粒子群算法优化扩展卡尔曼滤波器电机转速估计

摘要

The performance of sensorless vector control was directly affected by the accuracy of speed estimation. For it was hard to obtain the optimum value of the system and mesure matrix of rotating speed estimate under the induction motor using extended Kalman filter( EKF) , a speed estimate method based on EKF which was optimized by improved particle swarm optimization(IPSO) was proposed. Using optimized EKF to estimate the speed of induction motor, the proposed method could effectively improve the speed estimate accuracy comparing to those obtained by trial and error method, genetic algorithm(GA) and standard PSO algorithm.%转速估计的精度直接影响无速度传感器矢量控制的效果,针对感应电机扩展卡尔曼滤波器(EKF)转速估计中难以取得系统噪声矩阵和测量噪声矩阵最优值的问题,提出了一种基于改进粒子群算法优化的EKF转速估计方法.该方法利用改进的粒子群算法对EKF中的系统噪声矩阵和测量噪声矩阵进行优化处理,将优化后的EKF应用于感应电机转速估计.仿真试验表明,与试探法、标准粒子群算法及遗传算法比较,该方法能有效提高转速估计的精度,从而提高无速度传感器矢量控制系统的性能.