基于EMD和信息熵的往复压缩机气阀故障特征提取

摘要

Reciprocating compressor vibration signal is typical nonlinear and non-stationary.In the paper,the EMD is applied to analyze the compressor vibration signal,to get a series of IMF whose frequency is changing from high to low and to calculate information entropy of different IMF component.The IMF component containing main fault information is found out.Signal features which can reflect the working condition of compressor are extracted truly and accurately.It will provide the necessary basis to determine fault diagnosis of valves.%往复压缩机振动信号属于典型非线性和非平稳性信号.用EMD对压缩机气阀信号进行分解,得到一系列从高频到低频不断变化的IMF分量,并计算不同IMF分量的信息熵,找出包含有主要故障信息的IMF分量,从中提取出能够准确反映压缩机气阀工作状态的能量特征,为进一步确诊气阀故障提供了依据.