采用品质因子优化和子带重构的共振稀疏分解滚动轴承故障诊断方法

摘要

A fault diagnosis method based on the fusion of resonance-based sparse signal decomposition (RSSD) and wavelet transform is proposed to focus on the problem of early fault diagnosis of rolling bearing under strong background noise.The RSSD method is based on the optimization of quality factors with a genetic algorithm and the reconstruction of sub-bands.During the processing of RSSD for the early fault vibration signal of a rolling bearing,the kurtosis value of the low resonance component is used as the objective function,and the genetic algorithm is used to optimize the combination of high and low quality factors.The main sub-band is selected following the principle of energy dominant distribution to reconstruct the low resonance component,so as to reduce the noise interference and to enhance the impulse characteristic of the fault signal.It is based on the characteristics of local optimization and multiresolution of wavelet analysis that the multi-scale wavelet decomposition is applied to the reconstructed low resonance component to extract and deeply mine the fault features of the bearing.The effectiveness of the method is verified by two diagnosis cases of early different type faults of the rolling bearing.The results show higher ability of the proposed method on highlighting the weak fault feature compared with the ordinary RSSD method.%针对强背景噪声下的滚动轴承早期故障诊断问题,提出了一种融合遗传算法品质因子参数优化、子带重构共振稀疏分解(RSSD)和小波变换的故障诊断方法.该方法以滚动轴承早期故障振动信号RSSD低共振分量峭度指标作为目标函数,利用遗传算法对品质因子参数组合进行优化选择.依据能量分布占优原则对低共振分量进行主子带重构、减少噪声干扰且增强故障信号冲击特征.借助小波分析局部优化和多分辨的特性,对重构低共振分量进行多尺度小波分解,提取轴承故障特征且深度挖掘轴承故障特征信息.通过滚动轴承的2种不同类型故障诊断实例表明,与一般RSSD方法相比,该方法具有更强的削弱背景噪声影响、凸显微弱故障特征的能力.