A diagnosis method for diesel engine wear fault based on grey rough set and SOM neural network

The paper aims to establish a model to identify wear fault of marine diesel engine based on grey rough set and Self-Organizing Map (SOM) network with oil monitoring data analysis. The empirical data indicates the wear fault takes great proportion in fault types of diesel engine. Through oil monitoring, the change of parameters of lubricating oil and the information of wear particle can be obtained to analyze status of components. Firstly, the paper constructs the two-dimensional fault decision table. Subsequently, the grey relational analysis and rough set theory are used to reduce the fault decision table horizontally and longitudinally. Next, the fault diagnosis model is established by SOM network. Finally, the proposed model is validated by empirical research. The result suggests that the proposed model is feasible in wear fault diagnosis problem. Moreover, compared with the traditional SOM neural network, the model has less error and better diagnosis effect.