Long Short-Term Memory Network-Based HVDC Systems Fault Diagnosis under Knowledge Graph

To enhance the precision of fault diagnosis for high-voltage direct-current (HVDC) systems by effectively extracting various types characteristics, a method based on long short-term memory network (LSTM) is proposed in this paper. The relies knowledge graph platform and developed using measured data from four an HVDC substation located southwest China. Firstly, constructed, then waveform preprocessed divided into training set test set. Various optimizers are employed to train LSTM. strategy’s accuracy calculated compared with recurrent neural (RNN), eXtreme Gradient Boosting (XGBoost), support vector machine (SVM), Naive Bayes classifier, probabilistic networks (PNN), classification learner (CL), which commonly used diagnosis. Results indicate that achieves over 95%, 30% higher than RNN, 8% XGBoost, 4% SVM, 7% Bayes, 40% PNN, 42% respectively; also has minimum time cost, fully demonstrating its superiority effectiveness other methods.