An Identification Model of Health States of Machine Wear Based on Oil Analysis

This paper presents a modeling procedure for deriving a single value measure based on a regression model, and a method for determining a statistical threshold value as identification criterion of normal or abnormal states of machine wear. A real numerical example is examined by the method and identification criterion presented. The results indicate that the judgments by the presented methods are basically consistent with the real facts, and therefore the method and identification criterion are valuable for judging the normal or abnormal state of machine wear based on oil analysis.