Abstract: Beyasian network (BN) is applied to the field of rotor fault diagnosis for increasing the ability to handle uncertainty and inference quality. A general framework of BN for rotor fault diagnosis is proposed, which concerns more information both from the machine performance layer and the symptom layer. The network accords with the general diagnosis thinking of experts. The rotor synchronous fault diagnosis examples demonstrate its good recognition results.

Key words: Bayesian network, Fault diagnosis, Rotor