摘要
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。
By accurately sampling the shock signals during a pump valve shutting and based on the symptom of its vibration spectrum on the pedestal body in a reciprocating pump. this paper establishes a fault diagnosis system on pump valves with BP neural network. Computer simulation shows that the system or software can be used to identify pump valve fault accurately.
出处
《噪声与振动控制》
CSCD
1998年第3期36-39,共4页
Noise and Vibration Control
关键词
BP神经网络
往复泵
泵阀
故障识别
油田
BP neural network, reciprocating pump,pump valve,fault identification
