Permanent magnet synchronous motor parameter identification method based on improved eel-grouper algorithm

Abstract

The invention belongs to the technical field of parameter estimation of machines, and particularly relates to a permanent magnet synchronous motor parameter identification method based on an improved eel-grouper algorithm. The method comprises the following steps: establishing a full-rank discrete voltage equation set of a mathematical model of a PMSM under a dq coordinate system; constructing a fitness function reflecting an error between an actual parameter output value and a theoretical output value of the motor; performing parameter identification on the permanent magnet synchronous motor by using an improved eel-grouper optimization algorithm, and initializing a population by using an elite reverse learning mechanism; a Levy flight strategy is introduced in the algorithm development stage; introducing a reconnaissance alert strategy of a sparrow search algorithm; a chaotic Tent search strategy is introduced, a fitness function is calculated, parameter optimization is carried out through algorithm iteration, iteration is stopped after the maximum number of iterations is reached, and optimal identification parameters are output. By improving a search strategy and a parameter adjustment mechanism of the algorithm, the precision and robustness of parameter identification are further improved, and more reliable parameter support is provided for high-performance control of the permanent magnet synchronous motor.

Description

Technical Field

[0001] The present invention belongs to the technical field of parameter estimation of machines, and particularly relates to a parameter identification method for a permanent magnet synchronous motor based on an improved eel - grouper algorithm. Background Technique

[0002] Permanent Magnet Synchronous Motor (PMSM) has been widely used in many fields such as transportation, metallurgy, aerospace, etc. due to its advantages of high efficiency, high power factor, simple structure, etc. However, the parameters of the permanent magnet synchronous motor will be affected by factors such as temperature and magnetic circuit saturation during actual operation, resulting in a reduction in the effect of the motor control system and even possible permanent damage to the motor. Therefore, accurately obtaining the parameters of the permanent magnet synchronous motor is crucial for achieving high - performance control and reliable condition monitoring.

[0003] Currently, the parameter identification methods of permanent magnet synchronous motors are mainly divided into two categories: offline identification and online identification. Although the offline identification method is simple, it cannot reflect the changes of motor parameters in real time, while the online identification method requires accurate model input - output information and has high requirements for the hardware circuit and identification algorithm. Existing online identification methods include the variable information least - squares algorithm based on Gaussian modulation, the method based on sliding - mode observers, etc. These methods have improved the identification accuracy and speed - regulation stability to a certain extent, but still have problems such as large computational amount and insufficient identification accuracy.

[0004] In recent years, intelligent optimization algorithms have been widely used in the field of motor parameter identification. For example, parameter identification methods based on the Whale Optimization Algorithm (WOA), Dragonfly Optimization Algorithm (DA), etc. have achieved efficient identification of motor parameters by simulating the behavior patterns of natural organisms. In 2024, Ali Mohammadzadeh and Seyedali Mirjalili proposed a new meta - heuristic algorithm - the Eel and Grouper Optimizer (EGO). This algorithm is inspired by the symbiotic interaction and foraging strategies of eels and groupers in the marine ecosystem and has good global search ability and convergence speed. However, when the existing eel - grouper optimization algorithm is applied to the parameter identification of permanent magnet synchronous motors, there are still deficiencies such as the convergence speed and identification accuracy of the algorithm not being ideal enough, and it is easy to fall into local optima. Summary of the Invention

[0005] The present invention aims to provide a parameter identification method for a permanent magnet synchronous motor based on an improved eel - grouper algorithm, so as to improve the accuracy, convergence speed and robustness of the parameter identification of the permanent magnet synchronous motor through an online identification method, and provide more reliable parameter support for the high - performance control of the permanent magnet synchronous motor.

[0006] To achieve the above object, the solution of the present invention is: a permanent magnet synchronous motor parameter identification method based on an improved eel-grouper algorithm, comprising the following steps:

[0007] Step S1: Establish a full-rank discrete voltage equation set based on the mathematical model of the permanent magnet synchronous motor in the dq coordinate system;

[0008] Step S2: Construct a fitness function reflecting the error between the actual parameter output value and the theoretical output value of the motor;

[0009] Step S3: Use the improved eel-grouper optimization algorithm to identify the parameters of the permanent magnet synchronous motor, calculate the fitness function, perform parameter optimization through algorithm iteration, stop iteration after reaching the maximum number of iterations, and output the optimal identification parameters;

[0010] In step S3, the improved eel-grouper optimization algorithm includes: initializing the population using an elite reverse learning mechanism; introducing a levy flight strategy in the algorithm development stage; introducing a scouting and warning strategy of the sparrow search algorithm with an adaptive factor; introducing a chaotic Tent search strategy.

[0011] The working principle and beneficial effects of this solution are as follows: The present invention has made multiple optimization improvements to the original eel-grouper algorithm, including initializing the population using an elite reverse learning mechanism to improve the optimization efficiency of the algorithm and reduce the randomness and blindness of the initial population; introducing a levy flight strategy in the algorithm development stage to enhance the global search ability of the eel-grouper algorithm; introducing a warning and scouting strategy of the sparrow search algorithm with an adaptive factor to guide individuals to move towards a better solution based on the positions of a part of the individuals, the optimal position, and the worst position in the population, avoiding falling into a local optimum; introducing a chaotic Tent search strategy, which is beneficial for the algorithm to jump out of the local optimum, avoid premature convergence, and improve the optimization accuracy of the algorithm. The above improvement points enable the permanent magnet synchronous motor parameter identification method to have higher identification accuracy and faster convergence speed.

[0012] Optionally, step S1 further includes:

[0013] When the motor is operating stably, a strategy of injecting negative-sequence weak magnetic current on the d-axis is used to establish a discrete voltage equation for the permanent magnet synchronous motor identification model:

[0014]

[0015] In the above formula: i d , i q are the stator currents on the d-axis and q-axis respectively, u d0 (k), u q0 (k), ω e0 (k), i q0(k) are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d = 0; u d1 (k), u q1 (k), ω e1 (k), i q1 (k) are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d ≠ 0; R s is the stator resistance; L d , L q are the d-axis and q-axis stator winding inductances; ω e is the rotor electrical angular velocity, is the permanent magnet flux linkage.

[0016] Establish the stator voltage equation of a permanent magnet synchronous motor (PMSM) in the synchronous rotating coordinate system (dq coordinate system):

[0017]

[0018] In the above formula: u d is the d-axis stator voltage, u q is the q-axis stator voltage; i d , i q are the d-axis and q-axis stator currents respectively; R s is the stator resistance; L d , L q are the d-axis and q-axis stator winding inductances; ω e is the rotor electrical angular velocity, is the permanent magnet flux linkage; and are the change rates of the d-axis current and q-axis current respectively;

[0019] When the motor operates stably, it is approximately considered that And in the control strategy of the permanent magnet synchronous motor, usually i d = 0, then the discrete voltage equation of the permanent magnet synchronous motor in the simplified dq coordinate system is:

[0020]

[0021] In the above formula: u d0 , u q0 , ω e0 , i q0 are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d = 0;

[0022] When the motor operates stably, a strategy of injecting negative-sequence weak magnetic current into the d-axis is adopted, that is, id ≠0. Combining equations (1) and (2), a discrete voltage equation of a fourth-order full-rank permanent magnet synchronous motor identification model can be further obtained:

[0023]

[0024] In the above equation: u d0 (k), u q0 (k), ω e0 (k), i q0 (k) are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d = 0; u d1 (k), u q1 (k), ω e1 (k), i q1 (k) are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d ≠ 0; R s is the stator resistance; L d , L q are the d-axis and q-axis stator winding inductances; ω e is the rotor electrical angular velocity, is the permanent magnet flux linkage.

[0025] Optionally, in step S2, the fitness function formula is as follows:

[0026]

[0027] In the above equation, h1, h2, h3, and h4 are fitness weight coefficients, is the fitness function value, u d0 , u q0 are the d-axis and q-axis stator voltages when i d = 0, u d1 , u q1 are the d-axis and q-axis stator voltages when i d ≠ 0; are the d-axis and q-axis stator voltages predicted by the algorithm when i d = 0, are the d-axis and q-axis stator voltages predicted by the algorithm when i d ≠ 0.

[0028] Optionally, in step S3, a Levy flight strategy is introduced in the algorithm development stage. The formula used by the original eel-grouper algorithm in the development stage is as follows:

[0029] a = 2 - 2×(t / Max_t)

[0030] r3 = (a - 2)×r1 + 2

[0031]

[0032]

[0033] Among them, a is the contraction and enclosure coefficient, t is the current iteration number, Max_t is the maximum iteration number, r1 is a random number between [0, 1], and r3 is a random number whose value range changes with the number of iterations. is the coefficient vector, and X1 is the mathematical model of the eel hunting path. is the position of the i-th eel at the t-th iteration. is the position of the prey, and X2 is the mathematical model of the grouper hunting path. is the optimal position of the whole at the t-th iteration.

[0034] Optionally, in step S3, the formula adopted by the improved eel-grouper algorithm is:

[0035]

[0036] Among them, t is the current iteration number. is the coefficient vector, and X1 is the mathematical model of the eel hunting path. is the position of the i-th eel at the t-th iteration. is the position of the prey, and X2 is the mathematical model of the grouper hunting path. is the optimal position of the whole at the t-th iteration; Levy(dim) is the Levy distribution function, and dim is the dimension. The original eel-grouper optimization algorithm is prone to falling into local optimum in the exploitation stage because the predation behaviors of eels and groupers are limited to the search around the current optimal position. In the exploitation stage of the improved eel-grouper algorithm, when simulating the predation behaviors of eels and groupers to update the individual positions, the Levy flight strategy is introduced. The introduction of the Levy flight strategy enables the algorithm to also perform global exploration in the exploitation stage. The Levy flight can generate long-distance random jumps, enabling the individuals in the population to conduct large-scale exploration in the search space, thus having the opportunity to escape from the local optimum area and continue to search for the possible better solutions, breaking the original local search limitation, enhancing the global exploration ability, effectively avoiding the algorithm from falling into local optimum prematurely, and improving the global search ability and convergence accuracy.

[0037] Optionally, the formula for initializing the population using the elite opposition-based learning mechanism is:

[0038] X′ i = X i + r1(r2(u b + u l - X i ) - Xi ) (4)

[0039] In the above formula, u b is the upper bound, u l is the lower bound, X i is the position randomly generated within the range of [u b , u l for the current individual in the initial population, X i ′ is the dynamic reverse position of the current individual, and r1, r2 are random numbers between [0, 1];

[0040]

[0041] In the above formula, X i is the current individual in the initial population, X i ′ is the dynamic reverse position of the current individual; f(X i ′) is the fitness function of X i ′, and f(X i ) is the fitness function of X i . The elite reverse learning mechanism provides the population with a wider space exploration ability by generating the dynamic reverse position of the current individual. The reverse learning mechanism generates the reverse position through formula (4). When comparing the fitness of the reverse position and the original position in formula (5), the position with better fitness is selected as the final initial individual position. This mechanism introduces reverse thinking, enabling the initial population not only to be distributed within the existing random range but also to automatically judge whether the reverse position is more conducive to the optimization goal, reducing the number of iterations and increasing the convergence speed. Compared with the traditional completely random initialization, this mechanism can utilize the information of the search space more efficiently. Due to the strong randomness of the initial population in the original algorithm, it may lead to dense distribution near local regions and lack the ability to effectively explore the global space. The elite reverse learning mechanism increases the diversity of the population and expands the search range by introducing the reverse position, avoiding the initial population being too concentrated in local regions and falling into local optima. The diverse initial population can explore within a wider range, find new optimal parameters faster, make the algorithm have stronger adaptability when facing uncertainties, and improve the robustness.

[0042] Optionally, in step S3, introduce the scouting and warning strategy of the sparrow search algorithm with an adaptive factor. The formula for the adaptive factor is as follows:

[0043]

[0044] In the above formula, w is the adaptive factor, t is the current iteration number, Max_t is the maximum iteration number, wmax is the maximum value of the adaptive factor, and wmin is the minimum value of the adaptive factor; the formula for the scouting and warning strategy of the sparrow search algorithm is as follows:

[0045]

[0046] In the above formula, is the position of the i-th sparrow in the j-th dimension in the t-th iteration, is the overall optimal position of the tth iteration, This is the overall worst position at this time, w is the adaptive factor, and f i is the fitness of a sparrow randomly selected according to the proportion, f g This is the overall optimal fitness at this time, r1 is a random number between [0,1], f w This is the overall worst fitness at this time, and ε is a minimum constant. The reconnaissance and vigilance strategy of the sparrow search algorithm simulates the vigilance behavior in the sparrow population, so that the algorithm can maintain the ability to explore the global optimal solution during the search process. When some sparrows in the population find potential predators (i.e., local optimal solutions), they will alert other sparrows, prompting the entire population to move quickly to a new location, thereby avoiding falling into the local optimal solution. Based on the original eel-grouper algorithm, the reconnaissance and vigilance strategy of the sparrow search algorithm with an adaptive factor is introduced to enhance the overall global search capability and identification accuracy of the algorithm. In formula (7), when the individual fitness f i Greater than the global optimal fitness f g When the individual position is close to the global optimal position, the convergence speed of the algorithm is accelerated; when the individual fitness f i Equal to the global optimal fitness f g When , the individual position will be far away from the global worst position, thereby increasing the global search capability of the algorithm; the dynamic adjustment of the adaptive factor w enables the individual to automatically balance the relationship between global search and local search during the search process. In the early stage of the search, a larger w enables the individual to search over a large range and avoid falling into the local optimum. In the later stage of the search, a smaller w enables the individual to search more finely in the local area, improving the recognition accuracy; the introduction of r1 increases the randomness of the individual's movement, allowing the individual to explore more search space while moving away from the worst position. This randomness helps the algorithm to jump out of the local optimum and improve the global search capability.

[0047] Optionally, in step S3, a chaotic tent search strategy is introduced, and its formula is as follows:

[0048]

[0049] In the above formula, is the i-th individual in the d-th dimension, for The result after mapping to [0,1] is is the largest individual in the dth dimension, is the individual that is the smallest in the d-th dimension;

[0050]

[0051] In the above formula, is the result after the Tent chaotic mapping;

[0052]

[0053] In the above formula, is the result of the carrier to the original search space.

[0054]

[0055] In the above formula, is the original individual 's fitness value, is 's fitness value. In the chaotic Tent search strategy, first, the position of the individual is mapped between [0, 1], then a new chaotic sequence is generated through the Tent mapping, and then the chaotic sequence is transformed back to the original search space. Through formulas (8) to (10), the positions of the original population are mapped to the chaotic space. After the chaotic mapping, new individual positions are obtained. Finally, through formula (11), the fitness of the original population and the new population is compared, and the individual with a higher fitness is selected as the next-generation population. Introducing the chaotic Tent search strategy based on the original eel-groupers algorithm can guide the algorithm to jump out of the local optimum and enter a new search area, further improving the global search ability and identification accuracy of the algorithm. In addition, the chaotic Tent mapping can also enhance the anti-interference ability of the algorithm. When affected by external interference, the algorithm can, under the action of the chaotic search mechanism, quickly return to the correct search trajectory through re-mapping and adjustment, improving the robustness.

[0056] Optionally, the improved eel-groupers optimization algorithm is used to identify the parameters of the permanent magnet synchronous motor. The obtained identified parameters are substituted into the discrete voltage equation of the permanent magnet synchronous motor identification model in step S1 to calculate the fitness value. Through algorithm iteration, these four identified parameters are optimized and corrected to make them continuously approach the corresponding true parameters. After reaching the maximum number of iterations, the iteration is stopped, and the optimal identified parameters are output. Brief Description of the Drawings

[0057] Figure 1 is the system control block diagram of a permanent magnet synchronous motor parameter identification method based on an improved eel-groupers algorithm according to an embodiment of the present invention;

[0058] Figure 2It is the flow schematic diagram of a method for identifying the parameters of a permanent magnet synchronous motor based on an improved eel-grouper algorithm in an embodiment of the present invention;

[0059] Figure 3 It is the change curve of the stator resistance identified in an embodiment of the present invention;

[0060] Figure 4 It is the change curve of the d-axis inductance identified in an embodiment of the present invention;

[0061] Figure 5 It is the change curve of the q-axis inductance identified in an embodiment of the present invention;

[0062] Figure 6 It is the change curve of the permanent magnet flux linkage in an embodiment of the present invention. Specific embodiments

[0063] The following is a further detailed description through specific embodiments:

[0064] Embodiment

[0065] A method for identifying the parameters of a permanent magnet synchronous motor based on an improved eel-grouper algorithm, the system control block diagram of which is as shown in the appendix Figure 1 and includes the following steps:

[0066] Step S1: Establish a full-rank discrete voltage equation set based on the mathematical model of the permanent magnet synchronous motor in the dq coordinate system:

[0067] Establish the stator voltage equation of the permanent magnet synchronous motor (PMSM) in the synchronous rotating coordinate system (dq coordinate system):

[0068]

[0069] In the above formula: u d is the d-axis stator voltage, u q is the q-axis stator voltage; i d , i q are the d-axis and q-axis stator currents respectively; R s is the stator resistance; L d , L q are the d-axis and q-axis stator winding inductances; ω e is the rotor electrical angular velocity, is the permanent magnet flux linkage; and are the change rates of the d-axis current and the q-axis current respectively;

[0070] When the motor runs stably, it is approximately considered that And in the control strategy of the permanent magnet synchronous motor, i dIf it is equal to 0, the discrete voltage equation of the permanent magnet synchronous motor in the simplified dq coordinate system is:

[0071]

[0072] In the above formula: u d0 、u q0 、ω e0 、i q0 are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d = 0;

[0073] When the motor is operating stably, a strategy of injecting negative-sequence weak magnetic current into the d-axis is adopted, that is, i d ≠0. Combining equations (1) and (2), the discrete voltage equation of a fourth-order full-rank permanent magnet synchronous motor identification model can be obtained:

[0074]

[0075] In the above formula: u d0 (k), u q0 (k), ω e0 (k), i q0 (k) are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d = 0; u d1 (k), u q1 (k), ω e1 (k), i q1 (k) are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d ≠0; R s is the stator resistance; L d , L q are the stator winding inductances of the d-axis and q-axis; w e is the rotor electrical angular velocity, is the permanent magnet flux linkage.

[0076] Step S2: Construct a fitness function that reflects the error between the actual parameter output value and the theoretical output value of the motor:

[0077]

[0078] In the above formula, h1, h2, h3, and h4 are fitness weight coefficients, is the fitness function value, u d0 , u q0 are the d-axis and q-axis stator voltages when i d = 0, u d1 , u q1 are when id Stator voltages of the d-axis and q-axis when ≠ 0; is the d-axis and q-axis stator voltages predicted by the algorithm when i d = 0, is i d ≠ 0, the d-axis and q-axis stator voltages predicted by the algorithm.

[0079] Step S3: Use the improved eel-grouper optimization algorithm to identify the parameters of the permanent magnet synchronous motor, calculate the fitness function, perform parameter optimization through algorithm iteration, stop the iteration after reaching the maximum number of iterations, and output the optimal identified parameters.

[0080] Use the improved eel-grouper optimization algorithm to identify the parameters of the permanent magnet synchronous motor, substitute the obtained identified parameters into the discrete voltage equation of the permanent magnet synchronous motor identification model in Step S1, calculate the fitness value, optimize and correct these four identified parameters through algorithm iteration to make them continuously approach the corresponding true parameters, stop the iteration after reaching the maximum number of iterations, and output the optimal identified parameters.

[0081] In Step S3, as shown in the appendix Figure 2 The flow of the permanent magnet synchronous motor parameter identification method based on the improved eel-grouper algorithm is as follows:

[0082] Step S31: Set the population size, the maximum number of iterations Max_t, the search space dimension dim (since there are 4 identified parameters, dim is taken as 4), generate the upper limit u of the population position b and the lower limit u l ; Generate the initial population based on the elite opposition-based learning mechanism, set the current iteration number t to 0, and the formula for generating the initial population using the elite opposition-based learning mechanism is as follows:

[0083] X i ' = X i + r1(r2(u b + u l - X i ) - X i ) (4)

[0084]

[0085] In equation (4), u b is the upper bound, u l is the lower bound, that is, the value range of the parameter to be identified, X i is the current individual in the initial population (randomly generated within the range of [u b , u l ), X i ’ is the dynamic opposite position of the current individual, and r1, r2 are random numbers between [0, 1];

[0086] In equation (5), X i is the current individual position in the initial population, and X i ’ is the dynamic reverse position of the current individual; f(X i ’) is the fitness function of X i ’, and f(X i ) is the fitness function of X i . The reverse learning mechanism generates the reverse position through formula (4). When comparing the fitness of the reverse position and the original position in formula (5), the position with better fitness is selected as the final initial individual position. This mechanism introduces reverse thinking, enabling the initial population to not only be distributed within the existing random range but also automatically determine whether the reverse position is more conducive to the optimization goal, reducing the number of iterations, increasing the convergence speed, and enhancing the diversity of the population and the ability to jump out of local optima.

[0087] Step S32: Substitute each individual position into equation (3), calculate the fitness value of each individual, and update the best position of the population and the worst position of the population

[0088] Step S33: Based on the initial position and fitness value obtained from the above steps, use the improved eel-grouper algorithm EGO for motor parameter identification, which mainly includes two stages: the exploration stage and the exploitation stage:

[0089] In the exploration stage, the grouper searches for the best fitness based on the position of the prey. The position of the eel is randomly selected from the initial population. Two coefficients C1 and C2 are used, and their random values are greater than 1 or less than -1 to make the search individuals move away from the best position. The grouper shares its attack position and observations by shaking its head. As the starvation_rate increases, the probability that the grouper persuades the eel to attack the prey increases. The formula for the exploration stage is as follows:

[0090]

[0091] a = 2 - 2×(t / Max_t)

[0092]

[0093] r3 = (a - 2)×r1 + 2

[0094] r4 = 100×rand

[0095] starvation_rate = 100×(t / Max_t)

[0096]

[0097] In the above formula, is a random position vector, is the position of the i-th grouper at the t-th iteration, and are coefficient vectors, is the optimal position vector at the t-th iteration, fitness is the fitness value, a is the contraction and enclosure coefficient, t is the current iteration number, Max_t is the maximum iteration number, r1 and r2 are random numbers between [0, 1], r3 is a random number whose value range changes with the iteration number, r4 is a random number between [0, 100], starvation_rate is the starvation degree of the grouper, is the position of the i-th eel at the t-th iteration.

[0098] In the exploitation stage, the EGO algorithm simulates the behavior of moray eels preying on reefs and groupers preying on prey in the open water near coral reefs. The success rates of the two predation behaviors are equal. Therefore, the EGO algorithm uses the predation behavior of groupers or eels to update the individual position with a probability of 50%. The formula used in the exploitation stage is as follows:

[0099]

[0100] In the above formula, X1 is the mathematical model of the eel hunting path, is the position of the i-th eel at the t-th iteration, is the position of the prey, and X2 is the mathematical model of the grouper hunting path;

[0101] In this embodiment, the levy flight strategy is introduced in the exploitation stage to enhance the global search ability of the eel-grouper algorithm. The improved formula is as follows:

[0102]

[0103] In the above formula, Levy(dim) is the Levy distribution function, and dim is the dimension. The formula used for Levy(dim) is as follows:

[0104]

[0105] In the above formula, both μ and ν are random numbers between [0, 1], σ is the standard deviation, β is the distribution parameter, and β takes 1.5.

[0106] The original algorithm is more likely to fall into local optima during the development stage because the foraging behaviors of eels and groupers are limited to searching around the current optimal position. The improved eel-grouper algorithm can also perform global exploration during the development stage. Levy flight can generate long-distance random jumps, enabling individuals in the population to explore extensively in the search space, thus having the opportunity to escape from the local optimal region and continue to search for potentially better solutions, enhancing the global exploration ability, effectively avoiding the premature convergence of the algorithm to local optima, and improving the global search ability and convergence accuracy.

[0107] Step S34: Add a scouting strategy to the sparrow search algorithm with an adaptive factor. This strategy simulates the sparrow's perception and response to potential threats during foraging, that is, based on the positions of some individuals, the optimal position and the worst position in the population to guide individuals to move towards better solutions and avoid the algorithm falling into local optima. The formula is as follows:

[0108]

[0109] In equation (6), w is the adaptive factor, t is the current iteration number, Max_t is the maximum iteration number, wmax is the maximum value of the adaptive factor, and wmin is the minimum value of the adaptive factor; the adaptive factor w gradually decreases with the increase of the iteration number, which can achieve the automatic balance between global search and local search of individuals during the search process. In the initial stage of the search, a larger w enables individuals to conduct extensive global searches and avoid falling into local optima. In the later stage of the search, a smaller w enables individuals to conduct more refined searches in the local area, improving the identification accuracy.

[0110] In equation (7), is the position of the i-th sparrow in the j-th dimension at the t-th iteration, is the overall optimal position at the t-th iteration, is the overall worst position at this time, w is the adaptive factor, f i is the fitness of a randomly selected sparrow according to the ratio, f g is the overall optimal fitness at this time, r1 is a random number between [0, 1], f w is the overall worst fitness at this time, and ε is a very small constant. In formula (7), when the individual fitness f i is greater than the global optimal fitness f g the individual position will approach the global optimal position, thus accelerating the convergence speed of the algorithm; when the individual fitness f i is equal to the global optimal fitness f gWhen this happens, the individual position will move away from the global worst position, thus increasing the global search ability of the algorithm. At the same time, r1 increases the randomness of the individual movement, enabling the individual to explore more search spaces while moving away from the worst position. This randomness helps the algorithm jump out of the local optimum and improve the global search ability.

[0111] Step S35: Add the chaotic Tent search strategy, which is beneficial for the algorithm to jump out of the local optimum and improve the optimization accuracy. The Tent search strategy maps the original population position to the interval [0,1], generates a new chaotic sequence through the Tent chaotic mapping, and then carriers the new chaotic sequence back to the original search space. Compare the fitness values of the original population position and the new position, and select the position with better fitness for substitution. The formula is as follows:

[0112]

[0113] In equation (8), is the i-th individual in the d-th dimension, is the result after mapping to [0,1], is the largest individual in the d-th dimension, is the smallest individual in the d-th dimension;

[0114] In equation (9), is the result after Tent chaotic mapping;

[0115] In equation (10), is the result of carrier back to the original search space.

[0116] In equation (11), is the fitness value of the original individual , is the fitness value of.

[0117] In the chaotic Tent search strategy, first map the position of the individual to the interval [0,1], then generate a new chaotic sequence through the Tent mapping, then convert the chaotic sequence back to the original search space to obtain the new individual position. Finally, compare the fitness of the original population and the new population, and select the individual with higher fitness as the next generation population. Introducing the chaotic Tent search strategy on the basis of the original eel-groupers algorithm can guide the algorithm to jump out of the local optimum and enter a new search area, further improving the global search ability and identification accuracy of the algorithm.

[0118] Step S36: Determine whether the current iteration number t reaches the maximum iteration number Max_t. If the condition is satisfied, output the final identification parameter, that is, R sStator resistance, stator winding inductances L of the d-axis and q-axis d , L q , permanent magnet flux linkage φ f ; If the condition is not met, repeat steps two to six until the condition is met.

[0119] In an example adopting this embodiment, the identification change curves of four parameters to be identified by using the method of this application are as Figures 3 to 6 shown. The errors between the identified values and the true values of the four parameters to be identified finally obtained are shown in the following table:

[0120] Motor parameters Identified value True value Error <![CDATA[Stator resistance R s > 2.78417441 2.76 0.8768％ <![CDATA[d-axis inductance L d > 0.00642932 0.00642 0.1453％ <![CDATA[q-axis inductance L q > 0.00641763 0.00642 0.0307％ <![CDATA[Permanent magnet flux linkage φ f > 0.20379691 0.204 0.0996％

[0121] In this example, after convergence, the four parameters to be identified are all very close to the true values of the motor parameters, which proves that the method of this application has the ability to accurately identify motor parameters.

[0122] The above are only embodiments of the present invention. The invention is not limited to the fields involved in this embodiment case. Common general knowledge such as specific structures and characteristics known in the art are not described in detail here. Those of ordinary skill in the art know all the common general knowledge in the technical field to which the invention belongs before the application date or the priority date, can know all the existing technologies in this field, and have the ability to apply the conventional experimental means before this date. Those of ordinary skill in the art can, under the inspiration given by this application, combine their own abilities to improve and implement this solution. Some typical well-known structures or well-known methods should not become obstacles for those of ordinary skill in the art to implement this application. It should be noted that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can still be made, and these should also be regarded as the protection scope of the present invention, and these will not affect the implementation effect of the present invention and the practicality of the patent. The protection scope required by this application should be based on the content of its claims, and the specific implementation manners described in the specification can be used to interpret the content of the claims.

Claims

1. A parameter identification method for permanent magnet synchronous motors based on an improved eel-grouper algorithm, characterized in that: It includes the following steps: Step S1: Establish a full-rank discrete voltage equation set based on the mathematical model of the permanent magnet synchronous motor in the dq coordinate system; Step S2: Construct a fitness function that reflects the error between the actual parameter output value and the theoretical output value of the motor; Step S3: Use the improved eel-grouper optimization algorithm to identify the parameters of the permanent magnet synchronous motor, calculate the fitness function, perform parameter optimization through algorithm iteration, stop the iteration after reaching the maximum number of iterations, and output the optimal identified parameters; In Step S3, the improved eel-grouper optimization algorithm includes: initializing the population using the elite opposition-based learning mechanism; introducing the levy flight strategy in the algorithm development stage; introducing the scouting and warning strategy of the sparrow search algorithm with an adaptive factor; introducing the chaotic Tent search strategy.

2. A parameter identification method for a permanent magnet synchronous motor based on an improved eel-grouper algorithm according to claim 1, characterized in that: Step S1 further includes: When the motor is running stably, a strategy of injecting negative-sequence weak magnetic current on the d-axis is adopted to establish the discrete voltage equation of the permanent magnet synchronous motor identification model: In the above formula: i d , i q are the stator currents of the d-axis and q-axis respectively, u d0 (k), u q0 (k), ω e0 (k), i q0 (k) are respectively the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d = 0; u d1 (k), u q1 (k), ω e1 (k), i q1 (k) are the d-axis stator voltage, q-axis stator voltage, rotor electrical angular velocity, and q-axis stator current when i d ≠ 0; R s is the stator resistance; L d , L q are the stator winding inductances of the d-axis and q-axis; ω e is the rotor electrical angular velocity, is the permanent magnet flux linkage.

3. A permanent magnet synchronous motor parameter identification method based on an improved eel-grouper algorithm according to claim 1, characterized in that: In Step S2, the formula of the fitness function is as follows: In the above formula, h1, h2, h3, and h4 are fitness weight coefficients, is the fitness function value, u d0 , u q0 are the d-axis and q-axis stator voltages at i d = 0, and u d1 , u q1 are the d-axis and q-axis stator voltages at i d ≠0; are the d-axis and q-axis stator voltages predicted by the algorithm at i d = 0, and are the d-axis and q-axis stator voltages predicted by the algorithm at i d ≠0.​ 4. A permanent magnet synchronous motor parameter identification method based on an improved eel-grouper algorithm according to claim 3, characterized in that: In Step S3, when introducing the levy flight strategy in the algorithm development stage, the formula used by the original eel-grouper algorithm in the development stage is as follows: a = 2 - 2×(t / Max_t) r3＝(a-2)×r1+2 Among them, a is the contraction and enclosure coefficient, t is the current iteration number, Max_t is the maximum iteration number, r1 is a random number between [0, 1], and r3 is a random number whose value range changes with the iteration number. is the coefficient vector, and X1 is the mathematical model of the eel hunting path. is the position of the i-th eel at the t-th iteration. is the position of the prey, and X2 is the mathematical model of the grouper hunting path. is the optimal position of the whole at the t-th iteration.

5. A method for identifying the parameters of a permanent magnet synchronous motor based on an improved eel-grouper algorithm according to claim 4, characterized in that: In Step S3, the formula adopted by the improved eel-grouper algorithm is: where t is the current iteration number, is the coefficient vector, X1 is the mathematical model of the eel hunting path, is the position of the i-th eel at the t-th iteration, is the position of the prey, X2 is the mathematical model of the grouper hunting path, is the optimal position at the t-th iteration of the whole; Levy(dim) is the Levy distribution function, and dim is the dimension.

6. A permanent magnet synchronous motor parameter identification method based on an improved eel-grouper algorithm according to claim 5, characterized in that: The formula for initializing the population using the elite opposition-based learning mechanism is: X i ' = X i + r1(r2(u b + u l - X i ) - X i ) In the above formula, u b is the upper bound, u l is the lower bound, X i is the position randomly generated by the current individual in the initial population within the range of [u b , u l , X i ′ is the dynamic reverse position of the current individual, and r1, r2 are random numbers between [0, 1]; In the above formula, X i is the current individual in the initial population, and X i ′ is the dynamic reverse position of the current individual; f(X i ′) is the fitness function of X i ′, and f(X i ) is the fitness function of X i .

7. A method for identifying parameters of a permanent magnet synchronous motor based on an improved eel-grouper algorithm according to claim 6, characterized in that: In Step S3, when introducing the scouting and warning strategy of the sparrow search algorithm with an adaptive factor, the adaptive formula is as follows: In the above formula, w is the adaptive factor, t is the current iteration number, Max_t is the maximum number of iterations, wmax is the maximum value of the adaptive factor, and wmin is the minimum value of the adaptive factor; the formula of the scouting and warning strategy of the sparrow search algorithm is as follows: In the above formula, is the position of the i-th sparrow in the j-th dimension in the t-th iteration, is the overall optimal position of the t-th iteration, This is the overall worst position at this time, w is the adaptive factor, and f i is the fitness of a sparrow randomly selected according to the proportion, f g This is the overall optimal fitness at this time, r1 is a random number between [0,1], f w This is the overall worst fitness at this time, and ε is a very small constant.

8. A permanent magnet synchronous motor parameter identification method based on an improved eel - grouper algorithm according to claim 7, characterized in that: In Step S3, when introducing the chaotic Tent search strategy, its formula is as follows: In the above formula, is the i-th individual in the d-th dimension, is the result after mapping to [0, 1], is the largest individual in the d-th dimension, is the smallest individual in the d-th dimension; In the above formula, is the result after Tent chaotic mapping; In the above formula, W i d is the result of the carrier to the original search space. In the above formula, is the fitness value of the original individual , and f(W i d ) is the fitness value of W i d .

9. A parameter identification method for a permanent magnet synchronous motor based on an improved eel - grouper algorithm according to claim 1, characterized in that: Use the improved eel-grouper optimization algorithm to identify the parameters of the permanent magnet synchronous motor, substitute the obtained identified parameters into the discrete voltage equation of the permanent magnet synchronous motor identification model in Step S1, calculate the fitness value, perform optimization and correction on these four identified parameters through algorithm iteration to make them continuously approach the corresponding true parameters, stop the iteration after reaching the maximum number of iterations, and output the optimal identified parameters.

Images