Debris Material Identification Based on EMD - SVM

Different mechanical parts are often made of different metal materials, and obtaining information about the abrasive grain material can help to realize accurate positioning of the wear location. However, a weak point in oil monitoring has always been the distinct identification of abrasive particle material. As a result, a unique method based on EMD and SVM is proposed for the differentiation of metal materials. Through simulation and theoretical analysis, it is proven that metal particles of different materials have different signal characteristics due to their different electrical and magnetic conductivities. The approach described in this study can achieve precise differentiation of six distinct kinds of metal abrasive particle materials with an accuracy rate as high as 95.8% and has a high efficiency and steady performance, according to trials with various algorithmic models. For the evaluation of equipment's health and the diagnosis of faults, this study is extremely important.