Mechanism-guided few-sample polishing material removal rate prediction model construction method

Abstract

The invention discloses a low-cost and high-precision mechanism-guided few-sample polishing material removal rate prediction model construction method, which is based on a Preston equation, combines mechanism characteristics of a polishing technology and simulation modeling of a polishing process, deeply studies a relationship between polishing process parameters and a material removal rate, and provides a mechanism-guided small-sample polishing material removal rate prediction model construction method. The method comprises the following steps: establishing a mathematical model with certain precision for predicting the material removal rate, designing a neural network model structure based on a mathematical form of the mathematical model, and training the neural network model to approach the established mathematical model so as to obtain a pre-trained neural network model; and finally, based on a transfer learning strategy, continuing to train the neural network model on a small number of real samples to obtain a final material removal rate prediction model. The prediction model integrates the interpretability of a mathematical model and the nonlinear fitting capability of a neural network model, is high in prediction precision, and can provide reliable prediction support for the actual polishing process.

Description

Technical Field

[0001] The invention relates to the field of ultra-precision machining and artificial intelligence technology, and in particular to a mechanism-guided method for constructing a few-sample polishing material removal rate prediction model. Background Art

[0002] Shear thickening polishing technology is a non-contact ultra-precision polishing method that uses the nonlinear thickening rheological properties of non-Newtonian fluid polishing fluid under shear stress to achieve material removal. Deterministic shear thickening polishing means that the active rotation of the polishing head drives the flow of the polishing fluid, causing the flowing polishing fluid to produce a shear thickening effect, forming a particle cluster wrapped with abrasive particles, thereby performing deterministic removal processing on a small area of ​​the workpiece material.

[0003] In order to optimize the polishing process, it is usually necessary to predict the removal rate of the polishing material. In the prior art, some researchers have conducted relevant research based on mathematical physics theory and simulation methods. For example, in the prior art, someone designed the constitutive equation of shear thickening polishing based on the Preston equation, and then studied the velocity field and pressure field distribution during shear thickening deterministic polishing based on fluid simulation. Finally, the simulation coefficient was corrected through experiments to achieve the modeling of the prediction of shear thickening polishing material removal rate. Although it has strong theoretical support and good generalization based on principles and physical laws. However, due to the complexity of the shear thickening polishing process, the theoretical modeling method is often modeled under certain ideal conditions. It assumes that the material removal rate is only related to the polishing speed, gap height and polishing liquid concentration, and cannot take into account the various complex dynamic factors in the polishing process, such as changes in polishing liquid temperature, changes in rheological properties, etc., which makes it difficult to obtain high-precision prediction results in real scenarios.

[0004] In recent years, with the development of artificial intelligence technology, more and more researchers have begun to study modeling methods based on machine learning / deep learning. Deep learning models represented by neural networks have powerful nonlinear fitting capabilities and can update model parameters in continuous learning based on back propagation. This powerful nonlinear fitting ability can transform the potential laws between input and output into neural network parameters through continuous learning from the perspective of real data, without relying on various restrictive assumptions, thereby achieving high-precision prediction of real scenes. However, its modeling process relies on a large number of high-quality data samples for training, and the real high-quality experimental data samples must rely on professional experimental personnel to conduct polishing experiments, which requires a lot of manpower and material resources and is often difficult to obtain. Generally, only a small amount of experimental data is available. If modeling is directly based on deep learning in the case of few samples, it often leads to overfitting / underfitting problems, and a high-precision prediction model cannot be obtained. This makes it difficult for artificial intelligence technology to be effectively applied in the prediction modeling of shear thickening polishing material removal rate. Therefore, researchers proposed transfer learning technology, which is to first train the model on a large number of easily accessible data samples of similar scenes to obtain a pre-trained model, and then fine-tune the model on a small number of hard-to-obtain real target samples, so as to achieve better modeling effects using a small number of samples on the target scene. However, in the field of shear thickening polishing material removal rate prediction modeling, the above method cannot be copied, because on the one hand, the polishing field lacks such a public large-scale data set for pre-training, and on the other hand, the principles and characteristics of various polishing methods are significantly different, making it difficult to establish a public large-scale data set for unified modeling. In summary, how to use the nonlinear fitting ability of the neural network model and build a high-precision material removal rate prediction model based on a small amount of shear thickening polishing experimental sample data is a problem that needs to be solved urgently. Summary of the invention

[0005] The present invention provides a low-cost, high-precision mechanism-guided method for constructing a few-sample polishing material removal rate prediction model, which is used to solve the problem of lack of real experimental samples in the traditional neural network-based material removal rate modeling method, and the problem of low modeling accuracy caused by the inability of pure mathematical physics modeling methods to reflect the real complex processing process. The prediction model construction method of the present invention is first based on the Preston equation, and combines the mechanism characteristics of the polishing technology and the simulation modeling of the polishing process to construct a mathematical model for material removal rate prediction with a certain accuracy; then, based on the mathematical form of the mathematical model, a neural network model is designed and the neural network pre-training process is guided to obtain a pre-trained neural network model with a certain prediction accuracy; finally, based on the transfer learning strategy, the neural network model is continued to be trained on a small number of real samples to obtain the final polishing material removal rate prediction model; the prediction model integrates the interpretability of the mathematical model and the nonlinear fitting ability of the neural network model, has high prediction accuracy, and can provide reliable prediction support for the actual polishing process, thereby guiding production practice, optimizing the polishing process, and improving production efficiency and product quality.

[0006] The technical solution of the present invention is:

[0007] A mechanism-guided method for constructing a few-sample polishing material removal rate prediction model, the method comprising the following steps: S1, determining the polishing object, designing a polishing experiment, and constructing a real experimental data set D, each sample in the real experimental data set D contains a real polishing process parameter combination and a corresponding material removal rate, and then combining the polishing principle and Preston equation: , establish a mathematical model M for predicting polishing material removal rate; where MRR is the material removal rate, P is the polishing pressure, V is the polishing speed, and k is the coefficient; S2, set different polishing process parameter combinations to input the mathematical model M, and output the predicted results of the material removal rate to obtain a simulated experimental data set Dm, each experimental data contains a set of simulated polishing process parameter combinations and the corresponding material removal rate prediction results; S3, according to the mathematical form of the mathematical model M, design the initial state of the neural network model N, and train it on the simulated experimental data set Dm until convergence, so that the trained model N infinitely approaches the mathematical model M, and obtains the pre-trained model N_1; S4, fine-tune the pre-trained model N_1 on the real experimental data set D, and optimize the model parameters based on the real experimental data, so as to further improve the prediction accuracy of the pre-trained model N_1, and obtain the final polishing material removal rate prediction model N_2.

[0008] Compared with the prior art, the mechanism-guided few-sample polishing material removal rate prediction model construction method of this application organically combines theoretical modeling with neural network training, realizes high-precision prediction modeling of polishing material removal rate under few-sample conditions, and provides strong technical support for the optimization of polishing process. It has the following advantages:

[0009] (1) The present invention constructs a polishing material removal rate prediction mathematical model based on theoretical analysis and simulation modeling, and on this basis, guides the pre-training process of the neural network model so that it has a certain prediction accuracy before contacting real experimental data, and fine-tunes the training on a small number of real samples to further improve the prediction accuracy. The final prediction model not only has the high-precision prediction ability of the neural network, but also inherits the interpretability of the mathematical model. This interpretability is crucial for understanding the decision-making process of the model, verifying the rationality of the model, and establishing users' trust in the model's prediction results;

[0010] (2) Using simulated data for pre-training reduces the reliance on real experimental samples, thereby reducing experimental costs and time consumption. In addition, the model can learn a wider range of polishing conditions, enhancing the model's ability to adapt to unknown conditions;

[0011] (3) The present invention reduces the number of experiments and material consumption, conforms to the green and environmentally friendly production concept, is not only suitable for the prediction of shear thickening polishing material removal rate, but can also be extended to the construction of prediction models for other material processing processes, and has high practical value and market prospects.

[0012] As an optimization, in the aforementioned mechanism-guided method for constructing a few-sample polishing material removal rate prediction model, in step S1, the polishing experiment is a shear thickening polishing experiment, and the formula of the mathematical model M based on the expansion of the Preston equation is: ; Where Kc is the abrasive concentration correction coefficient, Ks is the simulation correction coefficient, K(w,h) is the correction coefficient related to the polishing speed and polishing gap, P(x,y) is the pressure distribution on the xy plane, and V(x,y) is the velocity distribution on the xy plane. Furthermore, the simulation correction coefficient Ks and the abrasive concentration correction coefficient Kc are obtained by modeling the pressure and velocity fields in the polishing area and fitting them in combination with the polishing experimental data.

[0013] As an optimization, in the aforementioned mechanism-guided method for constructing a few-sample polishing material removal rate prediction model, in step S3, the formula of the neural network model N in the initial state is: ; Wherein, X represents various polishing process parameter combinations; represents a custom activation function related to variables Xi, Xj, Xz designed according to the mathematical form of the model M; α represents a nonlinear activation function; F represents the forward calculation process of the multilayer perceptron.

[0014] Furthermore, in step S3, the simulation experiment data set Dm is divided into a simulation training set Dm_train and a simulation test set Dm_test; the neural network model N is first trained on the simulation training set Dm_train until the difference between the prediction result of the trained neural network model on the test set Dm_test and the prediction result of the mathematical model M is less than the set threshold T, and the pre-trained model N_1 is obtained, that is, At this time, the acquisition of the pre-trained model N_1 not only depends on the training effect of the model on the simulated data, but also needs to be verified through the test set, so as to ensure the generalization ability of the model and the prediction accuracy of the model.

[0015] Furthermore, the training process of the neural network model N includes forward propagation, loss calculation, back propagation and parameter update, and the model parameters are continuously optimized through iterative training, so that the neural network model N can gradually improve its performance and accuracy.

[0016] As an optimization, in the aforementioned mechanism-guided few-sample polishing material removal rate prediction model construction method, in the step S4, the real experimental data set D is divided into a real training set D_train and a real test set D_test; the pre-trained model N_1 is first fine-tuned on the real training set D_train until the prediction accuracy P_n of the pre-trained model N_1 on the real test set D_test is greater than the prediction accuracy P_m of the mathematical model M on the real test set D_test and the accuracy no longer increases, thereby obtaining the final material removal rate prediction model N_2.

[0017] The specific judgment process is as follows: T and D N The MSE of is used as the loss of the pre-trained model N_1 during fine-tuning training: ; Among them, D T represents the material removal rate label value on the real test set D_test, D N is the output value of the pre-trained model N_1, D M is the output value of the mathematical model M. MSE (Mean Square Error) represents the mean square error, which is a common indicator for measuring the prediction accuracy of the model. When it reaches When , it is considered that P_n is greater than P_m.

[0018] As an optimization, in the aforementioned mechanism-guided method for building a few-sample polishing material removal rate prediction model, when dividing the simulated experimental data set and the real experimental data set, the division ratio of the training set and the test set can be adjusted according to the actual situation, but it is usually maintained at 65-75% for training and 25-35% for testing. Specifically: the data ratio in the simulated training set Dm_train and the simulated test set Dm_test is 7:3; the data ratio in the real training set D_train and the real test set D_test is 7:3, so as to ensure the effectiveness of model training and the accuracy of testing.

[0019] As an optimization, in the aforementioned mechanism-guided method for constructing a small sample polishing material removal rate prediction model, in step S1, the shear thickening polishing experiment is implemented using a shear thickening deterministic polishing device; the deterministic polishing device includes a polishing tank, a polishing head disposed above the polishing tank, and a lifting mechanism for driving the polishing head to move up and down; a fixture is provided inside the polishing tank, and a pressure sensor is provided at the bottom; the polishing head is connected to a motor, and a level gauge is provided on the motor; during the experiment, the workpiece is fixed in the polishing tank by a fixture, and polishing liquid is filled in the polishing tank to immerse the workpiece, and the lower part of the polishing head is located in the polishing liquid; the motor then drives the polishing head to rotate to remove the material on the workpiece surface. The polishing experiment using the polishing device of the above structure is easy to implement and easy to obtain various polishing process parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a schematic diagram of a mechanism-guided method for constructing a few-sample polishing material removal rate prediction model of the present application;

[0021] Figure 2 is a schematic structural diagram of a deterministic polishing device in an embodiment of the present application;

[0022] Figure 3 It is a schematic diagram of the neural network model in the initial state designed according to the mathematical model.

[0023] The markings in the attached drawings are: 1-polishing tank; 2-polishing head; 3-clamp; 4-pressure sensor; 5-motor; 6-level measuring instrument; 7-workpiece; 8-polishing liquid. DETAILED DESCRIPTION

[0024] The present invention is further described below in conjunction with the accompanying drawings and embodiments; the contents not described in detail in the following embodiments are all common technical knowledge in the art.

[0025] In order to solve the problems of scarce sample size for material removal rate prediction modeling in the shear thickening polishing process and low prediction accuracy of modeling based on mathematical physics theory, the present invention proposes a mechanism-guided method for constructing a small-sample polishing material removal rate prediction model. The core of this method is to combine theoretical modeling with artificial intelligence algorithms. This method can achieve high-precision prediction modeling of polishing material removal rate with a small number of real samples, providing strong technical support for the optimization of polishing process.

[0026] This method is implemented through the following key steps:

[0027] Firstly, based on the Preston equation, the present invention combines the mechanism characteristics of the polishing technology and the simulation modeling of the polishing process to conduct an in-depth study on the relationship between the polishing process parameters and the material removal rate, and establishes a mathematical model M for predicting the polishing material removal rate with a certain accuracy, thereby providing a theoretical basis for the subsequent prediction model construction and enhancing the interpretability of the constructed prediction model.

[0028] Next, the present invention adopts simulation data generation technology to generate a large number of simulation experimental data sets based on the mathematical model M; the key to this step lies in the diversity and representativeness of the simulation data, which ensures that the simulation data sets can cover various possible polishing conditions and provide rich data resources for the training of the neural network model.

[0029] On this basis, the present invention designs and trains a neural network model N, which refers to the mathematical model M in terms of structural design, in order to achieve rapid convergence through training on simulated experimental data, so that the model can infinitely approach the prediction accuracy of the mathematical model M, thereby obtaining a pre-trained model N_1; this process makes full use of the advantages of neural networks in processing complex nonlinear relationships, and through training with simulated data, the pre-trained model N_1 has a prediction ability equivalent to that of the mathematical model.

[0030] Finally, the present invention fine-tunes the pre-trained model N_1 on a small amount of real experimental data; the purpose of this step is to improve the prediction accuracy of the model on real experimental data, and by fine-tuning the model parameters, the model can better fit the complex real processing process, and finally obtain a higher accuracy polishing material removal rate prediction model; this prediction model not only has the high interpretability of the mathematical model M, but also realizes the high-precision prediction capability based on neural network.

[0031] Example:

[0032] A mechanism-guided method for constructing a prediction model for a small number of sample polishing material removal rates includes the following steps.

[0033] S1, determine the polishing object and design the shear thickening polishing experiment. The experimental data are shown in the following table. Based on this, a real experimental data set D is constructed. Each sample in the real experimental data set D contains a set of real polishing process parameter combinations and corresponding material removal rates. Then, the shear thickening polishing principle and Preston equation are combined: , a mathematical model M of shear thickening polishing material removal rate is established; where MRR is the material removal rate, P is the polishing pressure, V is the polishing speed, and k is a coefficient related to the polishing object and the polishing liquid.

[0034] Experimental conditions parameter Glass 1.5 (inch) Abrasive <![CDATA[Al2O3]]> Abrasive grit size #3000 Abrasive Concentration 3, 5, 7, 9 (wt%) Polishing head speed 4, 8, 12, 16 (r / s) Polishing gap 0.6, 0.8, 1, 1.2 (mm)

[0035] When designing a polishing experiment, it is necessary to consider the characteristics of the polishing object, including but not limited to material hardness, surface roughness, chemical composition, etc. These characteristics will directly affect the parameter setting of the polishing experiment and the final material removal rate.

[0036] In this embodiment, the shear thickening polishing experiment is implemented by a shear thickening deterministic polishing device; the deterministic polishing device includes a polishing tank 1, a polishing head 2 arranged above the polishing tank 1, and a lifting mechanism for driving the polishing head 2 to move up and down; a clamp 3 is provided inside the polishing tank 1, and a pressure sensor 4 is provided at the bottom; the polishing head 2 is connected to a motor 5, and a level measuring instrument 6 is provided on the motor 5; during the experiment, the workpiece 7 is fixed in the polishing tank 1 by the clamp 3, and the polishing liquid 8 is filled in the polishing tank 1 to immerse the workpiece 7, and the lower part of the polishing head 2 is located in the polishing liquid 8, leaving a polishing gap between the workpiece 7; the motor 5 then drives the polishing head 2 to rotate, so that the polishing liquid 8 between the polishing head 2 and the workpiece 7 produces a shear rheological effect, thereby removing the surface material of the workpiece 7.

[0037] The polishing gap height between the workpiece 7 and the polishing head 2 and the rotation speed of the polishing head 2 respectively affect the pressure distribution and flow field distribution of the polishing liquid 8 on the surface of the workpiece 7, which are important factors affecting the material removal rate. In addition, the concentration of the polishing liquid 8 is also an important factor affecting the material removal rate. Based on the shear thickening polishing principle, the Preston equation is expanded and the constitutive equation of shear thickening polishing material removal rate is established as: ; Wherein, Kc is the abrasive concentration correction coefficient, Ks is the simulation correction coefficient, K(w,h) is the correction coefficient related to the polishing speed and polishing gap, P(x,y) is the pressure distribution on the xy plane, and V(x,y) is the velocity distribution on the xy plane.

[0038] The pressure and velocity fields in the polishing area are modeled, and the modeling results are fitted into a mathematical form and substituted into the above-mentioned material removal rate constitutive equation. Then, the simulation correction coefficient Ks and the abrasive concentration correction coefficient Kc are fitted in combination with the polishing experimental data. The final mathematical model M is: Wherein, C represents the abrasive concentration; w and h represent the rotation speed of the polishing head and the polishing gap height respectively; x and y represent the coordinates of the polishing area.

[0039] S2, set different polishing process parameter combinations to input mathematical model M, and output the predicted results of material removal rate to obtain the simulated experimental data set Dm. Each experimental data contains a set of simulated polishing process parameter combinations and the corresponding predicted results of material removal rate; the simulated experimental data set Dm is divided into a simulated training set Dm_train and a simulated test set Dm_test in a ratio of 7:3. The generation of the simulated experimental data set needs to consider the randomness and diversity of the polishing process parameters to ensure that the generated data set can cover a wide range of experimental conditions, thereby improving the generalization ability of the model.

[0040] S3, according to the mathematical form of the mathematical model M, design the neural network model N in the initial state. When designing, refer to the form of the mathematical model M to select and design the appropriate network architecture, activation function, loss function and optimization algorithm, etc., to ensure that the model can effectively learn and approximate the mathematical model M; the formula is: ; Wherein, X represents various polishing process parameter combinations; It represents the custom activation function related to variables Xi, Xj, Xz designed according to the mathematical form of model M; α represents a nonlinear activation function, such as relu, which is used to provide nonlinear capabilities and fit real experimental data; F represents the forward calculation process of the multilayer perceptron.

[0041] In this embodiment, according to the mathematical form of the mathematical model M, the activation functions θ and λ are customized to simulate the influence of factors such as the polishing head rotation speed, the polishing gap height, and the abrasive concentration on the material removal rate, so as to improve the fitting ability of the neural network model N to the mathematical model M. By observing the form of the mathematical model M, it can be found that w, h and x, y and the material removal rate all satisfy a certain Gaussian distribution. Therefore, the Gaussian surface equations can be set as the activation functions θ and λ respectively, and designed in the neural network model N to simulate the mathematical representation of the mathematical model M: ;

[0042]

[0043] Then, let the neural network model N be trained on the simulated training set Dm_train, with the model parameters W and the accuracy threshold T set; the training process includes forward propagation, loss calculation, back propagation and parameter update, and through iterative training, the model parameters are continuously optimized until the prediction performance of the model on the simulated experimental data is infinitely close to the mathematical model M, and the pre-trained model N_1 is obtained, that is, M≈N_1 in prediction accuracy; the infinitely close mathematical model M means that on the same batch of simulated test sets Dm_test, the difference between the prediction results of the neural network pre-trained model N_1 and the mathematical model M is less than the set threshold T, that is , the neural network model parameters are updated to W_1. At this time, the acquisition of the pre-trained model N_1 not only depends on the training effect of the model on the simulated data, but also needs to be verified by the test set to ensure the generalization ability and prediction accuracy of the model.

[0044] S4, divide the real experimental data set D into a real training set D_train and a real test set D_test in a ratio of 7:3; then, let the pre-trained model N_1 be fine-tuned on the real training set D_train, optimize the model parameters based on the real experimental data, and further improve the prediction accuracy of the pre-trained model N_1 until the prediction accuracy P_n of the pre-trained model on the real test set D_test is greater than the prediction accuracy P_m of the mathematical model M on the real test set D_test and no longer increases, then the final material removal rate prediction model N_2 is obtained, the neural network model parameters are updated from W_1 to W_2, and the modeling is completed.

[0045] The fine-tuning training process needs to be carefully designed, including choosing the appropriate learning rate, fine-tuning strategy and training cycle, to ensure that the model can quickly adapt to a small amount of real data and improve the prediction accuracy. T and D N The MSE is used as the loss of the predicted model during fine-tuning training: ; Among them, D T represents the material removal rate label value on the real test set D_test, D N is the output value of the prediction model (the prediction value obtained by inputting the data on the real test set into the prediction model), D M is the output value of the mathematical model M (the predicted value obtained by inputting the data on the real test set into the mathematical model M). MSE (Mean Square Error) represents the mean square error, which is a common indicator for measuring the prediction accuracy of the model. The smaller this value is, the higher the prediction accuracy of the model is. When it reaches When , it is believed that the prediction accuracy of the final trained model is further improved on the basis of the mathematical model constructed based on the polishing process mechanism theory.

[0046] In this embodiment, in the real test set D_test, the real values ​​of the material removal rate are [10.7, 5.5, 10.2, 5.9, 8.45, 8.27, 9.96, 10.57, 6.0] (μm / h); the predicted values ​​of the mathematical model M on the real test set D_test are [10.27656018, 5.54128245, 10.17580959, 5.94428481, 8.46304956, 7.40925152, 8.0060315, 11.38117627, 6.0450354]; the predicted values ​​of the pre-trained model N_1 on the real test set D_test are [10.15728069, 4.79993754, 9.72893041, 5.53493383, 8.34791813, 7.09903125, 8.96807892, 10.51678517, 5.19792818]; the prediction value of the final prediction model N_2 on the real test set D_test is [10.27388505, 5.00993888, 9.64422017, 6.07567224, 8.85241334, 7.85554219, 9.678 999, 10.15477847, 5.4139422]; MSE is used to measure the difference between the model prediction value and the true value, MSE (true value, predicted value of the mathematical model) = 0.6, MSE (true value, predicted value of the pre-trained model) = 0.461, MSE (true value, predicted value of the final prediction model) = 0.188; It can be seen that the final material removal rate prediction model obtained by the construction method of the present application has a higher prediction accuracy.

[0047] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A mechanism-guided method for constructing a few-sample polishing material removal rate prediction model, characterized in that: The method comprises the following steps: S1, determine the polishing object, design the polishing experiment, and construct a real experimental data set D. Each sample in the real experimental data set D contains a real polishing process parameter combination and the corresponding material removal rate, and then combines the polishing principle and Preston equation: , a mathematical model M for predicting polishing material removal rate is established; where MRR is the material removal rate, P is the polishing pressure, V is the polishing speed, and k is the coefficient; S2, setting different polishing process parameter combinations to input into the mathematical model M, and outputting the predicted result of the material removal rate, to obtain a simulated experimental data set Dm, where each experimental data contains a set of simulated polishing process parameter combinations and the corresponding predicted result of the material removal rate; S3, according to the mathematical form of the mathematical model M, design the neural network model N in the initial state, and train it on the simulated experimental data set Dm until convergence, so that the trained model N is infinitely close to the mathematical model M, and obtain the pre-trained model N_1; S4, fine-tune the pre-trained model N_1 on the real experimental data set D, optimize the model parameters based on the real experimental data, so as to further improve the prediction accuracy of the pre-trained model N_1, and obtain the final polishing material removal rate prediction model N_2, and the modeling is completed.

2. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 1, characterized in that: In step S1, the polishing experiment is a shear thickening polishing experiment, and the formula of the mathematical model M based on the expansion of the Preston equation is: ; Wherein, Kc is the abrasive concentration correction coefficient, Ks is the simulation correction coefficient, K(w,h) is the correction coefficient related to the polishing speed and polishing gap, P(x,y) is the pressure distribution on the xy plane, and V(x,y) is the velocity distribution on the xy plane.

3. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 2, characterized in that: The pressure and velocity fields in the polishing area are modeled, and the simulation correction coefficient Ks and abrasive concentration correction coefficient Kc are obtained based on the polishing experimental data.

4. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 1, characterized in that: In step S3, the formula of the neural network model N in the initial state is: ; Wherein, X represents various polishing process parameter combinations; represents a custom activation function related to variables Xi, Xj, Xz designed according to the mathematical form of the model M; α represents a nonlinear activation function; F represents the forward calculation process of the multilayer perceptron.

5. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 4, characterized in that: In step S3, the simulation experiment data set Dm is divided into a simulation training set Dm_train and a simulation test set Dm_test; the neural network model N is trained on the simulation training set Dm_train until the difference between the prediction result of the trained neural network model on the test set Dm_test and the prediction result of the mathematical model M is less than the set threshold T, and the pre-trained model N_1 is obtained, that is, .

6. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 5, characterized in that: The training process of the neural network model N includes forward propagation, loss calculation, back propagation and parameter update, and the model parameters are continuously optimized through iterative training.

7. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 5, characterized in that: In the step S4, the real experimental data set D is divided into a real training set D_train and a real test set D_test; the pre-trained model N_1 is first fine-tuned on the real training set D_train until the prediction accuracy P_n of the pre-trained model N_1 on the real test set D_test is greater than the prediction accuracy P_m of the mathematical model M on the real test set D_test and no longer increases, thereby obtaining the final material removal rate prediction model N_2.

8. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 7, characterized in that: Using D T and D N The MSE of is used as the loss of the pre-trained model N_1 during fine-tuning training: ;D T represents the material removal rate label value on the real test set D_test, D N is the output value of the pre-trained model N_1, D M is the output value of the mathematical model M, MSE represents the mean square error; when it is achieved on the real test set D_test , it is considered that P_n is greater than P_m.

9. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 7, characterized in that: The data ratio in the simulated training set Dm_train and the simulated test set Dm_test is 7:3; the data ratio in the real training set D_train and the real test set D_test is 7:

3.

10. The method for constructing a mechanism-guided few-sample polishing material removal rate prediction model according to claim 2, characterized in that: In step S1, the shear thickening polishing experiment is implemented by using a shear thickening deterministic polishing device; the deterministic polishing device comprises a polishing tank (1), a polishing head (2) arranged above the polishing tank (1), and a lifting mechanism for driving the polishing head (2) to move up and down; a fixture (3) is provided inside the polishing tank (1), and a pressure sensor (4) is provided at the bottom; the polishing head (2) is connected to a motor (5), and a level measuring instrument (6) is provided on the motor (5); during the experiment, a workpiece (7) is fixed in the polishing tank (1) by the fixture (3), and a polishing liquid (8) is placed in the polishing tank (1) to immerse the workpiece (7), and the lower part of the polishing head (2) is located in the polishing liquid (8); then the motor (5) drives the polishing head (2) to rotate to remove the surface material of the workpiece (7).

Images