Active regulation and control method for surface residual stress in numerical control milling process

Abstract

The invention discloses a method for actively regulating and controlling surface residual stress in a numerical control milling process, which comprises the following steps of: measuring the influence degree of different process parameters on the surface residual stress by performing range analysis on process parameters and surface residual stress data, and identifying key factors for controlling the surface residual stress as control objects; when the surface residual stress monitoring value exceeds an expected range, taking the current process parameters and a prediction result of the surface residual stress monitoring model as NMPC model input; compared with a traditional method, the dynamic characteristics of the milling surface residual stress are considered, and the optimal control strategy at the current moment can be output. Based on the optimal control strategy output by the NMPC controller, the adjustment amount of the technological parameters is converted into a control instruction which can be executed by a machine tool PLC, issuing of the control instruction is achieved through a numerical control machine tool communication module, and control over the surface residual stress is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of surface residual stress control, and specifically relates to an active regulation method for surface residual stress during CNC milling machining process. Background Art

[0002] During the milling process, the workpiece is subjected to the combined action of multiple physical fields such as force heat, mechanical load, and thermal load, generating self-balanced residual stress on the workpiece surface. The state of the residual stress affects the service performance of the component. If the cutting surface is in a residual tensile stress state, the fatigue strength of the component decreases, affecting the service life of the component; while when the cutting surface is in a residual compressive stress state, the fatigue strength of the component increases, the corrosion resistance is improved, and the service life is extended. Therefore, controlling the residual stress generated on the workpiece surface during the milling process is of great significance for controlling the surface integrity of the milled workpiece and improving the machining performance.

[0003] Currently, there are already methods and applications for machining quality control during the cutting process. For example, the Chinese patent application with the application number CN201410312703.8 discloses an adaptive roll variable-speed grinding intelligent control method, which takes into account the change of the roll axial stiffness and controls the grinding feed through a fuzzy logic system to improve the grinding surface roughness and ensure the machining quality. The Chinese patent application with the application number CN202310556185.3 discloses an adaptive regulation system for the surface roughness of a five-axis milling machine based on digital twin. By inputting the cutting depth, rotational speed, feed, and vibration into a neural network prediction model to predict the surface roughness, and then inputting the vibration and surface roughness into a fuzzy control model to obtain the rotational speed and feed, and using the obtained rotational speed and feed to control the machining parameters of the five-axis milling machine through the numerical control system, so as to realize machining feedback and improve the part quality.

[0004] In summary, most of the existing machining quality control methods adjust the process parameters by using methods such as fuzzy logic control, and then control the surface quality such as surface roughness by changing the working conditions. Therefore, the existing machining quality control methods have the following deficiencies: Since the milling surface residual stress has non-linear and time-varying characteristics, compared with the strong linear correlation between control objectives such as surface roughness and process parameters (control objects), it is difficult to establish control rules for process parameters such as spindle speed on residual stress; and traditional control methods such as traditional PID control and fuzzy logic control are difficult to consider or need to simplify the non-linear and dynamic characteristics of the milling process surface residual stress, resulting in deviations or performance degradation of the control strategies output by their control models during working condition adjustment. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide an active control method for surface residual stress in the process of CNC milling, so as to solve the problems that it is difficult to establish a control strategy and the lack of consideration of dynamic characteristics in the quality control methods in the existing cutting processing field, and finally realize the active control of surface residual stress in the milling process.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An active control method for surface residual stress in the process of CNC milling includes the following steps:

[0008] Step 1: Range analysis of the controlled object

[0009] Use range analysis to calculate the range of the surface residual stress data measured under the process parameters of different factor levels, measure the influence degree of different process parameters on the surface residual stress, and identify the key factors for surface residual stress control as the process parameters to be controlled and adjusted during the milling process.

[0010] Step 2: Nonlinear model predictive control

[0011] Based on the process parameters selected by the range analysis of the controlled object, use the NMPC controller to output the control strategy; the nonlinear model predictive control of the surface residual stress includes the following steps:

[0012] 21) Use the process parameters and obtain multi-directional cutting force data through sensors;

[0013] 22) Establish a data-driven surface residual stress monitoring model;

[0014] 23) Use the multi-directional cutting force data as the input, and continuously estimate the state of the residual stress in the current machining process during the cutting process to obtain the monitoring value of the surface residual stress;

[0015] 24) According to the monitoring value of the surface residual stress, judge whether the current surface residual stress is within the set target range: if so, do not trigger the control instruction, and loop to execute step 211), and continuously estimate the state of the residual stress in the machining process using the surface residual stress monitoring model; if not, trigger the control instruction and execute step 215);

[0016] 25) Use nonlinear model predictive control to perform rolling horizon optimization for the surface residual stress to obtain the optimal control strategy output by the NMPC controller;

[0017] Step 3: CNC machine tool communication module

[0018] Based on the optimal control strategy output by the NMPC controller, the adjustment amount of process parameters is converted into control instructions that can be executed by the machine tool PLC, and the control instructions are issued through the numerical control machine tool communication module to achieve the control of surface residual stress.

[0019] Further, in the first step, the process parameters include cutting speed v c , feed rate f, cutting depth a p and cutting width a e ;

[0020] The method steps of the range analysis of the control object are as follows:

[0021] 11) Set a four-factor and five-level orthogonal experiment according to the cutting speed v c , feed rate f, cutting depth a p and cutting width a e , and collect the milling surface residual stress values under different parameters;

[0022] 12) Conduct a range analysis on the experimental results

[0023] 121) Calculate the mean value of the experimental results for each level of each factor;

[0024] 122) Calculate the range of the mean values of each factor level;

[0025] 123) Compare the ranges of each factor. The factor with the largest range is the factor that has the most significant influence on the surface residual stress, and the key factor for surface residual stress control is identified.

[0026] Further, in step 22), the surface residual stress monitoring model is:

[0027] Rs t = f NN (v c , f, a p , a e , F t )

[0028] Where: Rs t represents the surface residual stress obtained by cutting at the current moment t of the control object; F t is the cutting force signal data collected by the sensor at the current moment t; v c is the cutting speed; f is the feed rate; a p is the cutting depth; a e is the cutting width.

[0029] Further, in step 23), the method for estimating the state of the residual stress to obtain the monitoring value of the surface residual stress is:

[0030] Rst+N = f NN (v c + Δμ(m), f, a p , a e , F t+N )

[0031] F t = f ANN (v c , f, a p , a e )

[0032] Where: Rs t+N represents the residual stress state of the controlled object at the next moment t + N; F t+N represents the cutting force state at the next moment t + N; Δμ(m) represents the process parameter adjustment amount; f ANN (·) represents the mapping relationship between process parameters and cutting force established based on the artificial neural network ANN; N is the time step.

[0033] Furthermore, in the step 25), the method steps for performing rolling horizon optimization and solution on the surface residual stress by using nonlinear model predictive control are as follows:

[0034] 251) Set the optimization objective

[0035] The optimization objective is set as the process parameter adjustment amount Δμ selected by range analysis, and minimize the adjustment amount of process parameters min(Δμ), so as to minimize the impact on the stability of the milling process while ensuring that the surface residual stress remains within the expected range during the machining process;

[0036] 252) Set the constraint conditions

[0037] Set the constraint conditions for the rolling optimization in the solution process of the NMPC controller for the milling surface residual stress as:

[0038] Rs t+N = f NN (v c + Δμ(m), f, a p , a e , F t+N ) ≤ 0

[0039] Where: Rs t+N represents the residual stress state of the controlled object at the next moment t + N; F t+N represents the cutting force state at the next moment t + N; Δμ(m) represents the process parameter adjustment amount; f ANN (·) represents the mapping relationship between process parameters and cutting force established based on the artificial neural network ANN; N is the time step;

[0040] 253) Optimization solution

[0041] Transform the constrained optimization process into an unconstrained optimization:

[0042] min(Δμ) s.t. f NN (v c +Δμ(m), f, a p , a e , F t+N ) ≤ 0

[0043] By introducing a barrier function into the above constrained optimization process and integrating the constraint conditions into the objective function, an unconstrained optimization problem is obtained:

[0044]

[0045] where: θ is the barrier parameter; log(―f NN (v c +Δμ(m), f, a p , a e , F t+N ) i ) is the barrier function;

[0046] 254) Iterative solution: Select an initial point and an initial barrier parameter θ, and gradually approach the constraint boundary by gradually reducing the barrier parameter, and finally obtain an optimal solution that satisfies the convergence condition.

[0047] Furthermore, in the step 254), since an optimization problem needs to be solved in each control cycle, the optimization problems of adjacent time steps are similar; the previous optimization result Δμ N―1 is used as the initial value Δμ N of the current optimization problem to reduce the number of iterations of the optimization algorithm.

[0048] Furthermore, in the third step, the control instruction is issued through the numerical control machine tool communication module, and the method steps for realizing the surface residual stress control are as follows:

[0049] 31) Assign IP addresses to the host computer and the numerical control machine tool PLC;

[0050] 32) Create a host computer client socket, establish a connection with the PLC of the numerical control machine tool based on the TCP protocol, and initiate a server socket connection request by specifying the IP address and port number of the PLC of the numerical control machine tool;

[0051] 33) Modify the process parameter magnification parameter stored in the PLC according to the cyclic sending of control instructions, and realize the surface residual stress control by continuously adjusting the process parameters.

[0052] The beneficial effects of the present invention are as follows:

[0053] The active regulation method for surface residual stress in the numerical control milling process of the present invention measures the influence degree of different process parameters on the surface residual stress through range analysis of the process parameters and surface residual stress data, and identifies the key factors for surface residual stress control as the control objects; when the monitored value of the surface residual stress exceeds the expected range, the current process parameters and the prediction results of the surface residual stress monitoring model are used as the input of the NMPC model; compared with the traditional method, it considers the dynamic characteristics of the milling surface residual stress and can output the optimal control strategy at the current moment. Based on the optimal control strategy output by the NMPC controller, the adjustment amount of the process parameters is converted into control instructions that can be executed by the machine tool PLC, and the control instructions are sent through the numerical control machine tool communication module to achieve the control of the surface residual stress. In summary, the active regulation method for surface residual stress in the numerical control milling process of the present invention analyzes the regulation objects of the surface residual stress, considers the dynamic changes of the cutting force in the milling process, actively regulates the surface residual stress of the workpiece in the milling process, solves the problems that it is difficult to establish a control strategy and the lack of consideration of dynamic characteristics in the quality control methods in the existing cutting processing field, and finally realizes the active regulation of the surface residual stress in the milling process.

[0054] In addition, the surface residual stress control strategy output by the present invention considers minimizing the adjustment amount of the process parameters, avoiding the milling stability problem caused by excessive adjustment of the process parameters.

[0055] The surface residual stress control method proposed by the present invention establishes communication with the machine tool PLC based on TCP / IP, and can realize the rolling issuance of control instructions and monitor the control feedback. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to make the objectives, technical solutions and beneficial effects of the present invention clearer, the following drawings are provided for illustration:

[0057] Figure 1 is the flow chart of the active regulation method for surface residual stress in the numerical control milling process of the present invention;

[0058] Figure 2 is the flow chart of the NMPC iterative solution process;

[0059] Figure 3 is the numerical control machine tool communication flow chart based on TCP / IP;

[0060] Figure 4 is the mean value of each level of the surface residual stress process parameters;

[0061] Figure 5 is the verification result of the milling surface residual stress control model. DETAILED DESCRIPTION OF THE INVENTION

[0062] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, so that those skilled in the art can better understand the present invention and be able to implement it, but the embodiments given are not intended to limit the present invention.

[0063] As Figure 1 shown, the active regulation method for the surface residual stress in the numerical control milling process of this embodiment mainly includes three stages: range analysis of the control object, nonlinear model predictive control (NMPC) control, and the numerical control machine tool communication module. Specifically, the active regulation method for the surface residual stress in the numerical control milling process of this embodiment includes the following steps.

[0064] Step 1: Range analysis of the control object

[0065] Range analysis is used to calculate the range of the surface residual stress data measured under the process parameters of different factor levels, measure the influence degree of different process parameters on the surface residual stress, and identify the key factors for surface residual stress control, which are used as the process parameters to be controlled and adjusted during the milling process.

[0066] Specifically, the surface residual stress of milling is affected by parameters such as cutting speed v c , feed rate f, cutting depth a p , cutting width a e , etc. That is, in this embodiment, the process parameters include cutting speed v c , feed rate f, cutting depth a p and cutting width a e . In order to select the active regulation object of the milling surface residual stress, range analysis is carried out on the control object.

[0067] In this embodiment, the method steps of range analysis of the control object are as follows.

[0068] 11) Set a four-factor five-level orthogonal experiment according to cutting speed v c , feed rate f, cutting depth a p and cutting width a e , and collect the milling surface residual stress values under different parameters.

[0069] 12) Carry out range analysis on the experimental results

[0070] 121) Calculate the mean value of the experimental results for each level of each factor.

[0071] 122) Calculate the range of the mean values of each factor level.

[0072] 123) Compare the ranges of each factor. The factor with the largest range is the factor that has the most significant influence on the surface residual stress. Identify the key factors for controlling the surface residual stress. The identified key factors are used as the process parameters for control and adjustment during milling to achieve effective control of the surface residual stress.

[0073] Step 2: Nonlinear model predictive control

[0074] Using the process parameters selected based on the range analysis of the controlled object as the controlled object, an NMPC controller outputs a control strategy. In this embodiment, the nonlinear model predictive control of the surface residual stress includes the following steps.

[0075] 21) Utilize the process parameters and obtain multi-directional cutting force data through sensors.

[0076] 22) Establish a data-driven surface residual stress monitoring model.

[0077] In this embodiment, the surface residual stress monitoring model is:

[0078] Rs t = f NN (v c , f, a p , a e , F t )

[0079] where: Rs t represents the surface residual stress obtained by cutting at the current time t of the controlled object; F t is the cutting force signal data collected by the sensor at the current time t; v c is the cutting speed; f is the feed rate; a p is the cutting depth; a e is the cutting width.

[0080] 23) Using the multi-directional cutting force data as the input, continuously estimate the state of the residual stress in the current machining process during cutting to obtain the monitored value of the surface residual stress.

[0081] In this embodiment, to estimate the state of the residual stress, it is necessary to obtain the process parameter adjustment amount Δμ(m) of the future control input result based on the prediction of the future state. In this embodiment, the method for obtaining the monitored value of the surface residual stress in the process of predicting the future state is:

[0082] Rs t+N = f NN (v c + Δμ(m), f, a p , a e , F t+N )

[0083] In order to obtain the modified spindle speed v c F in the future state t+N , an artificial neural network ANN is used to establish the mapping relationship between process parameters and cutting force:

[0084] F t = f ANN (v c , f, a p , a e )

[0085] where: Rs t+N represents the residual stress state of the controlled object at the next moment t + N; F t+N represents the cutting force state at the next moment t + N; Δμ(m) represents the process parameter adjustment amount; f ANN (·) represents the mapping relationship between process parameters and cutting force established based on the artificial neural network ANN; N is the time step. In this embodiment, N = 1s.

[0086] In the rolling horizon optimization process, the residual stress value at the current moment is obtained through the prediction model within each time step N. Based on the current process parameters and the multi-directional cutting force data obtained by the sensor as inputs, the surface residual stress value is predicted, and the state of the residual stress in the current machining process is continuously estimated during the cutting process.

[0087] 24) According to the monitored value of the surface residual stress, judge whether the current surface residual stress is within the set target range: if so, do not trigger the control instruction, and loop to execute step 211), and continuously estimate the state of the residual stress in the machining process using the surface residual stress monitoring model; if not, trigger the control instruction and execute step 215).

[0088] 25) Use nonlinear model predictive control to perform rolling horizon optimization on the surface residual stress to obtain the optimal control strategy output by the NMPC controller.

[0089] In this embodiment, the method steps for performing rolling horizon optimization on the surface residual stress using nonlinear model predictive control are as follows:

[0090] 251) Set the optimization objective

[0091] Since excessive adjustment of process parameters during the milling process will affect cutting stability, the optimization objective in the surface residual stress control process based on NMPC is set to the process parameter adjustment amount Δμ selected by range analysis, and the adjustment amount of the process parameters is minimized min(Δμ), so as to minimize the impact on the cutting stability of the milling process while ensuring that the surface residual stress during the machining process remains within the expected range.

[0092] 252) Set constraint conditions

[0093] To ensure that the residual stress on the milling surface is controlled within the expected range and to ensure that the surface residual stress is in a compressive stress state, taking the solution process of the NMPC controller for the milling surface residual stress as an example, the constraint conditions for the rolling optimization in the solution process are set as follows:

[0094] Rs t+N = f NN (v c +Δμ(m), f, a p , a e , F t+N ) ≤ 0

[0095] Where: Rs t+N represents the residual stress state of the controlled object at the next moment t + N; F t+N represents the cutting force state at the next moment t + N; Δμ(m) represents the adjustment amount of the process parameters; f ANN (·) represents the mapping relationship between the process parameters and the cutting force established based on the artificial neural network ANN; N is the time step.

[0096] 253) Optimize and solve

[0097] Solve the constrained problem based on the interior point method and use the solver IPOPT (Interior Point OPTimizer) to achieve accelerated calculation, and transform the constrained optimization process into an unconstrained optimization:

[0098] min(Δμ) s.t. f NN (v c +Δμ(m), f, a p , a e , F t+N ) ≤ 0

[0099] By introducing a barrier function into the above constrained optimization process and integrating the constraint conditions into the objective function, an unconstrained optimization problem is obtained:

[0100]

[0101] Where: θ is the barrier parameter; log(―f NN (v c +Δμ(m), f, a p , a e , F t+N ) i ) is the barrier function.

[0102] 254) Iterative solution: Select the initial point and the initial barrier parameter θ, and gradually approach the constraint boundary by gradually reducing the barrier parameter, and finally obtain the optimal solution that satisfies the convergence condition, such asFigure 2 as shown

[0103] Further, to quickly solve the optimization problem, since an optimization problem needs to be solved in each control cycle and the optimization problems at adjacent time steps are similar; using the previous optimization result Δμ N―1 as the initial value Δμ N of the current optimization problem to reduce the number of iterations of the optimization algorithm.

[0104] Step 3: CNC machine tool communication module

[0105] Based on the optimal control strategy output by the NMPC controller, convert the adjustment amount of the process parameters into control instructions that can be executed by the machine tool PLC, and realize the issuance of the control instructions through the CNC machine tool communication module to control the surface residual stress. The issuance of the CNC machine tool communication and control instructions includes processes such as IP allocation, TCP connection, data request and response, real-time transmission of control instructions, and status monitoring, as Figure 2 shown

[0106] In this embodiment, the method steps for realizing the issuance of control instructions through the CNC machine tool communication module to control the surface residual stress are as follows:

[0107] 31) Assign IP addresses to the host computer and the CNC machine tool PLC;

[0108] 32) Create a host computer client socket, establish a connection with the PLC of the CNC machine tool based on the TCP protocol, and initiate a server socket connection request by specifying the IP address and port number of the CNC machine tool PLC;

[0109] 33) Modify the process parameter magnification parameter stored in the PLC according to the cyclic transmission of control instructions, and realize the control of the surface residual stress by continuously adjusting the process parameters.

[0110] In this process, continuously monitor the variable values corresponding to the control target in the PLC to ensure the process of issuing control instructions.

[0111] Next, taking the milling of aluminum alloy as an example, the active control method for the surface residual stress in the CNC milling process of the present invention will be described.

[0112] 1. Aluminum alloy milling processing experiment environment

[0113] To prove the active control method for the surface residual stress in the CNC milling process proposed in this embodiment, this embodiment takes the milling processing experiment as an example, and the equipment used in the experiment is shown in Table 1.

[0114] Table 1 Main equipment used in the experiment

[0115]

[0116] The tool used in the aluminum alloy milling experiment is a solid carbide end mill with a diameter of 8 mm. The workpiece is 7075 aluminum alloy with dimensions of 50 mm × 20 mm × 20 mm. Dry cutting is used to mill the workpiece, and the cutting parameters including cutting speed v, feed rate f, cutting depth a p , and cutting width a e are taken as objects. The orthogonal experiment design method is used to design the milling experiment group for aluminum alloy components. The orthogonal experiment parameter levels are set as shown in Table 2. There are a total of 25 groups of milling experiments, and the specific milling experiment parameters are shown in Table 3. After the milling experiment, the residual stress on the milling surface is obtained through the stress test system.

[0117] Table 2 Orthogonal experiment parameter level settings

[0118]

[0119] 2. Range analysis of process parameters

[0120] Based on the above orthogonal test data, the range analysis method is used to analyze the residual stress under different process parameters by comparing the range of residual stress data. The analysis results are shown in Table 3 and Figure 3 as follows.

[0121] Table 3 Orthogonal experiment range analysis results

[0122]

[0123]

[0124] From the orthogonal experiment range analysis results in Table 3, it can be seen that among the spindle speed, feed rate, cutting width, and cutting depth, the cutting depth has the greatest influence on the surface residual stress, followed by the spindle speed. Considering that the machining stage with the greatest influence on the final surface residual stress of the workpiece is the finish machining stage, it is difficult to ensure the dimensional requirements of the workpiece by choosing the cutting depth as the control object. Therefore, the spindle speed is selected as the control object for active regulation of the surface residual stress.

[0125] 3. Performance of the milling surface residual stress control model based on NMPC

[0126] To verify the effectiveness of the milling surface residual stress control model, the established ANN model is used to simulate the CNC milling process. In order to improve the learning accuracy, neural network architectures with different hierarchical structures are designed, and the optimal model structure is selected with the lowest mean square error (MSE). A set of process parameters is selected for verification. The surface residual stress, spindle speed, and its adjustment amount are as Figure 5As shown. The surface residual stress control simulation can control and maintain the surface residual stress within the expected range, effectively achieving the control of the surface residual stress.

[0127] In order to further verify the effectiveness of the active regulation method for surface residual stress, a comparative experiment was set up. The experimental parameter table is shown in Table 4. The comparative experiment includes a processing method with fixed processing parameters once and a processing method with active regulation of processing parameters once.

[0128] Table 4 Parameter settings for the comparative experiment on the active regulation of surface residual stress

[0129]

[0130] The surface residual stress of the workpiece after milling was measured by an X-ray Proto residual stress test and analysis system. Considering the accuracy of the residual stress measurement, the final result is the average value obtained from multiple measurements. And in order to better compare the results of fixed processing parameters and active regulation, the surface residual stress of the same area on the surface of the milled workpiece was selected for comparison. The results of the comparative experiment are shown in Table 5. From the surface residual stress values of different processing methods in the table, it can be seen that the surface residual stress is regulated from residual tensile stress to residual compressive stress, and the proposed active regulation method can effectively control the surface residual stress within the set range.

[0131] Table 5 Results of the comparative experiment on the active regulation of surface residual stress

[0132]

[0133] The above-described embodiments are only preferred embodiments given to fully illustrate the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or transformations made by those skilled in the art on the basis of the present invention are all within the protection scope of the present invention. The protection scope of the present invention is subject to the claims.

Claims

1. A method for actively controlling surface residual stress during CNC milling, characterized in that: The steps include: Step 1: Range analysis of control objects The range analysis is used to calculate the range of the surface residual stress data measured under different factor levels of process parameters, measure the influence of different process parameters on the surface residual stress, and identify the key factors for controlling the surface residual stress as the process parameters to be controlled and adjusted during the milling process; Step 2: Nonlinear Model Predictive Control Based on the process parameters selected by the range analysis of the controlled object, the NMPC controller outputs the control strategy; the nonlinear model predictive control of the surface residual stress includes the following steps: 21) Using process parameters and sensors to obtain multi-directional cutting force data; 22) Establish a data-driven surface residual stress monitoring model; 23) Taking multi-directional cutting force data as input, the residual stress in the current machining process is continuously estimated during the cutting process to obtain the monitoring value of the surface residual stress; 24) According to the monitoring value of the surface residual stress, determine whether the current surface residual stress is within the set target range: if so, do not trigger the control instruction, loop through step 211), and use the surface residual stress monitoring model to continuously estimate the state of the residual stress during the processing; if not, trigger the control instruction, and execute step 215); 25) Using nonlinear model predictive control to perform rolling time domain optimization on the surface residual stress, the optimal control strategy output by the NMPC controller is obtained; Step 3: CNC machine tool communication module Based on the optimal control strategy output by the NMPC controller, the adjustment amount of the process parameters is converted into control instructions that can be executed by the machine tool PLC, and the control instructions are issued through the CNC machine tool communication module to achieve control of the surface residual stress.

2. The method for actively controlling surface residual stress during CNC milling according to claim 1, characterized in that: In step 1, the process parameters include cutting speed v c , feed rate f, cutting depth a p and cutting width a e ; The method steps for the range analysis of the control object are: 11) According to the cutting speed v c , feed rate f, cutting depth a p and cutting width a e A four-factor five-level orthogonal experiment was set up, and the residual stress values ​​of the milling surface under different parameters were collected; 12) Perform range analysis on the experimental results 121) Calculate the mean of the experimental results for each level of each factor; 122) Calculate the range of the mean level of each factor; 123) Compare the range of each factor. The factor with the largest range is the factor that has the most significant impact on the surface residual stress, and the key factor for controlling the surface residual stress is identified.

3. The method for actively controlling surface residual stress during CNC milling according to claim 1, characterized in that: In the step 22), the surface residual stress monitoring model is: Rs t =f NN (v c ,f,a p ,a e ,F t ) Of which: Rs t represents the surface residual stress obtained by cutting the control object at the current time t; F t is the cutting force signal data collected by the sensor at the current time t; v c is the cutting speed; f is the feed rate; e is the cutting speed; p is the cutting depth; a e is the cutting width.

4. The method for actively controlling surface residual stress during CNC milling according to claim 3, characterized in that: In the step 23), the method for performing state estimation on the residual stress and obtaining the monitoring value of the surface residual stress is: Rs t+N =f NN (v c +Δμ(m),f,a p ,a e ,F t+N ) F t =f ANN (v c ,f,a p ,a e ) Of which: Rs t+N Indicates the residual stress state of the control object at the next moment t+N; F t+N represents the cutting force state at the next moment t+N; Δμ(m) represents the process parameter adjustment; f ANN (·) represents the mapping relationship between process parameters and cutting forces based on artificial neural network ANN; N is the time step.

5. The method for actively controlling surface residual stress during CNC milling according to claim 1, characterized in that: In step 25), the method steps of using nonlinear model predictive control to perform rolling time domain optimization solution on the surface residual stress are as follows: 251) Set optimization goals The optimization target is set as the process parameter adjustment amount Δμ selected by the range analysis, and the process parameter adjustment amount is minimized min(Δμ) to minimize the impact on the stability of the milling process while ensuring that the surface residual stress during the machining process remains within the expected range; 252) Set constraints The constraints of rolling optimization in the process of solving the NMPC controller for residual stress on the milling surface are set as follows: Rs t+N =f NN (v c +Δμ(m),f,a p ,a e ,F t+N )≤0 Of which: Rs t+N Indicates the residual stress state of the control object at the next moment t+N; F t+N represents the cutting force state at the next moment t+N; Δμ(m) represents the process parameter adjustment; f ANN (·) represents the mapping relationship between process parameters and cutting forces based on artificial neural network ANN; N is the time step; 253) Optimization solution Convert a constrained optimization process into an unconstrained optimization: min(Dm)stf NN (v c +Δμ(m),f,a p ,a e ,F t+N )≤0 By introducing the barrier function into the above constrained optimization process and incorporating the constraints into the objective function, we can obtain the unconstrained optimization problem: Where: θ is the barrier parameter; log(―f NN (v c +Δμ(m),f,a p ,a e ,F t+N ) i ) is the barrier function; 254) Iterative solution: Select the initial point and the initial barrier parameter θ, and gradually reduce the barrier parameter so that the iterative point gradually approaches the constraint boundary, and finally obtain the optimal solution that meets the convergence conditions.

6. The method for actively controlling surface residual stress during CNC milling according to claim 5, characterized in that: In step 254), since an optimization problem needs to be solved in each control cycle, the optimization problems of adjacent time steps are similar; the previous optimization result Δμ N―1 As the initial value Δμ of the current optimization problem N , in order to reduce the number of iterations of the optimization algorithm.

7. The method for actively controlling surface residual stress during CNC milling according to claim 1, characterized in that: In the step 3, the control instructions are sent through the CNC machine tool communication module to achieve the method steps of controlling the surface residual stress: 31) Allocate IP addresses to the host computer and CNC machine tool PLC; 32) Create a client socket on the host computer, establish a connection with the PLC of the CNC machine tool based on the TCP protocol, and initiate a server socket connection request by specifying the IP address and port number of the PLC of the CNC machine tool; 33) Modify the process parameter multiplier parameters stored in the PLC according to the cyclically sent control instructions, and realize surface residual stress control by continuously adjusting the process parameters.

Images