Automatic cutting speed control method for precise stone cutting

Abstract

The invention discloses an automatic cutting speed control method for precise stone cutting, which belongs to the technical field of control and comprises the following steps: S1, constructing an automatic cutting speed control system; s2, improving a neural population dynamic optimization algorithm by improving an attractor trend strategy and providing a self-adaptive quantum guide disturbance strategy, and constructing an improved neural population dynamic optimization algorithm module; s3, optimizing Kp, Ki and Kd of a PID closed-loop control algorithm through an improved neural population dynamic optimization algorithm to realize optimization of PID closed-loop control; s4, the error value of the stone cutting speed at the t-th moment and the target cutting speed is input into a PID closed-loop control module, the cutting speed control increment at the t-th moment is output and input into a rotating speed driving module of the permanent magnet synchronous motor, and the stone cutting speed is output; and S5, returning to execute the step S4 until the stone cutting speed meets the condition that the error value between the stone cutting speed and the target cutting speed is zero, so that automatic cutting speed accurate control of stone cutting is realized.

Description

Technical Field

[0001] The invention belongs to the field of control technology optimization, and in particular relates to an automatic cutting speed control method for precision stone cutting. Background Art

[0002] In the process of stone cutting, precise cutting speed control plays a vital role in ensuring cutting quality, extending tool life, reducing energy consumption and improving production efficiency. Accurate cutting speed can ensure that the cutting surface is smooth, without cracks or excessive wear; too high or too low cutting speed will lead to rough cutting surface, cracks or uneven cutting, affecting the appearance and quality of the final product, especially for some high-value stones, the appearance and quality of the stone directly affect the value of the stone.

[0003] PID closed-loop control method is widely used in the adjustment of cutting speed and feed rate. In stone cutting, factors such as cutting force, cutting temperature or vibration will affect the cutting quality. The closed-loop control system adjusts the cutting speed process in real time through a feedback mechanism. A major disadvantage of PID closed-loop control is that it assumes that the system model is known and stable. However, in practical applications, especially in complex and variable processes such as stone cutting, factors such as cutting force, friction coefficient, material hardness, etc. may change continuously, which makes it difficult for PID closed-loop control to cope with the complexity and uncertainty of the system.

[0004] The mathematical model of the neural population dynamic optimization algorithm (NPDOA) is based on the theory of neural population dynamics, which simulates the changes in the neural state of the brain during cognition and decision-making. The algorithm contains three key dynamic strategies: attractor trend strategy, coupling interference strategy and information projection strategy. NPDOA optimizes problems by simulating the dynamic behavior of neural populations. However, in the optimization problem of precision cutting speed control of stone, the coupling and local attraction between the algorithm's agent individuals will cause the entire algorithm's population to fall into a local optimal solution. Especially during the optimization process, the attractors will be too concentrated, resulting in the loss of diversity in the search space, thereby losing the ability to explore globally, resulting in low accuracy of the optimal solution and affecting the precision cutting speed control of stone. Summary of the invention

[0005] Based on the above problems, the present invention improves and optimizes the traditional PID closed-loop control method and the neural population dynamic optimization algorithm, so that the neural population dynamic optimization algorithm and the traditional PID closed-loop control method are more suitable for the precision cutting speed control of stone cutting. The core is to use the improved neural population dynamic optimization algorithm to optimize the PID closed-loop control method of the precision cutting speed for stone cutting, so as to improve the precision cutting speed control accuracy of stone.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme: an automatic cutting speed control method for precision stone cutting, wherein the stone cutting speed is achieved by driving a permanent magnet synchronous motor, and the speed of the permanent magnet synchronous motor directly controls the stone cutting speed, and the specific steps are: S1. construct an automatic cutting speed control system, the control system comprising: a PID closed-loop control module, an improved neural population dynamic optimization algorithm module, and a speed drive module of a permanent magnet synchronous motor; S2. Improve the standard neural population dynamic optimization algorithm by improving the attractor tendency strategy and proposing an adaptive quantum guided perturbation strategy, and construct an improved neural population dynamic optimization algorithm module; S3, optimizing the PID closed-loop control method by optimizing Kp, Ki and Kd of the PID closed-loop control algorithm through an improved neural population dynamic optimization algorithm; S4: Input the error value between the stone cutting speed at the tth moment and the target cutting speed into the PID closed-loop control module, and output the cutting speed control increment at the tth moment. The cutting speed control increment is input into the speed drive module of the permanent magnet synchronous motor, and the output stone cutting speed is ; S5, return to execute S4 until the stone cutting speed reaches Meet the error value with the target cutting speed Zero, realizing automatic cutting speed control for precision stone cutting.

[0007] Preferably, the stone cutting speed Depends on the speed of the permanent magnet synchronous motor. The stone cutting speed is controlled by driving the permanent magnet synchronous motor at the speed of the tth moment. , the mathematical model is: ; in, is the speed of the permanent magnet synchronous motor at the tth moment, is the proportional constant, by collecting n groups and The value is constructed.

[0008] Preferably, the speed driving module of the permanent magnet synchronous motor is the cutting speed control increment at the tth moment The second-order conversion model between the speed of the permanent magnet synchronous motor and the speed of the permanent magnet synchronous motor is used to input the cutting speed control increment at the tth moment, and the speed drive module of the permanent magnet synchronous motor is used to obtain the speed of the permanent magnet synchronous motor at the tth moment. The mathematical model is: ; Where, J is the inertia of the permanent magnet synchronous motor, is the speed of the permanent magnet synchronous motor at the tth moment, is the first-order derivative of the speed of the permanent magnet synchronous motor at the tth moment, is the torque constant of the permanent magnet synchronous motor, is the cutting speed control value at the t-1th moment, where: , is the cutting speed control value at the tth moment.

[0009] Preferably, the PID closed-loop control module serves as the core of the automatic cutting speed control system, and controls the stone cutting speed through an incremental PID closed-loop control algorithm, wherein the mathematical model of the incremental PID closed-loop control algorithm is: ; in, is the stone cutting speed adjustment amount, , and are the proportional coefficient, integral coefficient and differential coefficient of the PID closed-loop control algorithm, and is the cutting speed error value at time t-1 and t-2, It is the error value between the speed of the permanent magnet synchronous motor at the tth moment and the set target speed.

[0010] Preferably, in order to improve the precision stone cutting effect, a material adaptability factor is introduced , adjust the stone cutting speed, the dynamic adjustment formula of cutting speed: ; in, is the cutting speed after dynamic adjustment. For stone cutting speed.

[0011] Preferably, the friction will cause the temperature between the teeth and the stone to rise, thus affecting the efficiency of the cutting process. The friction will not only increase with time, but also cause a temperature effect as the cutting time increases. This effect will affect the hardness of the material and the cutting speed. The present invention introduces a temperature factor , which will increase with the increase of friction and cutting time, the mathematical model is: ; in, is the temperature factor value at time t, is the initial temperature factor, i.e. the initial temperature value, γ is the temperature gain coefficient caused by friction, which controls the effect of friction on temperature, and T is the total operating time; As the temperature increases, the influence of friction will become more significant. The present invention introduces a nonlinear friction coefficient , which changes with increasing temperature, to express the effect of friction on cutting speed, which reflects the friction increase effect caused by high temperature. The mathematical model is: ; in, is the maximum value of the temperature factor, that is, the maximum value of the temperature factor, is the adjustment coefficient, which indicates the intensity of the effect of temperature on friction; According to the temperature factor and nonlinear friction coefficient Design material adaptability factor , the mathematical model is: ; in, is the material adaptability factor at the tth moment.

[0012] Preferably, in the process of optimizing the PID closed-loop control algorithm by the improved neural population dynamic optimization algorithm, the proxy individual position represents the solution of Kp, Ki and Kd of the PID closed-loop control algorithm, that is, the proxy individual position is a solution containing three parameters of the PID closed-loop control algorithm (proportional coefficient Kp, integral coefficient Ki and differential coefficient Kd), and each proxy individual corresponds to a parameter combination in the parameter space of the PID closed-loop control algorithm. The mathematical model is: , where is the position of the ith agent, , and are the proportional coefficient, integral coefficient and differential coefficient of the i-th agent.

[0013] Preferably, a fitness function is designed, and the improved neural population dynamic optimization algorithm is guided by the fitness function to optimize the proportional coefficient Kp, integral coefficient Ki and differential coefficient Kd of the PID closed-loop control algorithm. The input of the fitness function is the agent individual position, that is, a parameter combination in the parameter space of the PID closed-loop control algorithm is input into the automatic cutting speed control system, and the cutting speed control increment output by the automatic cutting speed control system acts on the speed state of the permanent magnet synchronous motor. The mathematical model of the fitness function is: ; in, and is the weight coefficient, T is the total running time, The speed at the tth moment controls the stone cutting speed. is the target cutting speed.

[0014] Preferably, the scaling factor of the current attractor-oriented strategy is Based on a fixed ratio, the present invention proposes an adaptive nonlinear scaling factor method based on local sensitivity and optimal solution information, wherein the scaling factor is adjusted according to the similarity between each neural population and its neighboring agent individuals and the dynamic change of the optimal solution, so that the convergence is more intelligent and dynamic. When the current agent individual is very similar to other agent individuals, the scaling factor becomes smaller, which means that the attractor trend of the population is weakened, encouraging more exploration; when the current agent individual is greatly different from other agent individuals, the scaling factor increases, enhancing the attractor trend and promoting faster convergence. The specific steps are: S201. Design a nonlinear scaling factor adjustment mechanism based on similarity and define local similarity is the position of the i-th agent Individual positions with other agents The similarity in the search space [lb, ub], the mathematical model is: ; Among them, iter is the current number of iterations, iter=1,2,...,Max_iter; Max_iter is the maximum number of iterations; S202, based on local similarity and the best fitness value Improved scaling factor of attractor-oriented strategy , the mathematical model is: ; in, is the scaling factor of the iter-th iteration of the ith agent, is the initial scaling factor, is the optimal fitness in the current population, is the fitness value of the position of the i-th agent at the iter-th iteration; S203, using the improved scaling factor to improve the attractor trending strategy of the neural population dynamic optimization algorithm; the present invention uses the agent individual corresponding to the minimum fitness value of each iteration as the attractor att, and the mathematical model is: ; In the formula, is the updated position of the i-th agent in the attractor-oriented strategy phase, is a random number between 0 and 1. is the attractor position of the iter-th iteration, represents the Gaussian noise of external input interference, The position of the i-th agent at the iter-th iteration.

[0015] Preferably, by introducing an improved attractor trend strategy, NPDOA can optimize the PID algorithm for automatic cutting speed control of stone cutting. During the dynamic adjustment process, the improved attractor trend strategy can improve the adaptability of the NPDOA algorithm to the changing environment, ensure local search while enhancing the global search capability, and refine the PID closed-loop control algorithm parameter adjustment, thereby achieving more efficient and precise automatic cutting control, which not only improves the cutting quality and speed, but also makes the control system more stable.

[0016] Preferably, the coupling interference simulates the interaction between neural populations, including additive coupling and diffusive coupling; the additive coupling is expressed by the following formula: ; in, This is the position of the agent individual after adjustment through diffusion and additive coupling under the coupled interference strategy; N is the total number of agent individuals, The position of the i-th agent individual in the iter-th iteration; A random number between 0 and 1; Diffusion coupling is expressed by the following formula: ; in, This is the position of the i-th agent individual after adjustment through differentiation under the coupled interference strategy; The position of the jth agent in the iterth iteration; A random number between 0 and 1; Combining additive coupling with diffusion coupling to update the individual agent position, the mathematical model is: ; in, is the updated position of the i-th agent in the coupled disturbance strategy phase, and d is the scaling factor of the coupled disturbance.

[0017] Preferably, the present invention abandons the information projection strategy of the original NPDOA algorithm and proposes an adaptive quantum guided perturbation strategy, so that the solutions in the Kp, Ki and Kd process of the PID closed-loop control algorithm of the automatic cutting speed control system are not limited to the existing solution space, but the superposition and coherence of quantum states are used to simultaneously explore and calculate in multiple possible solutions, and then the optimal solution is selected according to the interference effect; the solution is recorded as the proxy individual position, and the specific steps are: S301, each solution is represented as a quantum state ψ(x), and the fitness function value of each solution corresponds to the energy of the quantum state, and the mathematical model is: ; in, is a complex coefficient, i=1,2,...,N; it represents the amplitude of the solution in the quantum state, is the position of the i-th agent individual in the iter-th iteration, i=1,2,...,N; S302, simulate quantum tunneling, allowing the NPDOA algorithm to pass through the poor local optimal solution in the solution space and directly jump to a better solution. The mathematical model is: ; in, is the tunneling probability of the i-th agent from the current solution to the global optimal solution of the current iteration, is the position of the i-th agent individual in the iter-th iteration, i=1,2,...,N; is the optimal individual proxy position of the iter-th iteration, R is the tunneling coefficient, which controls the intensity of the tunneling effect; for The fitness value of for The fitness value of S303, a randomly generated number rand∈[0,1], if rand is less than , then the solution jump is allowed, that is, the position of the updated solution is close to the position of the best agent individual, otherwise S304 is executed, and the mathematical model is: ; in, is the updated agent individual position, is the step length, which controls the amplitude of the jump; S304. By simulating quantum interference, the NPDOA algorithm competes among multiple potential solutions during the optimization process and selects the optimal solution based on the interference results. The mathematical model is: ; in, To solve for the probability of interference under interference, h is Planck's constant, which determines the frequency of the interference effect; S305. Adaptively guide the perturbation process through the interference effect. In the optimization process, first select the most promising solution according to the interference result of the quantum state, and then apply the quantum tunneling effect to explore in the neighborhood of the solution. The mathematical model is: ; in, is the updated agent individual position, is the perturbation amplitude, which controls the size of the position update; is the disturbance direction of the i-th agent.

[0018] Preferably, by simulating quantum tunneling, it is described that a quantum particle can pass through an energy barrier that cannot be crossed in classical mechanics. In the process of optimizing Kp, Ki and Kd of the PID closed-loop control algorithm of the automatic cutting speed control system, the simulated quantum tunneling proposed by the present invention can be used to break through the local optimum, thereby exploring a broader solution space. Specifically, the search process of the solution space does not strictly follow the traditional local search, but allows the algorithm to "tunnel" through the poor local optimal solution in the solution space and directly jump to a better solution; quantum interference is when multiple quantum states interact with each other, constructive interference and destructive interference may occur, thereby affecting the solution search in the optimization process. Quantum interference usually determines the "phase" or "direction" of the quantum state, and in the optimization process, reflects the "importance" or "priority" of the solution. Through quantum interference, it is determined which solutions will affect the update of the current solution. According to the result of quantum interference, the present invention calculates the direction of solution update. , quantum interference is equivalent to the interaction between multiple solutions in the solution space.

[0019] Preferably, the PID closed-loop control method is optimized by optimizing Kp, Ki and Kd of the PID closed-loop control algorithm using an improved neural population dynamic optimization algorithm. The specific steps are as follows: S401, mapping the Kp, Ki and Kd parameters as a space vector to the agent individuals of the improved neural population dynamic optimization algorithm; initializing the maximum number of iterations Max_iter, the total number of agent individuals N, the upper bound ub and lower bound lb of the search space, and the problem dimension D; S402, using a random assignment method to initialize the agent population position of the improved neural population dynamic optimization algorithm, wherein the agent population includes N agent individuals; S403, using the fitness function to calculate the fitness value of each agent individual position, retaining the agent individual position corresponding to the current minimum fitness value as the best agent individual position for the iterth iteration; S404, determine whether the current iteration number iter satisfies iter greater than Max_iter, if so, output the best proxy individual position as the optimal solution, resolve the best proxy individual position into Kp, Ki and Kd parameter values, and record them as the best Kp, Ki and Kd parameter values; otherwise, execute S405 to S407 to update the proxy individual position; S405. Calculate the improved scaling factor , using the improved attractor tendency strategy to update the position of individual agents; S406, using a coupling interference strategy to update the position of the agent individual; S407, introduce an adaptive quantum guided perturbation strategy, guide the perturbation of the state of the neural population agent individual through the "quantum tunneling effect" and "quantum interference effect" in quantum mechanics, and update the agent individual position; S408, using the fitness function to calculate the current proxy individual fitness value, retaining the proxy individual position value corresponding to the minimum fitness value, which is the optimal proxy individual position for the iterth iteration , and the best agent individual position of the previous iteration , retain the smaller position value of the two, which is the global optimal population individual position value , return to execute S404.

[0020] Compared with the existing methods, the beneficial effects and innovations of the present invention are as follows: the present invention optimizes the parameter search process in the PID closed-loop control by simulating quantum tunneling and interference effects, so that the solution space is no longer limited to the traditional local search, but performs global search and jump through the principles of quantum mechanics to avoid falling into the local optimum; by combining the improved attractor tendency strategy, coupled interference strategy and quantum guided perturbation strategy, the adaptability of the neural population algorithm under complex cutting conditions is enhanced, and the accuracy and stability of the PID closed-loop control algorithm are improved; the improved neural population dynamic optimization algorithm (NPDOA) is combined with the traditional PID closed-loop control to optimize the precision cutting speed control method of stone cutting, so that the control system can effectively improve the stone cutting accuracy, reduce the error in cutting, and improve the cutting quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of the automatic cutting speed control method for precision stone cutting proposed by the present invention; Figure 2 Flow chart for optimizing PID closed-loop control method; Figure 3 The curve diagram of the change of fitness value during the optimization process of the standard NPDOA algorithm and the improved NPDOA algorithm Figure 4 This is the trend diagram of the three parameters of the PID closed-loop control algorithm during the optimization process of the standard NPDOA algorithm; Figure 5 This is the trend diagram of the three parameters of the PID closed-loop control algorithm during the optimization process of the improved NPDOA algorithm; Figure 6 The response curves of three different PID control methods in the automatic cutting speed control system. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] The present invention discloses an automatic cutting speed control method for precision stone cutting, wherein the stone cutting speed is realized by driving a permanent magnet synchronous motor, and the speed of the permanent magnet synchronous motor directly controls the stone cutting speed. The specific implementation method includes mathematical model establishment and software simulation; wherein the mathematical model establishment is specifically as follows Figure 1 As shown in S1 to S5.

[0024] S1. Construct an automatic cutting speed control system, wherein the control system includes: a PID closed-loop control module, an improved neural population dynamic optimization algorithm module, and a speed drive module of a permanent magnet synchronous motor.

[0025] Specifically, an automatic cutting speed control system simulation model is built in Siumlink, including an improved PID closed-loop control model, an objective function module, and a speed drive module of a permanent magnet synchronous motor. The input of the improved PID closed-loop control model in Simulink is the optimal Kp, Ki, and Kd parameter values; the objective function module adopts a fitness function, and the speed drive module of the permanent magnet synchronous motor is the Laplace transform form of the speed drive mathematical model of the permanent magnet synchronous motor; the time domain form of the speed drive mathematical model of the permanent magnet synchronous motor is: ; Where J is the inertia of the permanent magnet synchronous motor, which is set to 0.444 during the implementation process. is the speed of the permanent magnet synchronous motor at the tth moment, is the first-order derivative of the speed of the permanent magnet synchronous motor at the tth moment, is the torque constant of the permanent magnet synchronous motor, which is set to 2 during implementation. is the cutting speed control value at the t-1th moment, where: , is the cutting speed control value at the tth moment.

[0026] Specifically, the speed drive model of the permanent magnet synchronous motor in Siumlink adopts the complex frequency domain form, and the initial permanent magnet synchronous motor speed w(0)=0; the speed drive model of the permanent magnet synchronous motor in the complex frequency domain for: ; in, is the complex frequency domain form of the permanent magnet synchronous motor speed at the t moment, is the complex frequency domain form of the cutting speed control quantity at the tth moment, is a complex frequency domain variable.

[0027] S2. By improving the attractor tendency strategy and proposing an adaptive quantum-guided perturbation strategy, the standard neural population dynamic optimization algorithm is improved, and an improved neural population dynamic optimization algorithm module is constructed.

[0028] Specifically, the present invention proposes an adaptive nonlinear scaling factor method based on local sensitivity and optimal solution information, and the specific steps are: S201. Design a nonlinear scaling factor adjustment mechanism based on similarity and define local similarity is the position of the i-th agent Individual positions with other agents The similarity in the search space [lb, ub], the mathematical model is: ; Among them, iter is the current iteration number, iter=1,2,...,Max_iter; Max_iter is the maximum iteration number, which is set to 100 during the implementation; S202, based on local similarity and the best fitness value Improved scaling factor of attractor-oriented strategy , the mathematical model is: ; in, is the scaling factor of the iter-th iteration of the ith agent, is the initial scaling factor, the initial value is set to 0.5, and the value range is 0 to 1; is the optimal fitness in the current population, is the fitness value of the position of the i-th agent at the iter-th iteration; S203, using the improved scaling factor to improve the attractor trending strategy of the neural population dynamic optimization algorithm; the present invention uses the agent individual corresponding to the minimum fitness value of each iteration as the attractor att, and the mathematical model is: ; In the formula, is the updated position of the i-th agent in the attractor-oriented strategy phase, is a random number between 0 and 1. is the attractor position of the iter-th iteration, represents the Gaussian noise of external input interference, The position of the i-th agent at the iter-th iteration.

[0029] Specifically, the additive coupling mathematical model is constructed by the following formula: ; in, This is the position of the agent individual after adjustment through diffusion and additive coupling under the coupled interference strategy; N is the total number of agent individuals, The position of the i-th agent individual in the iter-th iteration; A random number between 0 and 1; Diffusion coupling is expressed by the following formula: ; in, This is the position of the i-th agent individual after adjustment through differentiation under the coupled interference strategy; The position of the jth agent in the iterth iteration; A random number between 0 and 1; Combining additive coupling with diffusion coupling to update the individual agent position, the mathematical model is: ; in, is the updated position of the i-th agent in the coupled disturbance strategy phase, and d is the scaling factor of the coupled disturbance.

[0030] Specifically, the present invention abandons the information projection strategy of the original NPDOA algorithm and proposes an adaptive quantum guided perturbation strategy, which uses the superposition and coherence of quantum states to simultaneously explore and calculate multiple possible solutions, and then selects the optimal solution based on the interference effect; the solution is recorded as the proxy individual position, and the specific steps are: S301, each solution is represented as a quantum state ψ(x), and the fitness function value of each solution corresponds to the energy of the quantum state, and the mathematical model is: ; in, is a complex coefficient, i=1,2,...,N; it represents the amplitude of the solution in the quantum state, is the position of the i-th agent individual in the iter-th iteration, i=1,2,...,N; N is the total number of agent individuals, which is set to 20 during the implementation; S302, simulate quantum tunneling, allowing the NPDOA algorithm to pass through the poor local optimal solution in the solution space and directly jump to a better solution. The mathematical model is: ; in, is the tunneling probability of the i-th agent from the current solution to the global optimal solution of the current iteration, is the position of the i-th agent individual in the iter-th iteration, i=1,2,...,N; is the optimal individual proxy position of the iter-th iteration, R is the tunneling coefficient, which controls the intensity of the tunneling effect; for The fitness value of for The fitness value of S303, a randomly generated number rand∈[0,1], if rand is less than , then the solution jump is allowed, that is, the position of the updated solution is close to the position of the best agent individual, otherwise S304 is executed, and the mathematical model is: ; in, is the updated agent individual position, is the step length, which controls the amplitude of the jump; S304. By simulating quantum interference, the NPDOA algorithm competes among multiple potential solutions during the optimization process and selects the optimal solution based on the interference results. The mathematical model is: ; in, To solve the probability of interference under interference, h is Planck's constant, which determines the frequency of the interference effect and is set to 0.5 during the implementation of the present invention; S305. Adaptively guide the perturbation process through the interference effect. In the optimization process, first select the most promising solution according to the interference result of the quantum state, and then apply the quantum tunneling effect to explore in the neighborhood of the solution. The mathematical model is: ; in, is the updated agent individual position, is the disturbance amplitude, which controls the size of the position update; during the implementation, the initial value is set to 0.3, and subsequent iterations are randomly generated within 0 to 1; is the disturbance direction of the i-th agent, which can be any number between 0 and 1.

[0031] More specifically, the tunneling coefficient R, the step length step, and the proportional factor d of the coupling disturbance in the present invention are designed in such a way that as the number of iterations increases, the tunneling coefficient and the step length gradually decrease, thereby reducing the exploratory and jumping characteristics and the coupling disturbance characteristics and enhancing the development characteristics. The mathematical models of R, step, and d are defined as follows: ; Among them, iter is the current number of iterations, iter=1,2,...,Max_iter; Max_iter is the maximum number of iterations, which is set to 100 during implementation.

[0032] S3. The PID closed-loop control method is optimized by optimizing Kp, Ki and Kd of the PID closed-loop control algorithm through an improved neural population dynamic optimization algorithm.

[0033] Specifically, in the process of optimizing the PID closed-loop control algorithm by the improved neural population dynamic optimization algorithm, the agent individual position represents the solution of Kp, Ki and Kd of the PID closed-loop control algorithm, that is, the agent individual position is a solution containing three parameters of the PID closed-loop control algorithm (proportional coefficient Kp, integral coefficient Ki and differential coefficient Kd). Each agent individual corresponds to a parameter combination in the parameter space of the PID closed-loop control algorithm. The mathematical model is: , where is the position of the ith agent, , and are the proportional coefficient, integral coefficient and differential coefficient of the i-th agent.

[0034] Specifically, the PID closed-loop control method is optimized by optimizing Kp, Ki and Kd of the PID closed-loop control algorithm using an improved neural population dynamic optimization algorithm. The specific steps are as follows: S401, mapping the Kp, Ki and Kd parameters as a space vector to the agent individuals of the improved neural population dynamic optimization algorithm; initializing the maximum number of iterations Max_iter, the total number of agent individuals N, the upper bound ub and lower bound lb of the search space, and the problem dimension D; S402, using a random assignment method to initialize the agent population position of the improved neural population dynamic optimization algorithm, wherein the agent population includes N agent individuals; S403, using the fitness function to calculate the fitness value of each agent individual position, retaining the agent individual position corresponding to the current minimum fitness value as the best agent individual position for the iterth iteration; S404, determine whether the current iteration number iter satisfies iter greater than Max_iter, if so, output the best proxy individual position as the optimal solution, resolve the best proxy individual position into Kp, Ki and Kd parameter values, and record them as the best Kp, Ki and Kd parameter values; otherwise, execute S405 to S407 to update the proxy individual position; S405. Calculate the improved scaling factor , using the improved attractor tendency strategy to update the position of individual agents; S406, using a coupling interference strategy to update the position of the agent individual; S407, introduce an adaptive quantum guided perturbation strategy, guide the perturbation of the state of the neural population agent individual through the "quantum tunneling effect" and "quantum interference effect" in quantum mechanics, and update the agent individual position; S408, using the fitness function to calculate the current proxy individual fitness value, retaining the proxy individual position value corresponding to the minimum fitness value, which is the optimal proxy individual position for the iterth iteration , and the best agent individual position of the previous iteration , retain the smaller position value of the two, which is the global optimal population individual position value , return to execute S404.

[0035] Specifically, the improved neural population dynamic optimization algorithm is guided by the fitness function to optimize the proportional coefficient Kp, integral coefficient Ki and differential coefficient Kd of the PID closed-loop control algorithm. The input of the fitness function is the agent individual position, that is, a parameter combination in the parameter space of the PID closed-loop control algorithm is input into the automatic cutting speed control system. The cutting speed control increment output by the automatic cutting speed control system acts on the speed state of the permanent magnet synchronous motor. The mathematical model of the fitness function is: ; in, and is the weight coefficient, which is set to 0.5 during the implementation process, T is the total running time, which is set to 20 seconds in the implementation of the present invention, The speed at the tth moment controls the stone cutting speed. is the target cutting speed.

[0036] S4: Input the error value between the stone cutting speed at the tth moment and the target cutting speed into the PID closed-loop control module, and output the cutting speed control increment at the tth moment. The cutting speed control increment is input into the speed drive module of the permanent magnet synchronous motor, and the output stone cutting speed is .

[0037] Specifically, the PID closed-loop control module adopts an incremental PID closed-loop control algorithm, and the mathematical model is: ; in, is the stone cutting speed adjustment amount, , and are the proportional coefficient, integral coefficient and differential coefficient of the PID closed-loop control algorithm, and is the cutting speed error value at time t-1 and t-2, It is the error value between the speed of the permanent magnet synchronous motor at the tth moment and the set target speed.

[0038] Specifically, the stone cutting speed is controlled by driving the permanent magnet synchronous motor at the speed of the first moment. , the mathematical model is: ; in, is the speed of the permanent magnet synchronous motor at the tth moment, is the proportional constant, by collecting n groups and The value is constructed and set to 1.

[0039] Specifically, a material adaptability factor is introduced , adjust the stone cutting speed, the dynamic adjustment formula of cutting speed: ; in, is the cutting speed after dynamic adjustment. For stone cutting speed, the specific implementation steps are: S501, introduce a temperature factor , which will increase with the increase of friction and cutting time, the mathematical model is: ; in, is the temperature factor value at time t, is the initial temperature factor, that is, the initial temperature value is set to 20, γ is the temperature gain coefficient caused by friction, which controls the effect of friction on temperature and is set to 0.5, and T is the total running time; S502, introduce a nonlinear friction coefficient , which changes with increasing temperature, to express the effect of friction on cutting speed, which reflects the friction increase effect caused by high temperature. The mathematical model is: ; in, is the maximum value of the temperature factor, that is, the maximum value of the temperature factor, set to 100. is the adjustment coefficient, which indicates the influence of temperature on friction and is set to 0.5; S503, according to the temperature factor and nonlinear friction coefficient Design material adaptability factor , the mathematical model is: ; in, is the material adaptability factor at the tth moment.

[0040] S5, return to execute S4 until the stone cutting speed reaches Meet the error value with the target cutting speed Zero, realizing automatic cutting speed control for precision stone cutting.

[0041] More specifically, the software design of the mathematical model of the method of the present invention is completed in Matlab. The final parameters of the program running in the Main function include the maximum number of iterations Max_iter set to 100, the total number of agent individuals N set to 20, the upper limit ub of the search space set to 90 and the lower limit lb set to 0.01, and the problem dimension D set to 3. The fitness function value is assigned to the objective function, and the code for parameter mapping in the optimization process of the PID control algorithm of the automatic cutting speed control system by the standard neural population dynamic optimization algorithm and the improved neural population dynamic optimization algorithm of the present invention is established; and the final optimized Kp, Ki and Kd parameter values ​​as well as the fitness function value and the control effect are plotted, and the program is run. The code is as follows: %% Clear the environment clear all; close all; clc; warning off; Max_iter = 100; N = 20; lb = 0.01; ub = 90; D = 3; f = @(x) Best_PID(x); [Best_score_1,Best_solution_1,K_PID1,fitness_1]=NPDOA(Max_iter,lb,ub,D,fobj); [best_score_2,best_solution_2,K_PID2,fitness_2]=GNPDOA(Max_iter,lb,ub,D,fobj); Get as Figures 3 to 6 .

[0042] More specifically, the fitness function measures the performance of the PID controller, with the goal of minimizing the control error. By continuously adjusting the parameters of the PID controller, the fitness function is optimized to reach its minimum value; e.g. Figure 3As shown in the figure, the improved NPDOA algorithm is significantly better than the standard NPDOA algorithm; the fitness function value of the standard NPDOA algorithm decreases slowly, and there is a long stagnation period in the middle stage, falling into the local optimum, and reaching the best solution after 46 iterations, with a fitness value of 6.57445; the fitness function value of the improved NPDOA algorithm decreases rapidly and stabilizes, showing that the method of the present invention can find the global optimal solution faster during the optimization process, and finds a better solution than the standard NPDOA algorithm after 11 iterations, with a fitness value of 4.08995; this shows that the improved NPDOA algorithm has higher efficiency and convergence speed when optimizing the parameters of the PID closed-loop control algorithm.

[0043] More specifically, if Figure 4 and Figure 5 As shown in Figure 3, the changing trends of the three parameters of the PID closed-loop control algorithm in the optimization process of the standard and improved NPDOA algorithms correspond to the changes in the fitness values. When optimizing the PID parameters, the convergence speed of the standard NPDOA algorithm is not as fast as expected. Although the adjustments of Kp, Ki, and Kd gradually tend to be stable, the convergence process is slow, especially in the initial stage of finding the optimal solution. Finally, Kp stabilizes after about 30 iterations without significant fluctuations. Ki stabilizes after about 35 iterations. Kd changes the slowest and has no obvious changes in the early stage until about 4 It started to decrease and stabilize after 6 iterations, and the final optimal control parameters were Kp=10.5216, Ki=0.559252, and Kd=1.27043. Compared with the standard NPDOA algorithm, the improved NPDOA showed a faster convergence speed in the optimization process of Kp, and the adjustment of Ki was more efficient. Although Kd did not change significantly, its stability was well reflected in the optimization process. The final optimal control parameters were Kp=86.7229, Ki=0.506707, and Kd=11.2386.

[0044] More specifically, if Figure 6 The response curves of the three different PID control methods in the automatic cutting speed control system are shown. The cutting speed control amount is set to 7. The standard PID method shows obvious overshoot and slow response speed. It takes longer for the system to stabilize to the target value. The overshoot phenomenon is more obvious, and the system needs time to stabilize. Compared with the standard PID, the ordinary optimization algorithm-PID method reduces the overshoot and improves the response speed of the system. However, although the system approaches the target value faster, there is still a certain oscillation, and the stability is improved. The optimization algorithm-PID method of the present invention has the best control effect, almost no overshoot, the system quickly stabilizes to the target value, and the response speed is the fastest. The improved NPDOA algorithm of the present invention can accurately reach the target in the shortest time without overshoot and oscillation.

Claims

1. An automatic cutting speed control method for precision stone cutting, characterized in that: The specific steps are: S1. construct an automatic cutting speed control system, the control system comprising: a PID closed-loop control module, an improved neural population dynamic optimization algorithm module, and a speed drive module of a permanent magnet synchronous motor; S2. Improve the standard neural population dynamic optimization algorithm by improving the attractor tendency strategy and proposing an adaptive quantum guided perturbation strategy, and construct an improved neural population dynamic optimization algorithm module; S3, optimizing the PID closed-loop control method by optimizing Kp, Ki and Kd of the PID closed-loop control algorithm through an improved neural population dynamic optimization algorithm; S4: Input the error value between the stone cutting speed at the tth moment and the target cutting speed into the PID closed-loop control module, and output the cutting speed control increment at the tth moment. The cutting speed control increment is input into the speed drive module of the permanent magnet synchronous motor, and the output stone cutting speed is ; S5, return to execute S4, until the stone cutting speed Meet the error value with the target cutting speed Zero, realizing automatic cutting speed control for precision stone cutting.

2. The automatic cutting speed control method for precision stone cutting according to claim 1, characterized in that: The speed drive module of the permanent magnet synchronous motor is the cutting speed control increment at the tth moment The second-order conversion model between the speed of the permanent magnet synchronous motor and the speed of the permanent magnet synchronous motor is used to input the cutting speed control increment at the tth moment, and the speed of the permanent magnet synchronous motor at the tth moment is obtained through the speed drive module of the permanent magnet synchronous motor. The mathematical model is: ； Where, J is the inertia of the permanent magnet synchronous motor, is the torque constant of the permanent magnet synchronous motor, It is the cutting speed control value at the moment t-1.

3. The automatic cutting speed control method for precision stone cutting according to claim 2 is characterized in that: The improved neural population dynamic optimization algorithm module includes: improving the standard neural population dynamic optimization algorithm by improving the attractor tendency strategy, specifically: S201. Design a nonlinear scaling factor adjustment mechanism based on similarity and define local similarity is the position of the i-th agent Individual positions with other agents The similarity in the search space [lb, ub], the mathematical model is: ； Among them, iter is the current number of iterations, iter=1,2,...,Max_iter; Max_iter is the maximum number of iterations; S202, based on local similarity and the best fitness value Improved scaling factor of attractor-oriented strategy , the mathematical model is: ； in, is the scaling factor of the iter-th iteration of the ith agent, is the initial scaling factor, is the optimal fitness in the current population, is the fitness value of the position of the i-th agent at the iter-th iteration; S203, using the improved scaling factor to improve the attractor trending strategy of the neural population dynamic optimization algorithm; the present invention uses the agent individual corresponding to the minimum fitness value of each iteration as the attractor att, and the mathematical model is: ； In the formula, is the updated position of the i-th agent in the attractor-oriented strategy phase, is a random number between 0 and 1. is the attractor position of the iter-th iteration, represents the Gaussian noise of external input interference, The position of the i-th agent at the iter-th iteration.

4. The automatic cutting speed control method for precision stone cutting according to claim 3 is characterized in that: An adaptive quantum guided perturbation strategy is proposed. It uses the superposition and coherence of quantum states to simultaneously explore and calculate multiple possible solutions, and then selects the optimal solution based on the interference effect. The solution is recorded as the individual position of the agent, specifically: S301, each solution is represented as a quantum state ψ(x), and the fitness function value of each solution corresponds to the energy of the quantum state, and the mathematical model is: ； in, is a complex coefficient, i=1,2,...,N; it represents the amplitude of the solution in the quantum state, is the position of the i-th agent individual in the iter-th iteration, i=1,2,...,N; S302, simulate quantum tunneling, allowing the NPDOA algorithm to pass through the poor local optimal solution in the solution space and directly jump to a better solution. The mathematical model is: ； in, is the tunneling probability of the i-th agent from the current solution to the global optimal solution of the current iteration, is the position of the i-th agent individual in the iter-th iteration, i=1,2,...,N; is the optimal individual proxy position of the iter-th iteration, R is the tunneling coefficient, which controls the intensity of the tunneling effect; for The fitness value of for The fitness value of S303, a randomly generated number rand∈[0,1], if rand is less than , then the solution jump is allowed, that is, the position of the updated solution is close to the position of the best agent individual, otherwise S304 is executed, and the mathematical model is: ； in, is the updated agent individual position, is the step length, which controls the amplitude of the jump; S304. By simulating quantum interference, the NPDOA algorithm competes among multiple potential solutions during the optimization process and selects the optimal solution based on the interference results. The mathematical model is: ； in, To solve for the probability of interference under interference, h is Planck's constant, which determines the frequency of the interference effect; S305. Adaptively guide the perturbation process through the interference effect. In the optimization process, first select the most promising solution according to the interference result of the quantum state, and then apply the quantum tunneling effect to explore in the neighborhood of the solution. The mathematical model is: ； in, is the updated agent individual position, is the perturbation amplitude, which controls the size of the position update; is the disturbance direction of the i-th agent.

5. The automatic cutting speed control method for precision stone cutting according to claim 4, characterized in that: The improved neural population dynamic optimization algorithm is used to optimize Kp, Ki and Kd of the PID closed-loop control algorithm. The specific steps are as follows: S401, mapping the Kp, Ki and Kd parameters as a space vector to the agent individuals of the improved neural population dynamic optimization algorithm; initializing the maximum number of iterations Max_iter, the total number of agent individuals N, the upper bound ub and lower bound lb of the search space, and the problem dimension D; S402, using a random assignment method to initialize the agent population position of the improved neural population dynamic optimization algorithm, wherein the agent population includes N agent individuals; S403, using the fitness function to calculate the fitness value of each agent individual position, retaining the agent individual position corresponding to the current minimum fitness value as the best agent individual position for the iterth iteration; S404, judging whether the current number of iterations iter satisfies iter greater than Max_iter, if so, outputting the best proxy individual position as the optimal solution, parsing the best proxy individual position into Kp, Ki and Kd parameter values, and recording them as the best Kp, Ki and Kd parameter values; Otherwise, execute S405 to S407 to update the agent individual position; S405. Calculate the improved scaling factor , using the improved attractor tendency strategy to update the position of individual agents; S406, using a coupling interference strategy to update the position of the agent individual; S407, introduce an adaptive quantum-guided perturbation strategy, guide the perturbation of the state of the neural population agent individual through the "quantum tunneling effect" and "quantum interference effect" in quantum mechanics, and update the agent individual position; S408, using the fitness function to calculate the current proxy individual fitness value, retaining the proxy individual position value corresponding to the minimum fitness value, which is the optimal proxy individual position for the iterth iteration , and the best agent individual position of the previous iteration , retain the smaller position value of the two, which is the global optimal population individual position value , return to execute S404.

6. The automatic cutting speed control method for precision stone cutting according to claim 1, characterized in that: The output stone cutting speed introduces a material adaptability factor , adjust the stone cutting speed, the dynamic adjustment formula of cutting speed: ； in, is the cutting speed after dynamic adjustment. is the stone cutting speed; the material adaptability factor design method steps are: S501, introduce a temperature factor , which will increase with the increase of friction and cutting time, the mathematical model is: ； in, is the temperature factor value at time t, is the initial temperature factor, i.e. the initial temperature value, γ is the temperature gain coefficient caused by friction, which controls the effect of friction on temperature, and T is the total operating time; S502, introduce a nonlinear friction coefficient , which changes with increasing temperature, to express the effect of friction on cutting speed, which reflects the friction increase effect caused by high temperature. The mathematical model is: ； in, is the maximum value of the temperature factor, that is, the maximum value of the temperature factor, is the adjustment coefficient, which indicates the intensity of the effect of temperature on friction; S503, according to the temperature factor and nonlinear friction coefficient Design material adaptability factor , the mathematical model is: ； in, is the material adaptability factor at the tth moment.

Images