阅读说明：本技术 一种永磁同步电机的分数阶模型辨识方法及装置 (Fractional order model identification method and device for permanent magnet synchronous motor ) 是由 余伟 梁恒辉 于 2020-06-24 设计创作，主要内容包括：本发明涉及电机建模技术领域,具体涉及一种永磁同步电机的分数阶模型辨识方法及装置,所述方法为：首先建立永磁同步电机的分数阶传递函数,并对永磁同步电机的分数阶传递函数进行简化,得到永磁同步电机的分数阶模型；接着将分数阶模型的全部参数作为一个个体,确定个体中每个参数的取值范围；初始化p个个体,形成父代种群,采用改进差分进化算法对父代种群进行训练,从而得到一个优选个体,最后将所述优选个体中的参数作为永磁同步电机的分数阶模型参数,本发明建立的分数阶模型能够方便对永磁同步电机进行更加精确的控制。(The invention relates to the technical field of motor modeling, in particular to a fractional order model identification method and a device of a permanent magnet synchronous motor, wherein the method comprises the following steps: firstly, establishing a fractional order transfer function of the permanent magnet synchronous motor, and simplifying the fractional order transfer function of the permanent magnet synchronous motor to obtain a fractional order model of the permanent magnet synchronous motor; then, all parameters of the fractional order model are used as an individual, and the value range of each parameter in the individual is determined; p individuals are initialized to form a parent population, the parent population is trained by adopting an improved differential evolution algorithm to obtain an optimal individual, and finally parameters in the optimal individual are used as fractional order model parameters of the permanent magnet synchronous motor.)

1. A method for identifying a fractional order model of a permanent magnet synchronous motor is characterized by comprising the following steps:

step S100, establishing a fractional order transfer function of the permanent magnet synchronous motor:

wherein G(s) is a fractional order transfer function of the permanent magnet synchronous motor, n is the rotating speed of the permanent magnet synchronous motor, Uq is the q-axis voltage of the permanent magnet synchronous motor, Ce is the electromotive force coefficient of the permanent magnet synchronous motor, and T₁Is the electromagnetic time constant, T, of the armature circuit_mFor the electromechanical time constant of the electric drive system, ξ is the fractional order of the electromagnetic link, theta is the fractional order of the mechanical link, and s is the operator of the Laplace transform;

step S200, simplifying a fractional order transfer function of the permanent magnet synchronous motor to obtain a fractional order model of the permanent magnet synchronous motor:

wherein, the parameter a is 1/C_e·T_m·T_lThe parameter α is + theta, and the parameter b is 1/T_lParameter β, parameter c 1/T_m·T_l；

Step S300, forming a parameter group from the parameters in the fractional order model, taking the parameter group as an individual, and determining the value range of each parameter in the individual;

s400, initializing p individuals to form a parent population, wherein the value of each parameter in the p individuals is randomly generated in a value range, and p is a positive integer;

and S500, training the parent population by adopting an improved differential evolution algorithm to obtain an optimal individual, and taking the parameters in the optimal individual as the parameters of the fractional order model to obtain the fractional order model of the permanent magnet synchronous motor.

2. The method according to claim 1, wherein the step S500 specifically includes:

step S510, acquiring p individuals x in the parent population_1，g，x_2，g，...，x_i，g，...，x_p，gWherein i is the serial number of the individual, i ∈ {1, 2.., p }, and g is the algebra of population evolution;

step S520, the individuals x in the parent population are subjected to the comparison according to the serial numbers of the individuals_i，gCarrying out mutation operations in sequence to obtain variant individuals v_i，gThe formula of the mutation operation is:

v_i，g＝x_pbest，g+F_i×(x_r1，g-x_r2，g)

wherein r1 ∈ {1,2, …, p }, r2 ∈ {1,2, …, p }, r1, r2 are not equal to each other, x_r1，g、x_r2，gAre all one individual randomly selected in the population, x_pbest，gIs an individual randomly selected from a preferred population, wherein the preferred population is that p individuals are increased from small to large according to fitness valueTop k individuals selected after ranking, F_iIs the scaling factor of the ith individual, the scaling factor F of the ith individual_iThe calculation formula of (2) is as follows:

F_i＝μF_g+t×rand(-1，1)

wherein the scaling factor F of the ith individual_iIs in the range of 0 to 1, μ F_gIs the average scaling factor of the g generation, the average scaling factor of the g +1 generation_g+1The calculation formula of (2) is as follows:

μF_g+1＝(1-e)·μF_g+e·mean_L(S_F)

wherein S is_FIs the set of scaling factors for successful variant individuals of the g-th generation, e ∈ [0.1, 0.5 ]]，mean_L(S_F) Is a set S_FLehmer mean, mean of_L(S_F) The calculation formula of (2) is as follows:

step S530, for the variant individual v according to the individual serial number_i，gSequentially carrying out cross operation to obtain an experimental individual u_i，gThe calculation formula of the cross operation is as follows:

wherein u is_i，gFor the experimental individual, x_i，gAs a parent individual, CR_iRand (0,1) is a random number between 0 and 1 for the cross probability factor;

cross probability factor CR for the ith individual_iThe calculation formula of (2) is as follows:

CR_i＝μCR_g+t×rand(-1，1)

wherein, CR_i∈(0，1)，μCR_gIs the average cross probability factor of the g generation and the average cross probability factor μ CR of the g +1 generation_g+1The calculation formula of (2) is as follows:

μCR_g+1＝(1-e)·μCR_g+e·mean_A(S_CR)

wherein mean is_A(S_CR) Is a set S_CRThe arithmetic mean of (a);

step S540, the experimental individuals u are sequentially subjected to the sequence number of the individuals_i，gAnd parent individual x_i，gPreferentially selecting individuals with smaller fitness value as next generation individuals to obtain p next generation individuals, and taking the p next generation individuals as next generation population;

s550, selecting an individual with the minimum fitness value from the next generation population as a preferred individual;

step S560, judging whether the fitness value of the preferred individual is lower than a specified threshold value, if not, taking the next generation population as the parent population and continuing to execute the step S520; and if so, finishing training, and taking the preferred individual with the minimum fitness value as the final preferred individual.

3. The method as claimed in claim 2, wherein in step S540, the experimental unit u is determined_i，gAnd parent individual x_i，gPreferentially selecting the individuals with smaller fitness values as the next generation of individuals, wherein the selection comprises the following steps:

the experimental subject u_i，gSubstituting into the fractional order model of the permanent magnet synchronous motor to obtain the fractional order model G (u) of the experimental individual_i，g) According to a fractional order model G (u)_i，g) Calculating to obtain the fitness value delta f (u) of the experimental individual_i，g)；

Will be the parent x_i，gSubstituting into the fractional order model of the permanent magnet synchronous motor to obtain the fractional order model G (x) of the parent individual_i，g) According to a fractional order model G (x)_i，g) Calculating to obtain the fitness value delta f (x) of the parent individuals_i，g)；

The fitness value Δ f (u) of the experimental individuals was compared according to the following function_i，g) And fitness value of parent individual Δ f (x)_i，g)：

Taking the individuals with smaller fitness value as next generation individuals x_i，g+1。

4. The method according to claim 2 or 3, wherein the fitness value is calculated according to a fitness function, and the fitness function is:

wherein f (hm) is the actual output rotating speed of the permanent magnet synchronous motor,

5. A device for identifying a fractional order model of a permanent magnet synchronous motor, the device comprising: memory, a processor and a computer program stored on the memory and executable on the processor, which computer program, when being executed by the processor, carries out the steps of the method of fractional order model identification of a permanent magnet synchronous machine according to any of claims 1 to 4.