基于超限学习机的柱塞泵故障诊断

Fault diagnosis of piston pump based on ELM

为了提高柱塞泵故障诊断的准确度,提出应用一种新的分类算法即超限学习机(ELM)对柱塞泵进行故障诊断与识别。首先,采集柱塞泵各工作状态下的信号,其次对采集到的信号进行预处理,选择出特征向量。最后,应用ELM与其它分类器分别对其进行诊断与识别。对比试验结果表明,新的方法故障诊断准确度高且诊断速度快。

In order to improve the accuracy of fault diagnosis of plunger pump,a new classification algorithm,the Extended Learning Machine (ELM),is proposed to diagnose and identify the fault of plunger pump.Firstly,the signals of the plunger pump in different working conditions are collected.Secondly,the collected signals are preprocessed and the eigenvectors are selected.Finally,ELM and other classifiers are used to diagnose and identify them respectively.The comparison test results show that the new method has high accuracy and fast diagnosis speed.