Research on Complex Power Quality Disturbance Identification and Classification Technology

Aiming at the problem that high frequency and complex power quality disturbances (PQDs) are difficult to be accurately identified in power systems, this paper proposes a method to solve them. An S transform based on Blackman window and window width ratio (BST) which controls the window width through the window width ratio and thus refrains from the situation of frequency controlling window width, is presented. The BST has a better time-domain accuracy than the traditional S transform based on Gaussian window and maintains good frequency domain accuracy when making high frequency PQD analysis. For complex PQD, the characteristic separation is carried out to extract characteristics to reduce the mutual interference between different characteristics. Through the BST, the modular matrix can be obtained. After that, it is segmented by principal component analysis in the frequency direction and energy accumulation and probability mean in the sampling point direction; in such way, the disturbance and noise interference are effectively ruled out, and the integrity of characteristic matrix information can be ensured and the characteristics of PQD are highlighted. PQD type is obtained through AlexNet and combination relationship. Finally, simulation data and field data are used to confirm the effectiveness of the proposed method.