Studies on denoising method of the stator current based on PCA Kalman filter and window processing

Abstract

Aiming at noise interference problem in detecting stator current of induction motor, denoising method based on principal component analysis (PCA) Kalman filtering is proposed in this paper. Firstly, the state equation is obtained by building five order mathematical model, in order to reduce the correlation of data, data space is mapped to low dimensional subspaces via orthogonal transformations. Then, the Kalman prediction value is taken as the center of the observation signal which is processed by adding window processing, and the prediction value is gotten based on the PCA in the window. There are good stability and simple calculation in the algorithm. Simulation results demonstrate that the mean square error based on PCA) Kalman filtering is lower than that based on the traditional Kalman filtering algorithm, and the filtering effect is obviously improved.