基于自适应峰值分解的广义局部频率时频分析方法

往复压缩机振动信号具有复杂的多源冲击特性,表现较强的非平稳性,传统的时频分析方法难以提取有效的故障特征.以傅氏变换为基础的传统频率概念和以希尔伯特变换为基础的瞬时频率概念存在固有缺陷,提出一种广义局部频率的概念,并结合自适应峰值分解方法,实现信号时频分布的构造途径;与HHT时频分析方法进行仿真对比,并应用到往复压缩机振动信号故障特征提取.结果表明,基于自适应峰值分解的广义局部频率方法有效揭示往复压缩机不同故障的多源冲击振动信号时频特征,为往复压缩机故障诊断提供一种新的手段.

Analysis of time-frquency method of general local frequency based on adaptive peak decomposition

Vibration signal of reciprocating compressor has complicated multi-source impact feature,which shows strong non-stationary,so traditional time frequency methods are difficult to effectively extract fault feature.Because raditional frequency concept based on FFT and instantaneous frequency based on Hilbert transform have inherent limitations,a notion of general local frequency is proposed,and combined with the adaptive peak-detection approach,which could achieve structured approach of time frequency distribution.Compared to HHT,and used to fault feature extraction of vibration signal of reciprocating compressor,the general local frequency based on adaptive peak-detection approach could effectively show time-frequency feature of multi-source impulse vibration signal for different fault in reciprocating compressor,and provide a new method for fault diagnosis of reciprocating compressor.