Research On Multi-objective Optimization Design Of Permanent Magnet Brushless DC Motor Based On Substitution Model

Permanent magnet brushless DC motor is a kind of permanent magnet synchronousmotor of electromechanical integration, it uses the electronic commutation replacemechanical commutation of traditional DC motor, its stator structure is similar to inductionmotor. So it not only has the mechanical characteristics of DC motor, and also has a lot ofinduction motor’s advantages, like simple structure, reliable operation, convenientmaintenance and so on. So it is widely used in areas such as the power system (electric carseries), high performance servo drive (air conditioner, refrigerator). But the permanentmagnet brushless DC motor also exposes some shortcomings and some performance need tobe improved in practical application, such as large torque ripple. This needs to optimize thedesign of motor body. But the structure optimization method of traditional motor is based ongeometric modeling or mathematical modeling. Geometry modeling is not conducive to themulti objective analysis of the motor, and inefficient. Mathematical modeling’s accuracycannot be guaranteed, and not practicality. So it is more effective to use multi-objectivesubstitution model for motor optimization and base on parameterized modeling and analysis.This paper analyzed the structure characteristic and operation principle of permanentmagnet brushless DC motor, adopted the method of magnetic circuit calculation and finiteelement numerical analysis to identify the size of each part of the brushless DC motor usingas mini electric vehicle permanent magnet, the parametric model of motor was establishedby using the motor design module RMxprt from software Ansoft Maxwell. It established apermanent magnet brushless DC motor of multi objective optimization model by the actualneeds of the multi-objective optimization design of the micro electric vehicle motor. itscontent mainly includes the selection of design variables, selection of the objective functionand constraint conditions, determined the relationship and through the parametric analysis ofand obtained a series of discrete between sample point array, the nonlinear fitting of thesesample points array by using neural network, gained multi-objective substitution model, andeventually established permanent magnet multi-objective optimization model of thebrushless DC electric machine. And then used an improved non dominated sorting geneticalgorithm (NSGA-II) and clone to select adaptive local search algorithm based on multiobjective Memetic (CSALS MOMA) to optimize the design of permanent magnet brushless DC motor of multi objective optimization model, and proposed an improved CSALSMOMA algorithm to solve the model and get the optimal solution set. Finally itseffectiveness was verified by using finite element simulation.The discussion in this paper can provide some guidance and reference for the multi-objective optimization design method of permanent magnet brushless DC motor.