A proposed approach to the classification of bearing condition using wavelets and random forests

This paper presents a proposed approach to the classification of rolling element bearing faults. The approach consists of vibration signal acquisition, digital signal processing, feature extraction from the vibration signal and classification into functional or defective rolling element bearing. Digital signal processing includes signal decomposition and de-nosing using wavelets. An 18-dimensional vector of the vibration signal feature is obtained as a result of feature extraction. Characterization of each recorded vibration signal is performed by a combination of signal's time varying statistical parameters and characteristic rolling element bearing fault frequency components. The classification is performed using random forests algorithm.