Motor Fault Detection and Diagnosis Based on Negative Selection Algorithm

Abstract

This paper presents a motor fault diagnosis method based on negative selection algorithm. It has the structure of two-level detectors, the first level detector detecting the presence of faults and the second level detector detecting the type of faults. Therefore the first level detectors are trained by using motor normal signals, and the second level detectors are trained by using several types of fault signals. During the process of detecting, only the test results of first level detectors are abnormal, the second level detectors are activated and implement fault detection to identify fault type. In this paper, normal vibration signals of motor bearing and three types of fault signals from American Case Western Reserve University bearing fault database are used to verify the fault diagnosis method. The experimental results show that the method can effectively detect early failure and can correctly identify the fault type.