新型建模及优化方法在BSRM中的应用

New methods of modeling and optimization in application of bearingless switched reluctance motor

针对电机参数优化设计问题,引入多支持向量机与混沌优化算法,优化设计磁悬浮开关磁阻电机的结构参数。采用有限元仿真建立样本空间,采用多支持向量机构建磁悬浮开关磁阻电机的非参数模型,基于该模型,以满足额定电磁转矩为条件,优化电机参数,优化目标为最大的悬浮力、最小的转矩脉动、最小的绕组间互感,最优的电机结构参数通过混沌优化算法得到。仿真结果表明,BSRM利用多支持向量机所建立的非参数模型高效且准确,采用此优化方法设计的BSRM转矩脉动小、悬浮力大、绕组间互感小。

For the parameters design of the motor, muhiple support vector machine and chaos algorithm are proposed to solve the problem of structure parameters＇ design. It obtained sample space by finite element simulation, and then built MultiSVM on pametric models of radial levitation force, torque, mutual inductance and torque ripple. Based on these models, it chosed rated torque reached as constraint condition, and selected maximal radial levitation force, minimal mutual inductance and minimal torque ripple as optimal objective. And then, the optimal parameters of BSRM were obtained bychaos algorithm searching. The simulation results prove that he MSVM nonparametric models have high accuracy and the optimized motor has strong suspension bearing capacity ,weak coupling and weak ripple.