OMCKD结合自互补Top-Hat变换的电机轴承故障诊断方法

摘要

针对电机轴承微弱故障识别困难这一问题,提出了优化最大相关峭度解卷积(optimized maxi-mum correlated kurtosis deconvolution,OMCKD)结合自互补Top-Hat变换的诊断方法.为解决MCKD关键影响参数难以设置的问题,提出利用人工鱼群算法(artificial fish swarm algorithm, AFSA)并行搜索MCKD参数全局最优解,实现关键影响参数的自动优化调节.首先利用OMCKD方法对原始信号进行预处理,提取被噪声所掩盖的微弱特征信息,继而对解卷积信号做自互补Top-Hat变换处理,进一步抑制背景噪声干扰,强化周期性冲击特征.最后对所得结果做频谱分析,并通过分析谱图中幅值突出的频率成分判定轴承的状态.两组实测信号分析结果表明所述方法可有效用于电机轴承故障诊断,具有一定可靠性及优越性.%To overcome the difficulty of weak fault identification for motor bearings, a diagnosis method based on optimized maximum correlated kurtosis deconvolution (OMCKD) and self-complementary Top-Hat transformation was proposed. In order to solve the problem of MCKD key influence parameters set,the method of parallel searching for the MCKD parameters global optimal solution using artificial fish swarm al-gorithm (AFSA) was proposed,and automatic optimization adjustment of key influence parameters could be achieved. Firstly,the original fault signal was preprocessed by the OMCKD method, and the weak feature information covered by the noise was extracted. Then, the deconvolution signal was processed using self-complementary Top-Hat transformation, the background noise interference could be further depressed, and the cyclical impact feature could be reinforced. Finally, the obtained result was analyzed using frequency spectrum,and the bearing condition could be judged by analyzing the frequency components with obvious amplitudes in the spectrum. The analysis results of two groups measured signals showed that the proposed method could be applied to motor bearing diagnosis effectively,and had a certain reliability and superiority.