Research On Permanent Magnet Synchronous Servo Motor Optimization Based On Response Surface Methods

Permanent magnet synchronous servo motor has high torque density and compact structure. It is widely used in machine tool and robot joint. This paper proposed a method of combining the response surface model and optimization of permanent magnet synchronous servo motor and has a study.Firstly, the basic theory of response surface model algorithm is summarized and the basic process of the response surface model is established with specific mathematical model. After that, the basic theory of genetic algorithm is described, and the multi-objective optimization process based on genetic algorithm is presented in combination with the specific mathematical model.According to the suppression of torque ripple of motors, the torque ripple of motors is analyzed theoretically. Considering the effect of no-load back EMF harmonic content to torque ripple is large, this paper proposed a structure optimization strategy based on optimization of no-load EMF harmonic compared with the results of finite element method. After the analytical expression of the no-load back EMF is obtained, the back EMF harmonics are optimized by genetic algorithm.According to the suppression of inhibition of cogging torque amplitude, the cogging torque motors is analyzed theoretically and get the solution of optimized structure. Compared with the finite element results, the error is large. Based on the above problems, the response surface model is established to solve the optimization problem and the optimization results of motors with two different pole slots are presented.At last, based on the above result, no-load EMF harmonics suppression target and cogging torque amplitude target for multi objective are optimized respectively with traditional optimization method and response surface modeling method with two different slot/pole motors. The traditional method is compared with the experimental and finite element method, and the finite element method is used to compare the response surface modeling method. The two indexes are optimized demonstrating the advantages of the new method by comparing the optimization efficiency.