A novel scheme for intelligent fault diagnosis of marine diesel engine using the multi-information fusion technology

Abstract

Various parameters in the process of diesel engine operation contain a lot of information. Through data mining, the inherent information of these parameters can be mined out to solve the problems of inaccurate diagnosis and time-consuming. In this paper, a fault diagnosis scheme for the diesel engine is proposed based on the K-means analysis and the back propagation (BP) neural network. K-means is used to cluster the data and BP neural network is designed to diagnose the running state of diesel engine. Then, the fault diagnosis scheme is optimized by principal component analysis (PCA) to simplify the raw data, which are clustered by K-means and set as the input of the BP neural network to establish the fault classification model. Through the analysis and comparison of the results of the two diagnosis algorithms, it shows that the optimized algorithm can extract data features more effectively, improve the diagnosis accuracy, and reduce the diagnosis time.