Permanent magnet synchronous motor control optimization method and system

Abstract

The invention discloses a permanent magnet synchronous motor control optimization method and system. The method comprises the following steps: modeling a permanent magnet synchronous motor by using a linear secondary regulator to obtain a mathematical model of the motor; performing discretization derivation on the voltage equation of the motor model under the two-phase rotating coordinate system to obtain a prediction model of the motor; designing a q-axis cost function by using the motor prediction model based on performance requirements of the motor; optimizing a q-axis current weight coefficient in the q-axis cost function by using the motor prediction model in combination with a model prediction-based control algorithm and a particle swarm optimization algorithm, and outputting the optimized weight coefficient; and according to the optimized weight coefficient, a gain parameter in a motor controller is adjusted in real time, so that the current control precision and the overall control performance of the motor are optimized. The method has good adaptability, can cope with complex working conditions, and is excellent in control precision and system stability.

Description

Technical Field

[0001] The present invention belongs to the field of motor control, and more specifically, relates to a method and system for optimizing the control of a permanent magnet synchronous motor. Background Art

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in high-efficiency drive systems due to their high power density, excellent dynamic response, and high efficiency. In the control of PMSMs, current control is a key part, which directly determines the operating performance and stability of the system. Traditional current control methods, such as PI controllers, although relatively simple, are prone to problems such as low control accuracy and poor system stability when the load disturbance is large or the motor operating conditions are complex. To improve the accuracy of current control and the robustness of the system, methods based on model predictive control (MPC) have been widely studied in recent years.

[0003] Model predictive control is to establish a mathematical model based on the dynamic model of the system to describe the behavior of the system, and use the current state information to predict the state of the system in the future for a period of time. At present, many researchers have proposed different PMSM control methods, and current control based on MPC has been widely used. Most of the existing implementation schemes simplify the model, ignore some non-linear characteristics or adopt traditional delay compensation techniques, and fail to completely solve common problems such as current fluctuations and harmonics in motor control.

[0004] The prior art patent with publication number CN112825468A proposes a system, method, and device for current control of a permanent magnet synchronous motor. The method includes: providing a control voltage to the motor; performing a motor current tracking process for tracking the observed motor current to the desired motor current; generating an error function in response to performing the motor current tracking process; tuning an adaptive control loop configured to minimize the error function using one or more estimated system parameters; and performing a control action in response to the tuned adaptive control loop. This solution lacks a real-time feedback and optimization mechanism and may be difficult to cope with dynamic changes in a complex operating environment. Summary of the Invention

[0005] In order to overcome the problems in the prior art that the motor control method has poor adaptability to complex working conditions, low control accuracy, and poor system stability, the present invention provides a method and system for optimizing the control of a permanent magnet synchronous motor.

[0006] The primary object of the present invention is to solve the above technical problems, and the technical solution of the present invention is as follows:

[0007] The first aspect of the present invention provides a method for optimizing the control of a permanent magnet synchronous motor, including the following steps:

[0008] Model the permanent magnet synchronous motor using a linear quadratic regulator to obtain a permanent magnet synchronous motor model;

[0009] Derive the discretization of the voltage equation of the permanent magnet synchronous motor model in the two-phase rotating coordinate system to obtain a motor prediction model;

[0010] Based on the performance requirements of the motor, design a q-axis cost function using the motor prediction model;

[0011] Optimize the q-axis current weight coefficient in the q-axis cost function by using the motor prediction model in combination with a model predictive control algorithm and a particle swarm optimization algorithm, and output the optimized q-axis current weight coefficient;

[0012] According to the optimized q-axis current weight coefficient, adjust the gain parameters in the motor controller in real time to optimize the motor current control accuracy and control performance.

[0013] Furthermore, model the permanent magnet synchronous motor using a linear quadratic regulator. The voltage equation expression of the permanent magnet synchronous motor is as follows:

[0014]

[0015] where d and q represent the d-axis and q-axis respectively, and u d 、u q represent the voltages of the d-axis and q-axis respectively, and β d 、β q are the non-physical constants corresponding to the voltages u d 、u q respectively, and their expressions are as follows:

[0016]

[0017] where ≡ means identically equal to, and L d 、L q are the equivalent inductance components on the d-axis and q-axis respectively;

[0018] i d 、i q are the current components corresponding to the d-axis and q-axis respectively, and α d 、α q are the non-physical constants corresponding to the currents i d 、i q respectively, and their expressions are as follows:

[0019]

[0020] where R s is the phase resistance of the three-phase winding;

[0021] F d, F q are the overall disturbances on the d-axis and q-axis of the system respectively, and the expressions are as follows:

[0022] F d ≡ ω e i q + f d

[0023]

[0024] where f d , f q are the unknown disturbances on the d-axis and q-axis respectively, ω e is the electrical angular velocity of the rotor, and ψ f is the magnetic flux generated by the rotor permanent magnet.

[0025] Furthermore, the discretization derivation is forward Euler discretization. The voltage equation of the permanent magnet synchronous motor model in the two-phase rotating coordinate system is discretized by forward Euler, and the expression is as follows:

[0026]

[0027] where the superscript p represents prediction and s represents sampling, represent the predicted current values of the permanent magnet synchronous motor on the d-axis and q-axis at the k + 1 moment respectively, i d (k), i q (k) represent the current values on the d-axis and q-axis at the k moment respectively, and T s represents the sampling period, is the voltage on the d-axis and q-axis corresponding to the jth switching sequence, and j is a positive integer representing the ordinal number of the voltage vector.

[0028] Furthermore, the expression of the q-axis cost function g is as follows:

[0029]

[0030] where are the given current values of the permanent magnet synchronous motor on the d-axis and q-axis respectively, and λ q represents the q-axis current weight coefficient, and λ q ≥ 1; are the predicted current values on the d-axis and q-axis respectively, is the unknown disturbance estimated on the q-axis.

[0031] Furthermore, a method for optimizing the q-axis current weight coefficient in the q-axis cost function by using the motor prediction model in combination with the model predictive control algorithm and the particle swarm optimization algorithm includes the following steps:

[0032] Optimize the control parameters of the prediction model using a model predictive control algorithm, and apply the optimized control parameters to a real motor system;

[0033] Compare the predicted value after optimizing the control parameters with the actual feedback value of the motor to obtain the error between the predicted q-axis current and the actual current;

[0034] Optimize the q-axis current weight coefficient in the q-axis cost function using the particle swarm optimization algorithm according to the error, and output the optimized q-axis current weight coefficient.

[0035] Furthermore, the method for optimizing the q-axis current weight coefficient in the q-axis cost function using the particle swarm optimization algorithm includes the following steps:

[0036] Randomly initialize each particle in the particle swarm. Each particle represents a candidate solution, which is the initial value of the q-axis current weight coefficient. The initial position of the particle is randomly selected within a predefined range, and the velocity of the particle is initialized;

[0037] For each particle, calculate the objective function value using its current position. The expression of the objective function J is as follows:

[0038] J = rms(Δi q ) + δ * rms(Δn r )

[0039] where, Δi q represents the steady-state error or instantaneous error of the q-axis current, Δn r represents the steady-state error or instantaneous error of the motor speed, rms(·) represents the root mean square value, μ represents the weight coefficient of the speed error in the objective function, and δ ∈ [0, 1];

[0040] For each particle, compare its current objective function value J with the previous optimal objective function value. If the current objective function value is smaller, update the historical optimal position of the particle to the current position;

[0041] Find the particle with the smallest objective function value in the particle swarm, and update the position of this particle to the global optimal position;

[0042] According to the update formula of the particle swarm optimization algorithm, calculate the new velocity and new position of the particle. The velocity pv k+1 (i) and position px k+1 (i) of the i-th particle in the (k + 1)-th generation are expressed as follows:

[0043] pν k+1 (i) = w * pv k (i) + c 1 r 1 [pBestk -px k (i)] + c 2 r 2 [gBest k -px k (i)]

[0044] px k+1 (i) = px k (i) + pv k+1 (i)

[0045] Among them, w is the inertia weight coefficient, w ∈ [0.5, 1], c 1 、c 2 are the local individual learning factor and the overall global learning factor respectively, r 1 、r 2 are random numbers between [0, 1];

[0046] Calculate the objective function value using the new positions of the updated particles, and update the optimal position according to the calculation results;

[0047] Repeat the above process until the maximum number of iterations is reached or the objective function converges to a predetermined threshold. If the stop condition is satisfied, stop the iteration and output the q-axis current weight coefficient corresponding to the global optimal position; if the stop condition is not satisfied, continue to update the particle positions and velocities.

[0048] Furthermore, set the inertia weight coefficient w to linearly decrease from the maximum inertia weight coefficient to the minimum inertia weight coefficient, and the expression is as follows:

[0049]

[0050] Among them, w max is the preset maximum inertia weight coefficient, w min is the minimum inertia weight coefficient, k is the number of iterations, G max is the maximum number of iterations.

[0051] Furthermore, set the local individual learning factor c 1 、the overall global learning factor c 2 in an asynchronous time-varying manner, and the expression is as follows:

[0052]

[0053] Among them, c 1max 、c 1min are the preset maximum and minimum values of the local individual learning factor respectively, c 2max 、c 2min are the preset maximum and minimum values of the overall global learning factor respectively, k is the number of iterations, Gmax is the maximum number of iterations.

[0054] The second aspect of the present invention provides a permanent magnet synchronous motor control optimization system, including a memory and a processor. The memory includes a program for the permanent magnet synchronous motor control optimization method. When the program for the permanent magnet synchronous motor control optimization method is executed by the processor, the steps of a permanent magnet synchronous motor control optimization method are implemented.

[0055] The third aspect of the present invention provides a computer-readable storage medium. The computer-readable storage medium includes a program for the permanent magnet synchronous motor control optimization method. When the program for the permanent magnet synchronous motor control optimization method is executed by a processor, the steps of the permanent magnet synchronous motor control optimization method are implemented.

[0056] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0057] This solution adopts a real-time optimization mechanism, establishes a permanent magnet synchronous motor (PMSM) model based on the linear quadratic regulator (LQR model), and performs modeling in combination with the two-phase rotating coordinate system, thereby achieving excellent responsiveness and accuracy of current control. By using the model predictive control method (MPCC) to optimize the control of the motor current, the current control accuracy is significantly improved, and the response ability and stability of the system under load disturbances are enhanced. In addition, the particle swarm optimization algorithm (PSO) is used to optimize the weight coefficients in the current control process, automatically adjust the optimal values according to different working conditions, achieve the balance between current tracking accuracy and speed fluctuation, improve the global search ability, and significantly improve the steady-state performance, effectively reducing the dependence of the control method on the accurate parameters of the motor and enhancing the adaptability of the model to load changes and external disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to make the objectives and technical solutions of the present invention clearer, the present invention provides the following drawings and descriptions:

[0059] Figure 1 is the flowchart of the method provided by the embodiment of the present invention;

[0060] Figure 2 is the comparison diagram of the selection results of the q-axis current weight coefficient of the particle swarm optimization algorithm provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] In order to be able to more clearly understand the above objectives, features and advantages of the present invention, the present invention will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited by the specific embodiments disclosed below.

[0063] Embodiment 1:

[0064] The present invention provides an optimization method for permanent magnet synchronous motor control, as Figure 1 shown in a flowchart of an optimization method for permanent magnet synchronous motor control. The specific steps are as follows:

[0065] S1: Use a linear quadratic regulator (LQR model) to model the permanent magnet synchronous motor (PMSM) to obtain a permanent magnet synchronous motor model.

[0066] More specifically, the voltage equation expression of the permanent magnet synchronous motor is as follows:

[0067]

[0068] where d and q represent the d-axis and q-axis respectively, and u d , u q represent the voltages of the d-axis and q-axis respectively, and β d , β q are the non-physical constants corresponding to the voltages u d , u q respectively, and the expressions are as follows:

[0069]

[0070] where L d , L q are the equivalent inductance components on the d-axis and q-axis respectively. In this embodiment, they are measured to be 9.6 mH;

[0071] i d , i q are the current components corresponding to the d-axis and q-axis respectively, and α d , α q are the non-physical constants corresponding to the currents i d , i q respectively, and the expressions are as follows:

[0072]

[0073]

[0074] where R s is the phase resistance of the three-phase winding. In this embodiment, it is measured to be 1.02 Ω;

[0075] F d 、Fq They are the overall disturbance of the system d-axis and the overall disturbance of the system q-axis respectively, and the expressions are as follows:

[0076] F d ≡ω e i q +f d

[0077]

[0078] Where f d and f q are the unknown disturbances on the d-axis and q-axis respectively, ω e is the electrical angular velocity of the rotor, which is equal to the mechanical angular velocity multiplied by the number of pole pairs, ψ f is the magnetic flux generated by the rotor permanent magnet. In this embodiment, the number of pole pairs is 4, and ψ f is 0.09 Wb.

[0079] S2: Discretize and derive the voltage equation of the permanent magnet synchronous motor model in the two-phase rotating coordinate system (d-q coordinate system) to obtain the motor prediction models of the permanent magnet synchronous motor on the d-axis and q-axis at the k+1 moment.

[0080] More specifically, the discretization derivation of the voltage equation of the permanent magnet synchronous motor model in the two-phase rotating coordinate system (d-q coordinate system) is the forward Euler discretization, and the expression is as follows:

[0081]

[0082] Where the superscript p represents prediction, s represents sampling, respectively represent the predicted current values of the permanent magnet synchronous motor on the d-axis and q-axis at the k+1 moment, i d (k) and i q (k) respectively represent the current values on the d-axis and q-axis at the k moment, T s represents the sampling period, is the voltage on the d-axis and q-axis corresponding to the jth switching sequence, and j is a positive integer representing the ordinal number of the voltage vector.

[0083] S3: Design the q-axis cost function using the motor prediction model based on the performance requirements of the motor.

[0084] More specifically, the expression of the q-axis cost function g is as follows:

[0085]

[0086] Where, are the given current values of the permanent magnet synchronous motor on the d-axis and q-axis respectively, are the predicted current values on the d-axis and q-axis respectively, Unknown disturbance estimated for the q-axis, λ q Represents the q-axis current weight coefficient, λ q ≥1, by designing λ q , the proportion of the q-axis current tracking performance is changed. The larger λ q is, the higher the priority of achieving the goal of accurate tracking of the q-axis current. When λ q is not appropriate, the control effect of the q-axis current decreases, resulting in torque mismatch, which may cause oscillations and breakdowns in the permanent magnet synchronous motor system, as Figure 2 shown in the comparison diagram of the selection of the q-axis current weight coefficient by the particle swarm optimization algorithm. It can be seen from the figure that when the weight coefficient λ q of the q-axis current i q as the torque current increases, the rapidity of the step response of the permanent magnet synchronous motor system will be improved to a certain extent, but it may also bring a large overshoot. Among them, when λ q = 3 and λ q = 7, the overshoot is the largest, reaching 6.41% and 6.82% respectively. Under different weight coefficients, the speed fluctuations of the permanent magnet synchronous motor system are not much different. When λ q = 7, the fluctuation is the largest, and the maximum value differs from the minimum value by about 2.79 r / min. When λ q = 10, the fluctuation is the smallest, about 1.91 r / min. Therefore, reasonably setting the q-axis current weight coefficient is an important factor in improving the stability of the permanent magnet synchronous motor system.

[0087] S4: Use the motor prediction model combined with the model predictive control algorithm (MPCC algorithm) and the particle swarm optimization algorithm (PSO) to optimize the q-axis current weight coefficient in the q-axis cost function, and output the optimized q-axis current weight coefficient.

[0088] The specific process is as follows:

[0089] Use the model predictive control algorithm (MPCC algorithm) to optimize the control parameters of the prediction model, and apply the optimized control parameters to the real motor system;

[0090] Compare the predicted value after optimizing the control parameters with the actual feedback value of the motor to obtain the error between the q-axis predicted current and the actual current;

[0091] According to the error, use the particle swarm optimization algorithm (PSO) to optimize the q-axis current weight coefficient in the q-axis cost function, and output the optimized q-axis current weight coefficient.

[0092] The specific process is as follows:

[0093] Randomly initialize each particle in the particle swarm. Each particle represents a candidate solution, which represents the initial value of the q-axis current weight coefficient. The initial position of the particle is randomly selected within a predefined range, and the velocity of the particle is initialized;

[0094] To reduce the steady-state error and speed ripple of the permanent magnet synchronous motor system, this patent uses the root mean square of the steady-state error Δi q of the q-axis current and the root mean square of the steady-state error Δn r of the motor speed to evaluate the weight coefficient. For each particle, the objective function value is calculated using its current position (i.e., the current q-axis current weight coefficient). The expression of the objective function J is as follows:

[0095] J = rms(Δi q ) + δ * rms(Δn r )

[0096] where, Δi q represents the steady-state error or instantaneous error of the q-axis current, Δn r represents the steady-state error or instantaneous error of the motor speed, rms(·) represents the root mean square value, δ represents the weight coefficient of the speed error in the objective function, which is used to balance the importance of the two. To make the q-axis current error Δi q as small as possible relative to the speed error Δn r , δ ∈ [0, 1];

[0097] For each particle, compare its current objective function value J with the previous optimal objective function value. If the current objective function value is smaller, then update the historical optimal position pBest k of this particle to the current position;

[0098] Find the particle with the smallest objective function value in the particle swarm, and update the position of this particle to the global optimal position gBest k ;

[0099] According to the update formula of the particle swarm optimization algorithm, calculate the new velocity and new position of the particle. The velocity pv k+1 (i) and position px k+1 (i) of the i-th particle in the (k + 1)-th generation are expressed as follows:

[0100] pv k+1 (i) = w * pν k (i) + c 1 r 1 [pBest k - px k (i)] + c 2 r 2 [gBest k - px k(i)]

[0101] px k+1 (i) = px k (i) + pv k+1 (i)

[0102] Among them, w is the inertia weight coefficient, reflecting the influence of an individual's historical performance on the present. w ∈ [0.5, 1], c 1 and c 2 are the local individual learning factor and the overall global learning factor respectively. Generally, their ranges are determined according to the value range of the independent variable. r 1 and r 2 are random numbers between [0, 1]. In this embodiment, r 1 = r 2 = 0.6;

[0103] When the flying speed of the particle is too fast, it may fly past the optimal solution. When the flying speed is too slow, it will lead to slow convergence. Therefore, the present invention places certain restrictions on the flying speed of the particle, giving the maximum and minimum boundary speeds. When the speed of the particle is greater than the maximum boundary speed or less than the minimum boundary speed, it is set to the boundary speed. To strengthen the global search in the initial stage and promote convergence in the later stage of the search, the inertia weight coefficient w is linearly decreased from the maximum inertia weight coefficient to the minimum inertia weight coefficient. The expression is as follows:

[0104]

[0105] Among them, w max is the preset maximum inertia weight coefficient, w min is the minimum inertia weight coefficient, k is the number of iterations, G max is the maximum number of iterations. In this embodiment, w max = 0.9, w min = 0.1, k = 50, G max = 100.

[0106] Similarly, the local individual learning factor c 1 and the overall global learning factor c 2 are set in an asynchronous time-varying manner. The expression is as follows:

[0107]

[0108] Among them, c 1max and c 1min are the maximum and minimum values of the preset local individual learning factor respectively, c 2max and c 2min are the maximum and minimum values of the preset overall global learning factor respectively, k is the number of iterations, G maxis the maximum number of iterations. In this embodiment, c 1max = c 2max = 0.5, c 1min = c 2min = 0.1.

[0109] Calculate the objective function value using the updated positions of the particles, and update the optimal position according to the calculation results;

[0110] Repeat the above process until the maximum number of iterations is reached or the objective function converges to a predetermined threshold. If the stopping condition is satisfied, stop the iteration and output the q-axis current weight coefficient corresponding to the global optimal position. In this embodiment, output the optimal q-axis current weight coefficient λ q = 7.5708; if the stopping condition is not satisfied, continue to update the particle positions and velocities.

[0111] S5: Adjust the gain parameters in the motor controller in real time according to the optimized q-axis current weight coefficient to optimize the motor current control accuracy and control performance.

[0112] The present invention adopts a real-time optimization mechanism, establishes a permanent magnet synchronous motor (PMSM) model based on the linear quadratic regulator (LQR model), and combines the two-phase rotating coordinate system for modeling, thereby achieving excellent responsiveness and accuracy of current control. By using the model predictive control method (MPCC) to optimize the control of the motor current, the current control accuracy is significantly improved, and the response ability and stability of the system under load disturbances are enhanced. The simulation results show that the dynamic tracking error is reduced by 30%, and the simulation calculation efficiency is improved by about 15% compared with the traditional MPCC method. In addition, the particle swarm optimization algorithm (PSO) is used to optimize the weight coefficient in the current control process, automatically adjust the optimal value according to different working conditions, achieve the balance between current tracking accuracy and speed fluctuation, improve the global search ability, and significantly improve the steady-state performance, effectively reduce the dependence of the control method on the accurate parameters of the motor, and enhance the adaptability of the model to load changes and external disturbances. Compared with the existing fixed weight coefficient design, the steady-state error is reduced by 15%, and the torque ripple is reduced by 10%; compared with the traditional PI control technology, the steady-state error is reduced by 20%, and the dynamic response speed is increased by 10%.

[0113] Embodiment 2:

[0114] This embodiment provides a permanent magnet synchronous motor control optimization system, including a memory and a processor. The memory includes a permanent magnet synchronous motor control optimization method program, and when the permanent magnet synchronous motor control optimization method program is executed by the processor, it realizes the steps of a permanent magnet synchronous motor control optimization method as described in Embodiment 1.

[0115] Embodiment 3:

[0116] This embodiment provides a computer-readable storage medium, which includes a program for optimizing the control of a permanent magnet synchronous motor. When the program for optimizing the control of the permanent magnet synchronous motor is executed by a processor, the steps of an optimization method for controlling a permanent magnet synchronous motor as described in Embodiment 1 are implemented.

[0117] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, rather than limitations on the implementation manners of the present invention. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the claims of the present invention.

Claims

1. A permanent magnet synchronous motor control optimization method, characterized in that: The steps include: The permanent magnet synchronous motor is modeled by using a linear quadratic regulator to obtain a permanent magnet synchronous motor model; The voltage equation of the permanent magnet synchronous motor model is discretized and derived in a two-phase rotating coordinate system to obtain the motor prediction model. Based on the performance requirements of the motor, the q-axis cost function is designed using the motor prediction model; The motor prediction model is combined with the control algorithm based on model prediction and the particle swarm optimization algorithm to optimize the q-axis current weight coefficient in the q-axis cost function, and the optimized q-axis current weight coefficient is output; The gain parameters in the motor controller are adjusted in real time according to the optimized q-axis current weight coefficient to optimize the motor current control accuracy and control performance.

2. A permanent magnet synchronous motor control optimization method according to claim 1, characterized in that: The permanent magnet synchronous motor is modeled using a linear quadratic regulator. The voltage equation of the permanent magnet synchronous motor is shown as follows: Where d and q represent the d-axis and q-axis respectively, u d 、u q Represent the voltage of d-axis and q-axis respectively, β d , β q The voltage u d 、u q The corresponding non-physical constants are expressed as follows: Among them, ≡ means it is equal to, L d , L q are the equivalent inductance components on the d-axis and q-axis respectively; i d 、i q are the current components corresponding to the d-axis and q-axis respectively, α d , α q is the current i d 、i q The corresponding non-physical constants are expressed as follows: Among them, R s is the phase resistance of the three-phase winding; F d 、F q They are the total disturbance of the system d-axis and the total disturbance of the system q-axis respectively, and the expressions are as follows: F d ≡ω e i q +f d Among them, f d 、f q are the unknown disturbances on the d and q axes, ω e is the electrical angular velocity of the rotor, ψ f is the flux generated by the rotor permanent magnet.

3. A permanent magnet synchronous motor control optimization method according to claim 1, characterized in that: The discretization is derived as forward Euler discretization. The voltage equation of the permanent magnet synchronous motor model in the two-phase rotating coordinate system is subjected to forward Euler discretization. The expression is as follows: Among them, the superscript p represents prediction, s represents sampling, They represent the predicted current values ​​of the permanent magnet synchronous motor on the d-axis and q-axis at time k+1, respectively. d (k), i q (k) represents the current value of d-axis and q-axis at time k, T s represents the sampling period, are the d and q axis voltages corresponding to the j-th switching sequence, where j is a positive integer representing the ordinal number of the voltage vector.

4. A permanent magnet synchronous motor control optimization method according to claim 1, characterized in that: The expression g of the q-axis cost function is as follows: in, are the given current values ​​of the permanent magnet synchronous motor on the d-axis and q-axis, λ q represents the q-axis current weight coefficient, λ q ≥1; are the predicted current values ​​of the d-axis and q-axis respectively, is the estimated unknown disturbance on the q-axis.

5. A permanent magnet synchronous motor control optimization method according to claim 1, characterized in that: The method for optimizing the q-axis current weight coefficient in the q-axis cost function by using a motor prediction model combined with a control algorithm based on model prediction and a particle swarm optimization algorithm comprises the following steps: Optimizing control parameters of the prediction model using a control algorithm based on model prediction, and applying the optimized control parameters to a real motor system; Compare the predicted value after optimizing the control parameters with the actual feedback value of the motor to obtain the error between the predicted current and the actual current of the q-axis; The q-axis current weight coefficient in the q-axis cost function is optimized using a particle swarm optimization algorithm according to the error, and the optimized q-axis current weight coefficient is output.

6. A permanent magnet synchronous motor control optimization method according to claim 5, characterized in that: The method for optimizing the q-axis current weight coefficient in the q-axis cost function by using a particle swarm optimization algorithm comprises the following steps: Randomly initialize each particle in the particle swarm, each particle represents a candidate solution, indicating the initial value of the q-axis current weight coefficient, the initial position of the particle is randomly selected within a predefined range, and the velocity of the particle is initialized; For each particle, the objective function value is calculated using its current position. The expression of the objective function J is as follows: J=rms(Δi q )+δ*rms(Δn r ) Among them, Δi q represents the steady-state error or instantaneous error of the q-axis current, Δnr represents the steady-state error or instantaneous error of the motor speed, rms(·) represents the root mean square value, δ represents the weight coefficient of the speed error in the objective function, δ∈[0,1]; For each particle, compare its current objective function value J with the previous optimal objective function value. If the current objective function value is smaller, update the historical optimal position of the particle to the current position. Find the particle with the smallest objective function value in the particle swarm and update the position of the particle to the global optimal position; According to the update formula of the particle swarm optimization algorithm, the new speed and new position of the particle are calculated. The speed pv of the i-th particle in the k+1 generation is k+1 (i) and position px k+1 (i) The expression is as follows: pv k+1 (i)=w*pv k (i)+c1r1[pBest k -px k (i)]+c2r2[gBest k -px k (i)] px k+1 (i)=px k (i)+pv k+1 (i) Where w is the inertia weight coefficient, w∈[0.5,1], c1 and c2 are local individual learning factors and overall global learning factors, respectively, and r1 and r2 are random numbers between [0,1]; The objective function value is calculated using the updated particle's new position, and the optimal position is updated based on the calculation result; The above process is repeated until the maximum number of iterations is reached or the objective function converges to a predetermined threshold. If the stopping condition is met, the iteration is stopped and the q-axis current weight coefficient corresponding to the global optimal position is output; if the stopping condition is not met, the particle position and velocity are continued to be updated.

7. A permanent magnet synchronous motor control optimization method according to claim 6, characterized in that: The inertia weight coefficient w is set to decrease linearly from the maximum inertia weight coefficient to the minimum inertia weight coefficient. The expression is as follows: Among them, w max is the preset maximum inertia weight coefficient, w min is the minimum inertia weight coefficient, k is the number of iterations, G max is the maximum number of iterations.

8. A permanent magnet synchronous motor control optimization method according to claim 6, characterized in that: The local individual learning factor c1 and the overall global learning factor c2 are set in an asynchronous time-varying manner, and the expressions are as follows: Among them, c 1max 、c 1min are the maximum and minimum values ​​of the preset local individual learning factors, c 2max 、c 2min are the maximum and minimum values ​​of the preset overall global learning factor, k is the number of iterations, G max is the maximum number of iterations.

9. A permanent magnet synchronous motor control optimization system, characterized in that: The system includes: a memory and a processor, wherein the memory includes a permanent magnet synchronous motor control optimization method program, and when the permanent magnet synchronous motor control optimization method program is executed by the processor, the steps of a permanent magnet synchronous motor control optimization method as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a permanent magnet synchronous motor control optimization method program. When the permanent magnet synchronous motor control optimization method program is executed by a processor, the steps of a permanent magnet synchronous motor control optimization method as described in any one of claims 1 to 8 are implemented.

Images