基于EMD包络谱分析的液压泵故障诊断方法

摘要

On the basis of failure mechanism of pump and aiming at the difficulty in early fault diagnosis of pump, that is, the fault characteristic signal is weak and the traditional envelope analysis needs to rely on experience for determining the analysed frequency bandwidth, a fault diagnosis approch based on empirical mode decomposition (EMD) and envelope spectrum analysis was proposed. According to the method, the vibration signal was decomposed into some intrinsic mode functions (IMF) using EMD. Then, envelop analysis was carried out on the first some IMFs and the envelope spectrums which include the main fault information were achieved. Finally, the working conditions and fault types of the pump were determined by the analysis of envelope spectrums. The experimental results show that the early fault characteristics of the pump can be extracted effectively and four modes of no fault, slipper loosing fault, swashplate wearing fault and valve plate wearing fault are correctly identified.%在对液压泵常见故障模式进行分析的基础上,针对液压泵早期故障诊断时故障特征信号微弱且传统的包络分析需要预先依靠经验确定分析频段的问题,提出了基于经验模式分解(EMD)和包络谱分析的诊断方法.该方法首先将采集到的液压泵壳体的振动信号利用EMD分解为有限个固有模式函数(IMF)之和:然后对前几个IMF分量进行包络分析,并求出包含主要故障信息的包络谱；最后通过对包络谱的分析,判断液压泵的工作状态和故障类型.实验结果表明,该方法有效提取了液压泵的早期故障特征,准确实现了液压泵正常、滑靴松动、滑靴与斜盘磨损及缸体与配流盘磨损四种状态的分类,是进行液压泵故障诊断行之有效的方法.