基于AWMMD的柴油机气缸故障特征提取方法研究

Fault feature extraction method of diesel engine cylinder based on AWMMD

针对柴油机气缸故障诊断时的噪声干扰问题,提出一种自适应加权多尺度形态分解(adaptive weighted multi-scale morphological decomposing, AWMMD)方法,从各个缸盖表面振动信号中提取故障特征。基于三种组合算子构造一种新的组合差值形态滤波器,用于对振动信号进行多尺度分解;以Teager能量峭度作为评判指标,设计基于遗传算法的各尺度形态模式分量(morphological mode component, MMC)权值自适应分配算法,提出加权多尺度形态分解方法;将自适应权值与多尺度分解的形态模式分量进行绑定,得到优化的故障特征提取结果。仿真信号测试与柴油机故障模拟信号分析结果表明,该方法能有效抑制噪声干扰并提取故障特征。

Here,aiming at the problem of noise interference in diesel engine cylinder fault diagnosis,an adaptive weighted multi-scale morphological decomposition(AWMMD)method was proposed to extract fault features from each cylinder head surface vibration signals.Firsdy,a new combined difference morphological filter was constructed based on 3 kinds of combination operators to do multi-scale decomposition for vibration signals.Secondly,Teager energy kurtosis was taken as the evaluation index,an adaptive weight assignment algorithm was designed based on genetic algorithm for morphological mode components(MMCs)of various scales,then a weighted multi-scale morphological decomposition method was proposed.Finally,adaptive weights were bound to MMCs of multi-scale decomposition to obtain optimized fault feature extraction results.The results of simulated signal testing and diesel engine fault simulation signal analysis showed that the proposed method can effectively suppress noise interference,and extract fault features.