基于自适应波形分解的峭度谱方法在往复压缩机气阀故障诊断中的应用

Application of Kurtogram in Fault Diagnosis of Reciprocating Compressor Gas Valve Based on Adaptive Waveform Decomposition

往复压缩机气阀振动信号具有非平稳非高斯的多源冲击特性,峭度谱方法是检测平稳信号冲击特征信息的有力工具。为此,本文通过自适应波形分解方法,在无需先验知识的情况下将气阀非平稳信号转换为一系列较为平稳波形分量,并利用峭度谱对包含典型冲击特征的分量进行特征提取。结果表明:所提方法不仅能够降低噪声对峭度指标的干扰,而且能够准确表征冲击的强弱和指示对应的频带,为有效提取不同气阀故障特征提供依据。

Vibration signals of reciprocating compressor gas valve usually appear with the multiple source impact characteristics of non-stationary and non-Gaussian. The kurtogram approach is a powerful tool of impact feature information detection for non-stationary signal. Therefore, in this paper, the non-stationary gas valve signal is firstly transformed into a series of more stationary waveform components, and then the impact features of each component are extracted. The results indicated that the proposed method not only can decrease the noise interference, but also can represent the impact strengthen and its corresponding frequency band. So it supply an effective basis for fault feature extraction of different gas valve.