LSTM-based Fault Prediction for Ship Power Systems

Abstract A ship power system refers to an isolated grid consisting of generation, electrical energy conversion, transmission, distribution, and consumption. As the heart ship’s route, station should try avoid malfunction, malfunction be quickly investigated recovered. In this paper, a stand-alone is built for simulation using Simulink software platform, fault voltage current parameters each phase are selected as source data set input Long Short Term Memory (LSTM) neural network algorithm training diagnostic prediction faults. By comparing results with those Back Propagation (BP) network, show that LSTM has accuracy 98.334%, which can diagnose failure modes stations more accurately higher explanatory predicted data.