A multi-parameter identification method for permanent magnet motor based on holographic improved sparrow algorithm

Abstract

The present application relates to permanent magnet synchronous motor multi-parameter identification technical field, and relates to a kind of permanent magnet motor multi-parameter identification method based on holographic improved sparrow algorithm, comprising:1: setting relevant parameters;2: chaos mapping initializes population;3: determine objective function, the fitness value of sparrow individual is calculated and compared, obtain initial best fitness and initial worst fitness and the corresponding position;4: select the sparrow of fitness as finder, the rest sparrow as follower, and a part of sparrow is randomly selected in sparrow population to carry out alarm to surrounding environment, and position is updated;5: using adaptive t distribution to disturb current position, update population, 6: the current optimal solution is disturbed by reverse learning strategy, to generate new solution;7: using greedy rule to judge whether it is global optimal position update.The present application is faster in identification speed, and higher in convergence precision.

Description

Technical Field

[0001] This invention relates to the field of multi-parameter identification technology for permanent magnet synchronous motors, and more specifically, to a multi-parameter identification method for permanent magnet motors based on a holographic improved sparrow algorithm. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in various fields such as home appliances and new energy electric vehicles due to their advantages such as robustness, reliability, high efficiency, energy saving, and wide speed range. Accurate mathematical description of the state parameters of automotive PMSMs under complex operating conditions is crucial, as the accuracy of parameter identification directly affects control performance. Understanding the parameter variation characteristics and influencing mechanisms of the motor under complex operating conditions is a key challenge in parameter identification.

[0003] Currently, motor parameter identification can be broadly categorized into offline identification and online identification. Offline identification methods mainly include experimental calibration, finite element analysis, deceleration, manual trajectory planning, and torque-limiting acceleration methods. These methods store the identified parameters in tabular form and obtain the required parameters for a specific operating condition through table lookup and interpolation. However, the parameters obtained from offline identification are static estimates and cannot reflect changes in motor parameters due to factors such as temperature.

[0004] Online parameter identification methods are essentially transformed into optimization algorithms, offering advantages such as low requirements on the objective function and high efficiency. Currently, various intelligent algorithms exist, including least squares, extended Kalman filtering, model reference adaptive methods, and artificial neural networks. Least squares methods involve large computations and are prone to data saturation. Extended Kalman filtering can effectively identify motor operating states and parameters, but its algorithm iteration requires extensive matrix operations, limiting its application areas. Model reference adaptive parameter identification has a simple structure and uses adaptive laws to ensure the convergence of parameters; however, limited by the accuracy of actual control systems, it often leads to identification results converging to incorrect values ​​or even diverging. Results show that it has high identification accuracy and fast convergence speed, but requires a large amount of computation, making it difficult to implement in practical engineering applications.

[0005] Intelligent optimization algorithms have been combined with system identification problems. One paper proposes a parameter identification algorithm for permanent magnet synchronous motors (PMSMs) based on the gray wolf optimization algorithm. This algorithm can simultaneously identify four parameters: stator resistance, dq-axis inductance, and permanent magnet flux linkage. However, its identification accuracy is low, and its convergence speed is relatively slow. Another paper uses the particle swarm optimization (PSO) algorithm for PMSM parameter identification. This algorithm is easy to understand and has a good convergence speed, but the parameters for PSO are difficult to select, and it is prone to getting trapped in local optima. A teaching and learning optimization algorithm has been proposed, combining a tutoring mechanism with step-by-step learning and reverse learning strategies for students, enhancing the accuracy and reliability of identification. In addition, there are also improved bacterial foraging algorithms, adaptive mayfly algorithms, improved immune algorithms, improved locust optimization algorithms, improved snake optimization algorithms, etc., all continuously researched to improve the convergence speed and accuracy of PMSM parameter identification.

[0006] However, the above identification algorithm has problems such as large computational load, high memory consumption, and high complexity. Therefore, a new multi-parameter identification method for permanent magnet motors is needed to solve the problem. Summary of the Invention

[0007] The present invention provides a multi-parameter identification method for permanent magnet motors based on a holographic improved sparrow algorithm, which can overcome some or all the defects of the prior art.

[0008] According to the present invention, a multi-parameter identification method for permanent magnet motors based on a holographic improved sparrow algorithm includes the following steps: Step 1: Set the relevant parameters; Step 2: Initialize the population using chaotic mapping, optimize the initial population, and improve the search space; Step 3: Determine the objective function, calculate and compare the fitness values ​​of individual sparrows, and obtain the initial best fitness and the initial worst fitness and their corresponding positions; Step 4: Select a sparrow with good adaptability as the discoverer, and the rest of the sparrows as followers. Randomly select a portion of the sparrows in the sparrow population to guard the surrounding environment and update their positions. Step 5: Utilize Adaptive Technology The distribution perturbs the current position and updates the population. Step 6: Perturb the current optimal solution using a reverse learning strategy to generate a new solution; Step 7: Use the greedy rule to determine if it is the global optimum before updating the position.

[0009] As a preferred option, relevant parameters include the maximum number of iterations, the warning threshold, and the proportion of discoverers.

[0010] As a preferred embodiment, the chaotic mapping expression in Step 2 is as follows: ; In the formula x n Represents the iteration value. , n For the number of iterations, µ It is an adjustment parameter between [0, 1].

[0011] As a preferred option, in Step 3, the set matrix of sparrow positions can be represented as: ; In the formula, X This represents the set of sparrow locations. X i Indicates the first i sparrows in rows i For the first i The number of individual sparrows. i =1, 2, ..., N , x i，d For the first i OK, d The dimension of the column; The individual fitness matrix of sparrows is as follows: ; The dominant foraging species are better adapted, prioritizing food acquisition and leading the entire population to move towards the food source area.

[0012] As a preferred option, in Step 4, the expression for updating the position of the foraging leader is: ; In the formula, t is the number of iterations. The current iteration number t sparrows in j Dimensional location information; iter max It is the maximum number of iterations. It is any random value within the range. R 2 and ST These are respectively represented as warning values ​​and safety thresholds, with ranges of... , ;exist R 2< ST Within the interval, the discoverer can expand the search scope to achieve a broad global search; if in R 2> ST Within the area, some sparrows sensed the approaching danger, alerted the entire population, and left and moved away from their current location to search other random safe areas. QIt is any random value belonging to (0, 1) that follows a normal distribution; L The matrix elements are all assigned the value 1. matrix; The follower's position update expression is: ; In the formula It is the position of the worst global result in the current iteration number t. The best position now occupied by the discoverer. A It is a 1'd vector in which elements are randomly assigned the value 1 or -1. This refers to sparrows that have poor adaptability among the followers. i Having failed to obtain food, they must fly to other areas in search of food sources; When sparrows are foraging, if they detect a threat in their surroundings and the area is below a safety threshold, they will alert the entire flock. The discoverer and its followers will abandon their current food location and randomly fly to a safer area. The expression for updating the location of the alert sparrows is: ; In the formula, This represents the globally optimal position in the current iteration number t. b It is a step size adjustment parameter that simultaneously satisfies the conditions of mean 0, variance 1, and normally distributed random numbers; k is a random number in [0, 1]. f i yes i The fitness of individual sparrows f b , f w These are the fitness scores for the globally best and globally worst results, respectively; to avoid a denominator of 0, we set... e It is a very small constant; when f i > f b This indicates that the individual sparrow is on the edge of the sparrow group; f i = f b This indicates that the sparrow at the center of the sparrow group has detected a threat and needs to move closer to other sparrows to reduce its danger.

[0013] As a preferred option, in Step 5, an adaptive mechanism is introduced to update the position of individual sparrows and mitigate perturbations in the incoming population. t The distribution variation strategy is as follows: ; In the formula, This indicates that after individual sparrows mutate iThe position of individual sparrows X i sparrow i Line individual position, t ( N () sets the number of iterations to degrees of freedom.

[0014] As a preferred option, the reverse learning strategy expression in Step 6 is as follows: ; In the formula, , yes The reverse solution, u b It is the upper boundary. l b It is the lower boundary. , b 1 is the information update parameter.

[0015] As a preferred option, in Step 7, a greedy rule is used to compare the fitness values ​​of the old and new positions to determine the update position. The greedy rule is as follows: ; If the new position has better fitness, replace the best fitness with the fitness of the new position; otherwise, the best fitness remains unchanged.

[0016] This invention addresses the shortcomings of traditional identification algorithms, such as low population diversity, potential for getting trapped in local optima, poor local search capability, and low algorithm accuracy, through a holistic improvement. Specifically, it improves the entire process, from population initialization and population position updating to algorithm search and the selection rule for the global optimum. Leveraging the advantages of the Sparrow Algorithm, such as strong optimization ability and fast convergence speed, it is applied to the multi-parameter identification of permanent magnet motors (PMSMs), resulting in faster identification speed and higher convergence accuracy. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a multi-parameter identification method for permanent magnet motors based on a holographic improved sparrow algorithm, as shown in the embodiment. Figure 2 This is a schematic diagram illustrating the PMSM parameter identification principle in the embodiment. Figure 3 This is a schematic diagram of the fitness curve of the Sphere function in the embodiment; Figure 4 This is a schematic diagram of the fitness curve of Rosenbrock's function in the embodiment; Figure 5 This is a schematic diagram of the fitness curve of the Griewank function in the embodiment; Figure 6 This is a simulation model for PMSM parameter identification based on the HISSA algorithm in the example. Figure 7 This is a schematic diagram of the fitness value curves of the three algorithms applied to PMSM parameter identification in the embodiment; Figure 8 This is a schematic diagram of the stator resistance identification curves of the three algorithms applied to PMSM parameter identification in the embodiment; Figure 9 This is a schematic diagram of the d-axis inductance identification curves of the three algorithms applied to PMSM parameter identification in the embodiment; Figure 10 This is a schematic diagram of the q-axis inductance identification curves of the three algorithms applied to PMSM parameter identification in the embodiment; Figure 11 This is a schematic diagram of the permanent magnet flux linkage identification curves for the three algorithms applied to PMSM parameter identification in the embodiment. Detailed Implementation

[0018] To further understand the content of this invention, a detailed description of the invention will be provided in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0019] Example

[0020] like Figure 1 As shown, this embodiment provides a multi-parameter identification method for permanent magnet motors based on the Holographic Improved Sparrow Search Algorithm (HISSA), which incorporates population initialization, adaptive distribution, reverse learning strategy, and greedy rules. The method includes the following steps: Step 1: Set the relevant parameters for HISSA, such as maximum number of iterations, warning threshold, and discoverer ratio. Step 2: Initialize the population using chaotic mapping, optimize the initial population, and improve the search space; Step 3: Determine the objective function, calculate and compare the fitness values ​​of individual sparrows, and obtain the initial best fitness and the initial worst fitness and their corresponding positions; Step 4: Select a sparrow with good adaptability as the discoverer, and the rest of the sparrows as followers. Randomly select a portion of the sparrows in the sparrow population to guard the surrounding environment and update their positions according to Equations (7), (8), and (9). Step 5: Utilize Adaptive Technology The distribution perturbs the current position and updates the population. Step 6: Perturb the current optimal solution using a reverse learning strategy to generate a new solution; Step 7: Use the greedy rule to determine if it is the global optimum before updating the position.

[0021] PMSM mathematical model To simplify the analysis and facilitate the study, the magnetic saturation effect of the permanent magnet synchronous motor and the eddy current and hysteresis loss of the iron core are ignored. The stator voltage equation (1) and flux linkage equation (2) of the PMSM are given in the synchronous rotating dq coordinate system. ; In the formula, u d and u q These are the stator direct-axis voltage and quadrature-axis voltage, respectively. i d and i q These are the stator direct-axis current and quadrature-axis current, respectively. R s It is the stator resistance. y d , y d For magnetic flux linkages on the direct and quadrature axes, L d , L q For stator direct-axis and quadrature-axis inductors, y f It is a magnetic flux generated by a permanent magnet. w e It is the electric angular velocity. (1) and (2) show that there are four unknown parameters. R s , L d , L q , y f The solution needs to be identified, and the rank of the system of equations is 2 < 4, indicating that there are infinitely many solutions.

[0022] When the motor is running stably, the dq-axis current changes little, which can be considered as... The value of the differential term is approximately 0, in i d In the =0 control strategy, the discrete voltage equation of PMSM in the dq coordinate system is (3). ; To ensure that the system of equations has a unique solution, an injection is made along the d-axis. i d The weak magnetic current ≠0, when combined with the voltage equation (3), yields the fourth-order PMSM identification model (4). ; In the formula u d0 ( k ), uq0 ( k ), i d0 ( k ), i q0 ( k )for i d The data collected when =0, similarly, the physical quantities marked with a subscript "1" are the data collected when a weak magnetic current is injected, and the physical quantities marked with "*" are the parameters to be identified.

[0023] General Sparrow Search Algorithm The sparrow search algorithm is a population intelligence optimization algorithm proposed based on the foraging and anti-predation behaviors of sparrow populations in nature. In a sparrow population, individuals with strong foraging abilities act as foraging leaders, leading the entire population to the location of food sources. The rest are followers. Among the followers, some highly alert individuals conduct reconnaissance and early warning, alerting the group when danger is detected, thus keeping the group away from dangerous areas.

[0024] The set matrix of sparrow positions can be represented as: ; In the formula, X This represents the set of sparrow locations. X i Indicates the first i The position of the sparrow in the row, i For the first i Number of individual sparrows i =1, 2, ..., N , x i，d For the first i OK, d Sparrow dimension of the column.

[0025] The individual fitness matrix of sparrows is matrix (6). ; The dominant foraging species are better adapted, prioritizing food acquisition and leading the entire population to move towards the food source area.

[0026] The foraging leader's position update expression is (7). ; In the formula, t is the number of iterations. The current iteration number t sparrows in j Location information of the dimension. iter max It is the maximum number of iterations. Any random value within the range, R 2 and ST These are respectively represented as warning values ​​and safety thresholds, with ranges of... , .exist R 2< ST Within the interval, the discoverer can expand the search scope to achieve a broad global search; if in R 2> ST Within the area, some sparrows sensed the approaching danger, alerted the entire population, and left and moved away from their current location to search other random safe areas. Q It is any random value belonging to (0, 1) that follows a normal distribution. L The matrix elements are all assigned the value 1. matrix.

[0027] The position update expression for the follower (8). ; In the formula It is the position of the worst global result in the current iteration number t. The best position now occupied by the discoverer. A It is a 1'd vector in which elements are randomly assigned the value 1 or -1. This refers to sparrows that have poor adaptability among the followers. i Having failed to obtain food, they must fly to other areas in search of food sources.

[0028] The expression for updating the location of the alert sparrows is (9). When a sparrow population is foraging, if 10-20% of the sparrows in the group sense a threat in their surroundings and the location is below a safe threshold, they will alert the entire group. The discoverer and the follower will abandon their current food location and randomly fly to a safe area. ; In the formula, This represents the globally optimal position in the current iteration number t. b It is a step size adjustment parameter that simultaneously satisfies the conditions of a mean of 0, a variance of 1, and a normally distributed random number. k is a random number in [0, 1]. f i yes i The fitness of individual sparrows f b , f w These represent the globally best and globally worst fitness values, respectively. To avoid a denominator of 0, we set... e It is a very small constant. When f i > f b This indicates that the individual sparrow is on the edge of the sparrow group; fi = f b This indicates that the sparrow at the center of the sparrow group has detected a threat and needs to move closer to other sparrows to reduce its danger.

[0029] Holographic Improved Sparrow Search Algorithm HISSA The Sparrow Search Algorithm (HISSA) has strong optimization capabilities and fast convergence speed, but it suffers from low population diversity, poor local search ability in the later stages of iteration, and low algorithm accuracy, making it prone to getting trapped in local optima. Therefore, this embodiment improves the Sparrow Search Algorithm (HISSA) by incorporating population initialization, adaptive distribution, a reverse learning strategy, and a greedy rule-based holographic approach to enhance the multi-parameter identification accuracy of permanent magnet synchronous motors.

[0030] (1) Chaotic mapping optimization of the initial population To optimize the initial sparrow population quality, increase sparrow population diversity, and improve the algorithm's early-stage global search capability, a chaotic mapping method is used to generate the initial population during initialization, ensuring a more uniform distribution in the solution space. The chaotic mapping expression is shown in (10). ; In the formula x n Represents the iteration value. , n For the number of iterations, µ It is an adjustment parameter between [0, 1].

[0031] (2) Adaptive location update t Distribution variation To further improve the optimization performance of the sparrow algorithm, an adaptive mechanism is introduced for updating the position of individual sparrows and handling perturbations in the incoming population. t Distribution variation strategy, as shown in equation (11). ; In the formula, This indicates that after individual sparrows mutate i The position of individual sparrows X i sparrow i Line individual position, t ( N () sets the number of iterations to degrees of freedom.

[0032] (3) Introduce a reverse learning strategy To further enable individual sparrows to approach the optimal solution faster, a reverse learning strategy is introduced to improve convergence efficiency and global optimization ability. The expression is (12). ; In the formula, , yes The reverse solution, u b It is the upper boundary. l b It is the lower boundary. , b 1 is the information update parameter.

[0033] (4) Greed Rule To address the uncertainty of whether the new position obtained after mutation is the global optimum and to avoid getting trapped in local optima, a greedy rule is used to compare the fitness values ​​of the old and new positions to determine the update position. The greedy rule is as shown in (13). ; If the new position has better fitness, replace the best fitness with the fitness of the new position; otherwise, the best fitness remains unchanged.

[0034] PMSM parameter identification principle The PMSM parameter identification principle is based on system identification theory. It calculates the difference between the output of the established theoretical identification model and the output of the actual system model, and uses an identification algorithm to repeatedly adjust the fitness value. The lower the fitness, the closer the theoretical model is to the actual system model. Finally, it obtains the parameters corresponding to the minimum fitness value. Essentially, it transforms the parameter identification problem into an optimization problem. Figure 2 As shown.

[0035] The fitness function used in this embodiment is shown in Equation 14: ; In the formula Represented as fitness value, These are the weighting coefficients.

[0036] Standard function test algorithm experiment This embodiment verifies the superiority of the improved HISSA algorithm in terms of convergence ability, accuracy, and global optimum through six different standard test functions. The standard test functions are shown in Table 1. To ensure fairness in the testing, all environments were set identically during the algorithm testing process. The SSA, CTSSA, and HISSA algorithms were each run independently 30 times on the three typical test functions, with 500 iterations per iteration. The logarithmic curves of iteration count versus fitness value and the three-dimensional graphs of each test function are shown in the figure. Figure 3-5 As shown.

[0037] Table 1 Standard Test Functions

[0038] Based on the testing characteristics of different test functions, algorithm performance can be compared more clearly. The Sphere test function is characterized by having one and only one unique minimum value. Figure 3 As can be seen from this, the global search minimum capability of the algorithm in this embodiment is stronger than that of the SSA and CASSA algorithms; the Rosenbrock test function can test the global search capability. Figure 4 It can be seen that the convergence speed and accuracy of the algorithm in this embodiment are higher than those of the other two algorithms. The Griewank test function is characterized by the existence of multiple local extrema, which can test the algorithm's ability to escape local optima. Figure 5 shows that the algorithm in this embodiment is significantly improved compared to the SSA algorithm, and has higher final accuracy than CTSSA. Figure 3-5 In summary, the HISSA algorithm in this embodiment performs better than the SSA and CTSSA algorithms in terms of convergence speed, accuracy, and global optimization capability.

[0039] To verify the performance of the HISSA algorithm in PMSM parameter identification, a multi-parameter identification model based on a PMSM vector control system was established in Matlab / Simulink, such as... Figure 6 As shown. Alternating currents of 0 and -2A are injected into the d-axis, respectively at... Control strategies and Collect 1000 sets of identification data under the control strategy As input to the identification model. To verify the performance of the algorithm, the identification range of the parameters to be identified is set far from the true value. In this embodiment, the range is set to [0,5]. The initial population size is set to 30 and the number of iterations is set to 100.

[0040] In the simulation, the motor speed is The load torque is Other parameters of the motor are shown in Table 2. To verify the superiority of the HISSA algorithm used in this embodiment, the HISSA algorithm was compared with SSA and CTSSA. The simulation results were taken as the average value of the three algorithms running independently for 10 times as the final identification output value.

[0041] Table 2 PMSM parameters

[0042] The simulation data is shown in Table 3. Table 3 shows that the HISSA algorithm in this embodiment has the smallest identification error and fitness value compared to the SSA and CTSSA algorithms, and its identification convergence speed and accuracy are the highest. The standard deviation reflects the identification stability; the standard deviation results show that the HISSA algorithm in this embodiment also has better stability. The simulation verifies the feasibility and correctness of the HISSA algorithm in PMSM parameter identification.

[0043] Table 3 Simulation Comparison of Parameter Identification for Three Algorithms

[0044] Experimental setup To further verify the feasibility and effectiveness of the HISSA algorithm used in this embodiment for PMSM parameter identification, a motor experimental platform was built based on simulation for verification. The experimental platform used for verification included a control platform with a PE-Exper4 control box and the accompanying PE-ViewX software as its core, a motor-to-drive platform, a host computer, an oscilloscope, and other equipment.

[0045] Under the same data, initial settings, and simulation conditions, the algorithm was run independently multiple times to verify its superiority in PMSM multi-parameter identification. Similar to the simulation, the three algorithms were run independently 10 times each, and the average value of the 10 runs was taken as the final output value.

[0046] Experimental Results and Analysis Figure 7 The fitness curves of three algorithms applied to PMSM parameter identification are shown below. Figure 7 It can be seen that all three algorithms can converge within 100 iterations. The HISSA algorithm gradually converges and tends to stabilize after 30 iterations, and its convergence speed is significantly faster than the other two algorithms.

[0047] Figure 8 For stator resistance identification curves, Figure 9 The d-axis inductance identification curve. Figure 10 The q-axis inductance identification curve. Figure 11 Permanent magnet flux linkage identification curve.

[0048] from Figures 7-11 It can be seen that the general Sparrow Search Algorithm (SSA) converges slowly in the early stages of iteration, accelerates in the middle stages, but is prone to local optima in the later stages. The final average fitness value is 7.63, which is not high for the actual multi-parameter identification of motors. The improved Sparrow Search Algorithm (CTSSA), which introduces a crossover mutation strategy and dynamic search, achieves better convergence accuracy than the SSA algorithm. In this embodiment, the holographic improved algorithm uses chaotic mapping to generate the initial population during initialization and introduces adaptive methods during the position update process of individual sparrows. The HISSA algorithm employs a distribution mutation strategy, combining the sparrow search algorithm with a back-learning strategy. It uses a greedy rule to compare the fitness values ​​of the old and new positions to determine the update position, avoiding getting trapped in local optima. Compared to SSA and CTSSA algorithms, the HISSA algorithm converges faster in the early stages of iteration and is less prone to getting trapped in local optima. In the middle stages of iteration, due to the introduction of adaptive... The distribution mutation strategy and the use of greedy rules to compare the fitness values ​​of the old and new positions to determine the update position enhance its ability to escape local optima. In the later stages of iteration, its final convergence speed and accuracy are higher than the other two algorithms.

[0049] A comparison of the identification results for the four parameters reveals that the identification result for flux linkage has the highest accuracy compared to the other three parameters, while the identification result for stator resistance has the lowest accuracy. The essence of using intelligent algorithms to identify multiple parameters of a PMSM is to find the solution vector with the minimum fitness value; the parameter with the most accurate identification result is the solution vector that has the greatest impact on the fitness value. Combined with... Figure 7 It can be seen that, in the fitness function, the change in flux linkage has the greatest impact, followed by inductance.

[0050] in conclusion This embodiment addresses the problem that general sparrow search algorithms have weak search capabilities and are prone to getting trapped in local optima in the later stages of multi-parameter identification in PMSM. It proposes a holographically improved sparrow algorithm, introducing chaotic mapping for population initialization, employing an adaptive t-distribution to optimize the population, and improving the convergence speed of all individuals in the population through a back-learning mechanism. A greedy rule is used to select the global optimum, avoiding getting trapped in local optima. First, the SSA, CTSSA, and HISSA algorithms are compared and tested using standard test functions, demonstrating the superior identification performance of the HISSA algorithm. Finally, experiments verify that this algorithm has a faster identification speed and higher convergence accuracy in multi-parameter identification of PMSM.

[0051] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A multi-parameter identification method for permanent magnet motors based on a holographic improved sparrow algorithm, characterized in that: Includes the following steps: Step 1: Set the relevant parameters; Step 2: Initialize the population using chaotic mapping, optimize the initial population, and improve the search space; In Step 2, the chaotic mapping expression is as follows: ； In the formula x n Represents the iteration value. , n Indicates the number of iterations. µ It is an adjustment parameter between [0, 1]; Step 3: Determine the objective function, calculate and compare the fitness values ​​of individual sparrows, and obtain the initial best fitness and the initial worst fitness and their corresponding positions; Step 4: Select a sparrow with good adaptability as the discoverer, and the rest of the sparrows as followers. Randomly select a portion of the sparrows in the sparrow population to guard the surrounding environment and update their positions. Step 5: Utilize Adaptive Technology t The distribution perturbs the current position and updates the population. In Step 5, adaptive methods are introduced to update the position of individual sparrows and address the perturbations in the new flock. t The distribution variation strategy is as follows: ； In the formula, This indicates that after individual sparrows mutate i The position of individual sparrows X i sparrow i The individual sparrow's position. t ( N This sets the number of iterations as the degrees of freedom. Step 6: Perturb the current optimal solution using a reverse learning strategy to generate a new solution; In Step 6, the reverse learning strategy expression is as follows: ； In the formula, , yes The reverse solution, u b It is the upper boundary. l b It is the lower boundary. , b 1 is the information update parameter. Step 7: Use the greedy rule to determine if it is the global optimum and update the position accordingly; In Step 7, a greedy rule is used to compare the fitness values ​​of the old and new positions to determine the update position. The greedy rule is as follows: ； If the new position has better fitness, replace the best fitness with the fitness of the new position; otherwise, the best fitness remains unchanged.

2. The method for multi-parameter identification of permanent magnet motors based on the holographic improved sparrow algorithm according to claim 1, characterized in that: Relevant parameters include the maximum number of iterations, the warning threshold, and the proportion of discoverers.

3. The method for multi-parameter identification of permanent magnet motors based on the holographic improved sparrow algorithm according to claim 2, characterized in that: In Step 3, the set matrix of sparrow positions is represented as follows: ； In the formula, X This represents the set of sparrow locations. X i Indicates the first i The position of the sparrow in the row, i For the first i Number of individual sparrows i =1, 2, ..., N , x i，d For the first i OK, d The sparrow dimension of the column; The individual fitness matrix of sparrows is as follows: ； The dominant foraging species are better adapted, prioritizing food acquisition and leading the entire population to move towards the food source area.

4. The method for multi-parameter identification of permanent magnet motors based on the holographic improved sparrow algorithm according to claim 3, characterized in that: In Step 4, the expression for updating the position of the foraging leader is: ； In the formula, t is the number of iterations. The current iteration number t sparrows in j Dimensional location information; iter max It is the maximum number of iterations. It is any random value within the range. R 2 and ST These are respectively represented as warning values ​​and safety thresholds, with ranges of... , ;exist R 2< ST Within the interval, the discoverer can expand the search scope to achieve a broad global search; if in R 2> ST Within the area, some sparrows sensed the approaching danger, alerted the entire population, and left and moved away from their current location to search other random safe areas; Q It is any random value belonging to (0, 1) that follows a normal distribution; L It is a 1× matrix where all elements are assigned the value 1. d matrix; The follower's position update expression is: ； In the formula It is the position of the worst global result in the current iteration number t. This is the best position the discoverer has now occupied. A It is a 1×d vector whose elements are randomly assigned the value of 1 or -1. ; This refers to sparrows that have poor adaptability among the followers. i Having failed to obtain food, they must fly to other areas in search of food sources; When sparrows are foraging, if they detect a threat in their surroundings and the area is below a safety threshold, they will alert the entire flock. The discoverer and its followers will abandon their current food location and randomly fly to a safer area. The expression for updating the location of the alert sparrows is: ； In the formula, This represents the globally optimal position in the current iteration number t. It is a step size adjustment parameter that simultaneously satisfies the conditions of mean 0, variance 1, and normally distributed random numbers; k is a random number in [0, 1]. f i yes i The fitness of individual sparrows f b , f w These are the fitness scores for the globally best and globally worst results, respectively; to avoid a denominator of 0, we set... It is a very small constant; when f i > f b This indicates that the individual sparrow is on the edge of the sparrow group; f i = f b This indicates that the sparrow at the center of the sparrow group has detected a threat and needs to move closer to other sparrows to reduce its danger.

Images