Intelligent operation and maintenance system of Marine equipment based on PHM

Abstract

In view of the immature status of complex ship system fault diagnosis, prediction and health management (PHM) technology and the urgent development needs of ship intelligence, the key technologies of ship equipment PHM were studied by means of artificial intelligence. In this paper, an intelligent fault diagnosis technology of Marine equipment based on long short-term memory network (LSTM) is proposed. The dynamic data of Marine equipment is studied by LSTM, and a multi-layer LSTM neural network is established to diagnose the type and degree of faults. A trend prediction technique of state parameters based on ARMA-BP hybrid prediction model is proposed, which combines ARMA and BP neural network to analyze the information in the sequence of state parameters, and effectively improves the prediction accuracy. This paper proposes a method of constructing intelligent reasoning knowledge base of Marine equipment based on production rules. The fault diagnosis and prediction results are associated with corresponding fault modes and impact analysis tables according to certain rules to form intelligent reasoning knowledge base and generate corresponding maintenance decisions intelligently.