Diagnosing Localized and Distributed Bearing Faults by Bearing Noise Signal Using Machine Learning and Kurstogram

Bearings are a common component and crucial to most rotating machinery. Their failures the causes for more than half of total machine failures, each with potential cause extreme damage, injury, downtime. Therefore, fault detection through condition monitoring has significant importance. Since initial cost standard techniques such as vibration signature analysis is high long payback period, via audio signal processing proposed both localized faults distributed/ generalized roughness in rolling bearing. It not appropriate analyze bearing using Fast Fourier Transform (FFT) noise since Amplitude Modulated (AM) mixed up background noises. Localized processed Kurstogram technique finding filtering band because faulty bearings produce impulsive signals