Power system transient stability assessment based on the multiple paralleled convolutional neural network and gated recurrent unit

Abstract In order to accurately evaluate power system stability in a timely manner after faults, and further improve the feature extraction ability of model, this paper presents an improved transient assessment (TSA) method CNN + GRU. This comprises convolutional neural network (CNN) gated recurrent unit (GRU). has capability for micro short-term time sequence, while GRU can extract characteristics contained macro long-term sequence. The two are integrated comprehensively high-order features that process. To overcome difficulty sample misclassification, multiple parallel (MP) GRU, with connected parallel, is created. Additionally, focal loss (FL) function which implement self-adaptive adjustment according training introduced guide model training. Finally, proposed methods verified on IEEE 39 145-bus systems. simulation results indicate have better TSA performance than other existing methods.