Power Quality Disturbances Recognition Based on PCA and BP Neural Network

This paper presents a new approach for recognizing disturbances signals in power quality (PQ) disturbances by principal-component analysis (PCA) and BP neural network. The new approach identifies most types of PQ disturbance, such as voltage sags, swells, interruptions, transients, harmonics and flickers. The new model mainly includes three steps. Firstly, S-transform is used to analyze power system disturbance signals, and 18 distinguishing features are extracted from the result of S-transform. Secondly, principal-component analysis (PCA) is used to reduce the dimensionality of features data set, mean while extract principal-components to describe nonstationary signals of power system. Finally, use the principal-components as the input vectors of BP neural network modified by adaptive learning factor, and classify the disturbances signals. The simulation results show the validity and efficiency of the proposed model.