A method and device for predicting residual stress on silicon carbide grinding surface

Abstract

The present application provides a method and device for predicting residual stress of a silicon carbide grinding surface, comprising: obtaining mechanical response data of a silicon carbide material to construct a silicon carbide simulation model; setting grinding parameters to simulate grinding forces on the silicon carbide simulation model to obtain preliminary surface residual stress prediction results; optimizing the model parameters of the silicon carbide simulation model based on the preliminary surface residual stress prediction results and grinding parameters using a Bayesian inference method and a Monte Carlo method to obtain optimal model parameters; performing grinding force simulation on the silicon carbide simulation model using the optimal model parameters and grinding parameters to obtain optimized surface residual stress prediction results. Using the silicon carbide simulation model after parameter optimization based on the Bayesian inference method and the Monte Carlo method to perform grinding force simulation, an optimized surface residual stress prediction result is obtained, thereby realizing the prediction of the surface residual stress of silicon carbide after grinding.

Description

Technical Field

[0001] The present invention relates to the technical field of material processing, and in particular to a method and device for predicting residual stress of a silicon carbide grinding surface. Background Art

[0002] Silicon carbide (SiC) is an important wide-bandgap semiconductor material. Due to its high hardness, excellent thermal stability, and chemical inertness, it is widely used in power electronics, aerospace, and advanced manufacturing. In the manufacturing process of silicon carbide substrates, the grinding process is a key surface processing step, and its quality directly affects the mechanical strength and service life of the device. However, due to the high brittleness and anisotropy of silicon carbide materials, residual stresses are easily introduced during the grinding process. These stresses may cause internal defects in the material, surface cracks, and even reduce its service reliability. Therefore, accurately predicting the residual stress after grinding and quantifying the uncertainty of the prediction are core issues in optimizing the grinding process and improving the quality of material processing.

[0003] Currently, research on residual stresses in silicon carbide during grinding primarily focuses on experimental measurement and numerical simulation. Experimental measurement methods, such as X-ray diffraction (XRD), confocal Raman spectroscopy, and nanoindentation, can provide direct stress data, but they suffer from limitations in test accuracy, time-consuming data acquisition, and high experimental costs. Finite element simulation (FEM), on the other hand, is widely used to predict residual stresses during grinding, with the Johnson-Holmquist-2 (JH-2) constitutive model considered particularly suitable for describing the high strain rate and high stress state of silicon carbide.

[0004] However, the traditional FEM method relies on empirically selected material parameters, and its prediction accuracy is affected by the errors of the material model and experimental data. It also lacks a systematic uncertainty quantification method and cannot provide reliable stress confidence intervals. Summary of the Invention

[0005] The present invention provides a method and device for predicting residual stress on a ground surface of silicon carbide, which are used to predict the residual stress on the surface of silicon carbide after grinding.

[0006] In a first aspect, the present invention provides a method for predicting residual stress on a polished silicon carbide surface, comprising:

[0007] Obtain mechanical response data of silicon carbide materials;

[0008] Constructing a silicon carbide simulation model based on the mechanical response data;

[0009] Setting grinding parameters to simulate grinding forces on the silicon carbide simulation model and obtain preliminary surface residual stress prediction results;

[0010] Optimizing the model parameters of the silicon carbide simulation model based on the preliminary surface residual stress prediction result and the grinding parameters by using a Bayesian inference method and a Monte Carlo method to obtain optimal model parameters;

[0011] The optimal model parameters and the grinding parameters are used to simulate the grinding force of the silicon carbide simulation model to obtain an optimized surface residual stress prediction result.

[0012] Optionally, constructing a silicon carbide simulation model based on the mechanical response data includes:

[0013] Constructing a preliminary silicon carbide simulation model based on the mechanical response data;

[0014] The preliminary silicon carbide simulation model is meshed, and corresponding material parameters and load boundary conditions are set to obtain a silicon carbide simulation model.

[0015] Optionally, setting grinding parameters to simulate grinding forces on the silicon carbide simulation model to obtain preliminary surface residual stress prediction results includes:

[0016] Setting grinding parameters; the grinding parameters include abrasive particle diameter, grinding depth, abrasive particle spacing and grinding conditions;

[0017] Based on the abrasive grain diameter, the grinding depth, the abrasive grain spacing and the grinding conditions, a grinding force simulation is performed on the silicon carbide simulation model to obtain a preliminary surface residual stress prediction result.

[0018] Optionally, the model parameters of the silicon carbide simulation model are optimized based on the preliminary surface residual stress prediction result and the grinding parameters by using a Bayesian inference method and a Monte Carlo method to obtain optimal model parameters, including:

[0019] Based on the preliminary surface residual stress prediction results, a Bayesian inference method is used to calculate the posterior distribution of model parameters of the silicon carbide simulation model;

[0020] Optimizing the posterior distribution by a Monte Carlo method to obtain an optimized posterior distribution;

[0021] According to the optimized posterior distribution, the model parameters of the silicon carbide simulation model are optimized to obtain optimal model parameters.

[0022] Optionally, using the optimal model parameters and the grinding parameters, performing grinding force simulation on the silicon carbide simulation model to obtain an optimized surface residual stress prediction result includes:

[0023] Optimizing the silicon carbide simulation model using the optimal model parameters to obtain an optimized silicon carbide simulation model;

[0024] Based on the grinding parameters, grinding force simulation is performed on the optimized silicon carbide simulation model to obtain an optimized surface residual stress prediction result.

[0025] Optionally, after performing grinding force simulation on the silicon carbide simulation model using the optimal model parameters and the grinding parameters to obtain an optimized surface residual stress prediction result, the method further includes:

[0026] Silicon carbide is ground based on the grinding parameters to obtain actual residual stress, and then the actual residual stress is compared with the optimized surface residual stress prediction result to determine the accuracy of the optimized surface residual stress prediction result.

[0027] In a second aspect, the present invention provides a device for predicting residual stress of a silicon carbide polished surface, comprising:

[0028] An acquisition module, used for acquiring mechanical response data of silicon carbide materials;

[0029] A construction module, configured to construct a silicon carbide simulation model based on the mechanical response data;

[0030] A simulation module, used to set grinding parameters to simulate the grinding force of the silicon carbide simulation model and obtain preliminary surface residual stress prediction results;

[0031] an optimization module, configured to optimize the model parameters of the silicon carbide simulation model based on the preliminary surface residual stress prediction result and the grinding parameters by using a Bayesian inference method and a Monte Carlo method to obtain optimal model parameters;

[0032] The prediction module is used to use the optimal model parameters and the grinding parameters to simulate the grinding force of the silicon carbide simulation model to obtain an optimized surface residual stress prediction result.

[0033] Optionally, the building blocks include:

[0034] A construction submodule is used to construct a preliminary silicon carbide simulation model based on the mechanical response data;

[0035] The first setting submodule is used to mesh the preliminary silicon carbide simulation model and set corresponding material parameters and load boundary conditions to obtain a silicon carbide simulation model.

[0036] Optionally, the simulation module includes:

[0037] The second setting submodule is used to set grinding parameters; the grinding parameters include abrasive particle diameter, grinding depth, abrasive particle spacing and grinding conditions.

[0038] The simulation submodule is used to simulate the grinding force of the silicon carbide simulation model based on the abrasive particle diameter, the grinding depth, the abrasive particle spacing and the grinding conditions to obtain a preliminary surface residual stress prediction result.

[0039] Optionally, the optimization module includes:

[0040] a calculation submodule, configured to calculate, based on the actual grinding residual stress data and the preliminary surface residual stress prediction result, a posterior distribution of model parameters of the silicon carbide simulation model using a Bayesian inference method;

[0041] A first optimization submodule is used to optimize the posterior distribution by using a Monte Carlo method to obtain an optimized posterior distribution;

[0042] The second optimization submodule is used to optimize the model parameters of the silicon carbide simulation model according to the optimized posterior distribution to obtain optimal model parameters.

[0043] Optionally, the prediction module includes:

[0044] A third optimization submodule is configured to optimize the silicon carbide simulation model using the optimal model parameters to obtain an optimized silicon carbide simulation model;

[0045] The prediction submodule is used to perform grinding force simulation on the optimized silicon carbide simulation model based on the grinding parameters to obtain an optimized surface residual stress prediction result.

[0046] Optionally, the device for predicting residual stress on a silicon carbide polished surface further comprises:

[0047] A verification module is used to perform silicon carbide grinding based on the grinding parameters, and after obtaining actual residual stress, compare the actual residual stress with the optimized surface residual stress prediction result to verify the accuracy of the optimized surface residual stress prediction result.

[0048] In a third aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in the method provided in the first aspect are executed.

[0049] In a fourth aspect, the present invention provides a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, runs the steps of the method provided in the first aspect above.

[0050] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, runs the steps of the method provided in the first aspect.

[0051] It can be seen from the above technical solutions that the present invention has the following advantages:

[0052] The present invention provides a method and device for predicting residual stress of a silicon carbide grinding surface. The method comprises: obtaining mechanical response data of a silicon carbide material; constructing a silicon carbide simulation model based on the mechanical response data; setting grinding parameters to simulate grinding forces on the silicon carbide simulation model to obtain a preliminary surface residual stress prediction result; optimizing the model parameters of the silicon carbide simulation model based on the preliminary surface residual stress prediction result and the grinding parameters using a Bayesian inference method and a Monte Carlo method to obtain optimal model parameters; performing grinding force simulation on the silicon carbide simulation model using the optimal model parameters and the grinding parameters to obtain an optimized surface residual stress prediction result. The model parameters of the silicon carbide simulation model are optimized using the Bayesian inference method and the Monte Carlo method, and the silicon carbide simulation model with optimized parameters is used to perform grinding force simulation to obtain an optimized surface residual stress prediction result, thereby realizing the prediction of the surface residual stress of silicon carbide after grinding. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 This is a flow chart of Example 1 of a method for predicting residual stress on a polished silicon carbide surface according to the present invention;

[0055] Figure 2 This is a flow chart of Example 2 of a method for predicting residual stress on a polished silicon carbide surface according to the present invention;

[0056] Figure 3 A comparison chart of the optimized surface residual stress prediction results and actual data;

[0057] Figure 4 The error distribution diagram between the optimized surface residual stress prediction results and the actual data;

[0058] Figure 5 This is a structural block diagram of an embodiment of a device for predicting residual stress on a silicon carbide polished surface according to the present invention. DETAILED DESCRIPTION

[0059] The embodiment of the present invention provides a method and device for predicting residual stress on the ground surface of silicon carbide, which are used to predict the surface residual stress after silicon carbide is ground.

[0060] In order to make the purpose, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0061] For example 1, please refer to Figure 1 , Figure 1 This is a flow chart of a first embodiment of a method for predicting residual stress on a polished silicon carbide surface according to the present invention. The method comprises:

[0062] S1, obtain mechanical response data of silicon carbide material;

[0063] In an embodiment of the present application, mechanical response data of silicon carbide material is obtained. The mechanical response data includes data such as hardness, elastic modulus, flexural strength, compressive strength and fracture toughness, which can be obtained through nanoindentation experiments and confocal Raman spectroscopy technology.

[0064] S2, constructing a silicon carbide simulation model based on the mechanical response data;

[0065] The embodiment of the present application constructs a preliminary silicon carbide simulation model based on the mechanical response data, meshes the preliminary silicon carbide simulation model, and sets corresponding material parameters and load boundary conditions to obtain the silicon carbide simulation model.

[0066] S3, setting grinding parameters to simulate grinding forces on the silicon carbide simulation model to obtain preliminary surface residual stress prediction results;

[0067] In an embodiment of the present application, grinding parameters are set, and the grinding parameters include abrasive diameter, grinding depth, abrasive spacing and grinding conditions. Based on the abrasive diameter, grinding depth, abrasive spacing and grinding conditions, the grinding force simulation of the silicon carbide simulation model is performed to obtain preliminary surface residual stress prediction results.

[0068] S4, optimizing the model parameters of the silicon carbide simulation model based on the preliminary surface residual stress prediction result and the grinding parameters through a Bayesian inference method and a Monte Carlo method to obtain optimal model parameters.

[0069] In an embodiment of the present application, based on the preliminary surface residual stress prediction results, the Bayesian inference method is used to calculate the posterior distribution of the model parameters of the silicon carbide simulation model. The posterior distribution is optimized by the Monte Carlo method to obtain the optimized posterior distribution. According to the optimized posterior distribution, the model parameters of the silicon carbide simulation model are optimized to obtain the optimal model parameters.

[0070] S5, performing grinding force simulation on the silicon carbide simulation model using the optimal model parameters and the grinding parameters to obtain an optimized surface residual stress prediction result;

[0071] In an embodiment of the present application, the silicon carbide simulation model is optimized using the optimal model parameters to obtain an optimized silicon carbide simulation model. Based on the grinding parameters, the optimized silicon carbide simulation model is subjected to grinding force simulation to obtain an optimized surface residual stress prediction result. After obtaining the optimized surface residual stress prediction result, silicon carbide is ground based on the grinding parameters. After obtaining the actual residual stress, the actual residual stress is compared with the optimized surface residual stress prediction result to determine the accuracy of the optimized surface residual stress prediction result.

[0072] An embodiment of the present invention provides a method for predicting residual stress on the surface of silicon carbide after grinding, which uses a Bayesian inference method and a Monte Carlo method to optimize the model parameters of a silicon carbide simulation model, and uses the silicon carbide simulation model with optimized parameters to perform grinding force simulation to obtain an optimized surface residual stress prediction result, thereby realizing the prediction of the surface residual stress of silicon carbide after grinding.

[0073] For example 2, please refer to Figure 2 , Figure 2 This is a flow chart of Example 2 of a method for predicting residual stress on a polished silicon carbide surface according to the present invention, the steps comprising:

[0074] Step S201, obtaining mechanical response data of silicon carbide material;

[0075] In an embodiment of the present application, mechanical response data of silicon carbide material is obtained. The mechanical response data includes data such as hardness, elastic modulus, flexural strength, compressive strength and fracture toughness, which can be obtained through nanoindentation experiments and confocal Raman spectroscopy technology.

[0076] Step S202: constructing a preliminary silicon carbide simulation model based on the mechanical response data;

[0077] In the embodiment of the present application, through simulation software such as Matlab, based on the mechanical response data of silicon carbide material, taking into account the hard and brittle characteristics of silicon carbide material itself, a Johnson-Holmquist-2 (JH-2) constitutive model is constructed with silicon carbide material as the substrate to obtain a preliminary silicon carbide simulation model.

[0078] Step S203, meshing the preliminary silicon carbide simulation model, and setting corresponding material parameters and load boundary conditions to obtain a silicon carbide simulation model;

[0079] In the embodiment of the present application, the preliminary silicon carbide simulation model is meshed to improve the calculation accuracy of the model, and the corresponding material parameters and load boundary conditions are set according to the mechanical response data.

[0080] Step S204, setting grinding parameters; the grinding parameters include abrasive particle diameter, grinding depth, abrasive particle spacing and grinding conditions;

[0081] In the embodiment of the present application, grinding parameters are set, which are simulation parameters during grinding, including abrasive particle diameter, grinding depth, abrasive particle spacing and grinding conditions.

[0082] Step S205, performing grinding force simulation on the silicon carbide simulation model based on the abrasive grain diameter, the grinding depth, the abrasive grain spacing, and the grinding conditions to obtain a preliminary surface residual stress prediction result;

[0083] In an embodiment of the present invention, simulation software is used to determine the motion trajectory of the abrasive particles during grinding based on the abrasive particle diameter, grinding depth, abrasive particle spacing, and grinding conditions. This simulates the grinding path of a real grinding wheel on the SiC wafer surface, realizes the grinding force simulation analysis of the silicon carbide simulation model, and obtains preliminary surface residual stress prediction results. The grinding force simulation analysis process of the silicon carbide simulation model is an automatic calculation process of the simulation software and is not described in detail here.

[0084] Among them, through simulation software, a typical dual-abrasive finite element simulation model can be established according to the abrasive diameter, grinding depth and abrasive spacing. Using the typical dual-abrasive finite element simulation model, according to the determined motion trajectory, the grinding process of the real grinding disk on the SiC wafer surface is simulated, which can effectively simulate the complex physical phenomenon of the interaction between abrasives in the real grinding process, thereby significantly improving the accuracy and calculation efficiency of the simulation model. Moreover, the use of the typical dual-abrasive finite element simulation model not only reduces the calculation cost, but also effectively optimizes the physical process of grinding processing and improves the prediction and control capabilities of the silicon carbide material grinding process.

[0085] Step S206, based on the preliminary surface residual stress prediction result, using the Bayesian inference method to calculate the posterior distribution of the model parameters of the silicon carbide simulation model;

[0086] In this example, Bayesian inference was used to optimize the model parameters of the JH-2 constitutive model. A Bayesian framework was used to fuse actual measurement data with finite element simulation data to establish a JH-2 parameter distribution that better reflects actual operating conditions. To improve optimization efficiency, the Metropolis-Hastings sampling (MH) strategy was selected to optimize the posterior distribution, resulting in more stable convergence of parameter optimization.

[0087] First, actual measurements are performed using grinding parameters to obtain actual SiC residual stress data. The Bayesian optimization framework for the JH-2 model is then constructed, with the prior distribution of the JH-2 model set to a uniform distribution. The actual measured data provide a reasonable parameter range, and the actual measured SiC residual stress data are used to construct a likelihood function as a Gaussian distribution (where the likelihood function is a normal distribution, and its mean and variance can be calculated based on the actual SiC residual stress data). This measures the degree of fit between the simulation prediction value (preliminary surface residual stress prediction result) and the actual data (actual SiC residual stress data). Using the likelihood function constructed using the Bayesian method, with the simulation prediction value as the prior data and the actual data as the likelihood data, the posterior distribution of the model parameters of the silicon carbide simulation model is calculated. The calculation formula for the posterior distribution is as follows:

[0088] ;

[0089] in, is the posterior distribution, which means that when observing data (ie actual data), the model parameters Updates; is the likelihood function, which means that given the parameters When the observation data Probability of occurrence; is the prior distribution (i.e., the simulated predicted value), which represents the model parameters before the observation data The initial value of is a normalization constant used to ensure the normalization of the posterior distribution. It is generally difficult to calculate and satisfies: ; Due to the normalization constant term The calculation is complex. MH sampling directly samples the posterior distribution by constructing a Markov chain, bypassing the explicit calculation of the integral term.

[0090] Bayesian inference is used to optimize the parameters of the JH-2 constitutive model. The experimental measurement data and finite element simulation data are fused through a Bayesian framework to establish a JH-2 parameter distribution that is more in line with actual working conditions.

[0091] Step S207, optimizing the posterior distribution by using the Monte Carlo method to obtain an optimized posterior distribution;

[0092] In the embodiment of the present application, the MH sampling method in the Monte Carlo method is used to optimize the JH-2 parameters, set the sampling initial value and the adaptive adjustment parameter step size, and generate candidate parameters in each iteration, while calculating the acceptance probability of the new parameters (updated model parameters). As acceptance criteria, among others, are the current model parameters, is the updated model parameter. If the new parameter meets the optimization conditions, the model parameter is updated, otherwise the original parameter value is retained. In summary, the MH sampling method is used to update the posterior distribution so that the posterior distribution converges and the optimized posterior distribution is obtained.

[0093] The steps for adaptively adjusting the parameter step size are as follows:

[0094] (1) Set the initial parameter step size.

[0095] (2) Calculate the optimal step size using the mean shift adjustment method:

[0096] ;

[0097] in, The new step size to be used in the next iteration; The step size used in the current iteration.

[0098] (3) Set the parameter step range to prevent outliers from affecting the optimization results:

[0099] ;

[0100] Therefore, during the MH sampling process, the sampling step size is dynamically adjusted according to the parameter acceptance rate calculated in real time, thereby effectively improving the optimization efficiency and robustness of the sampling process.

[0101] In the embodiments of the present application, Bayesian inference is used to optimize the JH-2 constitutive model parameters, and the MH sampling strategy is selected to optimize the posterior distribution of the JH-2 constitutive model parameters, so that the parameter optimization convergence is more stable, thereby realizing the automatic optimization of the finite element simulation model parameters based on experimental data, making the model more accurately reflect the mechanical behavior in the actual grinding process, significantly improving the accuracy of simulation prediction of residual stress, and ensuring the reliability of the model in actual engineering applications.

[0102] Step S208 : optimizing the model parameters of the silicon carbide simulation model according to the optimized posterior distribution to obtain optimal model parameters.

[0103] In the embodiment of the present application, the model parameters of the silicon carbide simulation model are optimized using the optimized posterior distribution to obtain the optimal model parameters, which are verified in the finite element simulation using the optimal model parameters. A new simulation is run to calculate the residual stress distribution of the SiC grinding surface after optimization, and compared with the actual data to evaluate the optimization effect. The model parameters before optimization have a large prediction error in the simulation, while the optimal model parameters significantly reduce the prediction error. At the same time, the uncertainty analysis calculates the 95% confidence interval, which improves the prediction reliability. The calculated 95% confidence interval obtained by calculation quantifies the uncertainty range of the prediction results, provides more comprehensive, accurate and reliable prediction data for practical engineering applications, and improves the accuracy of residual stress prediction.

[0104] Step S209, optimizing the silicon carbide simulation model using the optimal model parameters to obtain an optimized silicon carbide simulation model;

[0105] In an embodiment of the present application, the optimal model parameters are applied to the silicon carbide simulation model to optimize the silicon carbide simulation model, thereby obtaining an optimized silicon carbide simulation model.

[0106] Step S210, performing grinding force simulation on the optimized silicon carbide simulation model based on the grinding parameters to obtain an optimized surface residual stress prediction result;

[0107] In the examples of this application, grinding force simulation was performed using an optimized silicon carbide simulation model to obtain an optimized surface residual stress prediction result. The error between the optimized surface residual stress prediction result and the actual data was small, with a high accuracy rate. A comparative analysis found that the error of the optimized surface residual stress prediction result of this application was significantly reduced to within 2%, significantly improving the accuracy of residual stress prediction.

[0108] After calculating the optimized surface residual stress prediction results, compare the optimized surface residual stress prediction results with the actual data. Please refer to Figure 3 and Figure 4 , Figure 3 This is a comparison chart of the optimized surface residual stress prediction results and actual data. Figure 4 The error distribution diagram of the optimized surface residual stress prediction result and the actual data, where line segment a is the optimized surface residual stress prediction result, line segment b is the actual data, and line segment c is the error. Figure 3 and Figure 4 The horizontal coordinate x is the grinding trajectory point, Figure 3 and Figure 4The vertical coordinate y is stress, which is given by Figure 3 and Figure 4 It can be seen that the accuracy of the calculated optimized surface residual stress prediction results can reach 98.39%, confirming that the optimized surface residual stress prediction results have high reliability (generally speaking, an accuracy greater than 97% can be considered reliable). In addition, the process of calculating the optimized surface residual stress prediction results can provide a reasonable uncertainty assessment, making the prediction results more valuable for engineering applications.

[0109] The following is an application example of this embodiment. For example, the real mechanical response data of silicon carbide material is obtained through indentation experiments, including the indenter load corresponding to 50 displacement points. The residual stress on the surface of the silicon carbide material is measured using confocal Raman microscopy, which serves as a prior distribution and experimental data support. Based on the real mechanical response data, a finite element simulation model of silicon carbide material is constructed, including geometric model simplification, key unit division, reference point selection, and residual stress extraction. The model is then exported as an "inp" format file (a file with the suffix "inp"). Based on the established finite element model, an Abaqus Python script is used to quickly extract residual stresses from an "odb" format file: (1) Open the "odb" format file after the Abaqus simulation to ensure that the simulation data is available; (2) Initialize the "csv" format file and write the header information (step, frame index, node number, force RF2, displacement U2); (3) Iterate through all steps and frames in sequence to extract force and displacement data from the specified node set (such as "SET-RP"); (4) Match the extracted data by node number and write them line by line to the "csv" format file; (5) After completing the data extraction, close the "odb" format file to release resources and ensure the stability of subsequent calculations. Save the extracted residual stresses as a "csv" format file for further processing in the "Matlab" software. In Matlab, the "inp" file was modified and simulated using Abaqus. Automated data processing was achieved using Python scripts, and uncertainty quantification was performed based on a Bayesian framework. A Bayesian inference model was constructed to calculate the posterior distribution of the parameters. The parameters were optimized using the Metropolis-Hastings (MH) sampling method within the Monte Carlo simulation (MCMC) algorithm to achieve convergence. Once the posterior distribution converged, the optimized parameters were automatically modified using Matlab code to modify the "inp" file, and Abaqus was used for iterative simulation. The optimized material parameters from the "inp" file were extracted, and a simplified simulation model for silicon carbide grinding was constructed. Abaqus was run based on the optimized parameters to predict the surface residual stress during grinding. The resulting surface residual stress predictions were compared with experimental data to verify the effectiveness of the method.

[0110] In addition, the method for predicting residual stress of a silicon carbide grinding surface disclosed in an embodiment of the present invention can be widely used to predict residual stress of grinding of other materials, improve the controllability of the grinding process, optimize processing parameters, improve product quality, and provide important theoretical support and engineering application value for the high-end manufacturing industry.

[0111] An embodiment of the present invention discloses a method for predicting residual stress on the surface of silicon carbide after grinding. The method utilizes a Bayesian inference method and a Monte Carlo method to optimize the model parameters of a silicon carbide simulation model, and uses the silicon carbide simulation model with optimized parameters to perform grinding force simulation to obtain an optimized surface residual stress prediction result, thereby realizing the prediction of the surface residual stress of silicon carbide after grinding.

[0112] For example three, please refer to Figure 5 , Figure 5 This is a structural block diagram of an embodiment of a device for predicting residual stress on a silicon carbide polished surface according to the present invention, the device comprising:

[0113] An acquisition module 301 is used to acquire mechanical response data of silicon carbide material;

[0114] A construction module 302 is configured to construct a silicon carbide simulation model based on the mechanical response data;

[0115] A simulation module 303 is used to set grinding parameters to simulate the grinding force of the silicon carbide simulation model and obtain preliminary surface residual stress prediction results;

[0116] An optimization module 304 is configured to optimize the model parameters of the silicon carbide simulation model based on the preliminary surface residual stress prediction result and the grinding parameters by using a Bayesian inference method and a Monte Carlo method to obtain optimal model parameters;

[0117] The prediction module 305 is used to perform grinding force simulation on the silicon carbide simulation model using the optimal model parameters and the grinding parameters to obtain an optimized surface residual stress prediction result.

[0118] In an optional embodiment, the building module 302 includes:

[0119] A construction submodule is used to construct a preliminary silicon carbide simulation model based on the mechanical response data;

[0120] The first setting submodule is used to mesh the preliminary silicon carbide simulation model and set corresponding material parameters and load boundary conditions to obtain a silicon carbide simulation model.

[0121] In an optional embodiment, the simulation module 303 includes:

[0122] The second setting submodule is used to set grinding parameters; the grinding parameters include abrasive particle diameter, grinding depth, abrasive particle spacing and grinding conditions.

[0123] The simulation submodule is used to simulate the grinding force of the silicon carbide simulation model based on the abrasive particle diameter, the grinding depth, the abrasive particle spacing and the grinding conditions to obtain a preliminary surface residual stress prediction result.

[0124] In an optional embodiment, the optimization module 304 includes:

[0125] a calculation submodule, configured to calculate, based on the actual grinding residual stress data and the preliminary surface residual stress prediction result, a posterior distribution of model parameters of the silicon carbide simulation model using a Bayesian inference method;

[0126] A first optimization submodule is used to optimize the posterior distribution by using a Monte Carlo method to obtain an optimized posterior distribution;

[0127] The second optimization submodule is used to optimize the model parameters of the silicon carbide simulation model according to the optimized posterior distribution to obtain optimal model parameters.

[0128] In an optional embodiment, the prediction module 305 includes:

[0129] A third optimization submodule is configured to optimize the silicon carbide simulation model using the optimal model parameters to obtain an optimized silicon carbide simulation model;

[0130] The prediction submodule is used to perform grinding force simulation on the optimized silicon carbide simulation model based on the grinding parameters to obtain an optimized surface residual stress prediction result.

[0131] In an optional embodiment, the device for predicting residual stress of a silicon carbide grinding surface further comprises:

[0132] A verification module is used to perform silicon carbide grinding based on the grinding parameters, and after obtaining actual residual stress, compare the actual residual stress with the optimized surface residual stress prediction result to verify the accuracy of the optimized surface residual stress prediction result.

[0133] Embodiment 4. An embodiment of the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of a method for predicting residual stress of a silicon carbide grinding surface in any embodiment.

[0134] Embodiment 5. The embodiment of the present invention further provides a computer storage medium having a computer program stored thereon. When the computer program is executed by the processor, the steps of a method for predicting residual stress of a silicon carbide grinding surface in any embodiment are implemented.

[0135] Embodiment 6. The embodiment of the present invention further provides a computer program product having a computer program stored thereon, which implements the steps of a method for predicting residual stress of a silicon carbide grinding surface in any embodiment when the computer program is executed by the processor.

[0136] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0137] In the several embodiments provided in this application, it should be understood that the methods, devices, electronic devices and storage media disclosed in the present invention can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0138] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0139] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0140] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned readable storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0141] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting residual stress on a polished silicon carbide surface, characterized in that: include: Obtain mechanical response data of silicon carbide materials; Constructing a silicon carbide simulation model based on the mechanical response data; Setting grinding parameters to simulate grinding forces on the silicon carbide simulation model and obtain preliminary surface residual stress prediction results; Based on the preliminary surface residual stress prediction results, the Bayesian inference method is used to calculate the posterior distribution of the model parameters of the silicon carbide simulation model: by actually measuring the grinding parameters, the actual SiC residual stress data is obtained; a Bayesian optimization framework of the JH-2 model is constructed, the prior distribution of the JH-2 model is set to a uniform distribution, and a Gaussian distribution is constructed using the actual SiC residual stress data to measure the degree of fit between the preliminary surface residual stress prediction results and the actual SiC residual stress data. The likelihood function constructed using the Bayesian method is used, with the simulation prediction value as the prior data and the actual data as the likelihood data, to calculate the posterior distribution of the model parameters of the silicon carbide simulation model; The posterior distribution is optimized by the Monte Carlo method to obtain the optimized posterior distribution: the parameters of the JH-2 model are optimized by the MH sampling method in the Monte Carlo method, the sampling initial value and the adaptive parameter step size are set, and candidate parameters are generated in each iteration. At the same time, the parameter acceptance probability of the updated JH-2 model is calculated, and the parameters are obtained by the MH sampling method in the Monte Carlo method. As acceptance criteria, among others, are the current model parameters, is the updated model parameter. If the new parameter meets the optimization conditions, the model parameter is updated; otherwise, the original parameter value is retained. Optimizing the model parameters of the silicon carbide simulation model according to the optimized posterior distribution to obtain optimal model parameters; Optimizing the silicon carbide simulation model using the optimal model parameters to obtain an optimized silicon carbide simulation model; Based on the grinding parameters, grinding force simulation is performed on the optimized silicon carbide simulation model to obtain an optimized surface residual stress prediction result.

2. The method for predicting residual stress on a polished silicon carbide surface according to claim 1, wherein: Based on the mechanical response data, a silicon carbide simulation model is constructed, including: Constructing a preliminary silicon carbide simulation model based on the mechanical response data; The preliminary silicon carbide simulation model is meshed, and corresponding material parameters and load boundary conditions are set to obtain a silicon carbide simulation model.

3. The method for predicting residual stress on a polished silicon carbide surface according to claim 1, wherein: Grinding parameters are set to simulate the grinding force of the silicon carbide simulation model and obtain preliminary surface residual stress prediction results, including: Setting grinding parameters; the grinding parameters include abrasive particle diameter, grinding depth, abrasive particle spacing and grinding conditions; Based on the abrasive grain diameter, the grinding depth, the abrasive grain spacing and the grinding conditions, a grinding force simulation is performed on the silicon carbide simulation model to obtain a preliminary surface residual stress prediction result.

4. The method for predicting residual stress on a polished silicon carbide surface according to claim 1, wherein: After performing grinding force simulation on the silicon carbide simulation model using the optimal model parameters and the grinding parameters to obtain an optimized surface residual stress prediction result, the method further includes: Silicon carbide is ground based on the grinding parameters to obtain actual residual stress, and then the actual residual stress is compared with the optimized surface residual stress prediction result to determine the accuracy of the optimized surface residual stress prediction result.

5. A device for predicting residual stress on a silicon carbide polished surface, characterized in that: include: An acquisition module, used for acquiring mechanical response data of silicon carbide materials; A construction module, configured to construct a silicon carbide simulation model based on the mechanical response data; A simulation module, used to set grinding parameters to simulate the grinding force of the silicon carbide simulation model and obtain preliminary surface residual stress prediction results; The optimization module is used to calculate the posterior distribution of the model parameters of the silicon carbide simulation model based on the preliminary surface residual stress prediction result using the Bayesian inference method: the actual SiC residual stress data is obtained by actually measuring the grinding parameters; the Bayesian optimization framework of the JH-2 model is constructed, the prior distribution of the JH-2 model is set to a uniform distribution, and the actual SiC residual stress data is used to construct a Gaussian distribution to measure the degree of fit between the preliminary surface residual stress prediction result and the actual SiC residual stress data. The likelihood function constructed using the Bayesian method is used to calculate the posterior distribution of the model parameters of the silicon carbide simulation model with the simulation prediction value as the prior data and the actual data as the likelihood data; the posterior distribution is optimized by the Monte Carlo method to obtain the optimized posterior distribution: the MH sampling method in the Monte Carlo method is used to optimize the parameters of the JH-2 model, the sampling initial value and the adaptive adjustment parameter step size are set, and candidate parameters are generated in each iteration. At the same time, the parameter acceptance probability of the updated JH-2 model is calculated, and the updated parameter acceptance probability is calculated. As acceptance criteria, among others, are the current model parameters, The updated model parameters are updated if the new parameters meet the optimization conditions, otherwise the original parameter values ​​are retained; the model parameters of the silicon carbide simulation model are optimized according to the optimized posterior distribution to obtain the optimal model parameters; A prediction module is used to optimize the silicon carbide simulation model using the optimal model parameters to obtain an optimized silicon carbide simulation model; based on the grinding parameters, the optimized silicon carbide simulation model is subjected to grinding force simulation to obtain an optimized surface residual stress prediction result.

6. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 4 is executed.

7. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 4 is executed.

8. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images