本文基于BP神经网络模型，通过准确采集往复泵液力端各阀箱阀关闭激起的瞬态响应，并以响应的振动频谱为主要征兆，建立了基于BP网络往复泵泵阀故障的诊断系统。计算机模拟识别表明，该系统可较为准确地识别出各泵阀的故障。 Based on the BP neural network model, by accurately collecting the transient response of each valve closed at the hydraulic end of the reciprocating pump, and the vibration frequency spectrum of the response as the main indication, a diagnosis system based on the BP network reciprocating pump valve fault was established . Computer simulation identification shows that the system can more accurately identify the pump valve failure.