摘要:the running status of it can be more comprehensive judged through acquisition and analysis of vibration signal for water pump rotary machinery. Therefore, the collection of vibration signal is the key to monitor equipment running status and diagnose faults. This paper introduces water pump fault diagnosis system based on DSP and wavelet transform, and the system can realize water pump vibration signal acquisition and real-time fault diagnosis. In order to improve the vibration signal to noise ratio and get relatively pure vibration signal, the vibration signal that is collected is de-noised by the wavelet de-noising technology. Due to vibration signal often contain singularity component when the pump device occur fault, therefore, the fault diagnosis information of water pump can be derived from vibration signal singularity analysis, the vibration signal is processed and analyzed by combining the modulus maximum value of wavelet transform ( WTMM ) with index, and the fault characteristic values are obtained. The knowledge base of fault model is obtained by training a large number of fault characteristic values, and suspected fault sets are obtained according to model matching method, finally diagnostic results are received by the intersection of suspected fault sets. The experiment results show that the system can effectively find the faults, distinguish faults type and identify faults degree, a kind of effective method is provided for water pump unit faults diagnosis, and has the certain instruction significance for other mechanical fault diagnosis.