基于遗传算法优化的定子磁链扩展卡尔曼估计方法

定子磁链估计的精度直接关系到异步电机直接转矩控制(DTC)的控制效果,本文提出了一种基于扩展卡尔曼滤波器(EKF)的定子磁链观测方法,该方法选取DTC反馈通道中多个主要参数作为状态变量进行联合滤波估计,为保证EKF算法滤波参数的准确性,采用遗传算法(GA)对EKF中的系统噪声矩阵和测量噪声矩阵进行了优化处理.通过使用该方法设计的磁链观测器与DTC构成闭环系统进行仿真实验表明,滤波参数优化后的EKF算法更加有效地提高了磁链估计精度,从而提高DTC系统的低速控制性能.

A Method of the Stator Flux EKF Estimation for Induction Motors Based on Genetic Algorithm Optimizing

The performance of direct torque control (DTC) for induction motor is depended on the stator flux estimation in a large degree. Extended Kalman filter (EKF) is useful to estimate state variables when the measured signal is mixed with noise. A flux linkage flux estimation method for induction motor based on EKF theory is presented in this paper, where, the stator current, stator resistance as well as linkage flux are regarded as the state variables to be estimated by joint filtering. For the sake of improving the accuracy of the filtering, genetic algorithm (GA) is introduced to optimize the noise matrix in EKF. The closed loop system composed by DTC controller and the proposed linkage flux observer with optimized filtering parameter has better estimating accuracy and dynamic low speed performance than ordinary EKF-DTC scheme. The effectiveness is verified by simulation and experimental results.