摘要：The faults of rotating machinery monitored by fixing the sensors on rotor directly brings some problems such as difficulty in fixing the sensors, poor universality and so on, while it brings some advantages such as rapidness, convenience and good universality and so on by fixing the sensors on the base. Since the base is far away from the fault source, the collected vibration signals are relatively weak. Multi-sensor information fusion method can describe the diagnostic object completely because it extends spatial and temporal limits, and obtains more information of the diagnostic object. A rotor crack fault diagnosis method based on base and multi-sensor adaptive weighted information fusion is proposed in this paper. The arrangement of the sensors on rotating machinery base was designed by taking advantage of the vibration transitivity of rotor-bearing-base and was verified by experiments. The adaptive weighted information fusion method is investigated, and the fault characteristic value is detected with multi-layer decomposition of wavelet analysis, and the fault of the cracked rotor is diagnosed effectively by combining with the Bode diagram in the process of speeding up. The research will provide a new method for the fault diagnosis of rotating machinery.
