Controller optimization method and device for axial flux motor, and medium

Abstract

The embodiment of the invention discloses a controller optimization method and device for an axial flux motor and a medium, is applied to the technical field of control systems and is used for solving the problem that an existing optimization mode is low in efficiency and single. The method comprises the following steps: constructing an initial control state equation model according to electromechanical coupling characteristics of a to-be-controlled axial flux motor, correcting the initial control state equation model based on real-time operation data to obtain a control state equation model, identifying a key adjustment parameter set according to the control state equation model and a current optimization project demand, and setting the key adjustment parameter set according to the key adjustment parameter set. Calling historical tuning data to establish a constraint relation graph of the parameters, determining a multi-dimensional adjustable domain, performing parallel optimization in a solution space of the multi-dimensional adjustable domain according to a preset swarm intelligent optimization algorithm, generating a multi-target optimization solution set of dynamic weight distribution, and performing optimization in the multi-target optimization solution set based on an adaptive decision. And screening out an optimal control strategy combination of the controller to be optimized for different working conditions.

Description

Technical Field

[0001] This specification relates to the technical field of control optimization, and particularly to a method, device, and medium for optimizing the controller of an axial flux motor. Background Art

[0002] The axial flux motor is a motor with high power density, high efficiency, and high torque characteristics, and is widely used in fields such as electric vehicles, wind power generation, and aerospace. During the operation of the axial flux motor, there are various energy losses such as copper loss and iron loss. At this time, if the controller can reasonably control the operating parameters of the motor, the losses of the motor can be reduced and the efficiency of the motor can be improved. In addition, in different application scenarios of the axial flux motor, there are different requirements for the torque output of the axial flux motor. At this time, if it is necessary to make the motor output appropriate torque under various working conditions and meet the load requirements, it is also necessary for the controller to accurately adjust parameters such as the current and voltage of the motor. Therefore, in the application process of the axial flux motor, the optimization of the controller is an important link to ensure the efficient operation of the axial flux motor.

[0003] However, one of the current existing methods for optimizing the axial flux motor controller is to rely on the experience of engineers and the trial-and-error method, manually adjusting the controller parameters to achieve controller optimization. This manual adjustment method has low efficiency and is difficult to handle complex working conditions and multi-objective optimization requirements. The other is to optimize through a mathematical model. However, in the current process of parameter optimization based on the mathematical model, due to the difficulty in accurately modeling the electromechanical coupling characteristics of the axial flux motor, the optimization effect is limited. Moreover, the current method of optimizing for a single optimization objective such as efficiency or stability cannot meet the multi-objective optimization requirements for the controller, and may lead to a decline in other performance indicators. Summary of the Invention

[0004] To solve the above technical problems, one or more embodiments of this specification provide a method, device, and medium for optimizing the controller of an axial flux motor

[0005] One or more embodiments of this specification adopt the following technical solutions:

[0006] One or more embodiments of this specification provide a method for optimizing the controller of an axial flux motor, the method including:

[0007] Determine the electromechanical coupling characteristics of the axial flux motor to be controlled, so as to construct an initial control state equation model including the electromechanical coupling characteristics, and correct the initial control state equation model through the real-time operation data of the axial flux motor to be controlled to obtain the control state equation model;

[0008] Identify a set of key adjustment parameters that affect the control performance of the controller to be optimized according to the control state equation model and the current optimization project requirements;

[0009] Call historical tuning data matching the control state equation model to establish a constraint relationship graph of each parameter in the set of key adjustment parameters, and determine the multi-dimensional adjustable domain of each parameter in the set of key adjustment parameters based on the constraint relationship graph;

[0010] Parallelly search for the optimal solution in the solution space of each multi-dimensional adjustable domain based on a preset swarm intelligence optimization algorithm to generate a multi-objective optimization solution set with dynamic weight allocation;

[0011] Based on adaptive decision-making, screen the optimal control strategy combinations for the controller to be optimized for different working conditions in the multi-objective optimization solution set.

[0012] Optionally, in one or more embodiments of this specification, determine the electromechanical coupling characteristics of the axial flux motor to be controlled, construct an initial control state equation model including the electromechanical coupling characteristics, and correct the initial control state equation model through the real-time operation data of the axial flux motor to be controlled to obtain the control state equation model, specifically including: Specifically including:

[0013] Collect the axial flux motor to be controlled based on a preset multi-source sensing device to obtain the original data of the axial flux motor to be controlled; wherein, the original data includes: mechanical vibration spectrum data, current harmonic component data, and rotational speed fluctuation characteristic data;

[0014] Determine the mechanical end inertia parameters and electrical end impedance characteristics of the axial flux motor to be controlled according to the original data, and correlate the mechanical end inertia parameters and electrical end impedance characteristics to construct a coupling characteristic matrix of the axial flux motor to be controlled;

[0015] Input the coupling characteristic matrix into a preset finite element analysis tool to establish an initial control state equation model including the electromechanical coupling characteristics;

[0016] Perform online parameter identification on the real-time operation data through the recursive least squares algorithm to dynamically correct the corresponding parameters in the initial control state equation and obtain a control state equation model with time-varying characteristics.

[0017] Optionally, in one or more embodiments of this specification, identify a set of key adjustment parameters that affect the control performance of the controller to be optimized according to the control state equation model and the current optimization project requirements, specifically including:

[0018] Based on the control state equation model, the to-be-controlled axial flux motor and the to-be-optimized controller are simulated under a variety of preset working conditions, and the first operating parameters of the to-be-controlled axial flux motor under a variety of preset working conditions and the second operating parameters of the to-be-optimized controller corresponding to the first operating parameters are obtained;

[0019] Extract keywords from the current optimization project requirements, and match the keywords with the parameter tags corresponding to each of the operating parameters to determine the second operating parameter that matches the current optimization project requirements as the first key adjustment parameter; wherein, the first key adjustment parameter is directly related to the current optimization project requirements;

[0020] Based on the performance data of the to-be-optimized controller corresponding to the second operating parameters under each preset industrial control, determine the influence relationship between the second operating parameters and the performance of the to-be-optimized controller, and screen the second operating parameters based on the influence relationship to obtain the second key adjustment parameter;

[0021] Take the union of the first key adjustment parameter and the second key adjustment parameter as the initial key adjustment parameter;

[0022] Based on the preset simulation tool, sequentially input the initial key adjustment parameters into the control state equation model to obtain the output result of the control state equation model;

[0023] Compare the change trend of the performance of the to-be-optimized controller corresponding to the output result with the influence relationship between the performance of the to-be-optimized controller, and screen the initial key adjustment parameters according to the comparison result to obtain a set of key adjustment parameters that affect the control performance of the to-be-optimized controller.

[0024] Optionally, in one or more embodiments of this specification, call the historical tuning data that matches the control state equation model to establish a constraint relationship graph of each parameter in the set of key adjustment parameters, specifically including:

[0025] Based on the model data of the to-be-optimized controller and the set of key adjustment parameters of the control performance of the to-be-optimized controller, call the initial historical optimization data corresponding to the to-be-optimized controller;

[0026] Obtain the historical optimization requirement information corresponding to each of the initial historical optimization data, and screen the corresponding historical optimization data according to the matching relationship between the historical optimization requirement information and the current optimization project requirements;

[0027] Take each parameter in the set of key adjustment parameters as a node of the constraint relationship graph, and initialize the attribute values of each node according to the historical optimization data;

[0028] Calculate the attribute values of each of the nodes based on the Pearson correlation coefficient method, so as to determine the association relationship between each of the nodes according to the calculation result and a preset association threshold; wherein, the association relationship includes: strong association, weak association and no association;

[0029] Take the association relationship between each of the nodes as the edges of the constraint relationship graph, so as to obtain the constraint relationship graph of each parameter through the combination of the nodes and the edges.

[0030] Optionally, in one or more embodiments of this specification, determining the multi-dimensional adjustable domain of each parameter in the set of key adjustment parameters based on the constraint relationship graph specifically includes:

[0031] Obtain the parameter sequence corresponding to each of the key adjustment parameters according to the historical optimization data, so as to determine the initial adjustable range of the key adjustment parameters according to the parameter sequence; wherein, the initial adjustable range is determined based on the minimum threshold and the maximum threshold corresponding to the key adjustment parameter;

[0032] Divide the initial adjustable range of the key adjustment parameter based on a preset interval to obtain a plurality of adjustable sub-ranges;

[0033] Determine the optimization direction corresponding to each of the adjustable sub-ranges according to the parameter values within each of the adjustable sub-ranges and the current parameter value of the key adjustment parameter, and determine the corresponding optimization step size according to the parameter accuracy of the key adjustment parameter, so as to determine the sub-optimization path corresponding to each of the adjustable sub-ranges based on the optimization direction and the optimization step size;

[0034] According to the constraint relationship graph, combine the sub-optimization paths of the key adjustment parameters with an association relationship to construct an optimization path matrix, and use the optimization path matrix as the multi-dimensional adjustable domain of each parameter.

[0035] Optionally, in one or more embodiments of this specification, perform parallel optimization within the solution space of each of the multi-dimensional adjustable domains based on a preset swarm intelligence optimization algorithm to generate a multi-objective optimization solution set with dynamic weight allocation, specifically including:

[0036] Initialize the search parameters of the preset swarm intelligence optimization algorithm according to the number of each key adjustment parameter and the range of the multi-dimensional adjustable domain; wherein, the search parameters include: the number of particles, the inertia weight, the cognitive coefficient, and the social coefficient;

[0037] Take the value combinations of the key adjustment parameters within the multi-dimensional adjustable domain as particle positions, substitute each particle position into a preset fitness function for calculation to obtain the fitness values of each particle position, and update the particle velocities according to the fitness values, so as to update the particle positions based on the updated particle velocities;

[0038] Iteratively calculate the fitness values of the updated particle positions, compare the fitnesses corresponding to each particle position, and update the individual optimal position of each particle and the optimal position of the particle swarm;

[0039] Determine the initial region of the ant colony pheromone distribution according to the individual optimal position and the optimal position of the particle swarm;

[0040] Determine the ant colony pheromone concentration based on the distances between each particle position within the region and the individual optimal position and the optimal position of the particle swarm;

[0041] Calculate the transition probability according to the pheromone concentration and the heuristic factor in each direction of the current particle position, move to the next particle position based on the transition probability, and update the ant colony pheromone concentration from the current particle position to the next particle position; wherein, the heuristic factor is determined based on the distance of the individual optimal or swarm optimal position, and is used to guide the movement towards a direction with higher fitness;

[0042] Iteratively update the particle positions and the ant colony pheromone concentration to obtain the solution combination of the optimal path, and screen the optimal path within the solution combination according to the fast non-dominated sorting algorithm to obtain the multi-objective optimization solution set.

[0043] Optionally, in one or more embodiments of the present specification, after determining the ant colony pheromone concentration based on the distances between each particle position within the region and the individual optimal position and the optimal position of the particle swarm, the method further includes:

[0044] Determine the relaxation parameter corresponding to the current optimization project requirement based on the fault tolerance range corresponding to the current optimization project requirement; wherein, the relaxation parameter includes: relaxation range, relaxation stage, relaxation coefficient, relaxation adjustment step size;

[0045] Perform relaxation expansion on the constraint conditions corresponding to the current optimization project requirement based on the relaxation parameter, and synchronously perform relaxation expansion on the optimizable range of the key adjustment parameters according to the relaxed and expanded constraint conditions to obtain a relaxed and expanded optimizable path;

[0046] Adjust the heuristic factor according to the relaxed and expanded constraint conditions to obtain an updated heuristic factor;

[0047] Determine the fitness values of the particle positions corresponding to each optimizable path after relaxation expansion according to the relaxed and expanded constraint conditions and the current ant colony pheromone concentration, so as to determine the updated individual optimal position and the optimal position information of the population;

[0048] Update the current ant colony pheromone concentration based on the distances between each particle position and the updated individual optimal position and the optimal position of the updated particle swarm.

[0049] Optionally, in one or more embodiments of the present specification, based on an adaptive decision, screen the optimal control strategy combinations for different working conditions of the controller to be optimized within the multi-objective optimization solution set, specifically including:

[0050] Extract the characteristic parameters of the current working condition based on the real-time operation data of the controller to be optimized, and use a clustering algorithm to classify the working conditions into multiple categories based on the extracted characteristic parameters; wherein, the categories include: high-load working condition, low-load working condition, transient working condition;

[0051] Match the priority data corresponding to the categories of each working condition with the multi-objective optimization solution set to realize the sorting of each multi-objective optimization solution in the multi-objective optimization solution set, and determine the dynamic weight values of each multi-objective optimization solution;

[0052] Determine the current required fuzzy set of the controller to be optimized based on the key adjustment parameters, and determine the membership degree functions corresponding to each fuzzy set according to the membership degrees of each solution in the multi-objective optimization solution set to the fuzzy set;

[0053] Map the objective function value of each solution to the corresponding membership degree function, and calculate its membership degree value, so as to screen the multi-objective optimization solution set according to the membership degree value and the dynamic weight values of each multi-objective optimization solution for different working conditions, and determine the optimal control strategy combination for different working conditions of the controller to be optimized.

[0054] The above at least one technical solution adopted in the embodiments of the present specification can achieve the following beneficial effects:

[0055] Determine the electromechanical coupling characteristics of the axial-flux motor to be controlled and construct an initial control state equation model that includes these characteristics, which can accurately reflect the complex electromechanical interaction relationship of the motor. Then, use the real-time operation data to correct the initial control state equation model, enabling the model to adjust with the changes in the actual operating state of the motor, maintaining an accurate description of the motor's operating conditions, and improving the modeling accuracy. Determine the multi-dimensional adjustable domain of each parameter based on the constraint relationship graph, clarifying the value range of the parameters under mutual constraint conditions. This not only provides a reasonable search space for the optimization algorithm, avoiding the waste of computing resources and result deviation caused by blind search, but also ensures that the optimized parameter combination is feasible in actual operation, improving the practicality and reliability of the optimization results. Use the swarm intelligence optimization algorithm to perform parallel optimization within the multi-dimensional adjustable domain, generating a multi-objective optimization solution set with dynamic weight allocation, achieving a trade-off between multiple objectives. Based on the adaptive decision-making mechanism, screen the optimal control strategy combinations for different working conditions within the multi-objective optimization solution set, improving the adaptability of the controller. Brief Description of the Drawings

[0056] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in this specification. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings:

[0057] Figure 1 It is a schematic flowchart of a method for optimizing the controller of an axial-flux motor provided by an embodiment of this specification;

[0058] Figure 2 It is a schematic diagram of a constraint relationship graph in an application scenario provided by an embodiment of this specification;

[0059] Figure 3 It is a schematic diagram of the structure of a device for optimizing the controller of an axial-flux motor provided by an embodiment of this specification. Detailed Description of the Embodiments

[0060] An embodiment of this specification provides a method, device, and medium for optimizing the controller of an axial-flux motor.

[0061] To enable those skilled in the art to better understand the technical solutions in this specification, the following will clearly and completely describe the technical solutions in the embodiments of this specification in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only some embodiments of this specification, rather than all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this specification.

[0062] As shown Figure 1 in the figure, the embodiment of this specification provides a schematic flow chart of a method for optimizing the controller of an axial-flux motor. It can be seen Figure 1 from this that in one or more embodiments of this specification, a method for optimizing the controller of an axial-flux motor includes the following steps:

[0063] S101: Determine the electromechanical coupling characteristics of the axial-flux motor to be controlled, construct an initial control state equation model including the electromechanical coupling characteristics, and correct the initial control state equation model through the real-time operation data of the axial-flux motor to be controlled to obtain the control state equation model.

[0064] Due to its unique structure, the axial-flux motor has the advantages of high power density, high torque density, high efficiency, small axial size, etc., and has broad application prospects in many fields such as electric vehicles, aerospace, wind power generation, and industrial automation. However, its special structure also makes the electromagnetic and mechanical interactions inside the motor more complex, with significant electromechanical coupling characteristics. Different application scenarios have different performance requirements for the motor. For example, electric vehicles require the motor to achieve fast response and efficient operation under different working conditions; the aerospace field has extremely high requirements for the reliability and stability of the motor. Therefore, accurately controlling the axial-flux motor to meet diverse application requirements has become a key issue. Currently, traditional motor control methods often rely on simplified motor models and have limited capabilities in dealing with the dynamic changes and non-linear characteristics of the motor, making it difficult to achieve precise control of the motor. For example, when the motor load suddenly changes or the operating conditions change, traditional control methods may lead to a decline in the output performance of the motor, such as increased torque ripple and slower response speed, which cannot meet the requirements of actual applications. Therefore, in the embodiment of this specification, after determining the electromechanical coupling characteristics of the axial-flux motor to be controlled, an initial control state equation model including the electromechanical coupling characteristics is constructed to solve the problem of inaccurate description of the complex electromechanical coupling characteristics of the motor by the traditional model. Then, the initial control state equation model is corrected through the real-time operation data of the axial-flux motor to be controlled to obtain the control state equation model. That is, during the actual operation of the motor, its parameters will change with factors such as temperature and load. By correcting the model in real time, the model can always match the actual operating state of the motor, improving the accuracy and stability of control.

[0065] Specifically, in one or more embodiments of this specification, determining the electromechanical coupling characteristics of the axial-flux motor to be controlled, constructing an initial control state equation model including the electromechanical coupling characteristics, and correcting the initial control state equation model through the real-time operation data of the axial-flux motor to be controlled to obtain the control state equation model specifically includes the following process:

[0066] Collect the axial flux motor to be controlled according to the preset multi-source sensing device to obtain the original data of the axial flux motor to be controlled. The original data includes: mechanical vibration spectrum data, current harmonic component data, and rotational speed fluctuation characteristic data. Then, determine the mechanical end inertia parameter and electrical end impedance characteristic of the axial flux motor to be controlled according to the original data, that is, analyze the mechanical vibration spectrum data to extract the mechanical end inertia parameter, analyze the current harmonic component data to extract the electrical end impedance characteristic, so as to correlate the mechanical end inertia parameter and the electrical end impedance characteristic, and construct a coupling characteristic matrix of the axial flux motor to be controlled. As an energy conversion device, the core function of the motor is to convert electrical energy into mechanical energy. In this conversion process, there is a close relationship between electromagnetic torque, current, and back electromotive force coefficient. Specifically, the state space equation is a mathematical model used to describe the dynamic characteristics of a system, which represents the relationship between the state variables and input variables of the system with a system of first-order differential equations. That is, from the formula at the mechanical end and τ m =τ e , we can get where τ m represents the mechanical torque, ω is the mechanical angular velocity, J is the moment of inertia, and B is the mechanical damping coefficient. From the formula at the electrical end By transposing, we can get where R is the resistance, L is the inductance, V is the applied voltage, K e is the back electromotive force coefficient, I is the current, and τ e is the electromagnetic torque. In this way, the state space equation By writing the above state space equation in matrix form, let the state vector The input vector u = V, then we can obtain where the coupling characteristic matrix is the coefficient matrix of the state vector in the state space equation, which describes the dynamic coupling relationship between the mechanical end and the electrical end. The elements in the coupling characteristic matrix reflect the correlation between the mechanical end inertia parameter and the electrical end impedance characteristic.

[0067] After obtaining the coupling characteristic matrix of the axial flux motor to be controlled through the above process, the coupling characteristic matrix is ​​input into the preset finite element analysis tools such as ANSYS, COMSOL and other tools, so as to establish an initial control state equation model containing electromechanical coupling characteristics based on the finite element analysis tool. The initial control state equation established based on the preset finite element analysis tool can accurately describe the electromechanical coupling characteristics of the axial flux motor and avoid the errors of the traditional simplified model. Then, the real-time operating data is subjected to online parameter identification through the recursive least squares algorithm, so as to dynamically correct the corresponding parameters in the initial control state equation to obtain a control state equation model with time-varying characteristics. The real-time operating data is subjected to online parameter identification based on the recursive least squares algorithm. The model can dynamically adjust the parameters to adapt to the changes of the motor under different working conditions. This time-varying characteristic helps to deal with the uncertainty in the operation of the motor. Among them, it should be noted that

[0068] S102: According to the control state equation model and the current optimization project requirements, a set of key adjustment parameters that affect the control performance of the controller to be optimized is identified.

[0069] According to the control state equation model obtained in the above step S101 and the current optimization project requirements, a set of key adjustment parameters that affect the performance of the controller to be optimized is identified. It can be understood that the key adjustment parameters directly affect the error between the output of the controller and the expected output. By identifying these parameters, they can be adjusted in a targeted manner to make the output of the controller closer to the desired control target, thereby improving the control accuracy. In addition, traditional controller debugging often requires a lot of trials and errors to determine the appropriate parameters. After identifying the key adjustment parameters, the debugging work can be focused on these key parameters to avoid unnecessary adjustments to a large number of irrelevant parameters, thereby greatly reducing the debugging time and workload.

[0070] Specifically, in one or more embodiments of the present specification, according to the control state equation model and the current optimization project requirements, identifying a set of key adjustment parameters that affect the control performance of the controller to be optimized specifically includes the following processes:

[0071] First, different working conditions will cause the motor and controller to present different operating states. By simulating multiple preset working conditions, in order to fully understand the operating performance of the motor and controller under various conditions and provide a rich data basis for subsequent analysis, the axial flux motor to be controlled and the controller to be optimized will be simulated under multiple preset working conditions according to the control state equation model to obtain the first operating parameters of the axial flux motor to be controlled under multiple preset working conditions and the second operating parameters of the controller to be optimized corresponding to the first operating parameters.

[0072] Extract keywords for the requirements of the current optimization project, then match these keywords with the parameter tags corresponding to each operating parameter, and determine the second operating parameter that matches the requirements of the current optimization project as the first key adjustment parameter. These parameters are directly related to the requirements of the current optimization project. For example, if the optimization project requirement is to increase the output power of the controller, then through matching, the second operating parameter related to the output power can be determined as the first key adjustment parameter. Although some parameters may not be directly related to the optimization project requirements, they may have an important impact on the controller performance. By analyzing the influence relationship between the second operating parameter and the controller performance, these potential key parameters can be found. Therefore, based on the controller performance data to be optimized corresponding to the second operating parameter under each preset working condition, determine the influence relationship between the second operating parameter and the controller performance to be optimized, and then screen the second operating parameter based on this influence relationship to obtain the second key adjustment parameter. Take the union of the first key adjustment parameter and the second key adjustment parameter to obtain the initial key adjustment parameter.

[0073] Then use the preset simulation tool to input the initial key adjustment parameters into the control state equation model in sequence to obtain the output results of the control state equation model. It can be understood that by testing the initial key adjustment parameters in the control state equation model through the simulation tool, the influence of each parameter on the model output can be observed. This helps to further analyze the relationship between the parameters and the controller performance and provides actual data support for the subsequent screening. Then compare the trend of change in the controller performance to be optimized corresponding to the output results with the influence relationship between the second operating parameter and the controller performance to be optimized, and screen the initial key adjustment parameters according to the comparison results to finally obtain the set of key adjustment parameters that affect the control performance of the controller to be optimized. By comparing the performance change trend obtained from the actual simulation with the influence relationship obtained from the previous analysis, the effectiveness of the initial key adjustment parameter can be verified. If the performance change trend of a certain parameter in the simulation does not match the expected influence relationship, then this parameter may not be the key parameter that truly affects the controller performance and should be excluded.

[0074] In the above process, by simulating the operation of the axial-flux motor to be controlled and the controller to be optimized under various preset working conditions, the operation data of the motor and the controller under different conditions can be comprehensively obtained. Different working conditions may include different loads, different ambient temperatures, different rotational speeds, etc., which makes the analysis results closer to the actual application scenarios, avoids the problem of incomplete parameter identification caused by a single working condition, and ensures the comprehensiveness of the identification of key adjustment parameters. In addition, the first key adjustment parameter directly related to the optimization project requirements is determined by keyword matching, and the potential second key adjustment parameter is found by analyzing the influence relationship between the second operating parameter and the controller performance. Finally, the union is taken to obtain the initial key adjustment parameters. This method comprehensively considers the factors directly and indirectly affecting the controller performance, ensuring the systematicness and integrity of the key adjustment parameter set. And through the method of gradual screening and verification, a large number of parameters with little or no influence on the controller performance can be excluded, avoiding the cumbersome work of analyzing all operating parameters, greatly reducing the workload and time cost of the analysis, and improving the efficiency of identifying key adjustment parameters.

[0075] S103: Call the historical tuning data matching the control state equation model to establish a constraint relationship graph of each parameter in the key adjustment parameter set, and determine the multi-dimensional adjustable domain of each parameter in the key adjustment parameter set based on the constraint relationship graph.

[0076] In order to be able to adjust parameters within a known reasonable range and find the parameter combination that meets the requirements more quickly, in the embodiments of this specification, the historical tuning data matching the control state equation model is called to establish a constraint relationship graph of each parameter in the key adjustment parameter set, so as to determine the multi-dimensional adjustable domain of each parameter in the key adjustment parameter set according to the constraint relationship graph. For example, assume there are only two key adjustment parameters. According to the constraint relationship between them, a region can be determined, and all the points in this region corresponding to the parameter combinations are feasible values that meet the mutual constraint conditions.

[0077] Specifically, in one or more embodiments of this specification, calling the historical tuning data matching the control state equation model to establish a constraint relationship graph of each parameter in the key adjustment parameter set specifically includes the following process:

[0078] First, based on the model data of the controller to be optimized and the set of key adjustment parameters for the control performance of the controller to be optimized, the initial historical optimization data corresponding to the controller to be optimized is called. Among them, through the two pieces of information of the model data of the controller to be optimized and the control performance of the controller to be optimized, the historical optimization data related to the current controller to be optimized can be screened out from the historical data warehouse. For example, if the controller model is type A and the set of key adjustment parameters includes parameters a, b, and c, then all the optimization data of type A controllers that involve parameters a, b, and c are found from the historical data as the initial historical optimization data. Then, the historical optimization requirement information corresponding to each initial historical optimization data is obtained to screen the corresponding historical optimization data according to the matching relationship between the historical optimization requirement information and the current optimization project requirements. Then, each parameter in the set of key adjustment parameters is used as a node of the constraint relationship graph, and the attribute values of each node are initialized according to the historical optimization data. At the same time, according to the Pearson correlation coefficient method, the attribute values of each node are used to determine the association relationship between each node according to the calculation result and the preset association threshold; among them, it can be understood that the association relationship includes: strong association, weak association, and no association. Then, the association relationship between each node is used as the edge of the constraint relationship graph to obtain the constraint relationship graph of each parameter through the combination of nodes and edges. As Figure 2 shown in the constraint relationship graph in a certain application scenario of this specification, when the Pearson correlation coefficient is greater than or equal to 0.7, it is a strong association, and when the Pearson correlation coefficient is greater than or equal to 0.3 and less than or equal to 0.7, it is a weak association. Then, the association relationship schematic table shown in Table 1 below reflects the association relationship between parameters A, B, C, D, E, and F. When a solid line represents a strong association relationship and a dotted line represents a weak association relationship, and no line represents no association relationship, the association relationship corresponding to Table 1 can be obtained as Figure 2 shown in the constraint relationship schematic diagram.

[0079] Table 1. Schematic table of the association relationship of key adjustment parameters in a certain application scenario

[0080] Parameter pair Pearson correlation coefficient Association relationship A - B 0.85 Strong association B - C 0.45 Weak association A - D 0.20 No association C - E 0.75 Strong association D - E 0.60 Weak association E - F 0.10 No association

[0081] In this process, data is further screened through the matching between the historical optimization requirement information and the current optimization project requirements, realizing the utilization of past data and ensuring that the historical data used highly matches the current optimization goal. The establishment of the constraint relationship graph is based on a large amount of historical tuning data, which reflects the actual operation of the system under different conditions. Therefore, the parameter constraint relationship and adjustable range determined based on the graph are more in line with the actual system characteristics, and can effectively avoid the problem of system performance degradation caused by unreasonable parameter settings.

[0082] Specifically, in one or more embodiments of this specification, determining the multi-dimensional adjustable domains of the parameters in the set of key adjustment parameters based on the constraint relationship graph specifically includes the following process:

[0083] Obtain the parameter sequences corresponding to the key adjustment parameters according to the historical optimization data, so as to determine the initial adjustable range of the key adjustment parameters according to the parameter sequences. Among them, the initial adjustable range is determined based on the minimum threshold and the maximum threshold corresponding to the key adjustment parameter. Then, divide the initial adjustable range of the key adjustment parameter according to the preset interval to obtain a plurality of adjustable sub-ranges. Then, determine the optimization direction corresponding to each adjustable sub-range according to the parameter values within each adjustable sub-range and the current parameter value of the key adjustment parameter, and determine the corresponding optimization step size according to the parameter accuracy of the key adjustment parameter, so as to determine the sub-optimization paths corresponding to each adjustable sub-range according to the optimization direction and the optimization step size. For example: currently there are key adjustment parameters A and B, and their initial adjustable ranges and optimization paths are "the historical value range of parameter A is [0, 10], and the initial adjustable range is [0, 10]. The historical value range of parameter B is [5, 15], and the initial adjustable range is [5, 15]." If the preset interval of parameter A is 2, then the adjustable sub-ranges of A are [0, 2], [2, 4], [4, 6], [6, 8], [8, 10]. The preset interval of parameter B is 3, and the adjustable sub-ranges of B are [5, 8], [8, 11], [11, 14], [14, 15]. At this time, the current value of parameter A is 5, the optimization direction is to increase, the optimization step size is 1, and the sub-optimization path is [5 to 6], [6 to 7], [7 to 8], [8 to 9], [9 to 10]. The current value of parameter B is 10, the optimization direction is to decrease, the optimization step size is 1, and the sub-optimization path is [10 to 9], [9 to 8], [8 to 7], [7 to 6], [6 to 5]. According to the constraint relationship graph, combine the sub-optimization paths of the key adjustment parameters with an associated relationship to construct an optimization path matrix, and use the optimization path matrix as the multi-dimensional adjustable domain of each parameter.

[0084] In this process, the initial adjustable range is divided using a preset interval to obtain multiple adjustable sub-ranges, further refining the adjustment granularity of the parameters. This enables more precise exploration of the parameter space during the parameter optimization process, capturing the subtle impact of parameter changes on system performance and facilitating the finding of a better parameter combination. The optimization direction is determined based on the parameter values within each adjustable sub-range and the current parameter value of the key adjustment parameter, giving the optimization process a clear goal, avoiding blind parameter adjustment, and improving the optimization efficiency. In addition, by combining the parameter precision of the key adjustment parameter to determine the optimization step size, it is possible to ensure the optimization accuracy while avoiding missing the optimal solution due to an overly large step size or causing the optimization process to be too slow due to an overly small step size. Moreover, according to the constraint relationship graph, the sub-optimization paths of the key adjustment parameters with associated relationships are combined to construct an optimization path matrix as a multi-dimensional adjustable domain. This method fully considers the mutual constraints and influences between parameters, avoiding the problem of only focusing on individual parameters while ignoring the parameter associations during the optimization process.

[0085] S104: Based on a preset swarm intelligence optimization algorithm, perform parallel optimization within the solution space of each of the multi-dimensional adjustable domains to generate a multi-objective optimization solution set with dynamic weight allocation.

[0086] After obtaining the solution space as described above, when facing complex multi-objective optimization problems, the solution space may be extremely large and complex, and serial search may consume a large amount of time and computing resources. Therefore, in the embodiments of this specification, after obtaining the multi-dimensional adjustable domain based on the above steps, parallel optimization is performed within the solution space of each multi-dimensional adjustable domain according to a preset swarm intelligence optimization algorithm, thereby generating a multi-objective optimization solution set with dynamic weight allocation. Specifically, in one or more embodiments of this specification, performing parallel optimization within the solution space of each multi-dimensional adjustable domain based on a preset swarm intelligence optimization algorithm to generate a multi-objective optimization solution set with dynamic weight allocation specifically includes the following process:

[0087] First, to ensure that there are sufficient search points in the solution space, comprehensively explore all regions of the solution space, and enable fine-grained search within this region, in the embodiments of this specification, the search parameters of the preset swarm intelligence optimization algorithm will be initialized according to the number of each key adjustment parameter and the range of the multi-dimensional adjustable domain. Among them, the search parameters include: the number of particles, the inertia weight, the cognitive coefficient, and the social coefficient. The value combinations of each key adjustment parameter within the multi-dimensional adjustable domain are used as the particle positions, and each particle position is substituted into the preset fitness function for calculation to obtain the fitness value of each particle position, and the particle velocity is updated according to the fitness value, so as to update the particle position based on the updated particle velocity to determine the initial region of the ant colony pheromone distribution. That is, after obtaining the individual optimal position and the global optimal position through the particle swarm algorithm, the initial region of the ant colony pheromone distribution is determined accordingly. This enables the ant colony algorithm to focus the search on the region where the optimal solution may exist from the very beginning, avoiding blind search in the entire solution space, improving the pertinence of the search. In addition, the ant colony pheromone concentration is determined according to the distances between the particle positions and the individual optimal and global optimal positions. The closer the distance, the higher the pheromone concentration. Therefore, by continuously updating the pheromone concentration, this dynamic adjustment mechanism can guide the ant colony to gradually converge to the optimal path.

[0088] Then, by iteratively updating the particle positions and the ant colony pheromone concentration, the solution combination of the optimal path is finally obtained. That is, after determining the ant colony pheromone concentration based on the distances between each particle position within the region and the individual optimal position and the global optimal position of the particle swarm, the transfer probability is calculated according to the pheromone concentration and the heuristic factor in each direction of the current particle position, and thus move to the next particle position according to the transfer probability, and update the ant colony pheromone concentration from the current particle position to the next particle position. Among them, the heuristic factor is determined based on the distance to the individual optimal or global optimal position, and is used to guide the movement towards the direction with higher fitness. Then, iteratively update the particle positions and the ant colony pheromone concentration to obtain the solution combination of the optimal path, so as to screen the optimal path within the solution combination according to the fast non-dominated sorting algorithm to obtain the multi-objective optimization solution set. By continuously iteratively updating the particle positions and the ant colony pheromone concentration, the algorithm can gradually optimize the quality of the solution and adapt to the changes in the solution space. Even if a local optimal solution is encountered during the search process, it can jump out of the local optimal through subsequent iterations and continue to approach the global optimal solution.

[0089] Furthermore, in one or more embodiments of this specification, after determining the ant colony pheromone concentration based on the distances between each particle position within the region and the individual optimal position and the global optimal position of the particle swarm, the method further includes the following process:

[0090] In practical optimization problems, strict constraints may limit the search process and cause the algorithm to fall into a local optimal solution. Therefore, in order to increase the chance of finding the global optimal solution, a relaxation parameter corresponding to the current optimization project requirements is determined based on the fault tolerance range corresponding to the current optimization project requirements. Among them, it should be noted that the relaxation parameter includes: relaxation range, relaxation stage, relaxation coefficient, and relaxation adjustment step size. Then, the constraints corresponding to the current optimization project requirements are relaxed and extended according to the determined relaxation parameter, so as to synchronously relax and extend the optimizable range of the key adjustment parameters according to the relaxed and extended constraints, and obtain a relaxed and extended optimizable path. The constraints corresponding to the current optimization project requirements are relaxed and extended according to the determined relaxation parameter. This means that the originally strict constraints are relaxed to a certain extent, providing more feasible solution spaces for the search process. Then, the heuristic factor is adjusted according to the relaxed and extended constraints to obtain an updated heuristic factor. According to the relaxed and extended constraints and the current ant colony pheromone concentration, the fitness values of the particle positions corresponding to each optimizable path after relaxation are determined to determine the updated individual optimal position and the group optimal position information. Based on the distances between each particle position and the updated individual optimal position and the optimal position of the updated particle swarm, the current ant colony pheromone concentration is updated.

[0091] S105: Based on the adaptive decision, screen the optimal control strategy combinations for the controller to be optimized for different working conditions within the multi-objective optimization solution set.

[0092] In practical applications, the controller to be optimized will face multiple different working conditions, and the multi-objective optimization solution set contains multiple solutions for different objective trade-offs. Therefore, in order to find the most suitable control strategy combination for each working condition. In the embodiments of this specification, the optimal control strategy combinations for the controller to be optimized for different working conditions are screened within the multi-objective optimization solution set according to the adaptive decision.

[0093] Specifically, in one or more embodiments of this specification, screening the optimal control strategy combinations for the controller to be optimized for different working conditions within the multi-objective optimization solution set based on the adaptive decision specifically includes:

[0094] Extract the characteristic parameters of the current working condition based on the real-time operation data of the controller to be optimized, and then use a clustering algorithm to classify the working conditions according to the extracted characteristic parameters, dividing the working conditions into multiple categories; among them, the categories include: high-load working conditions, low-load working conditions, and transient working conditions. Match the priority data corresponding to the categories of each working condition with the multi-objective optimization solution set to realize the sorting of each multi-objective optimization solution in the multi-objective optimization solution set, and determine the dynamic weight values of each multi-objective optimization solution. In this step, based on the priority data corresponding to each working condition category, match with the multi-objective optimization solution set to realize the sorting of the multi-objective optimization solution and determine the dynamic weight value. This means that for the characteristics and requirements of different working conditions, each solution in the multi-objective optimization solution set can be reasonably evaluated and sorted, highlighting the importance of each objective under different working conditions. For example, in high-load working conditions, more attention may be paid to the output power and stability of the controller. Through this matching and sorting mechanism, the optimization solutions related to these objectives can have higher weights and are more likely to be selected as the optimal strategy under this working condition.

[0095] Then, determine the current required fuzzy set of the controller to be optimized based on the key adjustment parameters, and determine the membership function corresponding to each fuzzy set according to the membership degree of each solution in the multi-objective optimization solution set to the fuzzy set. Then map the objective function value of each solution to the corresponding membership function and calculate its membership value to screen the multi-objective optimization solution set according to the membership value and the dynamic weight values of each multi-objective optimization solution for different working conditions, and determine the optimal control strategy combination of the controller to be optimized for different working conditions. Map the objective function value of each solution to the corresponding membership function to calculate the membership value, and screen the multi-objective optimization solution set in combination with the dynamic weight value. This method comprehensively considers the weights of each objective under different working conditions and the satisfaction degree of each solution to the fuzzy requirements, and can comprehensively evaluate the multi-objective optimization solution from multiple dimensions, ensuring that the selected optimal control strategy combination can better balance multiple objectives under complex working conditions and realize the optimization of the overall performance. And because this process classifies the working conditions and screens the strategy based on the real-time operation data, it can track the changes of the working conditions in real time and make corresponding adjustments. When the working condition changes from one type to another, the system can quickly re-classify the working conditions, match the priority data, adjust the dynamic weight value, and screen the optimal strategy combination. By combining various information such as working condition classification, priority data matching, fuzzy set theory, and dynamic weight values for strategy screening, it provides rich and comprehensive basis for decision-making. Compared with the single-dimensional decision-making method, this multi-dimensional comprehensive decision-making method can more accurately reflect the actual needs of the system under different working conditions, thus improving the scientificity of decision-making.

[0096] As Figure 3 shown, in one or more embodiments of this specification, a structural schematic diagram of a controller optimization device for an axial-flux motor is provided. By Figure 3It can be seen that in one or more embodiments of this specification, an optimization device for a controller of an axial flux motor, the device includes:

[0097] at least one processor; and,

[0098] a memory communicatively connected to the at least one processor; wherein,

[0099] the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any one of the above-mentioned methods.

[0100] In addition, in one or more embodiments of this specification, a non-volatile storage medium is also provided, storing computer-executable instructions, and the computer-executable instructions can execute any one of the above-mentioned methods.

[0101] The various embodiments in this specification are all described in a progressive manner. The same or similar parts among the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the embodiments of the device, equipment, and non-volatile computer storage medium, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiments.

[0102] The above specifically describes certain embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be executed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain implementations, multitasking and parallel processing are also possible or may be advantageous.

[0103] The above is only one or more embodiments of this specification and is not used to limit this specification. For those skilled in the art, there can be various changes and modifications to one or more embodiments of this specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of this specification shall be included within the scope of the claims of this specification.

Claims

1. A method for optimizing the controller of an axial flux motor, characterized in that, The method includes: Determine the electromechanical coupling characteristics of the axial flux motor to be controlled, construct an initial control state equation model including the electromechanical coupling characteristics, and correct the initial control state equation model through the real-time operation data of the axial flux motor to be controlled to obtain the control state equation model; According to the control state equation model and the current optimization project requirements, identify the key adjustment parameter set that affects the control performance of the controller to be optimized; Call the historical tuning data matching the control state equation model to establish a constraint relationship graph of each parameter in the key adjustment parameter set, and determine the multi-dimensional adjustable domain of each parameter in the key adjustment parameter set based on the constraint relationship graph; Based on the preset swarm intelligence optimization algorithm, perform parallel optimization in the solution space of each multi-dimensional adjustable domain to generate a multi-objective optimization solution set with dynamic weight allocation; Based on adaptive decision-making, screen the optimal control strategy combination of the controller to be optimized for different working conditions in the multi-objective optimization solution set.

2. The controller optimization method of an axial flux motor according to claim 1, wherein Determine the electromechanical coupling characteristics of the axial flux motor to be controlled, construct an initial control state equation model including the electromechanical coupling characteristics, and correct the initial control state equation model through the real-time operation data of the axial flux motor to be controlled to obtain the control state equation model, which specifically includes: Collect the axial flux motor to be controlled based on the preset multi-source sensing device to obtain the original data of the axial flux motor to be controlled; wherein, the original data includes: mechanical vibration spectrum data, current harmonic component data, and rotational speed fluctuation characteristic data; Determine the mechanical end inertia parameter and electrical end impedance characteristic of the axial flux motor to be controlled according to the original data, associate the mechanical end inertia parameter and the electrical end impedance characteristic, and construct the coupling characteristic matrix of the axial flux motor to be controlled; Input the coupling characteristic matrix into the preset finite element analysis tool to establish an initial control state equation model including the electromechanical coupling characteristics; Perform online parameter identification on the real-time operation data through the recursive least squares algorithm to dynamically correct the corresponding parameters in the initial control state equation and obtain a control state equation model with time-varying characteristics.

3. The controller optimization method for an axial flux motor according to claim 1, wherein According to the control state equation model and the current optimization project requirements, identify the key adjustment parameter set that affects the control performance of the controller to be optimized, which specifically includes: Based on the control state equation model, simulate the operation of the axial flux motor to be controlled and the controller to be optimized under various preset working conditions to obtain the first operation parameters of the axial flux motor to be controlled under various preset working conditions and the second operation parameters corresponding to the first operation parameters of the controller to be optimized; Extract keywords from the current optimization project requirements, match the keywords with the parameter tags corresponding to each operation parameter, and determine the second operation parameter matching the current optimization project requirements as the first key adjustment parameter; wherein, the first key adjustment parameter is directly related to the current optimization project requirements; Based on the performance data of the controller to be optimized corresponding to the second operating parameters under each preset industrial control, determine the influence relationship between the second operating parameters and the performance of the controller to be optimized, and screen the second operating parameters based on the influence relationship to obtain the second key adjustment parameters; Use the union of the first key adjustment parameter and the second key adjustment parameter as the initial key adjustment parameter; Based on a preset simulation tool, sequentially input the initial key adjustment parameters into the control state equation model to obtain the output result of the control state equation model; Compare the change trend of the performance of the controller to be optimized corresponding to the output result with the influence relationship between the performance of the controller to be optimized, so as to screen the initial key adjustment parameters according to the comparison result to obtain a set of key adjustment parameters that affect the control performance of the controller to be optimized.

4. The controller optimization method of an axial flux motor according to claim 1, wherein Call historical tuning data matching the control state equation model to establish a constraint relationship graph of each parameter in the set of key adjustment parameters, specifically including: Based on the model data of the controller to be optimized and the set of key adjustment parameters for the control performance of the controller to be optimized, call the initial historical optimization data corresponding to the controller to be optimized; Obtain the historical optimization requirement information corresponding to each initial historical optimization data, and screen the corresponding historical optimization data according to the matching relationship between the historical optimization requirement information and the current optimization project requirements; Use each parameter in the set of key adjustment parameters as a node of the constraint relationship graph, and initialize the attribute values of each node according to the historical optimization data; Calculate the attribute values of each node based on the Pearson correlation coefficient method, and determine the association relationship between each node according to the calculation result and a preset association threshold; wherein, the association relationship includes: strong association, weak association, and no association; Use the association relationship between each node as the edge of the constraint relationship graph to obtain the constraint relationship graph of each parameter through the combination of the node and the edge.

5. The controller optimization method for an axial flux motor according to claim 4, wherein, Based on the constraint relationship graph, determine the multi-dimensional adjustable domain of each parameter in the set of key adjustment parameters, specifically including: Obtain the parameter sequence corresponding to each key adjustment parameter according to the historical optimization data, and determine the initial adjustable range of the key adjustment parameter according to the parameter sequence; wherein, the initial adjustable range is determined based on the minimum threshold and the maximum threshold corresponding to the key adjustment parameter; Divide the initial adjustable range of the key adjustment parameter based on a preset interval to obtain multiple adjustable sub-ranges; Determine the optimization direction corresponding to each adjustable sub-range according to the parameter value within each adjustable sub-range and the current parameter value of the key adjustment parameter, and determine the corresponding optimization step size according to the parameter accuracy of the key adjustment parameter, so as to determine the sub-optimization path corresponding to each adjustable sub-range based on the optimization direction and the optimization step size; According to the constraint relationship graph, each of the sub-optimization paths of the key adjustment parameters with an association relationship is combined to construct an optimization path matrix, and the optimization path matrix is used as the multi-dimensional adjustable domain of each parameter.

6. The controller optimization method for an axial flux motor according to claim 1, wherein Based on a preset swarm intelligence optimization algorithm, parallel optimization is performed in the solution space of each multi-dimensional adjustable domain to generate a multi-objective optimization solution set with dynamic weight allocation, specifically including: According to the number of each key adjustment parameter and the range of the multi-dimensional adjustable domain, the search parameters of the preset swarm intelligence optimization algorithm are initialized; wherein, the search parameters include: the number of particles, the inertia weight, the cognitive coefficient, and the social coefficient. The value combinations of each key adjustment parameter within the multi-dimensional adjustable domain are used as the particle positions, and each particle position is substituted into a preset fitness function for calculation to obtain the fitness value of each particle position, and the particle velocity is updated according to the fitness value, and the particle position is updated based on the updated particle velocity. Iteratively calculate the fitness value of the updated particle position to compare the fitness corresponding to each particle position, and update the individual optimal position of each particle and the optimal position of the particle swarm. According to the individual optimal position and the optimal position of the particle swarm, the initial area of the ant colony pheromone distribution is determined. Based on the distances between each particle position in the area and the individual optimal position and the optimal position of the particle swarm, the ant colony pheromone concentration is determined. Calculate the transfer probability according to the pheromone concentration and the heuristic factor in each direction of the current particle position, and move to the next particle position based on the transfer probability, and update the ant colony pheromone concentration from the current particle position to the next particle position; wherein, the heuristic factor is determined based on the distance of the individual optimal or group optimal position, and is used to guide the movement in the direction with higher fitness. Iteratively update the particle position and the ant colony pheromone concentration to obtain the solution combination of the optimal path, and screen the optimal path within the solution combination according to the fast non-dominated sorting algorithm to obtain the multi-objective optimization solution set.

7. The controller optimization method of an axial flux motor according to claim 6, wherein After determining the ant colony pheromone concentration based on the distances between each particle position in the area and the individual optimal position and the optimal position of the particle swarm, the method further includes: Based on the fault tolerance range corresponding to the current optimization project requirements, determine the relaxation parameters corresponding to the current optimization project requirements; wherein, the relaxation parameters include: the relaxation range, the relaxation stage, the relaxation coefficient, and the relaxation adjustment step size. Based on the relaxation parameters, relax and expand the constraint conditions corresponding to the current optimization project requirements, and synchronously relax and expand the optimizable range of the key adjustment parameters according to the relaxed and expanded constraint conditions to obtain a relaxed and expanded optimizable path. Adjust the heuristic factor according to the relaxed and expanded constraint conditions to obtain an updated heuristic factor. According to the relaxed and expanded constraint conditions and the current ant colony pheromone concentration, determine the fitness value of the particle positions corresponding to each optimizable path after relaxation and expansion, so as to determine the updated individual optimal position and group optimal position information. Update the current ant colony pheromone concentration based on the distances between the positions of each particle and the updated individual optimal position and the optimal position of the updated particle swarm.

8. The controller optimization method for an axial flux motor according to claim 1, characterized in that Based on the adaptive decision-making, screen the optimal control strategy combinations of the controller to be optimized for different working conditions within the multi-objective optimization solution set, specifically including: Extract the characteristic parameters of the current working condition based on the real-time operation data of the controller to be optimized, and classify the working conditions into multiple categories using a clustering algorithm based on the extracted characteristic parameters; wherein, the categories include: high-load working condition, low-load working condition, transient working condition. Match the priority data corresponding to the categories of each working condition with the multi-objective optimization solution set to sort each multi-objective optimization solution in the multi-objective optimization solution set and determine the dynamic weight values of each multi-objective optimization solution. Determine the current required fuzzy set of the controller to be optimized based on the key adjustment parameters, and determine the membership function corresponding to each fuzzy set according to the membership degree of each solution in the multi-objective optimization solution set to the fuzzy set. Map the objective function value of each solution to the corresponding membership function and calculate its membership degree value, so as to screen the multi-objective optimization solution set according to the membership degree value and the dynamic weight values of each multi-objective optimization solution for different working conditions, and determine the optimal control strategy combination of the controller to be optimized for different working conditions.

9. A controller optimization device for an axial flux motor, characterized in that, The device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1-8 above.

10. A non-volatile storage medium stores computer-executable instructions, characterized in that, The computer-executable instructions can: execute the method according to any one of claims 1-8 above.

Images