Multi-objective parameter intelligent optimization method and system for axial magnetic field permanent magnet motor

Abstract

This invention discloses a multi-objective parameter intelligent optimization method and system for axial magnetic field permanent magnet motors. The method includes generating a population based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor; generating a set of candidate design schemes based on the current population using an optimizer; calling a simulation evaluator to perform boundary clipping, integer variable rounding, and geometric constraint prediction; performing electromagnetic simulation and calculating the objective vector and total constraint violation degree of the candidate design schemes; backfilling the data to the NSGA-II algorithm; performing non-dominated sorting of the parent and child populations of the current population based on the Deb constraint dominance relationship and calculating the crowding distance within the same non-dominated layer; selecting individuals from the population according to the principle of prioritizing non-dominated levels and having larger crowding distances to obtain an updated population; and finally outputting the optimal candidate design scheme. This invention aims to reduce the proportion of invalid simulations and improve the optimization efficiency of multi-objective parameter optimization for axial magnetic field permanent magnet motor design.

Description

Technical Field

[0001] This invention relates to the field of motor design and intelligent optimization technology, specifically to a multi-objective parameter intelligent optimization method and system for an axial magnetic field permanent magnet motor. Background Technology

[0002] Axial flux permanent magnet motors (AFPMs) are characterized by their flat structure and high power density, but their performance is highly sensitive to multiple parameters, including stator and rotor geometry, slot tooth parameters, magnet dimensions, and mechanical structure. Traditional optimization methods for AFPMs involve manual parameter tuning / single-variable scanning. However, this experience-based approach struggles to achieve a global trade-off between multiple objectives such as weight, power density, and efficiency. Furthermore, high-precision electromagnetic calculations typically rely on simulation software, resulting in lengthy simulation runs. Direct searching across a wide variable domain can easily generate numerous geometric interferences, mesh failures, or solution failures, leading to low optimization efficiency and poor reproducibility. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a multi-objective parameter intelligent optimization method and system for axial magnetic field permanent magnet motors, which addresses the above-mentioned problems in the prior art. This invention aims to reduce the proportion of invalid simulations and improve the optimization efficiency of multi-objective parameters in the design of axial magnetic field permanent magnet motors.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A multi-objective parameter intelligent optimization method for an axial magnetic field permanent magnet motor includes the following steps: S101, an initial population is generated based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor, and the population is a set of candidate design schemes; S102, the NSGA-II algorithm optimizer generates a set of candidate design schemes based on the current population; S103 uses the NSGA-II algorithm optimizer to call the simulation evaluator to perform boundary trimming, integer rounding, and geometric constraint prediction on the candidate design scheme set. It then performs electromagnetic simulation on the candidate design schemes that pass the geometric constraint prediction and calculates the target vector and total constraint violation degree of the candidate design schemes. S104, the target vector and total constraint violation of the candidate design scheme obtained from electromagnetic simulation are backfilled into the NSGA-II algorithm through the NSGA-II algorithm optimizer; S105. Based on the Deb constraint dominance relationship, perform non-dominated sorting on the parent and offspring populations of the current population, calculate the crowding distance within the same non-dominated layer, and select individuals in the current population according to the principle of priority of non-dominated level and larger crowding distance to obtain the updated population. S106, determine whether the preset number of iterations has been reached. If the preset number of iterations has been reached, output the optimal candidate design scheme; otherwise, jump to step S102 to continue iterating.

[0005] Optionally, when generating the initial population in step S101 based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor, the design variables include part or all of the following: stator outer diameter, stator inner diameter, rotor outer diameter, shaft diameter, shaft hole diameter, stator axial length, rotor axial length, magnet length, winding inner extension length, winding outer extension length, slot depth, magnet thickness, magnet embedment depth, wrapping thickness, tooth tip depth, slot top width, slot bottom width, slot opening width, fillet radius, tooth tip angle, magnet arc angle, number of circumferential segments, and air gap.

[0006] Optionally, the functional expression for the geometric constraint prediction in step S103 is: ; in, Candidate design schemes Geometric constraints, A zero matrix, with geometric constraints less than or equal to a zero matrix. This indicates that the prediction is made based on geometric constraints, and the functional expression of the geometric constraints is: ; in, , , , and These are the five components of the geometric constraint. The superscript T denotes the transpose operation. The functional expressions for each component of the geometric constraint are as follows: ; ; ; ; ; ; in, For the width of the slot, The width of the top of the groove, Given the geometric margin, Where is the fillet radius, and min is the minimum value. The width of the trough bottom, Given the minimum allowable thickness of the tooth tip, For the depth of the groove, Given the available radial thickness, max is the maximum value. Stator outer diameter This is the inner diameter of the stator.

[0007] Optionally, in step S103, when performing electromagnetic simulation on the candidate design schemes predicted by geometric constraints and calculating the target vector and total constraint violation degree of the candidate design schemes, the calculation function expressions for the target vector and total constraint violation degree of the candidate design schemes are as follows:

[0008] ; in, Candidate design schemes The target vector is denoted by min, where min represents the minimum value. Candidate design schemes Total mass, Candidate design schemes power density, Candidate design schemes System efficiency; Candidate design schemes Total constraint violation The value must be greater than or equal to 0. A value of 0 indicates that the engineering constraint judgment has passed, while a value of 0 indicates that the engineering constraint judgment has failed. `max` represents the maximum value. This is the k-th constraint in terms of total constraint violation degree.

[0009] Optionally, the candidate design scheme The expression for calculating the power density is: ; in, Candidate design schemes power density, Candidate design schemes obtained from simulation calculations 'output power' Candidate design schemes The total mass; the constraints of the total constraint violation degree include engineering constraints. and failure penalty constraints The engineering constraints and failure penalty constraints The function expression is: ; ; ; in, This is the lower limit threshold for output power. Candidate design schemes 'output power' express The absolute value; This is the failure indicator function for electromagnetic simulation. The failure indicator function takes the value 1 or 0. A value of 1 indicates that the electromagnetic simulation has failed, and a value of 0 indicates that the electromagnetic simulation has succeeded. This is the penalty coefficient, and the penalty coefficient takes a value greater than 0.

[0010] Optionally, in step S105, when performing non-dominated sorting of the parent and offspring populations of the current population based on the Deb constraint dominance relationship, any two candidate design schemes The conditions for satisfying the Deb constraint dominance relationship include any of the following conditions: Condition (1): =0 and Greater than 0; Condition (2): Greater than 0 Greater than 0 and Less than Condition (3): and All are 0, candidate design schemes Target vector Non-inferior to candidate designs across all objectives Target vector And it is superior to the candidate design in at least one objective. Target vector ;in, Candidate design schemes Total constraint violation Candidate design schemes The total constraint violation rate; the calculation function expression for the congestion distance is: ; in, The crowding distance is the distance between the i-th individual and the i-th individual within the same non-dominated layer. Index the target dimension; and These are the target values ​​of individuals adjacent to individual i after sorting by the m-th target; and These are the maximum and minimum values ​​of the non-dominated layer on the m-th objective, respectively, and each individual is a single candidate design scheme.

[0011] Optionally, step S106 outputting the optimal candidate design scheme means selecting feasible non-dominated solutions from all historical samples of candidate design schemes to form a Pareto set output. The feasible non-dominated solution refers to a candidate design scheme with a total constraint violation of 0 and not associated with any other candidate design scheme with a total constraint violation of 0. Candidate design schemes that satisfy the Deb constraint dominance relationship .

[0012] The present invention also provides a multi-objective parameter intelligent optimization system for an axial magnetic field permanent magnet motor, comprising a microprocessor and a memory interconnected thereto, wherein the microprocessor is programmed or configured to execute the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor.

[0013] The present invention also provides a computer-readable storage medium storing a computer program or instructions that are programmed or configured to execute, via a processor, the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor.

[0014] The present invention also provides a computer program product, including a computer program or instructions, which are programmed or configured to execute, via a processor, the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor.

[0015] Compared with existing technologies, this invention mainly achieves the following beneficial effects: The method of this invention includes generating a population based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor; generating a set of candidate design schemes based on the current population using the NSGA-II algorithm optimizer; calling the simulation evaluator to perform boundary trimming, integer variable rounding, and geometric constraint prediction; performing electromagnetic simulation and calculating the target vector and total constraint violation degree of the candidate design schemes; backfilling the data into the NSGA-II algorithm; performing non-dominated sorting on the parent and child populations of the current population based on the Deb constraint dominance relationship and calculating the crowding distance within the same non-dominated layer; selecting individuals from the current population according to the principle of prioritizing non-dominated levels and having larger crowding distances to obtain an updated population; and finally outputting the optimal candidate design scheme. This invention aims to reduce the proportion of invalid simulations and improve the optimization efficiency of multi-objective parameter optimization for axial magnetic field permanent magnet motor design. This invention can unify and integrate variable domain constraint processing, electromagnetic simulation evaluation, and constrained multi-objective evolutionary search into a unified process, which can reduce the proportion of invalid simulations and improve the intelligent optimization efficiency of multi-objective parameters for axial magnetic field permanent magnet motors. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the basic process of the method in an embodiment of the present invention.

[0017] Figure 2This is a side view of the axial magnetic field permanent magnet motor in an embodiment of the present invention.

[0018] Figure 3 This is a schematic cross-sectional view of the stator structure of the axial magnetic field permanent magnet motor in an embodiment of the present invention.

[0019] Figure 4 This is a schematic diagram of the slot cross-sectional structure of the axial magnetic field permanent magnet motor in an embodiment of the present invention.

[0020] Figure 5 This is a schematic diagram of the shaft structure of the axial magnetic field permanent magnet motor in an embodiment of the present invention.

[0021] Figure 6 This is a schematic diagram showing the results of optimizing the Pareto front for the three objectives of weight, power density, and efficiency in an embodiment of the present invention.

[0022] Figure 7 This is a circular heatmap showing the grouped correlation coefficients of the design variables in this embodiment of the invention.

[0023] Legend: P1, Stator outer diameter; P2, Stator inner diameter; P3, Rotor outer diameter; P4, Shaft diameter; P5, Shaft hole diameter; P6, Stator axial length; P7, Rotor axial length; P8, Magnet length; P9, Winding inner extension length; P10, Winding outer extension length; P11, Slot depth; P12, Magnet thickness; P13, Magnet embedment depth; P14, Wrapping tape thickness; P15, Tooth tip depth; P16, Slot top width; P17, Slot bottom width; P18, Slot opening width; P19, Corner radius; P20, Tooth tip angle; P21, Magnet arc angle; P22, Number of circumferential segments; P23, Air gap. Detailed Implementation

[0024] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be further described in detail below with reference to the accompanying drawings in the embodiments of the present invention.

[0025] like Figure 1 As shown, the intelligent optimization method for multi-objective parameters of the axial magnetic field permanent magnet motor in this embodiment includes the following steps: S101, an initial population is generated based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor, and the population is a set of candidate design schemes; S102, the NSGA-II algorithm optimizer generates a set of candidate design schemes based on the current population; S103 uses the NSGA-II algorithm optimizer to call the simulation evaluator to perform boundary trimming, integer rounding, and geometric constraint prediction on the candidate design scheme set. It then performs electromagnetic simulation on the candidate design schemes that pass the geometric constraint prediction and calculates the target vector and total constraint violation degree of the candidate design schemes. S104, the target vector and total constraint violation of the candidate design scheme obtained from electromagnetic simulation are backfilled into the NSGA-II algorithm through the NSGA-II algorithm optimizer; S105. Based on the Deb constraint dominance relationship, perform non-dominated sorting on the parent and offspring populations of the current population, calculate the crowding distance within the same non-dominated layer, and select individuals in the current population according to the principle of priority of non-dominated level and larger crowding distance to obtain the updated population. S106, determine whether the preset number of iterations has been reached. If the preset number of iterations has been reached, output the optimal candidate design scheme; otherwise, jump to step S102 to continue iterating.

[0026] Figures 2-5 This is a schematic diagram of an axial magnetic field permanent magnet motor, where a is the rotor, b is the stator, c is the slot, and d is the shaft. Figure 2 This is a side view schematic diagram of an axial magnetic field permanent magnet motor. Figure 3 This is a schematic cross-sectional view of the stator structure of an axial magnetic field permanent magnet motor. Figure 4 This is a schematic diagram of the slot cross-sectional structure of an axial magnetic field permanent magnet motor. Figure 5 This is a schematic diagram of the shaft structure of an axial magnetic field permanent magnet motor. In step S101 of this embodiment, when generating the initial population based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor, the design variables include the stator outer diameter P1, stator inner diameter P2, rotor outer diameter P3, shaft diameter P4, shaft hole diameter P5, stator axial length P6, rotor axial length P7, magnet length P8, winding inner extension length P9, winding outer extension length P10, slot depth P11, magnet thickness P12, magnet embedment depth P13, wrapping thickness P14, tooth tip depth P15, slot top width P16, slot bottom width P17, slot opening width P18, fillet radius P19, tooth tip angle P20, magnet arc angle P21, number of circumferential segments P22, and air gap P23. For details, please refer to [link to relevant documentation]. Figures 2-5 In step S101 of this embodiment, when generating the initial population based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor, the candidate design schemes formed by the design variables of the axial magnetic field permanent magnet motor are defined as follows: ; in, As candidate design schemes, ~ These are the design variables listed as 1 to 23 above. The design variables are grouped according to their engineering meaning into geometric dimensions (8 items), tooth parameters (7 items), magnet parameters (4 items), and other parameters (4 items). The value range constraints for the given candidate design schemes are: ; in, and These represent the lower and upper boundaries of the vectors corresponding to the candidate design schemes, respectively. This embodiment also includes setting an integer subset. (At least including the number of circumferential segments) satisfies The number of slots and poles is fixed within a single optimization task, and the windings are written manually using the winding table.

[0027] The functional expression for the geometric constraint prediction in step S103 of this embodiment is: ; in, Candidate design schemes Geometric constraints, A zero matrix, with geometric constraints less than or equal to a zero matrix. This indicates that the prediction is made based on geometric constraints, and the functional expression of the geometric constraints is: ; in, , , , and These are the five components of the geometric constraint. The superscript T represents the transpose operation, which occurs when any component exists. When a candidate design is deemed geometrically infeasible, it is also treated as geometrically infeasible and directly enters the penalty / failure marking process. This reduces the unnecessary simulation overhead caused by geometric interference, mesh failure, or solution failure. Specifically, the functional expressions of the components of the five geometric constraints in this embodiment are as follows: ; ; ; ; ; ; in, The slot width is P18. The top width of the slot is P16. Given the geometric margin, Let P19 be the fillet radius, and min be the minimum value. The bottom width of the groove is P17. Given the minimum allowable thickness of the tooth tip, The groove depth is P11. Given the available radial thickness, max is the maximum value. The stator outer diameter is P1. P2 is the inner diameter of the stator.

[0028] In this embodiment, steps S102 to S104 involve decoupling the NSGA-II algorithm optimizer and the simulation evaluator. In practice, the NSGA-II algorithm optimizer iteratively executes steps S102 to S104 in batches, which can be represented as: Step S102 is the ask operation of the NSGA-II algorithm optimizer. , can be represented as: ; in, This is a set of candidate design schemes generated based on the current population. Step S103 is the evaluate operation of the NSGA-II algorithm optimizer, which includes boundary clipping, integer variable rounding, geometric constraint prediction, and calling the simulation evaluator. , can be represented as: ; in, and Let be the target vector and the vector of total constraint violation of the candidate design schemes in round t, where round t is the current round and round t+1 is the next round; Step S104 is the tell operation of the NSGA-II algorithm optimizer, which is used to backfill the target vector and total constraint violation of the candidate design scheme obtained from electromagnetic simulation into the NSGA-II algorithm, and can be expressed as: .

[0029] In this embodiment, the simulation evaluator For any candidate design scheme First, determine whether parallel evaluation is enabled: if enabled, then select candidate design schemes. Distribute to multiple isolated working directory instances for parallel execution; otherwise, execute sequentially. The evaluation process includes: selecting candidate design schemes. Write variables into the program by name; perform electromagnetic calculations and simulations under fixed operating conditions (speed, bus voltage, current definition method, phase extraction angle, etc.); read the output: total mass. System efficiency Output power The success of the simulation is determined based on whether the simulation was successful: if writing parameters fails, meshing / solving fails, readings fail, or the geometry is deemed infeasible in step two, the failure indicator function is set to [function name missing]. Record the original log / raw output for review purposes; otherwise, set it to... Then, proceed to the assembly of target and engineering constraints.

[0030] In step S103, when performing electromagnetic simulation on the candidate design schemes predicted by geometric constraints and calculating the target vector and total constraint violation degree of the candidate design schemes, the calculation function expressions for the target vector and total constraint violation degree of the candidate design schemes are as follows:

[0031] ; in, Candidate design schemes The target vector is denoted by min, where min represents the minimum value. Candidate design schemes Total mass, Candidate design schemes power density, Candidate design schemes System efficiency; Candidate design schemes Total constraint violation The value must be greater than or equal to 0. A value of 0 indicates that the engineering constraint judgment has passed, while a value of 0 indicates that the engineering constraint judgment has failed. `max` represents the maximum value. This represents the k-th constraint with the highest degree of constraint violation. Based on the above candidate design schemes... Target vector It can achieve weight minimization, power density maximization, and efficiency maximization within a unified minimization framework (through the... and Taking the negative sign achieves an equivalent transformation to maximize the objective. In engineering implementation, different objective combination modes can be configured as needed (such as optimizing only weight / efficiency), but all are input into the NSGA-II algorithm in the form of a unified minimized objective vector. This is equivalent to a candidate design scheme simultaneously satisfying both geometric and engineering constraints and being successfully evaluated; if any constraint is violated or any failure occurs, then... This is to allow for a priority comparison of the feasibility of subsequent Deb constraint dominance relationships.

[0032] In this embodiment, the candidate design scheme The expression for calculating the power density is: ; in, Candidate design schemes power density, Candidate design schemes obtained from simulation calculations 'output power' Candidate design schemes The total mass.

[0033] As an optional implementation, the constraints of total constraint violation in this embodiment include engineering constraints. and failure penalty constraints The engineering constraints and failure penalty constraints The function expression is: ; ; ; in, This is the lower limit threshold for output power. Candidate design schemes 'output power' express The absolute value; This is the failure indicator function for electromagnetic simulation. The failure indicator function takes the value 1 or 0. A value of 1 indicates that the electromagnetic simulation has failed, and a value of 0 indicates that the electromagnetic simulation has succeeded. This is a penalty coefficient, and the penalty coefficient must be greater than 0. In addition to the power lower limit, other engineering constraint components (such as temperature) can also be added in parallel according to design requirements. To handle exceptions such as "parameter writing failure / solution failure / reading failure", a failure penalty constraint is introduced. , A sufficiently large pre-set penalty coefficient is used to ensure that any failed candidate design is infeasible under the constraints; when the evaluation is successful... ,thereby When the evaluation fails ,thereby .

[0034] In step S105 of this embodiment, when performing non-dominated sorting of the parent and child populations of the current population based on the Deb constraint dominance relationship, any two candidate design schemes The conditions for satisfying the Deb constraint dominance relationship include any of the following: Condition (1): and ; Condition (2): , and ; Condition (3): and All are 0, candidate design schemes Target vector Non-inferior to candidate designs across all objectives Target vector And it is superior to the candidate design in at least one objective. Target vector ; in, Candidate design schemes Total constraint violation Candidate design schemes Total constraint violation rate.

[0035] When calculating congestion distance within the same non-dominated layer, the calculation function expression for congestion distance is: ; in, The crowding distance is the distance between the i-th individual and the i-th individual within the same non-dominated layer. Index the target dimension; and These are the target values ​​of individuals adjacent to individual i after sorting by the m-th target; and These represent the maximum and minimum values ​​of the non-dominated layer at the m-th objective, respectively, and each individual is a single candidate design. Individual selection prioritizes retaining individuals with higher-level and larger values, thus balancing convergence and diversity.

[0036] In this embodiment, step S106 outputting the optimal candidate design scheme refers to selecting feasible non-dominated solutions from all historical samples of candidate design schemes to form a Pareto set output, where the feasible non-dominated solution refers to the total constraint violation degree. Candidate design schemes with a value of 0 and that do not violate any other constraints with a total constraint violation value of 0. Candidate design schemes that satisfy the Deb constraint dominance relationship Among them, candidate design schemes The conditions include: (1) Total constraint violation (2) Candidate design schemes and candidate design schemes Satisfying the Deb constraint dominance relationship, i.e.: candidate design schemes Target vector Non-inferior to candidate designs across all objectives Target vector And it is superior to the candidate design in at least one objective. Target vector . Figure 6 This is the Pareto set in this embodiment. This diagram illustrates the Pareto front results for optimizing three objectives: weight, power density, and efficiency. Weight is measured in kg, power density in W / kg, and efficiency is dimensionless, ranging from 0 to 1. The optimal feasible solution represents the optimal Pareto front edge, infeasible solutions represent failed candidate designs, and non-optimal feasible solutions represent non-optimal candidate designs. Finally, the Pareto set can be... Candidate design schemes (Optimal candidate design scheme) and its objective vector Export as a log file. During playback, the parameters of the log file are read and written to the program. You can choose whether to solve again and keep it open for structure / field distribution verification.

[0037] To verify the importance of each design variable, this embodiment uses Pareto sets. The set of optimal candidate design schemes With target sample Based on this, calculate the correlation coefficient between the variable and the target: ; in, Let j be the design variable of the i-th candidate design scheme. Let m be the target value of the i-th candidate design scheme. , The value represents the sample mean. Further methods such as random forests can be used to calculate variable importance for ranking. Finally, a ring-shaped heatmap is drawn, grouped into four categories: geometric dimensions, tooth parameters, magnet parameters, and other parameters. This achieves a clear mapping between "variable category—variable—target," which is used for subsequent tolerance control and structural optimization. Figure 7 This embodiment presents a circular heatmap of the grouped correlation coefficients of the design variables, where the three rings from the inside out represent the Pareto sets. The set of optimal candidate design schemes With target sample The correlation coefficients for the three objectives of weight, power density, and efficiency are shown in the figure. Among the four groups of design variables, namely geometric dimensions (8 items), tooth parameters (7 items), magnet parameters (4 items), and other (4 items), the geometric dimensions (8 items) have the largest correlation coefficients, meaning they are most correlated with the three objectives of weight, power density, and efficiency.

[0038] Furthermore, this embodiment also provides a multi-objective parameter intelligent optimization system for an axial magnetic field permanent magnet motor, including a microprocessor and a memory interconnected, wherein the microprocessor is programmed or configured to execute the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor. This embodiment also provides a computer-readable storage medium storing a computer program or instructions programmed or configured to execute the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor via a processor. This embodiment also provides a computer program product, including a computer program or instructions programmed or configured to execute the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor via a processor.

[0039] Those skilled in the art will understand that the technical solutions provided by this invention may take the form of a method, system, or computer program product. Therefore, this invention may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this invention may take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. 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, produce an implementation of the flowchart... Figure One One or more processes and / or boxes Figure One The 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 operate 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 One One or more processes and / or boxes Figure One The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure One One or more processes and / or boxes Figure One The steps of the function specified in one or more boxes.

[0040] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A multi-objective parameter intelligent optimization method for an axial magnetic field permanent magnet motor, characterized in that, Includes the following steps: S101, an initial population is generated based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor, and the population is a set of candidate design schemes; S102, the NSGA-II algorithm optimizer generates a set of candidate design schemes based on the current population; S103 uses the NSGA-II algorithm optimizer to call the simulation evaluator to perform boundary trimming, integer rounding, and geometric constraint prediction on the candidate design scheme set. It then performs electromagnetic simulation on the candidate design schemes that pass the geometric constraint prediction and calculates the target vector and total constraint violation degree of the candidate design schemes. S104, the target vector and total constraint violation of the candidate design scheme obtained from electromagnetic simulation are backfilled into the NSGA-II algorithm through the NSGA-II algorithm optimizer; S105. Based on the Deb constraint dominance relationship, perform non-dominated sorting on the parent and offspring populations of the current population, calculate the crowding distance within the same non-dominated layer, and select individuals in the current population according to the principle of priority of non-dominated level and larger crowding distance to obtain the updated population. S106, determine whether the preset number of iterations has been reached. If the preset number of iterations has been reached, output the optimal candidate design scheme. Otherwise, proceed to step S102 and continue iterating.

2. The intelligent optimization method for multi-objective parameters of an axial magnetic field permanent magnet motor according to claim 1, characterized in that, In step S101, when generating the initial population based on the design variables and their value range constraints of the axial magnetic field permanent magnet motor, the design variables include part or all of the following: stator outer diameter, stator inner diameter, rotor outer diameter, shaft diameter, shaft hole diameter, stator axial length, rotor axial length, magnet length, winding inner extension length, winding outer extension length, slot depth, magnet thickness, magnet embedment depth, wrapping thickness, tooth tip depth, slot top width, slot bottom width, slot opening width, fillet radius, tooth tip angle, magnet arc angle, number of circumferential segments, and air gap.

3. The intelligent optimization method for multi-objective parameters of an axial magnetic field permanent magnet motor according to claim 2, characterized in that, The functional expression for the geometric constraint prediction in step S103 is: ； in, Candidate design schemes Geometric constraints, A zero matrix, with geometric constraints less than or equal to a zero matrix. This indicates that the prediction is made based on geometric constraints, and the functional expression of the geometric constraints is: ； in, , , , and These are the five components of the geometric constraint. The superscript T denotes the transpose operation. The functional expressions for each component of the geometric constraint are as follows: ； ； ； ； ； ； in, For the width of the slot, The width of the top of the groove, Given the geometric margin, Where is the fillet radius, and min is the minimum value. The width of the trough bottom, Given the minimum allowable thickness of the tooth tip, For the depth of the groove, Given the available radial thickness, max is the maximum value. Stator outer diameter This is the inner diameter of the stator.

4. The intelligent optimization method for multi-objective parameters of an axial magnetic field permanent magnet motor according to claim 1, characterized in that, In step S103, when performing electromagnetic simulation on the candidate design schemes predicted by geometric constraints and calculating the target vector and total constraint violation degree of the candidate design schemes, the calculation function expressions for the target vector and total constraint violation degree of the candidate design schemes are as follows: ； in, Candidate design schemes The target vector is denoted by min, where min represents the minimum value. Candidate design schemes Total mass, Candidate design schemes power density, Candidate design schemes System efficiency; Candidate design schemes Total constraint violation The value must be greater than or equal to 0. A value of 0 indicates that the engineering constraint judgment has passed, while a value of 0 indicates that the engineering constraint judgment has failed. `max` represents the maximum value. This is the k-th constraint in terms of total constraint violation degree.

5. The intelligent optimization method for multi-objective parameters of an axial magnetic field permanent magnet motor according to claim 4, characterized in that, The candidate design schemes The expression for calculating the power density is: ； in, Candidate design schemes power density, Candidate design schemes obtained from simulation calculations 'output power' Candidate design schemes The total mass; the constraints of the total constraint violation degree include engineering constraints. and failure penalty constraints The engineering constraints and failure penalty constraints The function expression is: ； ； ； in, This is the lower limit threshold for output power. Candidate design schemes 'output power' express The absolute value; This is the failure indicator function for electromagnetic simulation. The failure indicator function takes the value 1 or 0. A value of 1 indicates that the electromagnetic simulation has failed, and a value of 0 indicates that the electromagnetic simulation has succeeded. This is the penalty coefficient, and the penalty coefficient takes a value greater than 0.

6. The intelligent optimization method for multi-objective parameters of an axial magnetic field permanent magnet motor according to claim 1, characterized in that, In step S105, when performing non-dominated sorting of the parent and offspring populations of the current population based on the Deb constraint dominance relationship, any two candidate design schemes The conditions for satisfying the Deb constraint dominance relationship include any of the following conditions: Condition (1): =0 and Greater than 0; Condition (2): Greater than 0 Greater than 0 and Less than Condition (3): and All are 0, candidate design schemes Target vector Non-inferior to candidate designs across all objectives Target vector And it is superior to the candidate design in at least one objective. Target vector ;in, Candidate design schemes Total constraint violation Candidate design schemes The total constraint violation rate; the calculation function expression for the congestion distance is: ； in, The crowding distance is the distance between the i-th individual and the i-th individual within the same non-dominated layer. Index for the target dimension, Dimensions of candidate design schemes; and These are the target values ​​of individuals adjacent to individual i after sorting by the m-th target; and These are the maximum and minimum values ​​of the non-dominated layer on the m-th objective, respectively, and each individual is a single candidate design scheme.

7. The intelligent optimization method for multi-objective parameters of an axial magnetic field permanent magnet motor according to claim 6, characterized in that, Step S106, outputting the optimal candidate design scheme, refers to selecting feasible non-dominated solutions from all historical samples of candidate design schemes to form a Pareto set. The feasible non-dominated solution is defined as a candidate design scheme with a total constraint violation of 0 and not associated with any other candidate design scheme with a total constraint violation of 0. Candidate design schemes that satisfy the Deb constraint dominance relationship .

8. A multi-objective parameter intelligent optimization system for an axial magnetic field permanent magnet motor, comprising a microprocessor and a memory interconnected, characterized in that, The microprocessor is programmed or configured to execute the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor according to any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program or instructions, characterized in that, The computer program or instructions are programmed or configured to execute, via a processor, the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, The computer program or instructions are programmed or configured to execute, via a processor, the multi-objective parameter intelligent optimization method for the axial magnetic field permanent magnet motor according to any one of claims 1 to 7.

Images