A new condition monitoring and fault diagnosis method of engine based on spectrometric oil analysis

Abstract

According to statistics, wear fault is about sixty percent to eighty percent of all the machinery faults. Spectrometric oil analysis is an important condition monitoring technique for machinery maintenance and fault diagnosis. Now, there are two existing mathematics analysis models based on spectrometric oil analysis, namely concentration model and gradient model. However, the above two models have respective disadvantages in condition monitoring and fault diagnosis of the engine. Then in this paper a new mathematics model, proportional model, was put forward monitoring wear condition and diagnosing wear faults of the engine. Proportional model use the relationship and correlation among the elements in the lubricating oil to detect wear condition and occurring faults in the engine. The steps of establishment of proportional model were described firstly. Then we used the experiments data to verify the feasibility of proportional model and gave limit values of proportional model. In order to validate the feasibility of proportional model, proportional model was applied to monitor wear condition and diagnose wear faults of an engine. The results from this paper have proved that the method based on proportional model is applicable in condition monitoring and fault diagnosis of the engine. © 2011 Springer-Verlag Berlin Heidelberg.