Bearing Fault Diagnosis in Traction Motor Using the Features Extracted from Filtered Signals

Abstract

Motor bearing faults at an early stage may not cause critical problems, but they may cause motors to be damaged in advanced stages. So, detection of motor bearing faults at an early stage is crucial for preventing bigger problems. In this study, it is aimed to detect the bearing faults on traction motors from the vibration signals obtained by the sensors mounted on the motor. To analyze the signals, an intelligent filter is used to estimate the next healthy value from the previous values of the signal. With further analyses of the difference signal of actual and estimated signals, the defects are detected. The study focuses on the effects of the sensor positions, features chosen, and classifiers used on success of the method.