An optimization method for the unequal thickness magnetic pole structure of a surface mounted outer rotor permanent magnet motor

Abstract

The present invention proposes a method for optimizing the unequal thickness magnetic pole structure of a surface-mounted outer rotor permanent magnet motor. First, the magnetic pole arc coefficient, the pole width between the magnetic poles, the minimum air gap length and the eccentricity of the magnetic pole are determined as the permanent magnet structural parameters to be optimized; the required optimization parameters are sampled and analyzed in three stages according to the correlation degree to determine the value ranges of the four optimization parameters; the motor air gap flux density, cogging torque and the minimization of the harmonic distortion rate of the air gap flux density are selected as the optimization target of the motor; the response surface model of the motor is constructed using CCD; an improved multi-objective raccoon optimization algorithm is designed to perform multi-objective optimization on the four optimization parameters; and the final unequal thickness magnetic pole structure is determined. The present invention has a better global search capability for the motor optimization parameters and can accurately find the optimal value of the optimization target. At the same time, by selecting the optimization parameters and the optimization target and designing the objective function, the optimized outer rotor permanent magnet motor has better performance.

Description

Technical Field

[0001] The invention relates to a method for optimizing a magnetic pole structure, and in particular to a method for optimizing a magnetic pole structure of unequal thickness of a surface mounted outer rotor permanent magnet motor. Background Art

[0002] With the acceleration of industrialization, the motor industry has also developed rapidly, and various types of permanent magnet motors are becoming more and more abundant in the market. Surface-mounted permanent magnet motors use permanent magnets to provide excitation, and the motor structure is relatively simple, so the design optimization of permanent magnets can not only improve the performance of the motor, but also reduce losses and save energy.

[0003] At present, the permanent magnet materials used in permanent magnet motors are ferrite or neodymium iron boron. Among them, neodymium iron boron magnets are expensive and the manufacturing process requires the consumption of rare earth resources. Therefore, how to reasonably optimize the magnetic pole structure is also one of the important topics in motor structure optimization.

[0004] The existing methods for optimizing magnetic poles mainly include numerical analysis, statistical methods, machine learning and artificial intelligence algorithms. Among them, the numerical analysis method that relies on finite element analysis to determine the magnetic pole structure cannot effectively model and optimize the special magnetic pole structure. Statistical methods such as orthogonal experimental methods have a limited number of times to optimize the magnetic pole structure, and cannot accurately find the structural parameters that make the motor perform better. Ordinary multi-objective optimization algorithms have modeling difficulties and fall into local optimal solutions when optimizing the magnetic pole structure. In addition, the selection of optimization parameters and objectives and the design of objective functions will affect the performance of the final motor after optimization. Summary of the invention

[0005] Purpose of the invention: In view of the problems pointed out in the background technology, the present invention discloses a method for optimizing the unequal thickness pole structure of a surface-mounted outer rotor permanent magnet motor. By optimizing the pole structure parameters, a permanent magnet structure that improves the motor performance is obtained.

[0006] Technical solution: The present invention discloses a method for optimizing the unequal thickness magnetic pole structure of a surface mounted outer rotor permanent magnet motor, comprising:

[0007] Step (1) determining the magnetic pole arc coefficient l, the pole width w between the magnetic poles, the minimum air gap length g and the eccentricity d of the magnetic pole as the permanent magnet structural parameters to be optimized;

[0008] Step (2) selecting the minimum air gap flux density, cogging torque and air gap flux density harmonic distortion rate of the motor as the optimization target of the motor;

[0009] Step (3) Sampling multi-stage sampling method to determine the value ranges of four optimization parameters;

[0010] Step (4) Central Composite Design (CCD) is used to construct a response surface model of the motor's air gap flux density, cogging torque, and air gap flux density harmonic distortion rate;

[0011] Step (5) designs an improved multi-objective raccoon optimization algorithm, and uses the improved multi-objective raccoon optimization algorithm to perform multi-objective optimization on four optimization parameters; the improved multi-objective raccoon optimization algorithm uses Chebyshev chaotic mapping to initialize the population, and in the development stage, the new position is updated by combining the chaotic algorithm with the raccoon algorithm.

[0012] Furthermore, in step (3), the multi-stage sampling method is improved, including:

[0013] (3.1) According to the selected optimization parameters, it is divided into three layers for sampling. The first stage is a four-parameter layer that is independent of each other; according to the correlation between the four parameters, the second stage is divided into a pairwise combination sampling layer; the third stage is a lwgd combination sampling layer;

[0014] (3.2) Based on the sampling data of the first stage, the air gap flux density, cogging torque, air gap flux density harmonic distortion rate and air gap flux density harmonic distortion rate corresponding to the samples are calculated by random sampling method to judge the motor performance;

[0015] (3.3) Through the analysis of advantages and disadvantages, the better samples in the first stage are selected as the sampling data for the second stage. The data are randomly sampled, and the air gap flux density, cogging torque, and air gap flux density harmonic distortion rate and air gap flux density harmonic distortion rate corresponding to the samples are calculated to judge the motor performance;

[0016] (3.4) Through the pros and cons analysis, the better samples in the second stage are selected as the sampling data for the third stage; according to the range of the four optimization parameters determined in steps (3.1)(3.2)(3.3), the lwgd is arranged and combined by the method of permutation and combination, and the air gap flux density, cogging torque and air gap flux density harmonic distortion rate corresponding to the samples are calculated, the performance of the motor is analyzed, and the final parameter range is determined;

[0017] (3.5) When judging the motor performance, the CRITIC weighting method is used to make a more accurate judgment on the motor performance.

[0018] Furthermore, in step (3.5), the data is positively processed and then its standard deviation is analyzed:

[0019]

[0020] Among them, n=1,2,3,4, i=1,2,3,4, j=1,2,3, is the average value of air gap flux density, cogging torque and air gap flux density harmonic distortion rate corresponding to each parameter, S i Represents the standard deviation of the i-th data, i is the optimization parameter variable, and n is the number of optimization parameter variables;

[0021] The final objective weight is:

[0022]

[0023] Among them, R i is the correlation coefficient, r ij To evaluate the correlation coefficient between i and j.

[0024] Furthermore, the improved multi-objective raccoon optimization algorithm performs multi-objective optimization on four optimization parameters, and its objective function is:

[0025] According to the performance requirements of the outer rotor permanent magnet motor and the nonlinear relationship between the air gap flux density, cogging torque and air gap flux density harmonic distortion rate, the improved ideal point method is used to design the objective function:

[0026]

[0027] Among them, f hx (x) is the objective function of optimizing the raccoon algorithm, w i is the weight corresponding to the optimization target, Q ij is the measured value, Q j a is the optimal design value of the optimization objective, Q j b is the worst value of the optimization target. When the objective function is close to the optimal value, f hx (x) is smaller.

[0028] Furthermore, the specific process of the improved multi-objective raccoon optimization algorithm is as follows:

[0029] (4.1) The population is initialized using the improved Chebyshev chaotic map; the improved Chebyshev chaotic map is mathematically described as:

[0030] X i,j+1 =1-2k1(cos(2k2arccosx i,j )) 2

[0031] X i = l b +(u b -l b )x i

[0032] Among them, X i,j+1 is the j+1-dimensional position of the ith raccoon; K1 and K2 are constants whose values ​​can be adjusted to change the chaotic mapping effect, x i is the position of the i-th raccoon; u b , l b are the upper and lower limits of the search space;

[0033] (4.2) In the development stage, the raccoon will update the new position according to the current optimal value. By combining the chaos algorithm with the raccoon algorithm, the raccoon position update in the development stage is described as:

[0034] v i,j =V min +r(x)·(V max -V min ),V min ≤v i,j ≤V max

[0035] x i,j =X min +r(x)·(X max -X min ),X min ≤x i,j ≤X max

[0036] Among them, x ij is the position of the i-th raccoon in the k-th iteration; X max is the maximum border raccoon position; X min is the minimum boundary raccoon position; V ij is the speed of the ith raccoon in the kth iteration; V max is the maximum boundary raccoon speed; V max is the minimum boundary raccoon speed; r(x) is a function of random values ​​in [0,1];

[0037] (4.3) When the raccoon algorithm obtains the local optimal position solution, it is first mapped into a chaotic variable:

[0038]

[0039] The mapping f(x) is introduced to ensure the diversity of the positions searched by the raccoon and prevent the algorithm from falling into the local optimum when searching for the best result:

[0040]

[0041] Among them, λ is the control parameter; is a random number in [0, 1]; use f to map the chaotic variable, reverse the obtained chaotic sequence, and finally get its original solution space: Ck '=(C1', C2', ...C k ').

[0042] Beneficial effects:

[0043] The present invention proposes a method for optimizing the unequal thickness magnetic pole structure of a surface-mounted outer rotor permanent magnet motor. First, the optimization parameters and optimization targets are determined through an in-depth study of the motor poles. Then, the improved multi-stage sampling method is used to determine the sampling range of the optimization parameters. Then, the optimization target of the motor is modeled through the CCD method to optimize the problem of inaccurate model. Then, the improved ideal point method is introduced to design the objective function, and then the improved Chebyshev chaos mapping is used to initialize the raccoon population and the chaos-raccoon algorithm to solve the problem of local optimality of the algorithm. Compared with other optimization algorithms for motors, the global search capability for target parameters is improved. At the same time, through the selection of optimization parameters and optimization targets and the design of the objective function, the outer rotor permanent magnet motor has better performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is the magnetic pole optimization parameter selected in the embodiment of the present application.

[0045] Figure 2 It is a flow chart of the improved raccoon algorithm in the embodiment of the present application.

[0046] Figure 3 This is a flow chart for optimizing unequal thickness magnetic poles of an outer rotor permanent magnet motor in an embodiment of the present application.

[0047] Figure 4 It is a comparison diagram of the superposition field of magnetic lines of force and magnetic density cloud map before and after optimization in the embodiment of the present application.

[0048] Figure 5 It is a comparison diagram of the air gap magnetic density before and after optimization in the embodiment of the present application.

[0049] Figure 6 It is a comparison diagram of the Fourier analysis of the air gap magnetic flux before and after optimization in the embodiment of the present application.

[0050] Figure 7 It is a comparison diagram of the cogging torque before and after optimization in the embodiment of the present application.

[0051] Figure 8 It is a comparison diagram of the Fourier analysis of the cogging torque before and after optimization in the embodiment of the present application.

[0052] Fig. 9 It is an efficiency diagram after optimization in the embodiment of the present application. DETAILED DESCRIPTION

[0053] The present invention will be further described below in conjunction with the accompanying drawings.

[0054] The present invention discloses a method for optimizing the unequal thickness magnetic pole structure of a surface mounted outer rotor permanent magnet motor, comprising:

[0055] Step 1: Determine the pole arc coefficient l, pole width w between the poles, minimum air gap length g and pole eccentricity d as the permanent magnet structural parameters to be optimized. Figure 1 As shown, the four parameters that have the greatest impact on the permanent magnet structure and performance are selected as the optimization parameters of the unequal thickness pole structure of the surface mounted outer rotor permanent magnet motor: pole arc coefficient l, pole width w, minimum air gap length g and eccentricity d.

[0056] Step 2: Select the motor air gap flux density and cogging torque as the motor optimization targets.

[0057] Step 3: Sampling The improved multi-stage sampling method is used to determine the value ranges of the four optimization parameters.

[0058] According to the selected optimization parameters, it is divided into three layers for sampling. The first stage is a four-parameter layer that is independent of each other; according to the correlation between the four parameters, the second stage is divided into a two-by-two combination sampling layer; the third stage is a lwgd combination sampling layer.

[0059] Based on the sampling data of the first stage, the air gap flux density and cogging torque corresponding to the samples are calculated using a random sampling method to determine the motor performance.

[0060] Through the analysis of advantages and disadvantages, the better samples in the first stage are selected as the sampling data for pairwise combination in the second stage. The data are randomly sampled, and the air gap flux density and cogging torque corresponding to the samples are calculated to judge the motor performance.

[0061] Through the analysis of advantages and disadvantages, the better samples in the second stage are selected as the sampling data for the third stage. According to the range of the four optimized parameters determined, the lwgd is arranged and combined by permutation and combination, the air gap flux density and cogging torque corresponding to the samples are calculated, the motor performance is analyzed, and the final parameter range is determined.

[0062] Since the air gap flux density and cogging torque have different effects on motor performance, the CRITIC weighting method is used to make a more accurate judgment on motor performance.

[0063] Furthermore, the data is processed positively and then its standard deviation is analyzed:

[0064]

[0065] Among them, n=1,2,3,4, i=1,2,3,4, j=1,2,3, is the average value of air gap flux density and torque corresponding to each parameter, Si Represents the standard deviation of the i-th data, i is the optimization parameter variable, and n is the number of optimization parameter variables.

[0066] The final objective weight is:

[0067]

[0068] R i is the correlation coefficient, r ij To evaluate the correlation coefficient between i and j.

[0069] Step 4: Use Central Composite Design (CCD) to build the response surface model of the motor.

[0070] Step 5: Design an improved multi-objective raccoon optimization algorithm and use it to perform multi-objective optimization on the magnetic pole parameters.

[0071] Figure 2 The figure shows the flowchart of the improved raccoon algorithm. First, the improved Chebyshev chaotic mapping is used to make up for the shortcomings of the traditional raccoon algorithm in terms of insufficient initialization ability. Secondly, since the optimization parameters of the motor poles are relatively complex, in order to ensure the diversity of the population and solve the problem that the raccoon algorithm is prone to fall into local optimality when searching for the best, the chaotic algorithm is embedded in the raccoon algorithm to improve the search effect of the traditional raccoon algorithm. Finally, the optimal parameter combination is determined by the pros and cons.

[0072] According to the performance requirements of the outer rotor permanent magnet motor and the nonlinear relationship between the air gap flux density, cogging torque and air gap flux density harmonic distortion rate, the improved ideal point method is used to design the objective function:

[0073]

[0074] Among them, f hx (x) is the objective function of optimizing the raccoon algorithm, w i is the weight corresponding to the optimization target, Q ij is the measured value, Q j a is the optimal design value of the optimization objective, Q j b is the worst value of the optimization target. When the objective function is close to the optimal value, f hx (x) is smaller.

[0075] The improved Chebyshev chaotic map is used to initialize the population. The raccoon algorithm randomly generates the positions of the raccoon individuals in various dimensions during the initialization phase, which easily causes the raccoon individuals to cluster locally. Therefore, the improved Chebyshev chaotic map is introduced to initialize the raccoon population. The improved Chebyshev map shows the complex dynamic behavior in nonlinear systems and can be used to generate sequences with high randomness. The mathematical description of the Chebyshev chaotic map is:

[0076] X i,j+1 =1-2k1(cos(2k2arccosx i,j )) 2

[0077] X i = l b +(u b -l b )x i

[0078] Among them, X i,j+1 is the j+1-dimensional position of the ith raccoon; K1 and K2 are constants whose values ​​can be adjusted to change the chaotic mapping effect, x i is the position of the i-th raccoon; u b , l b are the upper and lower limits of the search space.

[0079] In the development stage, the raccoon will update the new position according to the current optimal value. This strategy can easily cause the algorithm to fall into the local optimum. This invention combines the chaos algorithm with the raccoon algorithm to solve the problem of the traditional raccoon algorithm falling into the local optimum. The raccoon position update in the development stage is described as:

[0080] v i,j =V min +r(x)·(V max -V min ),V min ≤v i,j ≤V max

[0081] x i,j =X min +r(x)·(X max -X min ),X min ≤x i,j ≤X max

[0082] Among them, x ij is the position of the i-th raccoon in the k-th iteration; X max is the maximum border raccoon position; X min is the minimum boundary raccoon position; Vij is the speed of the ith raccoon in the kth iteration; V max is the maximum boundary raccoon speed; V max is the minimum boundary raccoon speed; r(x) is a function of random values ​​in [0,1].

[0083] When the raccoon algorithm obtains the local optimal position solution, it first maps it into a chaotic variable:

[0084]

[0085] The mapping f(x) is introduced to ensure the diversity of the positions searched by the raccoon and prevent the algorithm from falling into the local optimum when searching for the best result:

[0086]

[0087] Among them, λ is the control parameter; is a random number in [0, 1].

[0088] Use f to map the chaotic variables, reverse the obtained chaotic sequence, and finally get its original solution space: C k '=(C1', C2', ...C k ').

[0089] Figure 3 This is a flowchart for optimizing the unequal thickness magnetic poles of an outer rotor permanent magnet motor. First, through the study of the magnetic poles of the outer rotor permanent magnet motor, the magnetic pole optimization parameters are determined to be the magnetic pole arc coefficient l, the pole width w between the magnetic poles, the minimum air gap length g and the eccentricity d of the magnetic pole; secondly, the important air gap flux density and cogging torque of the permanent magnet motor are selected as the optimization targets; then the parameter optimization range is reduced by the improved multi-stage sampling method; then the response surface modeling of the motor optimization is carried out by the CCD method; finally, the optimization parameters of the magnetic poles are optimized by the improved raccoon algorithm to obtain the optimal magnetic pole structure parameters.

[0090] Figure 4 , 5 , 6, 7, 8, and 9 are optimization result diagrams of the outer rotor permanent magnet motor using the present invention. P1 is the motor after optimization, and P2 is the motor before optimization.

[0091] Figure 4 This is a comparison diagram of the superposition field of magnetic lines of force and magnetic density cloud map before and after optimization in the embodiment of the present application. Compared with P2, the magnetic field saturation of P1 is significantly alleviated. The maximum magnetic density in the area near the stator slot of P2 reaches 2.8T, while the maximum magnetic density in the same position of P1 is only 2.0T, which is reduced by 28.57%. This optimization method effectively suppresses the magnetic field saturation phenomenon of the outer rotor permanent magnet motor.

[0092] Figure 5This is a comparison diagram of the air gap flux density before and after optimization in the embodiment of the present application. Compared with P2, the peak value of the air gap flux density of P1 dropped from 1.012T to 0.6501T, a decrease of 35.76%, indicating that the core magnetic field saturation condition of P2 has been alleviated.

[0093] Figure 6 This is a comparison chart of the air gap flux Fourier analysis before and after optimization in the embodiment of the present application. The air gap flux is normalized, and the main 3rd, 5th, 7th, 9th, 11th, 13th, and 15th odd harmonics are reduced by 49.39%, 0.97%, 76.56%, 74.23%, 92.40%, 74.82%, and 92.49% respectively. Among them, the 11th and 15th harmonics are basically eliminated. Therefore, combined with Figure 5 It can be seen that the present invention has a certain inhibitory effect on the higher harmonics of the optimized main harmonics of the motor, which is of great significance to the improvement of the motor efficiency.

[0094] Figure 7 This is a comparison chart of the cogging torque before and after optimization in the embodiment of the present application. The peak value of the cogging torque of P1 reaches 0.46N·m, and the maximum cogging torque of P2 is only 0.028N·m, and the cogging torque of P2 is reduced by 93.91% compared with P1.

[0095] Figure 8 This is a comparison chart of the Fourier analysis of the cogging torque before and after optimization in the embodiment of this application. Compared with P2, the second harmonic component of the cogging torque of P1 is reduced from 33% to 4%, a year-on-year decrease of 87.88%. Figure 7 It can be seen that the present invention can effectively weaken the cogging torque and its harmonic components of various orders.

[0096] Figure 7 This is the optimized efficiency diagram in the embodiment of the present application. P1 The motor speed is 1800r / min, and the motor torque is in the range of [300,700], and the motor efficiency can reach more than 91%, which meets the design requirements of the outer rotor permanent magnet motor, indicating that the present invention can improve the efficiency of the motor.

[0097] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the unequal thickness magnetic pole structure of a surface mounted outer rotor permanent magnet motor, characterized in that: include: Step (1) determining the magnetic pole arc coefficient l, the pole width w between the magnetic poles, the minimum air gap length g and the eccentricity d of the magnetic pole as the permanent magnet structural parameters to be optimized; Step (2) selecting the minimum air gap flux density, cogging torque and air gap flux density harmonic distortion rate of the motor as the optimization target of the motor; Step (3) Sampling multi-stage sampling method to determine the value ranges of four optimization parameters; Improvements to the multi-stage sampling method include: (3.1) According to the selected optimization parameters, it is divided into three layers for sampling. The first stage is a four-parameter layer that is independent of each other; according to the correlation between the four parameters, the second stage is divided into a pairwise combination sampling layer; the third stage is a lwgd combination sampling layer; (3.2) Based on the sampling data of the first stage, the air gap flux density, cogging torque and air gap flux density harmonic distortion rate corresponding to the samples are calculated by random sampling method to judge the motor performance; (3.3) Through the analysis of advantages and disadvantages, the better samples in the first stage are selected as the sampling data for the second stage. The data are randomly sampled, and the air gap flux density, cogging torque and air gap flux density harmonic distortion rate corresponding to the samples are calculated to judge the motor performance; (3.4) Through the pros and cons analysis, the better samples in the second stage are selected as the sampling data for the third stage; according to the range of the four optimization parameters determined in steps (3.1)(3.2)(3.3), the lwgd is arranged and combined by the method of permutation and combination, and the air gap flux density, cogging torque and air gap flux density harmonic distortion rate corresponding to the samples are calculated, the performance of the motor is analyzed, and the final parameter range is determined; (3.5) When judging the motor performance, the CRITIC weighting method is used to make a more accurate judgment on the motor performance; Step (4) using Central Composite Design (CCD) to construct a response surface model of the motor's air gap flux density, cogging torque, and air gap flux density harmonic distortion rate; Step (5) designs an improved multi-objective raccoon optimization algorithm, and uses the improved multi-objective raccoon optimization algorithm to perform multi-objective optimization on four optimization parameters; the improved multi-objective raccoon optimization algorithm uses Chebyshev chaotic mapping to initialize the population, and in the development stage, the new position is updated by combining the chaotic algorithm with the raccoon algorithm.

2. The method for optimizing the unequal thickness magnetic pole structure of a surface mounted outer rotor permanent magnet motor according to claim 1, characterized in that: The improved multi-objective raccoon optimization algorithm performs multi-objective optimization on four optimization parameters, and its objective function is: According to the performance requirements of the outer rotor permanent magnet motor and the nonlinear relationship between the air gap flux density, cogging torque and air gap flux density harmonic distortion rate, the improved ideal point method is used to design the objective function: Among them, f hx (x) is the objective function of optimizing the raccoon algorithm, w i is the weight corresponding to the optimization target, Q ij is the measured value, Q j a is the optimal design value of the optimization objective, Q j b is the worst value of the optimization target. When the objective function is close to the optimal value, f hx (x) is smaller.

3. The method for optimizing the unequal thickness magnetic pole structure of a surface mounted outer rotor permanent magnet motor according to claim 2, characterized in that: The specific process of the improved multi-objective raccoon optimization algorithm is as follows: (4.1) The population is initialized using the improved Chebyshev chaotic map; the improved Chebyshev chaotic map is mathematically described as: X i,j+1 =1-2k1(cos(2k2arccosx i,j )) 2 X i =l b +(u b -l b )x i Among them, X i,j+1 is the j+1-dimensional position of the ith raccoon; K1 and K2 are constants whose values ​​can be adjusted to change the chaotic mapping effect, x i is the position of the i-th raccoon; u b , l b are the upper and lower limits of the search space; (4.2) In the development stage, the raccoon will update the new position according to the current optimal value. By combining the chaos algorithm with the raccoon algorithm, the raccoon position update in the development stage is described as: v i,j =V min +r(x)·(V max -V min ),V min ≤v i,j ≤V max x i,j =X min +r(x)·(X max -X min ),X min ≤x i,j ≤X max Among them, x ij is the position of the i-th raccoon in the k-th iteration; X max is the maximum border raccoon position; X min is the minimum boundary raccoon position; V ij is the speed of the ith raccoon in the kth iteration; V max is the maximum boundary raccoon speed; V min is the minimum boundary raccoon speed; r(x) is a function of random values ​​in [0,1]; (4.3) When the raccoon algorithm obtains the local optimal position solution, it is first mapped into a chaotic variable: The mapping f(x) is introduced to ensure the diversity of the positions searched by the raccoon and prevent the algorithm from falling into the local optimum when searching for the best result: f=λ·ζ1(1-ζ1) Among them, λ is the control parameter; ζ1 is a random number in [0, 1]. Use f to map the chaotic variables, reverse the obtained chaotic sequence, and finally get its original solution space: C k '=(C1', C2', ... C k ').

Images