An intelligent diagnostic approach based on oil analysis for assessing the technical condition of diesel locomotive engines

Abstract

A new intelligent diagnostic method for assessing the technical condition of diesel locomotive engines using the lubricating oil condition monitoring system is presented in this paper. The proposed method incorporates key oil condition parameters such as kinematic viscosity, iron (Fe) and copper (Cu) wear particle concentrations, soot content and water contamination as indicators of engine health. These parameters are useful for understanding the lubrication performance, wear of the components and potential defects in the engine that may arise during operation. In order to increase the diagnostic accuracy, an Adaptive Neuro-Fuzzy Inference System (ANFIS) was designed to study the collected data and classify the engine condition based on the technical state. The proposed model has applied artificial neural network capability to learn the relationship and fuzzy logic to reason the relationship between oil properties and engine degradation, which makes it effective in identifying the complex nonlinear relations between them. The intelligent diagnostic system developed allows to identify early abnormal operating conditions, facilitates the planning of predictive maintenance actions and decreases the risk of unexpected engine failures. Consequently, maintenance expenses can be kept at low levels and locomotives can be made more available, reliable and service life extended. The proposed approach is also shown to be suitable for implementation at the real time condition monitoring system, and it is a powerful tool for maintaining and managing the assets of modern railways.