PMSM最优自适应CDKF估计方法

Adaptive CDKF Algorithm Research and Application on PMSM's Sensorless Vector Control

基于永磁同步电机(PMSM)无传感器矢量控制性能要求,提出一种无传感器PMSM非线性系统参数辨识最优自适应中心差分估计(Adaptive Central Divided Kalman Filtering,ACDKF)方法.该法基于Bayesian最优估计框架,利用Stirling多项式插值逼近确定Sigma采样点及其权值,构建CDKF估计算法;同时考虑系统噪声统计时变统计特性,基于估计信息和残差实现噪声自适应在线估计调整,面向永磁同步电机复杂工况条件下观测电流信号,实时估计转子转速和角位移.仿真结果表明该方法既能获得较高的估计精度,又能有效改善估计计算稳定性,满足永磁同步电机无传感器矢量控制性能要求.

To satisfy the sensorless vector control performance for permanent magnet synchronous motor （PMSM）, a sensorless control method based on optimal adaptive CDKF algorithm applied on nonlinear PMSM＇ s parameters estimation is presented, which is based on Bayesian optimal estimation frame and second- order Stirling polynomial interpolation approximation method to determine Sigma points and its weight coefficient, and meanwhile the statistical characteristics changing over time of the system and measurement noises calculated adaptively. With the ACDKF algorithm and the current signals of the PMSM under complicated con- ditions it calculates the speed and magnet angle displacement, then the speed and magnet angle displacement of vector control system are obtained. The simulation data demonstrate that the proposed method has better estimation precision and computation efficiency and numerical stability than others CDKF algorithms.