During machine operation, the debris shape feature is useful to diagnose the wear modes and types occur on a machine component surface. There are six basic shapes of debris: sphere, roll, platelet, chunky, long thin and cutting, that are normally visible during major machine components wear processes. These basic shapes are representing distinguished wear mode and type and can be utilized to predict them. In this paper an automatic debris shape classification technique is proposed and discussed. This technique can identify the basic shape of debris by using fuzzy logic reasoning. It uses debris physical attributes including antiroundness, width standard deviation, aspect ratio, edge regularity and shape regularity. The proposed technique was tested by performing shape analysis on debris samples of a gear wear failure test. The test details demonstrate the suitability and reliability of fuzzy logic reasoning to classify the debris shape.
展开▼