Cyclostationarity analysis and diagnosis method of bearing faults

The vibration signal of rotating machine is typical amplitude-modulated signal. In that case, the method of cyclostationary analysis is very effective for extracting and demodulating the modulating signal that usually contains some fault message. However, duo to a good many factors, such as the special construction, the amalgamation of multiple faults and the fluctuation of rotating speed etc., the cyclostationary analysis result of rolling bearing vibration signal become complex and is difficult to identify fault features with naked eyes. By means of the dynamic identification ability of CHMM, this paper proposes a new fault diagnosis scheme of rolling bearings based on cyclostationarity and CHMM. Experiment results show that single-fault and multi-fault can be recognized by this approach.