基于两段卡尔曼滤波的感应电机无位置传感器控制

Position Sensorless Control of Induction Machines Based on Two-Stage Kalman Filtering

传统的扩展卡尔曼滤波算法已经被广泛地应用在感应电机无位置传感器控制系统中,但其存在运算量大的问题,尤其是应用于感应电机这种多阶、强耦合的系统时.为了解决这一问题,文中引入一种与原算法数学模型上等效的两段式扩展卡尔曼滤波算法到感应电机无位置传感器控制中.在两相静止坐标系下,取定子电流和转子磁链为全阶状态量,以转子电角度及角速度为状态增广量,以此设计两段式扩展卡尔曼滤波算法.实验结果表明,相比传统卡尔曼滤波算法,该算法在保持与原算法相同参数辨识性能的情况下,能够有效地减少运算时间.

The conventional extended Kalman filtering( EKF) algorithm has been widely applied to the position sensorless control of induction machines( IM). However,heavy computational burden may accompany with EKF,especiallyfor the IM with a high-order and strong-coupling mathematical model. In order to solve this problem,a novel control algorithm using the two-stage extended Kalman filtering( TEKF),which is mathematically equivalent to the conventional EKF algorithm,is proposed. In this algorithm,the stator current and the rotor flux linkage in the stationary reference frame are taken as the full-order state variables,and the rotor electrical angle and angular speed are used as the augmented system state variables. Experimental results show that the proposed algorithm can effectively save the operation time without degrading the parameter identification performance of the conventional EKF algorithm.