(School of Electrical Engineering, Zhejiang University, Hangzhou 310027, China)

 Abstract: Aiming at the problems of traditional asynchronous motor fault diagnosis methods, such as characteristic frequency could be easily flooded by fundamental frequencies, poor anti interference ability and high false rate, a Bayesian network of motor fault diagnosis model based on stator current signal and its Hilbert marginal spectrum feature was introduced. Marginal spectrum feature was extracted from stator current signal by Hilbert Huang transform. In the process of building the model, the dependencies between the nodes were learned by cuckoo search algorithm. Levy flight mechanism was adopted to optimize the search path to improve the search efficiency. The competition mechanism of cuckoo search algorithm which improves the search speed was introduced. The validity of the diagnosis model was verified by the eccentric fault of induction motor. The results indicate that this fault diagnosis model is effective and accurate.

 Key words: asynchronous motor; bayesian network; marginal spectrum; fault diagnosis