Static Voltage Stability Evaluation of Large-Scale Renewable Energy Power System Based on Convolutional Neural Network

Abstract

The increase of the proportion of renewable energy and the change of operation mode will lead to voltage collapse. Short circuit ratio has become an effective means to evaluate the static voltage stability of new energy power systems, but the evaluation results are conservative and the accuracy is poor. Therefore, an evaluation method for static voltage stability of power system with multiple new energy stations is proposed. Firstly, considering the uncertainty of renewable energy output and topological structure, the initial feature set and short-circuit ratio of static voltage stability assessment are constructed with wide-area measurement data. Convolutional neural network is used to mine key features, and sample sets were constructed under different operating modes and the proportion of renewable energy. Then, the mapping relationship between power system steady-state data and short circuit ratio is established by using deep learning method. The purpose of evaluating static voltage stability of high proportion new energy power system is achieved. Finally, through the analysis of IEEE-39 node system with wind power equipment, the effectiveness and accuracy of the proposed static voltage stability evaluation method are verified.