Method for deterministic cutting compensation machining of infrared thin-wall optical elements

Abstract

The application discloses a method for machining infrared thin-wall optical elements by deterministic cutting compensation, which comprises the following steps: S1, obtaining the initial surface shape of a workpiece; the initial surface shape of a thin-wall optical element to be machined is measured by using an in-situ measurement technology of an ultra-precision machine tool, the surface profile p(x, y) is obtained, the material removal amount Δz(x, y) at each coordinate position is calculated according to the designed surface profile z(x, y), and Δz(x, y) = p(x, y) - z(x, y); S2, predicting the elastic deformation of clamping stress; based on a finite element simulation model, the three-dimensional spatial numerical distribution simulation of the clamping force distribution in the machining process is carried out, and the elastic deformation ε(x, y) of the element caused by the clamping stress is predicted; the elastic deformation amount is comprehensively combined with the initial material removal amount to obtain a new material removal amount z'(x, y); S3, predicting the deformation of the element caused by the residual stress in machining; S4, generating a non-uniform material removal compensation track; and S5, non-uniform cutting compensation machining based on slow tool servo.

Description

Technical Field

[0001] This invention belongs to the field of optical processing technology, specifically relating to a deterministic cutting compensation processing method for infrared thin-walled optical elements. Background Technology

[0002] Infrared thin-walled curved surface optical elements are widely used in high-end optical instruments, infrared detection systems, and other fields. The surface accuracy and quality of these elements directly affect their optical performance, thus requiring extremely high manufacturing precision. Typically, the surface roughness of infrared thin-walled optical elements needs to be at the nanometer level, and the surface accuracy needs to be controlled within the micrometer or even sub-micrometer range. During the manufacturing process, any minute errors introduced due to cutting, clamping, or other processes will affect the final optical performance. Furthermore, no impurities, cracks, or ripples can be introduced during manufacturing to ensure the accuracy and long-term stability of the optical system.

[0003] Currently, ultra-precision single-point diamond cutting technology is the primary method for manufacturing infrared thin-walled optical components. However, in the machining process of thin-walled structures, the sources of surface shape errors include not only conventional machining errors but also residual machining stress and elastic deformation caused by clamping. There is an urgent need to propose a non-uniform cutting compensation machining technology for infrared thin-walled optical components to compensate for surface shape errors, thereby improving machining accuracy and meeting the manufacturing requirements at the micron and even sub-micron levels. This technology aims to solve the surface shape errors caused by residual stress and clamping stress during the ultra-precision cutting process of infrared thin-walled optical components, ultimately achieving surface shape accuracy control and compensation.

[0004] The Influence of Residual Stress on Surface Shape Error: Residual stress is internal stress generated during machining, which typically affects the surface shape of components after material removal due to uneven deformation. Although the amplitude of residual stress can be reduced to some extent by optimizing machining parameters, completely eliminating residual stress is virtually impossible. In ultra-precision machining, the minute deformations caused by residual stress can lead to surface shape errors, particularly noticeable in thin-walled optical components. Current research mainly focuses on reducing residual stress through process optimization, but as the requirements for surface shape control accuracy continue to increase, the limitations of this approach become increasingly apparent.

[0005] The Influence of Clamping Stress on Facial Shape Error: In the machining of thin-walled components, clamping stress is another significant source of error. Existing methods for controlling clamping stress typically reduce facial shape error by optimizing the fixture support structure and minimizing uneven distribution of clamping force. However, unavoidable stress concentration during clamping can induce elastic deformation, causing changes in the component's facial shape during machining. Although flexible support fixture designs can mitigate this effect, they cannot completely eliminate the influence of clamping stress on facial shape error.

[0006] In the manufacturing of high-precision infrared optical components, especially thin-walled structures, effectively compensating for residual stress and clamping stress introduced during processing is a critical technical bottleneck that the industry urgently needs to address. These surface shape errors significantly limit the processing accuracy of the components and hinder the performance improvement of infrared optical systems. Therefore, proposing a technical solution capable of accurately compensating for these errors has significant theoretical value and practical application implications. Summary of the Invention

[0007] The purpose of this invention is to solve the above-mentioned problems and provide a deterministic cutting compensation machining method for infrared thin-walled optical elements by comprehensively analyzing and predicting residual stress and clamping stress during the machining process, establishing a corresponding compensation model, and combining deep learning and finite element simulation technology to achieve precise control of surface shape error.

[0008] To solve the above-mentioned technical problems, the technical solution of the present invention is: a deterministic cutting compensation machining method for infrared thin-walled optical elements, comprising the following steps:

[0009] S1. Obtaining the initial surface shape of the workpiece;

[0010] Using in-situ measurement technology on ultra-precision machine tools, the initial surface shape of the thin-walled optical element to be processed is measured to obtain its surface profile ρ(x,y). Based on the designed surface profile z(x,y), the material removal amount Δz(x,y) at each coordinate position is calculated.

[0011] Δz(x,y)=ρ(x,y)-z(x,y);

[0012] S2, Prediction of elastic deformation under clamping stress;

[0013] Based on the finite element simulation model, a three-dimensional numerical distribution simulation of the clamping force distribution during the machining process is performed, and the elastic deformation ε(x,y) of the component caused by clamping stress is predicted. This elastic deformation is combined with the initial material removal amount to obtain a new material removal amount z′(x,y), thereby eliminating the influence of clamping stress on the surface shape error of the component machining and providing a basis for the next compensation calculation.

[0014] z′(x,y)=Δz(x,y)+ε(x,y);

[0015] S3. Deformation prediction of components under residual stress during processing;

[0016] S4. Generation of compensation trajectory for non-uniform material removal;

[0017] S5. Non-uniform cutting compensation machining based on slow tool servo.

[0018] Furthermore, step S3 includes the following sub-steps:

[0019] S31. Construct a database T relating ultra-precision cutting parameters, tool geometry parameters, and residual stress; cutting parameters include cutting speed, depth of cut, and feed rate, while tool geometry parameters include nose radius, rake angle, and clearance angle.

[0020] S32. Based on the process parameter-residual stress correlation database T, a deep learning model is trained using this database to predict the residual stress under different process parameters.

[0021] S33. Using the residual stress prediction model trained in step S32, combined with the finite element analysis method, calculate the component deformation δ(x,y) caused by the residual stress under the given processing parameters.

[0022] Furthermore, step S32 also includes the following sub-steps:

[0023] S321. Data Collection and Preparation:

[0024] Database T should contain various parameters related to the cutting process, such as cutting speed, depth of cut, feed rate, and tool geometry parameters, including nose radius, rake angle, and clearance angle. The residual stress data of the material under these machining conditions should be obtained experimentally, or the correspondence between these parameters and residual stress should be calculated through theoretical models.

[0025] S322. Selection of Deep Learning Models:

[0026] Choose a deep learning model suitable for nonlinear mapping relationships, such as a multilayer perceptron (MLP), convolutional neural network (CNN), or long short-term memory network (LSTM), depending on the complexity of the data; the model input layer corresponds to process parameters, such as cutting speed, cutting depth, and feed rate, and the output layer is the predicted residual stress;

[0027] S333, Model Training:

[0028] To train the model using data from database T, the data needs to be preprocessed, including normalization and standardization, to improve the training efficiency of the model. A loss function, such as mean squared error (MSE), should be selected, and backpropagation algorithms, such as Adam or RMSprop, should be used to optimize the model. Cross-validation should be used during the training process to prevent overfitting, and the predictive performance of the model can be optimized by adjusting hyperparameters such as learning rate, number of layers, and number of neurons.

[0029] S334, Model Validation:

[0030] The model is validated using data not used in training to ensure its accuracy and robustness in predicting residual stress. The model's performance is evaluated by calculating the error between the predicted and actual residual stress values, such as mean square error or mean absolute error.

[0031] Furthermore, the error between the predicted value and the actual residual stress value calculated in step S334 is specifically as follows:

[0032] Suppose the trained deep learning model is represented as f θ Where θ represents the model parameters (weights and biases), then for a given set of processing parameters (p1, p2, ..., p...), n (e.g., cutting speed, depth of cut), the residual stress prediction of the model can be expressed as:

[0033]

[0034] in, The predicted residual stress;

[0035] The optimization objective during training is to minimize the loss function L, which is typically the mean square error (MSE) between the predicted and actual residual stress values.

[0036]

[0037] Where m is the number of training samples, R (i) Let be the actual residual stress of the i-th sample.

[0038] Furthermore, step S33 also includes the following sub-steps:

[0039] S331. Residual stress prediction:

[0040] Based on the deep learning model obtained from S02, the corresponding residual stress distribution is obtained by inputting the cutting process parameters used in the actual machining process.

[0041]

[0042] in, This is the predicted value of residual stress at the workpiece position (x, y); the distribution of residual stress may vary at different positions on the workpiece, so it is necessary to calculate the corresponding residual stress at different coordinate positions (x, y);

[0043] S332, Finite Element Analysis Calculation:

[0044] The predicted residual stress is input into the finite element model as a boundary condition or load condition. The influence of the residual stress on the deformation of the thin-walled optical element is calculated through finite element analysis. The finite element model should consider material properties (such as elastic modulus, Poisson's ratio, etc.) and geometry, and calculate the elastic deformation of the material through static analysis. Through the above methods, the influence of residual stress on the surface deformation of the element under given processing parameters can be accurately predicted and calculated, providing basic data for subsequent compensation processing.

[0045] Furthermore, through finite element analysis, the finite element model can obtain the deformation δ(x,y) caused by residual stress at different locations (x,y), that is, the deformation of the material during processing. This process depends on the results of the finite element analysis software and is usually expressed as:

[0046]

[0047] Here, FEA represents the finite element analysis process. For the predicted residual stress, material properties include the elastic modulus E and Poisson's ratio ν.

[0048] Furthermore, in step S4, the generation of the non-uniform material removal compensation trajectory is based on the combined effects of the aforementioned clamping stress and residual stress, using a genetic algorithm to iteratively optimize the material removal amount, and calculating the final compensated material removal amount Δz″(x,y), such that:

[0049] Δz″(x,y)=Δz′(x,y)+δ(x,y)

[0050] This result will be used as a compensation trajectory for non-uniform cutting and input into the machining equipment to ensure the achievement of final surface accuracy.

[0051] Furthermore, the non-uniform cutting compensation machining based on slow tool servo in step S5 is performed by using the freeform surface cutting technology of slow tool servo to perform non-uniform cutting along the compensation trajectory calculated above; the slow tool servo system can accurately control the movement trajectory of the tool, remove material layer by layer, and accurately realize the surface shape compensation of thin-walled optical elements.

[0052] The beneficial effects of this invention are:

[0053] 1. The deterministic cutting compensation machining method for infrared thin-walled optical elements provided by this invention first analyzes the main sources of surface shape errors in infrared thin-walled optical elements, including machining residual stress and clamping stress. By decomposing the surface shape error, these two types of stress are modeled and predicted separately, serving as the basis for subsequent compensation processes. Through comprehensive analysis and prediction of residual stress and clamping stress during machining, corresponding compensation models are established, and combined with deep learning and finite element simulation techniques, precise control of surface shape errors is achieved.

[0054] 2. By accurately predicting and compensating for residual stress and clamping stress, this invention can effectively reduce surface shape errors introduced during processing, thereby significantly improving the final surface shape accuracy of infrared thin-walled optical elements and meeting the accuracy requirements at the micron or even sub-micron level.

[0055] 3. By establishing an accurate surface shape error compensation model, this invention reduces the need for repeated measurement and correction of surface shape errors in traditional processes, greatly improving processing efficiency and reducing production costs.

[0056] 4. This invention can adapt to complex surface requirements. The compensation machining method based on slow tool servo can adapt to the machining needs of any complex surface. It is not only suitable for regular curved surfaces, but also for ultra-precision machining of free curved surfaces.

[0057] 5. The invention reduces costs. While achieving high-precision machining, the invention reduces material waste and the time spent on multiple correction processes, further reducing production costs.

[0058] 6. This invention can improve product reliability. The compensation technology of this invention reduces surface deformation caused by residual stress and clamping stress, so that the final processed infrared thin-walled optical element has higher surface quality and more stable optical performance, and is suitable for key components in high-precision optical systems. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the deterministic compensation cutting process of the deterministic cutting compensation processing method for infrared thin-walled optical elements of the present invention;

[0060] Figure 2 This invention is based on the actual surface shape error contour diagram calculated from the measured surface shape error and simulated deformation of a planar workpiece.

[0061] Figure 3 This is a comparison chart showing the surface shape error results when machining the same workpiece using the original tool path and the compensated tool path according to the present invention. Detailed Implementation

[0062] The present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0063] like Figure 1 , Figure 2 and Figure 3 As shown, the deterministic cutting compensation machining method for infrared thin-walled optical elements provided by the present invention includes the following steps:

[0064] S1. Obtain the initial surface shape of the workpiece.

[0065] Using in-situ measurement technology on ultra-precision machine tools, the initial surface shape of the thin-walled optical element to be processed is measured to obtain its surface profile ρ(x,y), where x,y represent the coordinate positions in the xOy plane in the Cartesian coordinate system. Based on the designed surface profile z(x,y), the material removal amount Δz(x,y) at each coordinate position is calculated.

[0066] Δz(x,y)=ρ(x,y)-z(x,y).

[0067] In practical application, this invention is used to measure the surface shape error of a φ38×2mm single-crystal silicon thin-walled planar optical element by using a vacuum chuck clamp for adsorption and clamping. The error is ρ(x,y).

[0068] S2. Prediction of elastic deformation under clamping stress.

[0069] Based on the finite element simulation model, a three-dimensional numerical distribution simulation of the clamping force distribution during the machining process is performed, and the elastic deformation ε(x,y) of the component caused by clamping stress is predicted. This elastic deformation is combined with the initial material removal amount to obtain a new material removal amount z′(x,y), thereby eliminating the influence of clamping stress on the surface shape error of the component machining and providing a basis for the next compensation calculation.

[0070] z′(x,y)=Δz(x,y)+ε(x,y).

[0071] Based on the workpiece dimensions, suction cup geometry, and suction pressure values, a static model of the clamping force and deformation of a single-crystal silicon optical element is constructed using finite element analysis software, and the clamping deformation of the planar element is accurately calculated as ε(x,y).

[0072] Calculated The actual surface shape error of the single-crystal silicon planar element before processing is:

[0073] z′(x,y)=ρ(x,y)+ε(x,y).

[0074] The calculated actual surface shape error point cloud data is plotted as a 3D image, and the result is as follows: Figure 2 As shown.

[0075] S3. Deformation prediction of components under residual stress during processing.

[0076] Step S3 includes the following sub-steps:

[0077] S31. Construct a database T relating ultra-precision cutting parameters and tool geometry parameters to residual stress; cutting parameters include cutting speed, depth of cut, and feed rate, and tool geometry parameters include nose radius, rake angle, and clearance angle.

[0078] S32. Based on the process parameter-residual stress correlation database T, a deep learning model is trained using this database to predict the residual stress under different process parameters.

[0079] Step S32 also includes the following sub-steps:

[0080] S321. Data Collection and Preparation:

[0081] Database T should contain various parameters related to the cutting process, such as cutting speed, depth of cut, feed rate, and tool geometry parameters, including nose radius, rake angle, and clearance angle. Residual stress data of the material under these machining conditions should be obtained experimentally, or the correspondence between these parameters and residual stress should be calculated using theoretical models.

[0082] S322. Selection of Deep Learning Models:

[0083] Choose a deep learning model suitable for nonlinear mapping relationships, such as a multilayer perceptron (MLP), convolutional neural network (CNN), or long short-term memory network (LSTM), depending on the complexity of the data; the input layer of the model corresponds to process parameters, such as cutting speed, cutting depth, and feed rate, and the output layer is the predicted residual stress.

[0084] S333, Model Training:

[0085] To train the model using data from database T, the data needs to be preprocessed, including normalization and standardization, to improve training efficiency. A loss function, such as mean squared error (MSE), should be selected, and backpropagation algorithms, such as Adam or RMSprop, should be used to optimize the model. Cross-validation should be used during training to prevent overfitting, and the predictive performance of the model can be optimized by adjusting hyperparameters such as learning rate, number of layers, and number of neurons.

[0086] S334, Model Validation:

[0087] The model is validated using data not used in training to ensure its accuracy and robustness in predicting residual stress. The model's performance is evaluated by calculating the error between the predicted and actual residual stress values, such as mean square error or mean absolute error.

[0088] The error between the predicted value and the actual residual stress value calculated in step S334 is as follows:

[0089] Suppose the trained deep learning model is represented as f θ Where θ represents the model parameters (weights and biases), then for a given set of processing parameters (p1, p2, ..., p...), n (e.g., cutting speed, depth of cut), the residual stress prediction of the model can be expressed as:

[0090]

[0091] in, This represents the predicted residual stress.

[0092] The optimization objective during training is to minimize the loss function L, which is typically the mean square error (MSE) between the predicted and actual residual stress values.

[0093]

[0094] Where m is the number of training samples, R (i) Let be the actual residual stress of the i-th sample.

[0095] S33. Using the residual stress prediction model trained in step S32, combined with the finite element analysis method, calculate the component deformation δ(x,y) caused by the residual stress under the given processing parameters.

[0096] Step S33 also includes the following sub-steps:

[0097] S331. Residual stress prediction:

[0098] Based on the deep learning model obtained from S02, the corresponding residual stress distribution is obtained by inputting the cutting process parameters used in the actual machining process.

[0099]

[0100] in, This represents the predicted value of residual stress at workpiece position (x, y). The distribution of residual stress may vary at different locations on the workpiece, so it is necessary to calculate the corresponding residual stress at different coordinate positions (x, y).

[0101] S332, Finite Element Analysis Calculation:

[0102] The predicted residual stress is input into the finite element model as a boundary condition or load condition. The influence of the residual stress on the deformation of the thin-walled optical element is calculated through finite element analysis. The finite element model should consider material properties (such as elastic modulus, Poisson's ratio, etc.) and geometry, and calculate the elastic deformation of the material through static analysis. Through the above methods, the influence of residual stress on the surface deformation of the element under given processing parameters can be accurately predicted and calculated, providing basic data for subsequent compensation processing.

[0103] The finite element model, through finite element analysis, can obtain the deformation δ(x,y) at different locations (x,y) caused by residual stress, that is, the deformation of the material during processing. This process depends on the results of finite element analysis software and is usually expressed as:

[0104]

[0105] Here, FEA represents the finite element analysis process. For the predicted residual stress, material properties include the elastic modulus E and Poisson's ratio ν.

[0106] S4. Generation of compensation trajectory for non-uniform material removal.

[0107] In step S4, the generation of the non-uniform material removal compensation trajectory combines the effects of the aforementioned clamping stress and residual stress. Based on a genetic algorithm, the material removal amount is iteratively optimized to calculate the final compensated material removal amount Δz″(x,y), such that:

[0108] Δz″(x,y)=Δz′(x,y)+δ(x,y)

[0109] This result will be used as a compensation trajectory for non-uniform cutting and input into the machining equipment to ensure the achievement of final surface accuracy.

[0110] S5. Non-uniform cutting compensation machining based on slow tool servo.

[0111] The non-uniform cutting compensation machining based on slow tool servo in step S5 is performed by using the freeform surface cutting technology of slow tool servo to perform non-uniform cutting along the compensation trajectory calculated above. The slow tool servo system can accurately control the movement trajectory of the tool, remove material layer by layer, and accurately realize the surface shape compensation of thin-walled optical elements.

[0112] In practical application, this invention constructs a correlation database T between ultra-precision cutting parameters (workpiece rotation speed = 2500 RPM; depth of cut = 4 μm; feed rate = 1 μm / rev) and tool geometry parameters (nose radius = 0.5 mm; rake angle = -25°; clearance angle = 10°) and residual stress for single-crystal silicon planar optical elements. Based on this correlation database, a residual stress prediction model for ultra-precision machining of single-crystal silicon is trained using a deep learning network. Furthermore, using finite element analysis, the deformation of the element under the residual stress conditions is calculated as δ(x,y).

[0113] If the processing of the single-crystal silicon planar is directly compensated based on the actual surface shape error z′(x,y) of the component, then the theoretical surface shape error of the processed single-crystal silicon wafer should be δ(x,y). The processing result in this embodiment is as follows: Figure 3As shown in the left figure, the machining surface shape error PV is 1.21μm.

[0114] If the residual stress deformation is directly compensated by setting the material removal amount as z′(x,y)+δ(x,y), then the residual stress value and deformation amount will change due to the change in the cutting depth machining parameters, and the residual stress deformation cannot be effectively eliminated.

[0115] Therefore, based on the aforementioned residual stress prediction model for ultra-precision machining of single-crystal silicon, a genetic algorithm is used to iteratively optimize the material removal amount in different regions, so that the final compensated material removal amount Δz″(x,y) satisfies:

[0116] Δz″(x,y)=z′(x,y)+δ′(x,y)

[0117] Where δ′(x,y) represents the residual stress deformation under the current cutting parameters Δz″(x,y).

[0118] Based on the above analysis and calculation results, a deterministic compensation cutting trajectory is obtained. This machining trajectory is input into the machining equipment, and a slow-tool servo turning method is used to perform non-uniform cutting along the calculated compensation trajectory, accurately achieving surface shape compensation for thin-walled optical elements. The surface shape measurement results of the workpiece processed using this method are as follows: Figure 3 As shown in the right figure, the machining surface shape error PV is 0.39μm.

[0119] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A deterministic cutting compensation machining method for infrared thin-wall optical elements, characterized in that, The method comprises the following steps: S1, obtaining the initial surface shape of the workpiece; The initial surface shape of the thin-wall optical element to be machined is measured by an in-situ measurement technology of an ultra-precision machine tool, the surface profile p(x, y) of the thin-wall optical element to be machined is obtained, the material removal amount Δz(x, y) at each coordinate position is calculated according to the designed surface profile z(x, y), and Δz(x, y) = p(x, y) - z(x, y); S2, elastic deformation prediction of clamping stress; Based on a finite element simulation model, the three-dimensional spatial numerical distribution of the clamping force distribution in the machining process is simulated, and the elastic deformation ε(x, y) of the element caused by the clamping stress is predicted; the elastic deformation and the initial material removal amount are integrated to obtain a new material removal amount z'(x, y), so as to eliminate the influence of the clamping stress on the machining surface error of the element and provide a basis for the next compensation calculation: z'(x, y) = Δz(x, y) + ε(x, y); S3, residual stress element deformation prediction; S4, generation of non-uniform material removal compensation track; S5, non-uniform cutting compensation machining based on slow tool servo; The step S3 comprises the following sub-steps: S31, constructing a correlation database T between the ultra-precision cutting parameters and the tool geometric parameters and the residual stress; the cutting parameters include the cutting linear speed, the cutting depth and the feed amount, and the tool geometric parameters include the tool nose radius, the rake angle and the relief angle; S32, based on the process parameter-residual stress correlation database T, a deep learning model is trained using the database to predict the residual stress under different process parameters; S33, using the residual stress prediction model trained in step S32, the element deformation δ(x, y) caused by the machining residual stress under the given machining parameter condition is calculated by combining the finite element analysis method; The S32 further comprises the following sub-steps: S321, data collection and preparation: The database T should contain various parameters related to the cutting process, including the cutting speed, the cutting depth, the feed amount and the tool geometric parameters, and the tool geometric parameters include the tool nose radius, the rake angle and the relief angle; the residual stress data of the material under these machining conditions are obtained through experiments, or the corresponding relationship between the parameters and the residual stress is calculated through a theoretical model; S322, selection of a deep learning model: A deep learning model suitable for a nonlinear mapping relationship is selected, and the deep learning model includes a multilayer perceptron (MLP), a convolutional neural network (CNN) or a long short-term memory network (LSTM), which depends on the complexity of the data; the input layer of the model corresponds to the process parameters; and the output layer is the predicted residual stress; S323, model training: The data in the database T is used for training, and the data needs to be preprocessed, including normalization and standardization operations, to improve the training efficiency of the model; a loss function is selected, the loss function includes a mean square error (MSE), and a back propagation algorithm including Adam or RMSprop is used to optimize the model; the training process needs to use a cross-validation technique to prevent overfitting, and the prediction performance of the model can be optimized by adjusting the learning rate, the number of layers and the number of neurons; ​ S324, model verification: The model is verified by using data not involved in training to ensure the accuracy and robustness of the model in predicting residual stress; the performance of the model is evaluated by calculating the error between the predicted value and the actual residual stress value, including mean square error or mean absolute error; The generation of the non-uniform material removal compensation trajectory in step S4 is based on genetic algorithm iterative optimization of material removal amount, taking into account the influence of the aforementioned clamping stress and residual stress, to calculate the final compensated material removal amount Δz''(x, y), so that: Δz''(x, y)=Δz'(x, y)+δ(x, y) This result will be input into the machining equipment as the compensation trajectory of non-uniform cutting to ensure the realization of the final surface accuracy.

2. The deterministic cutting compensation machining method of an infrared thin-wall optical element according to claim 1, characterized in that, The error between the predicted value and the actual residual stress value in step S324 is specifically: Assume the trained deep learning model is represented as f θ where θ are the parameters of the model, including weights and biases, then for a given set of processing parameters the model's residual stress prediction can be represented as: ； wherein, is the predicted residual stress; The objective of the training process is to minimize the loss function L, which is usually the mean square error (MSE) of the predicted value and the actual residual stress value: ； where m is the number of training samples, R (i) is the actual residual stress of the i-th sample.

3. The deterministic cutting compensation machining method of an infrared thin-wall optical element according to claim 1, characterized in that, The step S33 further includes the following sub-steps: S331, residual stress prediction: Based on the deep learning model obtained from S02, input the cutting process parameters used in the actual machining process to obtain the corresponding residual stress distribution , ； wherein, is the predicted value of the residual stress at the position (x, y) of the workpiece; the distribution of the residual stress can change at different positions of the workpiece, so it is necessary to calculate the corresponding residual stress at different coordinate positions (x, y); S332, finite element analysis calculation: The predicted residual stress is input as a boundary condition or load condition into the finite element model, and the influence of the residual stress on the deformation of the thin-walled optical element is calculated by finite element analysis; the finite element model should consider material properties and geometric structure, including elastic modulus, Poisson's ratio, and calculate the elastic deformation of the material through statics analysis; through the above method, the influence of residual stress on the surface deformation of the element under the given machining parameters can be accurately predicted and calculated, providing basic data for subsequent compensation machining.

4. The deterministic cutting compensation machining method of an infrared thin-wall optical element according to claim 3, characterized in that: The finite element model can obtain the deformation amount δ(x, y) caused by residual stress at different positions (x, y) through finite element analysis, i.e. the deformation of the material during machining, which depends on the results of the finite element analysis software, usually represented as: ； FEA stands for Finite Element Analysis process, For the predicted residual stresses, the material properties include the elastic modulus E and the Poisson's ratio v.

5. The deterministic cutting compensation machining method of an infrared thin-wall optical element according to claim 1, characterized in that: The non-uniform cutting compensation machining based on slow tool servo in step S5 is carried out by slow tool servo free-form surface cutting technology along the above calculated compensation trajectory; the slow tool servo system can accurately control the motion trajectory of the tool, remove the material layer by layer, and accurately realize the surface compensation of the thin-walled optical element.

Images