神经网络在往复泵活塞磨损故障诊断中的应用

FAILURE DIAGNOSIS OF WORN PISTON IN RECIPROCATING PUMP WITH ARTIFICIAL NEURAL NETWORK

对用人工神经网络识别往复泵活塞磨损故障进行了研究。以在往复泵缸套压盖上采集的振动信号的振动功率谱为主要征兆，建立了基于人工神经网络的往复泵活塞磨损故障诊断系统。计算机模拟识别表明，运用该系统可较为准确地判别出活塞的磨损（破损）故障。

This paper makes an approach on application of artificial neural network to the failure diagnosis of the worn piston in reciprocating pump. By sampling the vibration signals on the fixed platen of the pump cylinder, and based on the vibration spectrum symptom, a failure diagnosis system for worn piston in reciprocating pump is developed with artificial neural network. Computer simulation shows that the system and software can be used to identify the failure of worn piston accurately.