Research On Key Technology Of Rolling Bearing Failure Prediction

Rolling bearing is one of the most widely used in rotating machinery parts, and iseasily damaged parts, it has important practical significance to carry out rolling bearingsfor failure prediction. In the process of failure prediction, extract the failure feature fromvibration signal of rolling bearing is the key of failure prediction technology.In this thesis, based on the rolling bearing structure, studies the form of failure andvibration signal generation mechanism of the outer surface, as well as due to thedifferent forms of vibration caused by different failure, proved the feasibility of usingexternal rolling bearing vibration signal. In the failure feature extraction process,because rolling bearing vibration signal is non-stationary and non-linear, the traditionaltime domain or frequency domain analysis techniques exist a variety of limitations, andcombined with the time-frequency analysis technique have had become the most widelyused technique, this paper utilizes the classic wavelet technology, wavelet packettechnology and Hilbert-Huang Transform to extract the failure frequencies feature oftwo kinds of rolling bearing vibration signals, according to calculated the rollingbearing inner ring, outer ring, rolling body three kings of failure characteristicfrequency by theory, to verify the feasibility and effectiveness of these three methods infailure feature extraction.