-undulations and uniformity of the workpiece surface were analyzed.

[0048] The beneficial effects of this invention are:

[0049] This invention establishes a theoretical model for the cutting depth of a single abrasive grain based on the microscopic action mechanism of a single abrasive grain. It comprehensively considers the interface effects such as the geometric morphology of the abrasive grain, its spatial statistical distribution, the random distribution of its exposed height, the plastic deformation of the workpiece, and the elastic deformation of the resin matrix. This model overcomes the shortcomings of the traditional Preston macroscopic empirical model, which cannot reveal the physical nature of microscopic material removal. It can more accurately describe the microscopic contact state of the interface and improve the accuracy of predicting material removal behavior.

[0050] This invention selects five typical abrasive shapes: cone, tetrahedron, pentahedron, dodecahedron, and frustum. It sets a 15° tilt angle for each type of abrasive and establishes a pressure field template database under various typical working conditions through finite element numerical simulation. Compared with the modeling method of using only ideal spheres or single regular geometric shapes in the prior art, this invention more realistically reflects the influence of the irregular polyhedral shape of the abrasive on the contact pressure distribution in actual processing.

[0051] To address the limitations of existing material removal models in systematically characterizing the randomness of abrasive grain exposure height, the elastic yielding effect of the resin matrix, and the spatial distribution characteristics of abrasive grains, this invention constructs a material removal model that couples abrasive grain-workpiece contact characteristics, abrasive grain geometry, random distribution patterns, and kinematic trajectories. By employing an average height constraint mechanism to suppress numerical overcutting caused by the superposition of multiple abrasive grains, it achieves dynamic simulation of the evolution of material removal on the grinding surface, overcoming the inherent deficiency of existing methods in decoupling the analysis of abrasive grain contact mechanics and kinematics.

[0052] This invention introduces a pressure exponent (optimized value u=1.3) and a velocity exponent (optimized value z=1.1) to the traditional linear Preston model for nonlinear correction, and further employs a weighted fusion method (optimal weight coefficient λ=0.6) to post-process and calibrate the model prediction results, thereby reducing the average relative error of the material removal rate prediction to about 6.02%. The verification experiment error is controlled within 11%, and the prediction accuracy is significantly improved.

[0053] This invention establishes an integrated method system for predicting material removal rate, encompassing single-abrasive theoretical modeling, finite element pressure template database construction, grinding pad geometry and abrasive distribution modeling, motion trajectory field simulation, multi-factor coupled material removal modeling, and nonlinear correction and experimental calibration. This method is not only applicable to the bonded abrasive grinding process of optoelectronic functional materials such as polycrystalline zinc sulfide, but can also be extended to the ultra-precision machining of other brittle materials such as single-crystal silicon and sapphire, demonstrating promising engineering application prospects and general applicability. Attached Figure Description

[0054] Figure 1 This is a flowchart of the steps of the present invention;

[0055] Figure 2 This is a diagram showing the abrasive geometry of the present invention;

[0056] Figure 3 This is a schematic diagram showing the relationship between the abrasive protrusion height and the workpiece contact height according to the present invention;

[0057] Figure 4 This is a schematic diagram of the single abrasive pressure distribution template of the present invention;

[0058] Figure 5 This is a diagram showing the geometry of the abrasive pad and the abrasive distribution configuration of the present invention;

[0059] Figure 6 This is a schematic diagram of the abrasive trajectory movement according to the present invention;

[0060] Figure 7 This is a field diagram of the abrasive motion trajectory on the workpiece surface according to the present invention;

[0061] Figure 8 This is a diagram illustrating the material removal process of the present invention;

[0062] Figure 9 This is a schematic diagram of the average height constraint mechanism of the present invention;

[0063] Figure 10 This is a graph showing the average relative error results of the pressure index and velocity index model of the present invention;

[0064] Figure 11 This is a comparison chart of the material removal rates in the grinding experiment and simulation of the zinc sulfide bonded abrasive of the present invention. Detailed Implementation

[0065] The present invention will be further described in detail below with reference to specific embodiments.

[0066] like Figure 1 As shown, a method for predicting material removal rate based on bonded abrasive grinding includes the following steps:

[0067] S1. Construct a theoretical model of single abrasive cutting depth and establish a model of abrasive-workpiece contact pressure distribution;

[0068] To construct a theoretical model of single abrasive cutting depth, the macroscopic Preston equation is first introduced as the basic theory. Then, the theoretical modeling is developed from the microscopic action of single abrasive to the macroscopic material removal characterization, providing a theoretical basis for subsequent numerical simulation and process parameter analysis.

[0069] According to the Preston equation, the material removal rate can be expressed as:

[0070]

[0071] Where g is the Preston coefficient, whose value comprehensively reflects the influence of various factors such as the characteristics of the polishing fluid, the geometry and mechanical properties of the abrasive, the stiffness of the polishing pad, and the material properties of the workpiece; P is the pressure applied to the workpiece; and V is the relative velocity between the wafer and the polishing pad.

[0072] The theoretical model for single abrasive cutting depth and the abrasive-workpiece contact pressure distribution model employ a cylindrical workpiece model made of optoelectronic functional material and a single ideal conical abrasive grain model with a semi-apex angle of 30°. To reduce the complexity of theoretical modeling, irregular diamond abrasive grains are equated to ideal cones with defined geometric parameters. It is assumed that the diamond abrasive grains are statistically uniformly distributed within the raised areas of the resin matrix surface, and the number of effective abrasive grains is theoretically estimated using probabilistic statistical methods. It is also assumed that the exposure height of all effective cutting abrasive grains follows a uniform distribution within their working range. Furthermore, it is assumed that the workpiece material undergoes only ideal plastic deformation under the action of the abrasive grains, neglecting elastic recovery after unloading. Finally, it is assumed that the compression process of the abrasive grains on the resin matrix is ​​a completely elastic deformation. Finally, it is assumed that the diamond abrasive grains are ideal rigid bodies that do not wear or detach.

[0073] like Figure 2 As shown, based on the geometric assumption of a conical abrasive, the expression for the actual contact cross-sectional area of ​​the abrasive at the protrusion of the grinding pad can be obtained as follows:

[0074]

[0075] Where A(y) is the actual contact cross-sectional area of ​​the abrasive at the protrusion of the grinding pad, y is the exposed height of the abrasive, and θ is the angle between the conical abrasive and the plane of the grinding pad; Figure 2 In this context, l represents the total height of the abrasive, and a represents the radius of the cross-sectional circle corresponding to the maximum height of the abrasive.

[0076] To reasonably describe the distribution characteristics of the abrasive exposure height on the grinding pad surface in theoretical modeling, it is assumed that the diamond abrasive on the grinding pad surface has a uniform exposure height distribution. Under this assumption, the contact projection area formed by a single abrasive grain on the grinding pad surface is no longer a fixed value, but a random variable that varies with the exposure height.

[0077]

[0078] Where E(A) is the contact projection area formed by a single abrasive particle on the surface of the grinding pad, and f(y)=1 / l is the probability density function of the abrasive particle exposure height y.

[0079] The volume fraction is used to characterize the spatial proportion of abrasive in the grinding pad and is further used to describe the effective distribution characteristics of abrasive on the surface of the grinding pad. The volume fraction can be expressed as:

[0080]

[0081] in, M represents the volume fraction of abrasive in the grinding pad. a and M r ρ represents the mass of abrasive and resin added during the preparation of the grinding pad, respectively. a and ρ r These are the densities of the abrasive and the resin, respectively.

[0082] Based on the random distribution characteristics of abrasive particles and combined with the average contact state between the raised areas of the grinding pad and the workpiece surface, the expression for the effective number of abrasive particles participating in the actual grinding process can be obtained as follows:

[0083]

[0084] Where N0 is the effective number of abrasive particles participating in the actual grinding, κ is the maximum exposed height coefficient, usually taken as 1.5, and s is the gap between the grinding pad protrusion and the workpiece surface (e.g., ...). Figure 3 As shown in the figure, R is the equivalent volume radius of the abrasive, D is the equivalent volume particle size of the abrasive, R0 is the workpiece radius, and γ represents the average contact ratio between the raised area of ​​the grinding pad and the surface area of ​​the workpiece, which is used to reflect the proportion of the actual contact area in the overall grinding area.

[0085] Based on a mechanical model of complete plastic deformation caused by the indentation of an ideal conical abrasive into the workpiece surface, a normal load is established when the abrasive is embedded in the workpiece surface. Relationship;

[0086]

[0087] in, The indentation loading coefficient of the conical abrasive is given. The normal load when the abrasive grains embed into the workpiece surface is δ(y), where δ(y) is the actual cutting depth of the abrasive, and n is the normal load. p ≈1.2 represents the relative contact depth coefficient under fully plastic deformation conditions, used to characterize the relationship between indentation geometry and contact state. H M The value is the Mayer hardness of the workpiece material.

[0088] Based on the contact mechanics model of a conical indenter pressing into an elastic substrate, a relationship is established between the normal load of the abrasive elastically yielding to the resin matrix and the elastic indentation depth.

[0089]

[0090] Among them, F matrix The normal load of the conical abrasive acting on the resin matrix is ​​E, which represents the elastic modulus of the resin matrix of the abrasive pad. tIt is the elastic indentation depth of the abrasive into the grinding pad under external load, used to characterize the elastic yielding degree of the resin matrix.

[0091] Under the first-order approximation, the following relationship exists between the D-type Shore hardness (cone indenter) and the elastic modulus of the tested material:

[0092]

[0093] Among them, S D The D-type Shore hardness of the material.

[0094] According to Newton's second law, the load on the abrasive in the normal direction satisfies the equilibrium relationship:

[0095]

[0096] in, F is the normal load when abrasive grains embed into the workpiece surface. matrix This refers to the normal load exerted by the conical abrasive on the resin matrix.

[0097] The elastic indentation depth of the abrasive into the grinding pad under external load can be obtained.

[0098]

[0099] The depth of cut of a workpiece by a conical abrasive can be expressed as:

[0100]

[0101] During the grinding process, not all abrasive particles embedded in the grinding pad surface can effectively participate in material removal. Only when the abrasive particle's exposed height is within the range of [s, κR] can its cutting edge contact and penetrate the workpiece surface, thus resulting in actual material removal. Therefore, the probability density of the actual effective exposed height for material removal by a single abrasive particle is:

[0102]

[0103] Therefore, the expected normal pressure exerted on the workpiece by a single abrasive grain can be obtained by integrating its load with the probability density function over the effective exposure height range:

[0104]

[0105] During the grinding process, the total load applied to the workpiece can be expressed as the product of the effective number of abrasive particles involved in the processing and the expected normal pressure of each abrasive particle:

[0106]

[0107] Where P0 is the grinding pressure, The expected normal pressure exerted on the workpiece by a single abrasive grain, δ(y). ave This represents the desired depth of cut.

[0108] The distance between the grinding pad protrusions and the workpiece under a given grinding pressure is obtained as follows:

[0109]

[0110] Based on the desired depth of cut of the abrasive:

[0111]

[0112] The above formula allows for the statistical calculation of the average expected depth of cut of a single abrasive grain under a given process pressure. This average expected depth of cut comprehensively considers the cutting mechanical properties of a single abrasive grain, the probability distribution of the effective exposed height, and the elastic rebound effect of the grinding pad, providing a direct basis for the theoretical calculation of material removal. In subsequent simulations and experimental verifications of abrasive material removal, the material properties and geometric parameters of the grinding pad and workpiece need to be used; their specific values ​​are listed in Table 1.

[0113] Table 1 Grinding Experiment Parameters

[0114]

[0115] In the abrasive-workpiece contact pressure distribution model, a three-dimensional indentation contact model of abrasive-workpiece is established through ANSYS simulation. Various typical abrasive shapes are selected, and tilt angles are set for various abrasives to construct different contact boundary conditions and local stress distribution characteristics.

[0116] The processing materials and abrasives are assigned corresponding material parameters. The physical properties of the processing materials are set as shown in Table 2.

[0117] Table 2 Physical parameters of processed materials

[0118]

[0119] The model conditions are set as follows: a workpiece model in the form of a cube with a side length of 750 nm, the contact body is set as the workpiece surface, the target body is set as the abrasive surface that will come into contact with the workpiece during the pressing process, the contact type is set as Frictionless, and the bottom of the workpiece is set as a fixed support.

[0120] As orthogonal experiments were subsequently conducted to verify the material removal rate, the theoretical cutting depth of a single abrasive grain in the orthogonal experiments is shown in Table 3.

[0121] Table 3 Theoretical Depth of Cut for Single Abrasive Particle in Orthogonal Experiments

[0122]

[0123] The step control settings for the step loading solution are as follows: 1 loading step with a total time of 1 second, used to smoothly load the abrasive indentation displacement from 0 to the target cutting depth. To improve the convergence of the nonlinear contact solution, automatic time step is enabled, with both the initial and minimum step sizes set to 0.2 seconds to ensure a stable and uniform indentation process.

[0124] S2. Based on the abrasive-workpiece contact pressure distribution model, establish an abrasive-workpiece contact pressure distribution database to provide contact characteristic data for subsequent material removal modeling;

[0125] After completing the finite element solution and exporting the data, the nodal pressure data under various working conditions were read and processed using the MATLAB program.

[0126] The node pressure data is normalized according to the following formula to eliminate amplitude differences under different operating conditions.

[0127]

[0128] in, Let p be the pressure at a node under the k-th pressure distribution. max With p min These represent the maximum and minimum pressure values ​​in all abrasive-workpiece contact pressure distributions. It represents the pressure at a node under the normalized k-th pressure distribution.

[0129] Subsequently, a pressure threshold η was introduced. c After removing low-voltage noise nodes at the edges and obtaining only the effective contact nodes, the maximum radius of the contact area is calculated based on these nodes, and its expression is:

[0130]

[0131] in, Let x and y be the coordinates of node i under the k-th pressure distribution. Let x and y be the coordinates of the pressure center under the k-th pressure distribution, respectively. Since the pressure center has already been set at the center of the circle in the previous modeling, ... η c To introduce a pressure threshold.

[0132] The maximum radius of the contact area is calculated based on the effective contact nodes, and the discrete node pressure is mapped to a regular two-dimensional mesh using a linear interpolation algorithm. The pressure field interpolation is then expressed as:

[0133]

[0134] Where Interp represents a two-dimensional interpolation operation based on a linear interpolation algorithm. This maps the discrete point pressure to the x and y coordinates in a regular two-dimensional grid after two-dimensional interpolation.

[0135] The null value regions generated by interpolation are assigned zeros while retaining the data within the contact circle domain. Finally, a physically valid two-dimensional pressure matrix is ​​obtained, and a standardized single abrasive pressure field template is constructed and stored in the abrasive-workpiece contact pressure distribution database. Figure 4 This study demonstrates the two-dimensional pressure distribution of a single abrasive grain and its corresponding effective contact radius after normalization and interpolation under a grinding force of 10 kg. Through similar processing, this research team developed 40 pressure templates under four different grinding pressures. These templates not only reflect the contact pressure characteristics under different abrasive conditions but also provide repeatable and standardized input data for subsequent single-abrasive material removal calculations.

[0136] S3. Perform geometric modeling on the grinding pad to obtain the surface geometric features of the grinding pad and the spatial distribution information of the abrasive.

[0137] The spiral groove structure of the grinding pad is generated based on the involute equation. A groove band with a finite width is constructed by normal offset, and a periodic spiral structure is formed by circumferential array.

[0138] Abrasive points are randomly generated within the effective area of ​​the grinding pad, and minimum spacing constraints are applied to avoid excessive overlap. Based on the structural characteristics of the protrusions and grooves of the grinding pad, abrasive points located in the groove area are removed to ensure that the abrasive is only distributed in the effective processing protrusion area, making the simulation model conform to physical reality. After screening, the abrasive is displayed as scattered points in the grinding pad coordinate system to complete the abrasive field initialization (e.g., ...). Figure 5 As shown in the figure, a trajectory storage structure is assigned to each abrasive to record its motion path during the simulation process.

[0139] S4. Based on the motion parameters of the grinding process, establish a trajectory field model of the grinding motion process, and calculate the motion trajectory and coverage characteristics of the abrasive relative to the workpiece surface.

[0140] To establish the kinematic trajectory of a single abrasive particle on the workpiece surface, an absolute coordinate system with the center of the grinding pad as the origin and a workpiece coordinate system with the center of the workpiece as the origin were established. The absolute coordinate system {XOY} has the center of the grinding pad as the center, and the X and Y axes are fixed and do not move with the grinding pad. c Let the center of the carrier be e1, and the eccentricity between the center of the grinding pad and the center of the workpiece coordinate system be e1; w O w Y w}With workpiece center O w Centered on the circle, the X-axis direction is parallel to the vector. Perpendicular to the vector, the Y-axis direction is perpendicular to the vector. The grinding pads are collinear, with an eccentricity of e2 relative to the center of the vehicle. Assume the grinding pads rotate about its center with an angular velocity ω1, and the vehicle rotates about its center with an angular velocity ω2.

[0141] Based on the rigid body planar motion description method in robot kinematics, at time t, let the position of the center of a certain abrasive grain on the grinding pad in the absolute coordinate system be... Its initial position is The position of the abrasive grain under the rotation of the grinding pad can be represented as:

[0142]

[0143] in, , Let x and y be the x and y coordinates of the abrasive grain center in the absolute coordinate system at time t, respectively. , These are the x and y coordinates of the abrasive grain center at its initial position in the absolute coordinate system.

[0144] Assuming that at t=0, the center of the workpiece is located at the rightmost end of the circular trajectory of the carrier, then according to geometric relationships, the absolute coordinates of the workpiece center at any time t are:

[0145]

[0146] in, , Let x and y be the x and y coordinates of the workpiece center in the absolute coordinate system at time t, respectively.

[0147] Based on the coordinate transformation relationship, the vector of the abrasive grain position p in the workpiece coordinate system can be expressed as:

[0148]

[0149] The coordinates of the workpiece center in the absolute coordinate system can be obtained as follows: The coordinates of the abrasive center in the absolute coordinate system can be obtained.

[0150]

[0151] By using the coordinate transformation relationship between the absolute coordinate system and the workpiece coordinate system, the coordinates of the abrasive center in the workpiece coordinate system {X} at any time t can be obtained. w O w Y w Positional expressions in}

[0152]

[0153] in, This represents the rotation transformation matrix of the workpiece coordinate system relative to the absolute coordinate system, used to describe the change in the coordinate system's attitude caused by the rotation of the carrier. During the grinding process, the workpiece rotates with the carrier around its center at an angular velocity ω2. Therefore, the workpiece coordinate system has a time-varying rotation relationship with the absolute coordinate system, and its rotation matrix can be expressed as:

[0154]

[0155] The position vector of the abrasive in the absolute coordinate system can be obtained using the rotation matrix described above. Transform to the workpiece coordinate system. Substitute this into the coordinate transformation relationship described above to obtain the motion trajectory expression of any abrasive in the workpiece coordinate system:

[0156]

[0157] By applying a motion expression for each abrasive grain in the workpiece coordinate system and discretizing the time, its instantaneous position at any given moment can be obtained. When the time step is Δt, the abrasive grain will move from the position pi(t) to the position pi(t+Δt) (e.g., ...). Figure 6 (as shown)

[0158] When the abrasive enters the effective area of ​​the workpiece, the condition is met. (in , These are the x and y coordinates of the abrasive in the workpiece coordinate system at the working moment, R. workpiece If the abrasive is considered to have entered the workpiece surface area (where the radius is equal to the workpiece radius), then the instantaneous position of the abrasive in the workpiece coordinate system at that moment is recorded, and the trajectory storage structure is updated; otherwise, if the abrasive is outside the effective area of ​​the workpiece, trajectory recording is not performed.

[0159] As the abrasive passes over the workpiece surface, its corresponding two-dimensional coordinates are statistically analyzed, and the coordinate points at consecutive moments are connected to form the abrasive's motion trajectory. Through the above screening and recording process, the actual motion trajectory field of the abrasive participating in the processing on the workpiece surface can be obtained, realizing the simulation of the abrasive trajectory under the coupled action of the grinding pad's rotation and the workpiece's revolution. Based on the simulation results, the overall motion trajectory field of the abrasive on the workpiece surface can be further obtained (e.g., ...). Figure 7 (as shown)

[0160] S5. By integrating the single abrasive cutting depth, abrasive-workpiece contact pressure, grinding pad geometry, abrasive distribution characteristics, and motion trajectory field, a material removal process model is established to dynamically simulate the evolution of material removal from the workpiece surface during the grinding process.

[0161] Based on obtaining the abrasive's trajectory, further analysis of the velocity at the contact point is needed. During the contact process between a single abrasive grain and the workpiece surface, the relative velocity at the contact point mainly consists of two parts: one part is the linear velocity generated by the abrasive rotating with the grinding pad, and the other part is the linear velocity generated by the workpiece rotating with the carrier. Since the abrasive undergoes rigid rotational motion with the grinding pad throughout the grinding process, and the workpiece also continuously rotates around its own center, the relative motion state between the abrasive and the workpiece changes continuously over time. To accurately describe this motion characteristic, velocity expressions need to be established for the corresponding points on the abrasive surface and the workpiece surface, respectively.

[0162] According to the rotational motion relationship of a rigid body, the abrasive on the grinding pad will generate a linear velocity during rotation. For any point b(x) on the surface of the grinding pad... b ,y b Its velocity is determined by the rotational motion of the grinding pad around the center, therefore the linear velocity of this point at time t can be expressed as:

[0163]

[0164] in, Let be the linear velocity of point b on the surface of the grinding pad in the x-direction. Let be the linear velocity in the y-direction of point b on the surface of the grinding pad.

[0165] During the grinding process, the workpiece rotates around its center along with the carrier. Within the contact area, the rotation of the workpiece causes corresponding linear velocities at various points on its surface. For any point on the workpiece surface, its linear velocity can be expressed as:

[0166]

[0167] in, The linear velocity in the x-direction at any point on the surface of the workpiece. The linear velocity in the y-direction at any point on the surface of the workpiece.

[0168] Therefore, the relative velocity between the abrasive and the workpiece at the contact point can be expressed as the difference between their velocities, i.e.:

[0169]

[0170] in, Let be the relative velocity of the abrasive and the workpiece in the x-direction at the point of contact. Let be the relative velocity in the y-direction between the abrasive and the workpiece at the point of contact.

[0171] Based on the component relationships of the velocity vector, the magnitude of the relative velocity is:

[0172]

[0173] To achieve spatial distribution calculation of material removal, the workpiece surface is discretized into a two-dimensional regular grid. The height matrix corresponding to each grid node will be used as the object for subsequent material removal updates. The initial height is uniformly set to a constant H0 (the initial height matrix of the workpiece). The effective area of ​​the workpiece satisfies:

[0174]

[0175] in, , These are the x and y coordinates of the nodes in the workpiece coordinate system within the workpiece radius.

[0176] Outside the designated area, the value is set to invalid. When the time step is Δt, the amount of material removed in a single step is:

[0177]

[0178] in, This refers to the amount of material removed in a single step. Δt represents the relative velocity between the abrasive and the workpiece at the contact point, and Δt represents the time step.

[0179] To convert the template pressure into actual pressure, the abrasive-workpiece contact pressure is introduced:

[0180]

[0181] Where F is the grinding force, and A = πR 2 workpiece This represents the contact area of ​​the workpiece.

[0182] In the numerical implementation, the local pressure is obtained by finite element template interpolation:

[0183]

[0184] in, Obtained by finite element template interpolation.

[0185] The final amount of material removed at a single sampling point is expressed as follows:

[0186]

[0187] in, This represents the amount of material removed at the final single sampling point.

[0188] After sampling the material removal amount for all mesh nodes, it is mapped to the corresponding mesh cells to update the height field:

[0189]

[0190] in, for The amount of material removed after a certain time. for Material removal amount at any given time This represents the number of discrete points on the workpiece surface.

[0191] The material removal process is as follows: Figure 8 As shown, within each time step, the program randomly assigns a pressure template to each abrasive grain. Once the abrasive grain's center enters the effective area of ​​the workpiece, local pressure is applied according to the template distribution, and the material removal amount at each sampling point is calculated using the Preston equation. Subsequently, the removal contributions of all abrasive grains to all mesh nodes within the current time step are accumulated, and the workpiece surface height field is updated synchronously. This process iterates continuously along the abrasive grain's trajectory until the full-process simulation is completed.

[0192] In actual machining processes, due to the mutual shielding of abrasive particles and the local smoothness of contact pressure distribution, overcutting is usually naturally limited. To improve the uniformity of material removal and enhance the reliability of simulation predictions, an average height constraint mechanism (such as...) is introduced. Figure 9 As shown): When the height of a certain grid point is lower than a certain threshold of the current workpiece's average height, the condition is met. At that time, (among which) The average height of the workpiece surface. Pause material removal updates at this point (to maximize the allowable removal height);

[0193] Finally, based on the updated workpiece height matrix H, the material removal rate can be calculated:

[0194]

[0195] Where MRR is the material removal rate, T is the simulation time, H is the updated workpiece height matrix, and H0 is the initial workpiece height matrix. The average height of material removed from the workpiece surface;

[0196] At this point, the numerical simulation model for material removal has been established.

[0197] S6. Measure the material removal rate data through grinding experiments, correct the parameters in the material removal process model, and obtain the corrected empirical material removal rate prediction model.

[0198] The experiment used zinc sulfide as the workpiece material, a self-made involute groove bonded abrasive pad with a diameter of 300 mm, diamond (particle size 3 μm, concentration 59%) as the abrasive, and deionized water as the grinding fluid with a constant flow rate of 30 mL / min.

[0199] Table 4 shows the fixed experimental conditions;

[0200] Table 4 Fixed Experimental Conditions

[0201]

[0202] We designed and implemented a multi-factor, multi-level orthogonal grinding experiment to obtain experimental values ​​of material removal rate and surface roughness data under different combinations of process parameters. Based on the orthogonal experimental data and the magnitude of K values ​​at each level in the range analysis, we can obtain the optimal parameter combination corresponding to different performance indicators.

[0203] The orthogonal array L16(4³) is designed as shown in Table 5, where factor A represents the workpiece eccentricity, factor B represents the grinding force, and factor C represents the grinding pad rotation speed.

[0204] Table 5 Orthogonal Table for Grinding Experiments with Zinc Sulfide Bonded Abrasives

[0205]

[0206] Based on the orthogonal experimental design shown in Table 5, the surface of the zinc sulfide workpiece for each experimental case was ground, and the MRR, Sa, and Sq indices were measured. The experimental results are shown in Table 6.

[0207] Table 6. Results of orthogonal experiments on grinding with zinc sulfide bonded abrasives

[0208]

[0209] The data obtained in Table 6 were analyzed using the range method, and K was denoted as K. jm Let Rj be the sum of experimental indicators corresponding to level m of factor j, and let Rj be the range of factor j, which is the difference between the maximum and minimum values ​​of the average indicator at each level of the factor. See Tables 7, 8 and 9 for specific data.

[0210] Table 7 MRR Range Analysis of Zinc Sulfide Grinding

[0211]

[0212] Table 8. Range Analysis of Zinc Sulfide Grinding (Sa)

[0213]

[0214] Table 9. Range Analysis of Zinc Sulfide Grinding (Sq)

[0215]

[0216] Based on the magnitude of the K values ​​at each level in the range analysis, the optimal parameter combinations for different performance indices can be obtained. For the material removal rate, the combination corresponding to the maximum value is A4B4C1, indicating that higher material removal efficiency can be obtained under conditions of larger workpiece eccentricity, higher grinding pressure, and higher grinding pad rotation speed. For the surface roughness index Sa, its optimal combination is A1B1C4; the optimal combination for Sq is also A1B1C4. The results show that lower grinding pressure and higher grinding pad rotation speed are more conducive to obtaining better surface quality.

[0217] Based on the material removal rate data obtained from orthogonal experiments, the Preston formula was used to fit the experimental results, thereby determining the Preston coefficient g for the workpiece grinding process. The Preston coefficient g was obtained by linearly fitting the experimental data to the corresponding PV, with units of nm·s / (Pa·m·min).

[0218] To improve the prediction accuracy of the model, this study introduces exponential correction terms for pressure and velocity based on the original Preston formula, expressing the material removal amount as a nonlinear function:

[0219]

[0220] Where u is the pressure exponent and z is the relative velocity exponent, both of which are parameters to be determined and are used to describe the nonlinear effects of pressure and velocity on the amount of material removed.

[0221] During the parameter search process, a total of 36 index combinations were traversed with a step size of 0.2 to ensure that the search range fully covers all experimental conditions and to systematically evaluate the prediction accuracy of the model.

[0222] For each (u,z) combination, the material removal rate (MRRsim) under 16 experimental conditions was calculated using a simulation model, and compared with the experimentally measured MRRexp. To quantify the deviation between the simulation results and experimental data, an average relative error index was introduced:

[0223]

[0224] Where N=16 is the total number of experimental cases. and Let be the material removal rates in the simulation and experiment, respectively, for the i-th experimental condition. By traversing all (u,z) combinations, Figure 10 The distribution of the model's average relative error under different combinations of pressure index u and relative velocity index z is shown. The combination that minimizes the average relative error is selected as the optimal Preston index, thereby improving the prediction accuracy and reliability of the simulation model.

[0225] By substituting the optimized pressure index and relative velocity index into the material removal model, the original model is modified, thus obtaining a material removal relationship model that is more consistent with the experimental results.

[0226] Table 10 presents the simulated material removal rates obtained by substituting the optimized pressure index u=1.3 and velocity index z=1.1 into the modified model, and compares them with the removal rates measured by orthogonal experiments. In the 16 experimental groups, the relative errors between the simulated and experimental values ​​were all controlled within 19%, with the minimum error being 11.59% and the maximum error being 18.08%, showing relatively small overall fluctuations. Further calculations show that the average relative error of the model is approximately 15%, but some deviation still exists. This is mainly because, with the index parameters u=1.3 and z=1.1 already determined, the original model assigned equal weights to all experimental data, failing to fully reflect the inherent fluctuations in material removal rates under different working conditions. This deviation arises from the complex microscopic interactions during the grinding process, specifically manifested in changes in local contact conditions, differences in abrasive morphology, abrasive wear and shedding, fluctuations in the lubrication state of the grinding fluid, and the chemical removal effect of the material, causing some experimental data to deviate from the overall trend, thus affecting the model's prediction accuracy.

[0227] Table 10 Comparison of Predicted Material Removal Rate from Modified Model with Experimental Results

[0228]

[0229] To further improve the consistency between model results and experimental data, a weighted fusion method is introduced for post-processing calibration of the model prediction results. This method assigns weights to the simulation prediction results and experimental measurement results to achieve a weighted correction of the predicted material removal rate, thereby compensating for errors in the prediction results while maintaining the original physical meaning of the model. The corrected material removal rate can be expressed as:

[0230]

[0231] Among them, MRR final,i This represents the final corrected predicted material removal rate. The weighting coefficient λ is used to adjust the contribution ratio of the simulation prediction and the experimental measurement in the corrected model. When λ is small, the corrected result is closer to the original simulation prediction; when λ increases, the corrected result is closer to the experimental measurement.

[0232] To determine the optimal weighting coefficients, λ is solved using the principle of minimizing the average relative error, based on orthogonal experimental data. The relative error between the fusion result and the experimental measurement value under the i-th working condition is defined as:

[0233]

[0234] The average relative error under all experimental conditions can then be expressed as:

[0235]

[0236] Where n is the number of experimental data sets, and by iterating through the average relative error under different λ conditions, a set is selected that makes the data set more accurate and accurate. The solution with the smallest weight coefficient is the optimal solution, i.e.:

[0237]

[0238] After determining the optimal weighting coefficients, the material removal rates under 16 orthogonal experimental conditions were fused and corrected to obtain the corrected material removal rate results and their relative errors, as follows: Figure 11 As shown, the results indicate that after weighted fusion correction, the deviation between the material removal rate results and experimental measurements under each working condition is significantly reduced. The relative error of each experimental group is controlled within 10%, with the minimum error being approximately 4.63%, the maximum error approximately 7.23%, and the average relative error approximately 6.02%. Compared with the exponential correction model alone, the fusion results after incorporating experimental information show higher consistency with the experimental results, indicating that this method can effectively correct the systematic bias of the model under different combinations of processing parameters.

[0239] By introducing a weighted fusion correction method based on the exponentially modified Preston model, the deviation between the material removal rate results and experimental measurements can be effectively reduced, improving the accuracy of the model results in representing the actual grinding process. This method, while retaining the original model's mechanistic foundation, effectively compensates for complex factors that are difficult to fully and explicitly characterize in actual processing. Therefore, it can provide a reference for analyzing material removal patterns, evaluating model errors, and optimizing subsequent process parameters in the grinding process of polycrystalline zinc sulfide bonded abrasives.

[0240] S7. Input the grinding process parameters into the modified empirical material removal rate prediction model, calculate the predicted value of the material removal rate during the grinding process, and analyze the influence of process parameters on material removal rate and surface quality based on the prediction results.

[0241] Using a single-factor variable control method, while keeping other process parameters constant, the grinding pressure, grinding pad speed, and workpiece disk speed were varied. Numerical simulations were performed using a modified empirical material removal rate prediction model to calculate the material removal rate and the variation law of workpiece surface morphology under different parameter levels. Based on the simulation results, the mapping relationship between grinding pressure and material removal rate, and between grinding speed and material removal rate were established, and the influence trend of each process parameter on the micro-undulations and uniformity of the workpiece surface was analyzed.

[0242] Simulation results reveal that grinding pressure not only determines material removal efficiency but also significantly affects the changes in the microstructure of the workpiece surface. This reflects the trade-off between pressure, surface uniformity, and local peak-valley height. A moderate rotational speed ratio can optimize the abrasive particle trajectory and contact time, thereby improving removal efficiency. However, when the rotational speed ratio is too high, the relative speeds of the two particles are close, shortening the effective interaction time between the workpiece and the abrasive particles, which is detrimental to material removal. It is evident that the material removal rate follows the same evolutionary pattern as the surface morphology, and the degree of surface undulation is positively correlated with the magnitude of the MRR (Mean Removal Rate). This indicates that while pursuing high-efficiency removal, surface quality must also be considered. In actual processing, process parameters should be selected based on a trade-off according to requirements.

[0243] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A method for predicting material removal rate based on bonded abrasive grinding, characterized in that, Includes the following steps: S1. Construct a theoretical model of single abrasive cutting depth and establish a model of abrasive-workpiece contact pressure distribution; S2. Based on the abrasive-workpiece contact pressure distribution model, establish an abrasive-workpiece contact pressure distribution database; S3. Perform geometric modeling on the grinding pad to obtain the surface geometric features of the grinding pad and the spatial distribution information of the abrasive. S4. Establish a trajectory field model of the grinding motion process based on the motion parameters of the grinding process; S5. By integrating the single abrasive cutting depth, abrasive-workpiece contact pressure, grinding pad geometry, abrasive distribution characteristics, and motion trajectory field, a material removal process model is established. S6. Correct the model parameters by measuring the material removal rate data using grinding experiments; S7. Calculate the predicted material removal rate during the grinding process using the modified empirical material removal rate prediction model.

2. The method for predicting material removal rate based on bonded abrasive grinding according to claim 1, characterized in that, In step S1, the theoretical model for the single abrasive cutting depth is as follows: In the formula, δ(y) ave Let P0 be the desired depth of cut, R0 be the grinding pressure, D be the workpiece radius, D be the volumetric equivalent particle size of the abrasive, and E(A) be the contact projection area formed by a single abrasive particle on the surface of the grinding pad. H represents the volume fraction of abrasive in the grinding pad, γ represents the average contact ratio between the surface area of ​​the grinding pad protrusions and the surface area of ​​the workpiece, and H represents the volume fraction of abrasive in the grinding pad. M Where E represents the Mayer hardness of the workpiece material, E represents the elastic modulus of the resin matrix of the abrasive pad, and n p ≈1.2 represents the relative contact depth coefficient under fully plastic deformation conditions, used to characterize the relationship between indentation geometry and contact state, where θ is the angle between the conical abrasive and the grinding pad plane.

3. The method for predicting material removal rate based on bonded abrasive grinding according to claim 2, characterized in that, In step S1, the abrasive-workpiece contact pressure distribution model is established through ANSYS simulation. The specific steps are as follows: Select the abrasive shape and set the tilt angle of the abrasive, construct the contact boundary conditions and local stress distribution characteristics, and set the parameters for step-by-step loading solution control.

4. The method for predicting material removal rate based on bonded abrasive grinding according to claim 3, characterized in that, The steps for establishing the abrasive-workpiece contact pressure distribution database in step S2 are as follows: MATLAB was used to read nodal pressure data under various working conditions, and the discrete nodal pressure data under different pressure conditions were uniformly normalized to reduce the number of nodes to be extracted for subsequent material removal pressure. The maximum radius of the contact area was calculated based on the effective contact nodes, and a linear interpolation algorithm was used to map the discrete nodal pressure onto a regular two-dimensional mesh. The pressure field interpolation is expressed as: in, This represents a two-dimensional interpolation operation based on a linear interpolation algorithm. Let be the pressure at a node under the k-th pressure distribution. Let x and y be the coordinates of node i under the k-th pressure distribution. The pressure of discrete points after two-dimensional interpolation is mapped to the x and y coordinates in a regular two-dimensional grid. The null value regions generated by interpolation are assigned zeros while retaining the data within the contact circle domain to obtain a physically valid two-dimensional pressure matrix. A standardized single abrasive pressure field template is then constructed and stored in the abrasive-workpiece contact pressure distribution database.

5. The method for predicting material removal rate based on bonded abrasive grinding according to claim 4, characterized in that, The geometric modeling step for the abrasive pad described in step S3 includes: The spiral groove structure of the grinding pad is generated based on the involute equation. A groove band with a finite width is constructed by normal offset, and a periodic spiral structure is formed by circumferential array. Abrasive points are randomly generated within the effective area of ​​the grinding pad and a minimum spacing constraint is applied. Based on the structural characteristics of the protrusions and grooves of the grinding pad, abrasive points located in the groove area are eliminated, so that the abrasive is only distributed in the effective processing protrusion area. The abrasive is displayed in the form of scattered points in the grinding pad coordinate system to complete the initialization of the abrasive field. A trajectory storage structure is assigned to each abrasive to record its motion path during the simulation process.

6. The method for predicting material removal rate based on bonded abrasive grinding according to claim 5, characterized in that, The steps for establishing the trajectory field model of the grinding motion process described in step S4 are as follows: Establish an absolute coordinate system with the center of the grinding pad as the origin and a workpiece coordinate system with the center of the workpiece as the origin. The absolute coordinate system {XOY} has the center of the grinding pad as the center, and the X and Y axes are fixed and do not move with the grinding pad. c Let the center of the carrier be e1, and the eccentricity between the center of the grinding pad and the center of the workpiece coordinate system be e1; w O w Y w }With workpiece center O w Centered on the circle, the X-axis direction is parallel to the vector. Perpendicular to the vector, the Y-axis direction is perpendicular to the vector. The grinding pads are collinear, with an eccentricity of e2 relative to the center of the carrier. The grinding pad rotates about its center with an angular velocity ω1, and the carrier rotates about its center with an angular velocity ω2. At time t, the initial position of a certain abrasive grain p on the grinding pad in the absolute coordinate system is... ; The motion trajectory expression of any abrasive in the workpiece coordinate system is obtained as follows: Apply a motion expression of a single abrasive grain in the workpiece coordinate system to each abrasive grain, and discretize the time to obtain its instantaneous position at any time. When the abrasive grain passes through the workpiece surface area, statistically analyze its corresponding two-dimensional coordinates, and connect the coordinate points at consecutive times to form the abrasive grain motion trajectory, thereby obtaining the motion trajectory field of the abrasive grain actually participating in the processing on the workpiece surface.

7. The method for predicting material removal rate based on bonded abrasive grinding according to claim 6, characterized in that, The steps for establishing the material removal process model described in step S5 are as follows: The workpiece surface is discretized into a two-dimensional regular mesh. A pressure template is randomly assigned to each abrasive grain from the abrasive-workpiece contact pressure distribution database. When the abrasive center enters the effective area of ​​the workpiece, the single-step material removal amount of a single abrasive grain on a single mesh node within the current time step Δt is calculated based on the Preston equation: In the formula, g is the Preston coefficient. This represents the maximum pressure value across all pressure distributions. For the normalized first The pressure at a node under a certain pressure distribution. The relative velocity between the abrasive and the workpiece at the point of contact; After sampling the material removal amount for all mesh nodes, it is mapped to the corresponding mesh cells to update the height field: in, for The amount of material removed after a certain time. Let n be the amount of material removed at time t. k The number of discrete points on the workpiece surface. This represents the final amount of material removed at a single sampling point. Based on the updated workpiece height matrix The calculable material removal rate is: Where MRR is the material removal rate. H represents the simulation time, H is the updated workpiece height matrix, and H0 is the initial workpiece height matrix. This represents the average height of material removed from the workpiece surface.

8. The method for predicting material removal rate based on bonded abrasive grinding according to claim 7, characterized in that, The model parameter correction steps described in step S6 are as follows: Orthogonal grinding experiments were designed and implemented to obtain experimental values ​​of material removal rate and surface roughness data under different combinations of process parameters. Based on the orthogonal experimental data, range analysis was used to determine the primary and secondary relationships of the influence of each process parameter on material removal rate and surface quality. Based on the material removal rate data measured by the orthogonal experiments and the corresponding simulation input parameters, the Preston coefficients were calibrated by linear fitting. Pressure and velocity indices were introduced into the traditional linear Preston model to construct a nonlinear modified material removal rate prediction model. With the goal of minimizing the average relative error, different combinations of pressure and velocity indices were traversed within a set range, and the combination of indices that minimizes the average relative error between the simulation prediction value and the experimental measurement value was selected to complete the model correction and obtain the corrected empirical material removal rate prediction model.

9. The method for predicting material removal rate based on bonded abrasive grinding according to claim 8, characterized in that, After completing the nonlinear correction, the model prediction results are post-processed and calibrated using a weighted fusion method. The specific steps are as follows: A weighted function is constructed to integrate simulation predictions and experimental measurements. The contribution ratio of simulation values ​​and experimental values ​​to the final prediction result is adjusted by weighting coefficients. The optimal weighting coefficients are determined with the goal of minimizing the average relative error under all experimental conditions. By substituting the optimal weighting coefficients into the weighting function, the final corrected material removal rate prediction model is obtained.

10. The method for predicting material removal rate based on bonded abrasive grinding according to claim 9, characterized in that, The calculation steps for the predicted material removal rate in the grinding process described in step S7 are as follows: Using a single-factor variable control method, while keeping other process parameters constant, the grinding pressure, grinding pad speed, and workpiece disk speed were changed respectively. Numerical simulation was performed using a modified empirical material removal rate prediction model to calculate the material removal rate and workpiece surface morphology evolution results under different parameter levels. Based on the simulation results, the mapping relationships between grinding pressure and material removal rate, and between grinding speed and material removal rate were established respectively, and the influence trends of each process parameter on the micro-undulations and uniformity of the workpiece surface were analyzed.