基于BP神经网络左逆的无轴承永磁同步电机无位移传感器运行控制  

Displacement Sensorless Operation Control of Bearingless Permanent Magnet Synchronous Motor Based on BP Neural Network Left Inverse

作　　者：朱熀秋[1] 颜磊 刁小燕[1] ZHU Huangqiu;YAN Lei;DIAO Xiaoyan(School of Electrical and Information Engineering,Jiangsu University,Zhenjiang 212013,Jiangsu Province,China)

机构地区：[1]江苏大学电气信息工程学院,江苏省镇江市212013

出　　处：《中国电机工程学报》2020年第11期3673-3680,共8页Proceedings of the CSEE

基　　金：国家自然科学基金项目(61973144,51675244);江苏省重点研发计划(BE2016150);江苏省高校优势学科建设工程(三期)资助项目(PAPD-2018-87)。

摘　　要：针对无轴承永磁同步电机使用机械式位移传感器存在体积大、成本高和可靠性下降等问题,提出一种基于BP神经网络左逆的转子径向位移观测方法。首先,建立径向位移与悬浮力绕组磁链的子系统,证明其左可逆性,利用BP神经网络的拟合能力构建该子系统的左逆模型,实现对径向位移的观测。其次,为提高位移观测的精度,采用二阶广义积分器观测磁链。然后,在位移观测和磁链观测的基础上,基于直接转矩和直接悬浮力控制策略,构建无轴承永磁同步电机无位移传感器的控制系统,最后进行仿真和实验研究。实验结果表明,该方法能有效观测转子径向位移,验证了其准确性和可行性。In order to solve the problems of large volume,high cost and low reliability in the use of mechanical displacement sensors for bearingless permanent magnet synchronous motor(BPMSM),a rotor radial displacement observation method based on BP neural network was proposed.Firstly,the subsystem of radial displacement and suspension windings flux linkage was established,and its left reversibility was proved.The left inverse model of the subsystem was constructed by using the fitting ability of BP neural network,and the observation of radial displacements was realized.Secondly,in order to improve the accuracy of displacement observation,the second order generalized integrator was used to observe the flux linkage.Then,on the basis of displacement observation and flux linkage observation,based on the direct torque and direct suspension force control strategy,the displacement sensorless control system of the BPMSM was constructed.Finally,the simulation and experimental research were carried out.The results show that the method can effectively observe the radial displacements of the rotor,and its accuracy and feasibility were verified.

关 键 词：无轴承永磁同步电机 神经网络 左逆 无位移传感器 磁链观测器 

分 类 号：TM341[电气工程—电机]