基于随机森林的永磁球形电机优化通电策略

Optimization Electrifying Strategy of Permanent Magnet Spherical Motor Based on Random Forests

针对永磁球形电机的复杂转矩建模问题,提出了一种基于随机森林建模方法,同时提出了基于粒子群优化方案来解决永磁球形电机的优化通电策略问题。首先,通过电磁分析得到永磁球形电机转子与定子线圈在不同角度下驱动线圈产生的转矩值。然后,采用随机森林法建立了位置-转矩的黑箱模型。并在此基础上结合倾斜轨迹,采用粒子群算法对其通电策略进行优化设计,使得通电损耗降低了5%~10%。仿真和实验结果均表明,该方法能有效地实现永磁球形电机的稳定控制且通电损耗较小,可为下一步实现永磁球形电机高精度的轨迹跟踪控制提供参考和借鉴。

Aiming at the complex torque modeling problem of permanent magnet spherical motor(PMSM),a modeling method based on random forests(RF)is proposed,and a scheme based on particle swarm optimization(PSO)is proposed to obtain the optimization electrifying strategy of PMSM.Firstly,when the rotor and stator coils of the PMSM are at different angles,the torque value generated by the driven coils is obtained through electromagnetic analysis.Secondly,a position-torque black box model is established by using RF method.On this basis,combined with the tilt trajectory,the PSO algorithm is used to optimize the electrifying strategy,which reduces the electrifying loss by 5%~10%.The simulation and experimental results show that this method can effectively realize the stable control of PMSM,and reduce the electrifying loss,which can lay a foundation for the next step to achieve the high-precision tracking control of PMSM.