基于稀疏分解和包络解调的齿轮故障诊断  

Fault Diagnosis for Gear Based on the Sparse Decomposition and Envelope Demodulation

作　　者：雷璐娟 段腾飞[1] 

机构地区：[1]西安石油大学,陕西西安710065

出　　处：《机械研究与应用》2017年第4期107-109,共3页Mechanical Research & Application

摘　　要：齿轮发生局部损伤,其振动信号中存在瞬态冲击成分,而冲击成分往往被谐波和强噪声所掩盖。为提取瞬态冲击信号,构造了冗余的级联字典,建立了将谐波和瞬态冲击在级联字典上进行稀疏分解的数学模型,然后将块坐标松弛法应用于齿轮振动信号的稀疏分解模型上,将谐波和瞬态冲击成分进行分离,并且有效提高了振动信号的信噪比。最后应用Hilbert包络解调从瞬态冲击成分中提取出齿轮的故障特征频率,表明此方法在齿轮故障诊断中的有效性。The localized fault in gear would induce a series of transient impulsive waveforms that are usually distorted by har-monic waveforms and strong noises. In order to diagnose the localized fault of gear, the model of sparse decomposition for har-monic and impulsive waveforms in redundant concatenate dictionary is established in this paper. The block coordinate relaxa-tion is employed to solve the model. Hence the harmonic transient impulsive waveforms are separated and the noise is removedeffectively. And then, the feature frequencies of localized fault are extracted by using the Hilbert envelope demodulation fromimpulsive waveforms. It is demonstrated that the sparse decomposition is effective for gear fault diagnosis.

关 键 词：稀疏分解 故障诊断 包络解调 齿轮 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]