A robustness optimization method and system for permanent magnet synchronous motor

Abstract

The application relates to a permanent magnet synchronous motor robustness optimization method and system, and belongs to the technical field of motors. The method comprises the following steps: a fast demagnetization risk prediction model is established according to state variables of a permanent magnet synchronous motor in a healthy state and a demagnetization degree of the permanent magnet synchronous motor in a demagnetization fault; a demagnetization rate constraint condition is established according to the working condition of the permanent magnet synchronous motor; a multi-objective optimization function is constructed according to the constraint condition and a target state of the permanent magnet synchronous motor; the fast demagnetization risk prediction model and the multi-objective optimization function are used to search in a preset parameter search space of the permanent magnet synchronous motor by adopting a hierarchical guiding type optimization search mechanism, so that optimal structure parameters of the permanent magnet synchronous motor are obtained; and the robustness of the permanent magnet synchronous motor is optimized according to the optimal structure parameters. The application effectively improves the operation reliability and service life of the motor under complex working conditions, and realizes the optimization balance between the motor performance and the anti-demagnetization reliability.

Description

Technical Field

[0001] This invention relates to the field of motor technology, and in particular to a robust optimization method and system for permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors, with their outstanding advantages such as high efficiency, high power density, and wide speed range, have been widely used in key fields such as new energy vehicles, industrial servo systems, rail transportation, and smart home appliances, becoming a core power component for achieving energy-saving and intelligent development in these fields. As these applications place increasingly higher demands on motor performance, motors not only need excellent efficiency and torque density under rated operating conditions, but also need to maintain long-term reliability under complex operating conditions (such as high-temperature environments and sudden short-circuit faults). The risk of permanent magnet demagnetization is the core bottleneck restricting its reliability.

[0003] Currently, in the field of permanent magnet synchronous motor (PMSM) optimization design, related technologies mostly employ traditional intelligent optimization methods such as genetic algorithms, particle swarm optimization algorithms, or basic spikeless bee optimization algorithms. However, these existing traditional intelligent optimization methods have significant technical limitations, making it difficult to achieve a balanced optimization of PMSM performance and reliability. Specifically, these limitations are as follows: First, the initial population generation is not reasonable enough. Existing methods often use random sampling or simple Latin hypercube sampling to generate the initial population, which can easily lead to uneven distribution of initial solutions and local clustering. This causes the optimization algorithm to get stuck in local optima during the iteration process, making it difficult to fully explore the design space and obtain a globally optimal design scheme that balances performance and reliability.

[0004] Secondly, the search mechanism lacks dynamic adaptability. Taking the basic stingless bee optimization algorithm as an example, its honeypot construction stage only searches within a limited range around randomly selected honeypot queen bees. However, the design space of permanent magnet synchronous motors is a high-dimensional design space (usually including more than 10 design variables such as the number of winding turns, magnet size, and stator and rotor structural parameters). In this type of high-dimensional design space, the above-mentioned limited search method easily leads to a narrow search range of the algorithm, and at the same time causes the population diversity to decay rapidly, thereby reducing the convergence efficiency of the algorithm and increasing the time cost of optimization design.

[0005] Third, there is a lack of multi-objective collaborative optimization mechanisms. Existing traditional optimization algorithms do not construct targeted multi-objective collaborative optimization mechanisms to address the risk of permanent magnet demagnetization. In the design of permanent magnet synchronous motors, there is an inherent coupling conflict between demagnetization constraints and performance objectives such as efficiency and torque density (for example, reducing the thickness of the magnet can increase the motor's torque density, but it will reduce the demagnetization resistance of the permanent magnet). When faced with the above-mentioned conflict between objectives, traditional optimization algorithms cannot achieve a dynamic balance between multiple objectives, and are prone to optimization bias that prioritizes performance over reliability. As a result, although the optimized motor design can meet the performance requirements under rated operating conditions, its demagnetization resistance reliability is insufficient, and it cannot meet the long-term stable operation requirements under complex operating conditions.

[0006] In summary, existing optimization design technologies for permanent magnet synchronous motors based on traditional intelligent optimization algorithms suffer from defects such as unreasonable initial population generation, poor adaptability of the search mechanism, and lack of multi-objective collaborative optimization. These defects make it difficult to overcome the reliability bottleneck caused by the risk of permanent magnet demagnetization and fail to meet the dual requirements of high performance and high reliability for motors in current application scenarios. Summary of the Invention

[0007] Therefore, the technical problem to be solved by the present invention is to overcome the defects in the prior art, such as unreasonable initial population generation, poor adaptability of search mechanism, lack of multi-objective collaborative optimization, and difficulty in achieving balanced optimization of motor performance and anti-demagnetization reliability.

[0008] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a robust optimization method for permanent magnet synchronous motors, comprising: S1. Based on the state variables of the permanent magnet synchronous motor under healthy conditions and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault, establish a rapid demagnetization risk prediction model; S2. Based on the operating conditions of the permanent magnet synchronous motor, establish demagnetization rate constraints. S3. Construct a multi-objective optimization function based on the constraints and the target state of the permanent magnet synchronous motor; S4. Based on the rapid demagnetization risk prediction model and the multi-objective optimization function, within the preset permanent magnet synchronous motor parameter search space, a hierarchical guided optimization search mechanism is used to search for the optimal structural parameters of the permanent magnet synchronous motor; based on the optimal structural parameters, the robustness optimization of the permanent magnet synchronous motor is completed.

[0009] In one embodiment of the present invention, step S1, which involves establishing a rapid demagnetization risk prediction model based on the state variables of the permanent magnet synchronous motor in a healthy state and the degree of demagnetization of the permanent magnet synchronous motor under a demagnetization fault, is as follows: Based on the state variables of the permanent magnet synchronous motor under healthy conditions, establish the voltage equation in the natural coordinate system; By performing coordinate transformation on the voltage equation, we obtain the voltage component equations in a rotating coordinate system. Based on the voltage component equation, the electromagnetic torque equation of the permanent magnet synchronous motor is obtained. Based on the electromagnetic torque equation and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault, a rapid demagnetization risk prediction model is obtained.

[0010] In one embodiment of the present invention, the expression of the rapid demagnetization risk prediction model is: ; in, This indicates the predicted risk value for rapid demagnetization. This indicates the degree of demagnetization in a permanent magnet synchronous motor under demagnetization fault conditions. This indicates the magnetic flux linkage of the permanent magnet under demagnetization fault. and These represent the d-axis and q-axis components of the stator current, respectively. and These represent the d-axis inductance and q-axis inductance of the permanent magnet synchronous motor, respectively.

[0011] In one embodiment of the present invention, step S2, which establishes the demagnetization rate constraint condition based on the operating conditions of the permanent magnet synchronous motor, is as follows: A coercivity fitting model is established based on the operating temperature of the permanent magnet of the permanent magnet synchronous motor. Based on the healthy permanent magnet flux linkage of the permanent magnet synchronous motor and the operating temperature of the permanent magnet, the demagnetization rate of the permanent magnet flux linkage under high temperature conditions is calculated to obtain the high temperature demagnetization rate. Calculate the peak value of the three-phase short-circuit current based on the equivalent circuit of the permanent magnet synchronous motor. Based on the peak value of the three-phase short-circuit current and the number of series turns per phase of the stator of the permanent magnet synchronous motor, calculate the peak value of the armature reaction magnetomotive force of the permanent magnet synchronous motor during a short circuit; The magnetomotive force margin coefficient is calculated based on the coercivity fitting model and the peak value of the armature reaction magnetomotive force. Based on the magnetomotive force margin coefficient, the short-circuit demagnetization rate of the permanent magnet flux under short-circuit conditions is obtained. Based on the high-temperature demagnetization rate and the short-circuit demagnetization rate, a coupling demagnetization rate is established; Based on the high-temperature demagnetization rate, the short-circuit demagnetization rate, and the coupling demagnetization rate, the demagnetization rate constraint conditions under different operating conditions are obtained.

[0012] In one embodiment of the present invention, in step S3, the expression of the multi-objective optimization function is: ; in, Represents a multi-objective optimization function. , and Indicates the adjusted weights; This indicates the efficiency of a permanent magnet synchronous motor. This represents the target efficiency of the permanent magnet synchronous motor. Indicates the target torque density. Indicates torque density, Indicates the demagnetization risk index; and These represent the maximum and minimum permissible demagnetization risk indices, respectively.

[0013] In one embodiment of the present invention, step S4, which involves searching within a preset permanent magnet synchronous motor parameter search space using a hierarchical guided optimization search mechanism based on the rapid demagnetization risk prediction model and the multi-objective optimization function to obtain the optimal structural parameters of the permanent magnet synchronous motor, is as follows: An initial population is generated within a preset dimension of the permanent magnet synchronous motor parameter search space; each individual in the initial population is screened to obtain the core guiding individual; Using the core guided individual as the center, a local search is performed to generate new candidate solutions, and a local new solution generation strategy is determined; Individuals are randomly selected from the initial population as exploration guide individuals. A global search is performed with the exploration guide individuals as the center to generate new candidate solutions, and a global new solution generation strategy is determined. Using the multi-objective optimization function as the optimization objective, the local new solution generation strategy and the global new solution generation strategy are iteratively updated to obtain the optimal structural parameters of the permanent magnet synchronous motor.

[0014] In one embodiment of the present invention, the step of generating an initial population within a preset-dimensional search space for permanent magnet synchronous motor parameters is as follows: For each dimension of the permanent magnet synchronous motor parameter search space, it is divided into multiple equidistant sub-intervals according to the upper and lower bounds; a sample point is randomly selected from each of the sub-intervals to obtain multiple initial skeleton points; A chaotic sequence is generated, and the chaotic sequence is used to perturb the multiple initial skeleton points to obtain an initial population.

[0015] In one embodiment of the present invention, the local new solution generation strategy includes a first local strategy; the first local strategy is: the new candidate solution includes the core guiding individual, randomly selected elite individuals and random individuals, and the generation process introduces a temperature factor to adjust the search step size, the temperature factor being changed according to the current iteration number.

[0016] In one embodiment of the present invention, the process of randomly selecting individuals from the initial population as exploration guide individuals and generating new candidate solutions through a global search centered on the exploration guide individuals includes dividing the initial population and sorting it according to fitness.

[0017] Secondly, to solve the above-mentioned technical problems, the present invention provides a robust optimization system for permanent magnet synchronous motors, comprising: The model building module is used to establish a rapid demagnetization risk prediction model based on the state variables of the permanent magnet synchronous motor in a healthy state and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault. The constraint generation module is used to establish demagnetization rate constraints based on the operating conditions of the permanent magnet synchronous motor. The objective optimization module is used to construct a multi-objective optimization function based on the constraints and the target state of the permanent magnet synchronous motor. The search module is used to search within the preset permanent magnet synchronous motor parameter search space using a hierarchical guided optimization search mechanism based on the rapid demagnetization risk prediction model and the multi-objective optimization function, so as to obtain the optimal structural parameters of the permanent magnet synchronous motor. The execution module is used to perform robust optimization of the permanent magnet synchronous motor based on the optimal structural parameters.

[0018] Compared with the prior art, the above-described technical solution of the present invention has the following advantages: (1) The robust optimization method and system for permanent magnet synchronous motors described in this invention establishes a rapid demagnetization risk prediction model based on motor state parameters under healthy conditions and demagnetization faults. This model can accurately quantify the demagnetization risk of permanent magnets under different operating conditions, providing a reliable theoretical basis and data support for optimization. This effectively solves the problems of fuzzy demagnetization risk assessment and unreasonable constraint settings in traditional optimization methods. This invention establishes demagnetization rate constraints based on the actual operating conditions of the motor, and constructs a multi-objective optimization function based on these constraints and motor performance objectives. This avoids the bias of traditional optimization algorithms that emphasize performance over reliability, and achieves a balance between efficiency, torque density, and anti-demagnetization capability. This ensures that the optimized motor can not only meet the performance requirements under rated operating conditions, but also adapt to long-term stable operation under complex operating conditions such as high temperature and short circuit.

[0019] (2) This invention employs a hierarchical guided optimization search mechanism within a preset parameter search space. Through the hierarchical guided search logic, it balances local fine-grained exploration with global wide-area search, effectively improving population diversity and algorithm convergence efficiency, avoiding the optimization process from getting stuck in local optima, and enabling the rapid and accurate acquisition of the optimal structural parameters of the motor. Ultimately, it achieves robust optimization of the permanent magnet synchronous motor, significantly improving the motor's operational reliability and service life under complex operating conditions. Attached Figure Description

[0020] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0021] Figure 1 This is a flowchart of a robustness optimization method for a permanent magnet synchronous motor according to a preferred embodiment of the present invention; Figure 2 This is a structural diagram of a robust optimization system for a permanent magnet synchronous motor according to a preferred embodiment of the present invention. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0023] Example 1: Reference Figure 1 As shown, this embodiment of the invention provides a robust optimization method for a permanent magnet synchronous motor, including but not limited to the following steps: S1. Based on the state variables of the permanent magnet synchronous motor under healthy conditions and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault, establish a rapid demagnetization risk prediction model. S2. Based on the operating conditions of the permanent magnet synchronous motor, establish demagnetization rate constraints. S3. Based on the constraints and the target state of the permanent magnet synchronous motor, construct a multi-objective optimization function; S4. Based on the rapid demagnetization risk prediction model and multi-objective optimization function, within the preset permanent magnet synchronous motor parameter search space, a hierarchical guided optimization search mechanism is used to search for the optimal structural parameters of the permanent magnet synchronous motor; based on the optimal structural parameters, the robustness optimization of the permanent magnet synchronous motor is completed.

[0024] The robust optimization method for permanent magnet synchronous motors described in this invention achieves synergistic optimization of motor performance and demagnetization resistance reliability. A rapid demagnetization risk prediction model is established based on motor state parameters under healthy and demagnetization fault conditions. This model accurately quantifies the demagnetization risk of permanent magnets under different operating conditions, providing a reliable theoretical basis and data support for subsequent optimization. It effectively solves the problems of fuzzy demagnetization risk assessment and unreasonable constraint settings in traditional optimization design. Secondly, a demagnetization rate constraint is established based on the actual operating conditions of the motor, and a multi-objective optimization function is constructed based on this constraint and the motor performance objective. This breaks the bias of traditional optimization algorithms that prioritize performance over reliability, achieving a balance between efficiency, torque density, and demagnetization resistance. This ensures that the optimized motor meets the performance requirements under rated operating conditions and can adapt to long-term stable operation under complex conditions such as high temperature and short circuits. Within the preset parameter search space, a hierarchical guided optimization search mechanism is adopted for optimization. The hierarchical guided search logic takes into account both local fine exploration and global wide-area search, effectively improving population diversity and algorithm convergence efficiency, avoiding getting stuck in local optima during the optimization process, and enabling the rapid and accurate acquisition of the optimal structural parameters of the motor. Ultimately, robust optimization of the permanent magnet synchronous motor is achieved, significantly improving the motor's operational reliability and service life under complex working conditions.

[0025] Specifically, in step S1, an analytical calculation model of the permanent magnet synchronous motor under different operating conditions (such as high temperature and short circuit) is established. This process includes establishing a rapid demagnetization risk prediction model based on the state variables of the permanent magnet synchronous motor in a healthy state and the degree of demagnetization of the permanent magnet synchronous motor under a demagnetization fault.

[0026] Furthermore, the specific steps for establishing a rapid demagnetization risk prediction model are as follows: S110. Based on the state variables of the permanent magnet synchronous motor under healthy conditions, establish the voltage equation in the natural coordinate system. Specifically, when the permanent magnet is in a healthy state, establish the voltage equation of the permanent magnet synchronous motor in the natural coordinate system. The specific expression of this equation is as follows: (1); in, This represents the three-phase voltage of the motor under healthy conditions. This represents the three-phase current of the motor under healthy conditions. This represents the magnetic flux generated by the three-phase windings of a motor in a healthy state. Indicates stator resistance; Indicates different self-perceptions and mutual perceptions; Indicates the rotor position. Indicates time.

[0027] Furthermore, the magnetic flux generated by the three-phase windings of the motor in a healthy state The expression is: (2); in, Indicates a healthy permanent magnet flux linkage. This indicates the magnetic angle of the motor rotor relative to the stator windings. The symbol for pi.

[0028] Furthermore, based on formulas (1) and (2), the three-phase current of the motor under healthy conditions is obtained. The expression is: (3); S120. By transforming the voltage equation using Park coordinates, we obtain the voltage component equations in the rotated coordinate system, specifically expressed as follows: (4); in, Represents the d-axis and q-axis components of the stator voltage; Represents the d-axis and q-axis components of the stator current; Represents the d-axis and q-axis components of the stator flux linkage; This indicates the synchronous angular velocity.

[0029] Furthermore, the d-axis component of the stator flux linkage q-axis components The expression is: (5); in, and These represent the d-axis inductance and q-axis inductance of the permanent magnet synchronous motor, respectively.

[0030] S130. Based on the voltage component equation and equation (5), the electromagnetic torque equation of the permanent magnet synchronous motor is obtained, and its specific expression is as follows; (6); in, This represents the electromagnetic torque of a permanent magnet synchronous motor.

[0031] S140. In the mathematical model of a permanent magnet synchronous motor under demagnetization, after a uniform demagnetization fault occurs, the change trend of the air gap magnetic flux density amplitude of each magnetic pole is roughly the same, and the magnetic field lines and magnetic flux density remain uniformly distributed. Therefore, after a uniform demagnetization fault occurs in the permanent magnet synchronous motor, the new expression for the permanent magnet flux linkage is: (7); in, The permanent magnet flux linkage representing the effect of uniform demagnetization fault. Indicates the degree of demagnetization.

[0032] Furthermore, based on the electromagnetic torque equation (6) and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault, a rapid demagnetization risk prediction model is obtained. The expression for the rapid demagnetization risk prediction model is: ; in, This indicates the predicted risk value for rapid demagnetization. This indicates the degree of demagnetization in a permanent magnet synchronous motor under demagnetization fault conditions. This indicates the magnetic flux linkage of the permanent magnet under demagnetization fault. and These represent the d-axis and q-axis components of the stator current, respectively. and These represent the d-axis inductance and q-axis inductance of the permanent magnet synchronous motor, respectively.

[0033] Specifically, in step S2, the demagnetization risk of the permanent magnet synchronous motor under various operating conditions is quantitatively assessed. Specifically, a magnetic flux density-temperature coupling model is used to quantify the demagnetization risk under high-temperature conditions; an armature reaction magnetomotive force-short-circuit current model is used to quantify the demagnetization risk under short-circuit conditions; and the demagnetization risk under combined operating conditions (high temperature and short circuit) is also quantitatively assessed. Through the above quantitative assessments, corresponding demagnetization rate constraints are established. The specific steps are as follows: S210. Based on the operating temperature of the permanent magnet in a permanent magnet synchronous motor, establish a coercivity fitting model. Specifically, high temperatures can lead to increased coercivity in permanent magnets (such as neodymium iron boron). Nonlinear decay and the risk of demagnetization increase exponentially with increasing temperature. Therefore, a coercivity-temperature fitting model is first established. Based on experimental data or magnetic performance curves provided by the manufacturer, the coercivity fitting model expression obtained by quadratic polynomial fitting is as follows: ; in, This represents the residual coercivity at the corresponding temperature, and its unit is kA / m; Indicates the operating temperature of the permanent magnet (unit: °C); , and These represent the fitting coefficients, which were obtained by fitting the data at typical temperature points such as 25℃, 80℃, 120℃, and 150℃. The values ​​were obtained by calibration based on the measured values.

[0034] S220. Under high-temperature conditions, the core of demagnetization risk quantification is calculating the permanent magnet flux attenuation rate, which requires considering the coupling effect of armature reaction magnetomotive force and temperature under the actual load of the motor. Therefore, in this embodiment, based on the healthy permanent magnet flux and the permanent magnet operating temperature of the permanent magnet synchronous motor, the demagnetization rate of the permanent magnet flux under high-temperature conditions is calculated to obtain the high-temperature demagnetization rate. Among them, high-temperature demagnetization rate The expression is: ; in, This indicates a healthy permanent magnet flux at room temperature (25℃). Indicates high temperature Load current After demagnetization, the magnetic flux is reduced.

[0035] Furthermore, after demagnetization, the magnetic flux... The expression is: ; in, This indicates the standard temperature (25℃). Indicates the Curie temperature of the permanent magnet; Indicates the rated current of the motor; , This represents a correction factor, which can be calibrated through a high-temperature load test. In this embodiment, The value range is 0.3-0.8. The value range is 0.1-0.4.

[0036] The formula for calculating magnetic flux linkage after demagnetization reflects that the higher the temperature and the greater the load current, the more significant the magnetic flux linkage attenuation, i.e., the high-temperature demagnetization rate. The larger.

[0037] Furthermore, based on the high-temperature demagnetization rate The risk of high-temperature demagnetization is divided into three levels to provide a basis for subsequent dynamic constraints. The specific classification criteria are as follows: If high temperature demagnetization rate At that time, the risk level was classified as low risk, and the operating conditions were characterized by normal temperature and light load. ; like At that time, the risk level was classified as medium risk, and the operating conditions were characterized by medium temperature and medium load. ; If high temperature demagnetization rate At that time, the risk level was classified as high risk, and the operating conditions were characterized by high temperature and heavy load, such as... .

[0038] S230. Under short-circuit conditions, the armature reaction magnetomotive force increases sharply, which is the core factor leading to irreversible demagnetization of the permanent magnet. Therefore, in this embodiment of the invention, the peak value of the three-phase short-circuit current is calculated based on the equivalent circuit of the permanent magnet synchronous motor. The specific expression is: ; in, This represents the back electromotive force of the motor. Indicates stator resistance. This represents the direct-axis transient reactance.

[0039] S240. The demagnetization risk under short-circuit conditions is quantified by the magnetomotive force margin coefficient, which reflects the permanent magnet's ability to withstand short-circuit magnetomotive force. Therefore, based on the peak value of the three-phase short-circuit current and the number of series turns per phase of the stator of the permanent magnet synchronous motor, the peak value of the armature reaction magnetomotive force during a short circuit is calculated. Its specific expression is: ; in, Indicates the number of turns in series per phase of the stator; This represents the magnetic potential waveform correction coefficient, which is preferably 1.35 in this embodiment.

[0040] S250, Based on the coercivity fitting model and armature reaction magnetomotive force peak Calculate the magnetopotential margin coefficient The specific expression is: ; in, This indicates the thickness of the permanent magnet.

[0041] S260, Based on the magnetomotive force margin The short-circuit demagnetization rate of the permanent magnet flux linkage under short-circuit conditions is obtained. Specifically, if When , it indicates that the permanent magnet has no irreversible demagnetization; if 1 When the value is 0, it indicates that there is slight irreversible demagnetization (at which point the short-circuit demagnetization rate is 0). );like When this occurs, it indicates severe irreversible demagnetization (at which point the short-circuit demagnetization rate is...). ).

[0042] Furthermore, the magnetic potential margin coefficient With short-circuit demagnetization rate Establishing a mapping relationship facilitates the unification of indicator dimensions, specifically as follows: .

[0043] S270. In actual operating conditions, high temperature and short circuit often coexist, therefore a coupling demagnetization rate calculation model needs to be established. This embodiment uses a weighted summation method to integrate the effects of the two operating conditions. Specifically, based on the high temperature demagnetization rate and the short circuit demagnetization rate, a coupling demagnetization rate is established. Its expression is: ; in, and This represents the weighting coefficients under high temperature and short-circuit conditions, satisfying... .

[0044] Furthermore, the weighting coefficients for high-temperature and short-circuit conditions are dynamically adjusted according to the priority of the operating conditions, specifically as follows: If a short circuit is transient (e.g., duration <100ms) and a high temperature is steady-state (e.g.) If it lasts for more than 10 minutes, then take ; If the short circuit is a continuous short circuit (e.g., duration > 1s), and the high temperature is a medium temperature ( ), then take .

[0045] S280, based on high-temperature demagnetization rate Short-circuit demagnetization rate and coupling demagnetization rate This yields the demagnetization rate constraint conditions under different operating conditions. Specifically, the demagnetization rate ( , , (Integrated into a demagnetization risk index) As the core constraint index of the optimization algorithm, its specific expression is: .

[0046] Demagnetization risk index The maximum demagnetization rate under different operating conditions is taken to ensure coverage of the risks of the most severe operating conditions. For example, if a certain operating condition is not involved (such as only evaluating high-temperature operating conditions), then... In this embodiment of the invention, the demagnetization risk index is used as a constraint on the demagnetization rate under different operating conditions and is embedded in the implementation path of the optimization algorithm.

[0047] In this embodiment of the invention, an improved thornless bee optimization algorithm is used to iteratively optimize the mathematical model of a permanent magnet synchronous motor under the influence of demagnetization. The design variables include the number of winding turns, magnet width, magnet thickness, magnet included angle, pole arc coefficient, auxiliary slot size, magnetic bridge width, slot opening width, and stator tooth width; the optimization objectives include efficiency and torque density; and the constraints are the range of variation of the design variables and the constrained extreme values ​​of the optimization objectives.

[0048] Furthermore, the structural variables for optimizing the permanent magnet synchronous motor are expressed as follows: The various technical performance indicators of the motor are functions of these structural variables. These technical performance indicators are treated as constraints in the optimal design of the motor and can be expressed in the following form: ; in, Indicates the number of constraint functions.

[0049] Furthermore, the objective function in motor optimization design is expressed as a function of structural variables, with the following expression: .

[0050] Specifically, in step S3, the original optimization objectives of the permanent magnet synchronous motor include efficiency and torque density. In this embodiment, the demagnetization rate constraint (i.e., the demagnetization risk index) is included. This is embedded as a penalty term in the objective function, thus avoiding sacrificing demagnetization resistance for efficiency during optimization. Therefore, the expression of the modified multi-objective optimization function is: ; in, Represents a multi-objective optimization function. Indicates design variables (number of winding turns, magnet dimensions, etc.); , and This indicates the adjustment weight (e.g., 0.3, 0.4, 0.3), which can be adjusted according to design requirements; This indicates the efficiency of a permanent magnet synchronous motor; This indicates the target efficiency of the permanent magnet synchronous motor, such as 97%. This indicates the target torque density, such as 50 N·m / kg; Indicates torque density, Indicates the demagnetization risk index; and These represent the maximum allowable demagnetization risk index (e.g., 5%) and the minimum allowable demagnetization risk index (e.g., 0%), respectively. In this implementation, the target state of the permanent magnet synchronous motor includes the target efficiency and the target torque density.

[0051] Specifically, the optimization algorithm for stingless bees is inspired by the process of stingless bees building and regulating hive temperature. A mathematical model simulates three key behaviors of worker bees during hive construction: brood chamber construction, honeypot construction, and wax source foraging. These three behaviors correspond to the algorithm's mechanisms of utilization, exploration, and avoiding local optima. Through the synergistic effect of these three strategies, the algorithm can effectively balance global search and local exploitation in a complex search space, thereby finding the optimal solution. The specific steps of the stingless bee optimization algorithm are as follows: Step 1: Randomly generate an initial population of N individuals (bees) in a D-dimensional search space. Each individual represents a potential solution to the problem. Population initialization is performed within the predefined boundaries of the problem, with the upper and lower bounds defined by vectors ub and lb, respectively. The initial position of the i-th bee in the j-th dimension is... ; Generated by the following formula: ; in, This represents a random number between [0,1]. After initialization, the algorithm evaluates the fitness of each individual and designates the individual with the best fitness as the brood mother bee.

[0052] Step 2: Constructing Brood Chambers. This stage simulates the behavior of worker bees tightly constructing brood chambers in the central area of ​​the hive (brood zone), corresponding to the utilization phase of the algorithm, i.e., performing a fine-grained local search near the current optimal solution. In this stage, the algorithm generates new candidate solutions around the "brooding mother bee" (the current optimal solution). This process involves two strategies, with a probability p determining which one to execute, as follows: First local strategy (p<0.5): When the probability p is less than 0.5, a new candidate solution (brooder room) is proposed. It is a brood mother bee A bee randomly selected from an elite colony And another randomly selected bee It is generated by a combination of factors. Its position update formula is: ; in, Indicates a value based on the current temperature and maximum temperature The calculated temperature factor is used to adjust the search step size; It is a random number between -1 and 1. Temperature factor The calculation formula is: .

[0053] Current temperature It will increase with the number of iterations Linear decrease, current temperature The expression is: ; in, It represents the maximum number of iterations.

[0054] Second local strategy (p≥0.5): when probability When the value is greater than or equal to 0.5, a new candidate solution is obtained. From the brood mother bee And two other randomly selected bees , Together they are generated, and their expression is: ; in, It is a random number that follows a standard normal distribution (mean 0, standard deviation 1).

[0055] Step 3: Building Honeypots. This stage simulates the behavior of worker bees building honeypots far from the center of the hive to store nectar. This corresponds to the algorithm's exploration phase, which involves a broad global search across the entire search space. In this phase, the algorithm randomly selects a bee as the "honeypot queen" and generates new candidate solutions around it. This process also involves two strategies, selected by probability p, as follows: First global strategy (p<0.5): New candidate solutions From honeypot queen bees Current bees And another random bee Generate new candidate solutions through joint action The expression is: .

[0056] Second global strategy (p≥0.5): New candidate solutions From honeypot queen bees And two other random bees , Together they are generated, and the expression is: .

[0057] New candidate solutions are generated in these two stages. Then, the algorithm will perform a greedy selection. If the new solution... Its fitness is better than the current solution. If the current solution is not found, the new solution will replace it.

[0058] Step 4: Foraging for Wax Sources to Avoid Local Optimum. This stage simulates the behavior of some worker bees searching for new wax sources when there is insufficient wax in the hive. This corresponds to the mechanism in the algorithm to avoid getting trapped in local optima. The algorithm sets a trial counter `tria1_i` for each bee. If a bee's position does not improve after the brood chamber or honeypot construction stage, its corresponding `tria1_i` is incremented by one. When `tria1_i` exceeds a preset limit `imit`, the bee is considered "idle" and will be randomly re-initialized to a new position, just like when it goes out to find new wax sources. The position update formula is the same as the initialization formula: .

[0059] This mechanism helps increase the diversity of the population, enabling the algorithm to escape potential local optima traps and thus increasing the likelihood of finding the global optimum.

[0060] Specifically, in step S4, the hierarchical guided optimization search mechanism described in this embodiment of the invention improves upon the stingless bee optimization algorithm. Specifically, the hierarchical guided optimization search mechanism includes chaotic enhanced Latin hypercube initialization, dynamic population hierarchical-Levy flight exploration mechanism, elite pool guidance and adaptive temperature decay strategy, and directional mutation backtracking strategy, achieving improved optimization search performance through multi-strategy collaboration. In this embodiment of the invention, the hierarchical guided optimization search mechanism combines an analytical calculation model of a permanent magnet synchronous motor (PMSM) under multiple operating conditions and demagnetization effects. It uses key motor design parameters such as the number of winding turns and stator / rotor structural parameters as optimization variables, and efficiency, torque density, and demagnetization index as optimization objectives to perform multi-objective robustness optimization calculations for the PMSM, ultimately outputting optimal motor structural parameters that meet the balance requirements of performance and demagnetization resistance reliability.

[0061] Furthermore, based on the rapid demagnetization risk prediction model and the multi-objective optimization function, within the preset permanent magnet synchronous motor parameter search space, a hierarchical guided optimization search mechanism is used to search for the optimal structural parameters of the permanent magnet synchronous motor. The specific steps to obtain these parameters are as follows: S410. Generate an initial population within the preset-dimensional search space for permanent magnet synchronous motor parameters. Each individual in the initial population represents a potential solution to the problem. In this embodiment, the introduction of Chaotic Enhanced Latin Hypercube Initialization (CELHS) effectively solves the problem of uneven distribution of initial solutions. Furthermore, the uniformity of Latin Hypercube Sampling (LHS) and the randomness of the chaotic mapping can achieve the dual characteristics of global coverage and local dispersion of initial solutions. The specific steps for generating the initial population are as follows: S411. For each dimension of the permanent magnet synchronous motor parameter search space, divide it into multiple equidistant sub-intervals according to the upper and lower bounds; randomly select a sample point from each sub-interval to obtain multiple initial skeleton points.

[0062] For example, for a D-dimensional search space, the upper and lower bounds of each dimension j (j=1,2,...,D) are: , Divide it into N equidistant sub-intervals. Randomly select one sample point from each sub-interval to obtain N D-dimensional initial skeleton points, the expression of which is: .

[0063] S412. The initial solution generated by Latin hypercube sampling is subjected to chaotic perturbation to break the correlation between dimensions, further improve population diversity, and avoid initial clustering. Therefore, in this embodiment, a chaotic sequence is generated, and the chaotic sequence is used to perturb multiple initial skeleton points to obtain the initial population. The expression for generating the Logistic chaotic sequence is: ; in, Let z_0 represent the chaos parameter, preferably 4; z_0∈(0,1), and not 0.25, 0.5, 0.75. The j-th dimension of each skeleton point is perturbed, and the specific expression is as follows: ; in, The first chaotic sequence represents the... The perturbation amplitude is controlled within 1 / 2 of the sub-interval length to ensure that the solution remains within the effective search space.

[0064] S420. The individuals in the initial population generated in step S412 are screened to obtain the core guiding individual. A local search is performed centered on the core guiding individual to generate new candidate solutions, and a local new solution generation strategy is determined. In this embodiment, the local new solution generation strategy includes a first local strategy and a second local strategy.

[0065] Furthermore, the first local strategy is as follows: new candidate solutions include core guiding individuals, randomly selected elite individuals, and random individuals. The generation process introduces a temperature factor to adjust the search step size, and the temperature factor changes according to the current iteration number.

[0066] It should be noted that step S420 is similar to the brooding room construction logic in step 2 above, and will not be repeated here.

[0067] S430. In step 3 above, the search only revolves around random honeypot queen bees, resulting in a limited exploration scope and a tendency for population diversity to decline with iteration. To address this issue, this embodiment of the invention introduces dynamic population stratification and Levy flight, dividing the population into functional zones to achieve collaborative search for core utilization, mid-range exploration, and long-range novelty expansion. This embodiment of the invention randomly selects individuals from the initial population as exploration guide individuals, uses these guide individuals as the center for global search to generate new candidate solutions, and determines the specific steps for the global new solution generation strategy as follows: S431. Divide the initial population into 3 layers, sort them by fitness, and dynamically allocate the proportions (higher proportion in the early iteration layer, higher proportion in the later iteration layer), as follows: Core layer (percentage) ): Before fitness The brooding mother bee from step 2 (corresponding to the core guiding individual in step S420) is utilized locally, inheriting the original brooding room construction logic.

[0068] Exploration layer (percentage) ): Fitness Intermediate The original honeypot construction logic was improved by conducting mid-range exploration around the honeypot queen bee in step 3 (corresponding to the exploration guide individual in step S430).

[0069] New Layer (Percentage) Fitness end Long-range exploration can be conducted via levy flight, expanding the search area. , It increases linearly from 0.2 to 0.5. (It decreases linearly from 0.5 to 0.2).

[0070] Levy flight expansion: Utilizing the short-step local movement and long-step long-range jump characteristics of levy flight, it provides individuals in the expansion layer with long-distance search capabilities, avoiding the search range being limited to the current area.

[0071] Dynamic stratification triggering: Every 5 generations, the population fitness is recalculated, and stratification proportions are allocated after sorting by fitness in descending order, such as: , , .

[0072] Furthermore, the update strategies for each layer are as follows: Core layer: The original brooding room construction strategy is used, but elite weighting is introduced, that is, the brooding mother bees are replaced by the elite pool Top-1, where the elite pool consists of all the core layer bees; Exploration Layer: Improve the original honeypot construction strategy. Honeypot queen bees are randomly selected from the exploration layer. The update formula incorporates in-layer fitness weights, i.e., the weighted average fitness of the exploration layer. ; New Layer: The position is updated using Levy flight, with the specific formula as follows: ; in, This represents the levy flight characteristic parameter, preferably 1.5; , Represents a random number that follows a standard normal distribution; This represents the j-th dimension of the current globally optimal solution, ensuring that the new direction does not deviate from the high-quality region.

[0073] S432. Utilizing an elite pool for guidance and adaptive temperature decay, new solutions are rapidly moved towards high-quality regions, while dynamically balancing exploration and utilization. The specific steps are as follows: S4321, Elite Pool Construction and Weighting. Specifically, the elite pool is updated in each iteration (retaining the current top-5 individuals in terms of fitness), and the weights of individuals in the elite pool are calculated. The calculation formula is: ; in, This represents the fitness of the k-th elite. This represents the optimal fitness of the elite pool. Weight The higher the level of elites, the greater their influence on the new solution.

[0074] S4322, Elite-Guided New Solution Generation. Specifically, in the new solution generation formulas for the core layer and exploration layer, the original single... or Replace with the weighted average of the elite pool Its specific expression is: .

[0075] Furthermore, the first local policy update formula of the core layer is changed to: .

[0076] S4323, Adaptive temperature decay. Specifically, temperature factor. The computation depends on the population convergence. ,and Defined as the ratio of the standard deviation of population fitness to the initial standard deviation, its expression is: .

[0077] Furthermore, the temperature update formula is: .

[0078] Furthermore, temperature factor The original logic will still be used, but based on new... Calculations are performed to ensure that the step size dynamically adapts to the population state.

[0079] This invention integrates high-quality information from multiple sources in an elite pool, and when generating a new solution, it directly approaches the intersection of multiple high-quality regions, reducing invalid searches. The adaptive temperature decay avoids the "one-size-fits-all" problem of linear cooling, retains exploration ability when the population is dispersed, and converges quickly when the population is concentrated, significantly improving convergence efficiency.

[0080] S434. Perform multi-index local optimum detection, directional mutation backtracking mechanism, and precise escape from local optima. Specific steps are as follows: S4341. Implement a targeted mutation backtracking strategy. Specifically, for individuals (experimental counters) trapped in local optima... When needed, update the position using the following formula: ; in, The j-th dimension represents the historical optimal solution and is used to retain high-quality information; The j-th dimension represents the randomly selected nonlocally optimal individual, used to introduce the dispersion direction; express Random numbers that follow a uniform distribution are used to control the amplitude of variation; This represents the global disturbance term, used to ensure that the local region is escaped.

[0081] S4341. Simulating non-blind foraging by stingless bees. Specifically, when the wax source in the hive is insufficient (i.e., a local optimum occurs), worker bees will not fly randomly, but will search for new wax sources based on past foraging memories (historical optimal solutions) and information from their companions (random individual directions), which is consistent with the intelligent decision-making behavior of the bee colony.

[0082] S440. Using the multi-objective optimization function as the optimization objective, repeat the above steps to iteratively update the local new solution generation strategy and the global new solution generation strategy until the iteration termination condition is met, and obtain the optimal structural parameters of the permanent magnet synchronous motor.

[0083] This invention first constructs an analytical calculation model for a permanent magnet synchronous motor (PMSM) considering the demagnetization effects under multiple operating conditions, such as high temperature and short circuits. This model can dynamically adapt to demagnetization risk constraints according to different operating conditions, effectively improving the targeting and computational efficiency of the optimization process. Secondly, the basic stingless bee optimization algorithm is improved by innovatively introducing chaotic enhanced Latin hypercube initialization, dynamic population hierarchical-Levy flight exploration mechanism, elite pool guidance and adaptive temperature decay strategy, and directional mutation backtracking strategy, forming an improved stingless bee optimization algorithm. This algorithm is then used to perform optimization calculations on the analytical calculation model of the PMSM under the multi-condition demagnetization effects. Furthermore, this method quantifies demagnetization risk into a quantitative indicator that can be directly used in optimization calculations, and uses it together with efficiency and torque density as core optimization objectives. By improving the stingless bee optimization algorithm, it not only effectively solves the problem of uneven distribution of initial solutions in traditional algorithms, but also significantly improves population diversity, expands the effective search range, guides new solutions to converge quickly to the high-quality solution region, greatly improves the algorithm's convergence efficiency, and helps the optimization process escape the local optimum dilemma. Ultimately, it achieves precise optimization of key motor parameters and constructs a dual-objective optimization system that takes into account both motor performance and demagnetization resistance reliability.

[0084] The method described in this embodiment of the invention, through the above design, has the following significant advantages: First, it breaks through the limitations of traditional empirical demagnetization protection by constructing a quantitative model of demagnetization risk under high temperature, short circuit, and combined operating conditions. The high-temperature demagnetization rate is calculated using a coercivity model, and the short-circuit demagnetization risk coefficient is derived based on the armature reaction magnetomotive force and short-circuit current model. A weighted summation method is then used to integrate the effects of combined operating conditions, ultimately consolidating the demagnetization risk into a calculable demagnetization risk index. This quantitative approach can dynamically match different operating conditions, such as the high-temperature heavy load of new energy vehicle drive motors and the short-circuit faults of industrial servo motors, solving the problem of the disconnect between demagnetization constraints and actual operating conditions in traditional designs. It ensures that the motor can still avoid irreversible demagnetization under extreme conditions, significantly improving operational reliability.

[0085] Secondly, the method described in this embodiment of the invention addresses the shortcomings of traditional optimization algorithms by making four improvements to the stingless bee optimization algorithm, comprehensively optimizing its optimization performance. By combining the uniformity of Latin hypercube sampling and the randomness of chaotic mapping, global coverage and local dispersion of initial solutions are achieved, solving the problems of uneven initial solution distribution and local clustering, laying a solid foundation for subsequent optimization. Dynamic population stratification and the Levy flight exploration mechanism effectively improve population diversity and avoid getting trapped in local optima. Elite pool guidance and adaptive temperature decay mechanism improve convergence efficiency. The directed mutation backtracking strategy targets individuals trapped in local optima, combining historical best solutions and random individual directions for directed mutation, rather than blindly re-initializing, ensuring rapid escape from local optima, thereby improving the algorithm's optimization capability and efficiency.

[0086] Example 2: Based on the same inventive concept, this embodiment provides a robust optimization system for permanent magnet synchronous motors. The principle of solving the problem is similar to that of the robust optimization method for permanent magnet synchronous motors provided in Embodiment 1, and the repeated parts will not be described again.

[0087] Reference Figure 2 As shown, this embodiment provides a robust optimization system for a permanent magnet synchronous motor, including: The model building module is used to establish a rapid demagnetization risk prediction model based on the state variables of the permanent magnet synchronous motor under healthy conditions and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault. The constraint generation module is used to establish demagnetization rate constraints based on the operating conditions of the permanent magnet synchronous motor. The objective optimization module is used to construct a multi-objective optimization function based on the constraints and the target state of the permanent magnet synchronous motor. The search module is used to search within the preset permanent magnet synchronous motor parameter search space using a hierarchical guided optimization search mechanism based on the rapid demagnetization risk prediction model and multi-objective optimization function, in order to obtain the optimal structural parameters of the permanent magnet synchronous motor. The execution module is used to perform robust optimization of the permanent magnet synchronous motor based on the optimal structural parameters.

[0088] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0089] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0090] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0091] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0092] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A robust optimization method for a permanent magnet synchronous motor, characterized in that, include: S1. Based on the state variables of the permanent magnet synchronous motor under healthy conditions and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault, establish a rapid demagnetization risk prediction model; S2. Based on the operating conditions of the permanent magnet synchronous motor, establish demagnetization rate constraints. S3. Construct a multi-objective optimization function based on the constraints and the target state of the permanent magnet synchronous motor; S4. Based on the rapid demagnetization risk prediction model and the multi-objective optimization function, within the preset permanent magnet synchronous motor parameter search space, a hierarchical guided optimization search mechanism is used to search for the optimal structural parameters of the permanent magnet synchronous motor; based on the optimal structural parameters, the robustness optimization of the permanent magnet synchronous motor is completed.

2. The robustness optimization method for a permanent magnet synchronous motor according to claim 1, characterized in that, The steps in S1, which involve establishing a rapid demagnetization risk prediction model based on the state variables of the permanent magnet synchronous motor in a healthy state and the degree of demagnetization of the permanent magnet synchronous motor under a demagnetization fault, are as follows: Based on the state variables of the permanent magnet synchronous motor under healthy conditions, establish the voltage equation in the natural coordinate system; By performing coordinate transformation on the voltage equation, we obtain the voltage component equations in a rotating coordinate system. Based on the voltage component equation, the electromagnetic torque equation of the permanent magnet synchronous motor is obtained. Based on the electromagnetic torque equation and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault, a rapid demagnetization risk prediction model is obtained.

3. A robust optimization method for a permanent magnet synchronous motor according to claim 1 or 2, characterized in that, The expression for the rapid demagnetization risk prediction model is: ； in, This indicates the predicted risk value for rapid demagnetization. This indicates the degree of demagnetization in a permanent magnet synchronous motor under demagnetization fault conditions. This indicates the magnetic flux linkage of the permanent magnet under demagnetization fault. and These represent the d-axis and q-axis components of the stator current, respectively. and These represent the d-axis inductance and q-axis inductance of the permanent magnet synchronous motor, respectively.

4. The robustness optimization method for a permanent magnet synchronous motor according to claim 1, characterized in that, The step S2, establishing the demagnetization rate constraint condition based on the operating conditions of the permanent magnet synchronous motor, is as follows: A coercivity fitting model is established based on the operating temperature of the permanent magnet of the permanent magnet synchronous motor. Based on the healthy permanent magnet flux linkage of the permanent magnet synchronous motor and the operating temperature of the permanent magnet, the demagnetization rate of the permanent magnet flux linkage under high temperature conditions is calculated to obtain the high temperature demagnetization rate. Calculate the peak value of the three-phase short-circuit current based on the equivalent circuit of the permanent magnet synchronous motor. Based on the peak value of the three-phase short-circuit current and the number of series turns per phase of the stator of the permanent magnet synchronous motor, calculate the peak value of the armature reaction magnetomotive force of the permanent magnet synchronous motor during a short circuit; The magnetomotive force margin coefficient is calculated based on the coercivity fitting model and the peak value of the armature reaction magnetomotive force. Based on the magnetomotive force margin coefficient, the short-circuit demagnetization rate of the permanent magnet flux under short-circuit conditions is obtained. Based on the high-temperature demagnetization rate and the short-circuit demagnetization rate, a coupling demagnetization rate is established; Based on the high-temperature demagnetization rate, the short-circuit demagnetization rate, and the coupling demagnetization rate, the demagnetization rate constraint conditions under different operating conditions are obtained.

5. The robustness optimization method for a permanent magnet synchronous motor according to claim 1, characterized in that, In step S3, the expression for the multi-objective optimization function is: ； in, Represents a multi-objective optimization function. , and Indicates the adjusted weights; This indicates the efficiency of a permanent magnet synchronous motor. This represents the target efficiency of the permanent magnet synchronous motor. Indicates the target torque density. Indicates torque density, Indicates the demagnetization risk index; and These represent the maximum and minimum permissible demagnetization risk indices, respectively.

6. The robustness optimization method for a permanent magnet synchronous motor according to claim 1, characterized in that, In step S4, based on the rapid demagnetization risk prediction model and the multi-objective optimization function, the optimal structural parameters of the permanent magnet synchronous motor are obtained by searching within the preset permanent magnet synchronous motor parameter search space using a hierarchical guided optimization search mechanism. An initial population is generated within a preset dimension of the permanent magnet synchronous motor parameter search space; each individual in the initial population is screened to obtain the core guiding individual; Using the core guided individual as the center, a local search is performed to generate new candidate solutions, and a local new solution generation strategy is determined; Individuals are randomly selected from the initial population as exploration guide individuals. A global search is performed with the exploration guide individuals as the center to generate new candidate solutions, and a global new solution generation strategy is determined. Using the multi-objective optimization function as the optimization objective, the local new solution generation strategy and the global new solution generation strategy are iteratively updated to obtain the optimal structural parameters of the permanent magnet synchronous motor.

7. The robustness optimization method for a permanent magnet synchronous motor according to claim 6, characterized in that, The steps for generating the initial population within the pre-defined search space for permanent magnet synchronous motor parameters are as follows: For each dimension of the parameter search space of the permanent magnet synchronous motor, it is divided into multiple equidistant sub-intervals according to the upper and lower bounds; a sample point is randomly selected from each of the sub-intervals to obtain multiple initial skeleton points; A chaotic sequence is generated, and the chaotic sequence is used to perturb the multiple initial skeleton points to obtain an initial population.

8. The robustness optimization method for a permanent magnet synchronous motor according to claim 6, characterized in that, The local new solution generation strategy includes a first local strategy; the first local strategy is: the new candidate solution includes the core guiding individual, randomly selected elite individuals and random individuals, and the generation process introduces a temperature factor to adjust the search step size, the temperature factor is changed according to the current iteration number.

9. The robustness optimization method for a permanent magnet synchronous motor according to claim 6, characterized in that, The process of randomly selecting individuals from the initial population as exploration guide individuals and generating new candidate solutions through a global search centered on the exploration guide individuals includes dividing the initial population and sorting it according to fitness.

10. A robust optimization system for a permanent magnet synchronous motor, characterized in that, include: The model building module is used to establish a rapid demagnetization risk prediction model based on the state variables of the permanent magnet synchronous motor in a healthy state and the degree of demagnetization of the permanent magnet synchronous motor under demagnetization fault. The constraint generation module is used to establish demagnetization rate constraints based on the operating conditions of the permanent magnet synchronous motor. The objective optimization module is used to construct a multi-objective optimization function based on the constraints and the target state of the permanent magnet synchronous motor. The search module is used to search within the preset permanent magnet synchronous motor parameter search space using a hierarchical guided optimization search mechanism based on the rapid demagnetization risk prediction model and the multi-objective optimization function, so as to obtain the optimal structural parameters of the permanent magnet synchronous motor. The execution module is used to perform robust optimization of the permanent magnet synchronous motor based on the optimal structural parameters.

Images