时域径向基函数网络诊断方法在往复泵故障诊断中的应用

往复机械是工程中广泛应用的一种机械设备 ,由于其动力学和运动学形态比旋转机械复杂 ,对其进行故障诊断存在较大难度 ,有效提取往复机械运动中非平稳时变信号中的故障特征和将故障特征准确分类是解决往复机械故障诊断问题的两个关键。本文利用时域数字特征分析方法完成故障特征信息提取 ;通过径向基神经网络对特征信息分类识别 ,实现故障的自动诊断。利用以上原理建立的油田往复塞泵故障监测与诊断系统 ,通过在大庆油田的实际应用表明 ,系统能够比较准确的识别往复柱塞泵多种常见故障 ,且具有较高的运算速度。

Application of Radial Basis Function Neural Network on Fault Diagnosis for Reciprocating Pump

Reciprocating machinery is a widely used machine equipment in engineering. It is more difficult to detect and diagnose the fault in reciprocating machinery than in rotating machinery, because of its more complex dynamics behaviour. Two important issues of reciprocating machine falut diagnose, are extract the fault features of stationary process efficiently and classify these features correctly in reciprocating machinery motion. This paper uses the method of time domain digital signal analysis to extract the fault information, then classifies and identifies the fault using radial base function in order to realize fault auto diagnosing. this paper sets up a fault detecting and diagnosing system of oil field plunger pump based on above approuth. The practical application in Daqing oil field indicates that this system can identify several typical faults of reciprocating plunger pump correctly and has fast responding speed.