TSPM-CNN based power system transient stability assessment using a new data extraction method

Abstract

The increasing deployment of Wide Area Measurement Systems (WAMS) and Phasor Measurement Units (PMUs) in power grids has significantly improved the situational awareness capability of the system. Together with the increasing computing power of computers, data-driven Transient Stability Assessment (TSA) has become a very popular tool for system operator. In order to improve the accuracy of data-driven transient stability prediction, a new TSPM-CNN TSA method based on a new data extraction method is proposed in this paper. The method makes full use of the excellent feature self-extraction capability of convolutional neural network (CNN) in order to effectively extract the time-series information of individual features and the correlation information of different features, and effectively solve the problems of insufficient datasets and loss of time-series information in traditional data-driven TSA methods. Furthermore, this paper proposes a TSA method based on a combination of time-series prediction model (TSPM) and CNN. The method uses the features of the TSPM to predict the subsequent data of the PMU of the power system and combines it with CNN to predict the transient stability status of the power systems. The effectiveness and accuracy of the method have been successfully verified by simulation experiments in this paper.