滚动轴承自适应特征提取的包络谱多点峭度多级降噪方法

滚动轴承故障信号主要包含高品质因子振动分量和低品质因子瞬态冲击分量。采用多点最优最小熵解卷积方法初步削弱传输路径等干扰影响，使微弱瞬态冲击成分得到初步增强,然后针对共振稀疏分解(RSSD)方法存在的品质因子选择困难问题，同时考虑包络谱中故障频率成分的严格周期性，提出包络谱多点峭度(ESMK)概念并将其作为优化指标，采用粒子群优化算法(PSO)对品质因子进行选择，得到一种自适应稀疏分解方法（PSO-RSSD）用于瞬态冲击信号的提取，以消除信号中高幅值干扰冲击和背景噪声的影响。轴承仿真与实测信号分析结果表明，与最小熵解卷积信号共振稀疏分解方法相比，在强冲击干扰下ESMK能够有效度量周期性瞬态冲击，PSO-RSSD方法能自适应分离最优低品质共振分量，验证了该方法的有效性和优越性。

Multi-stage Noise Reduction Method with ESMK for Adaptive Feature Extraction of Rolling Bearings

The rolling bearing fault signals contained both of the high quality factor oscillation components and low quality factor periodic transient impact components. The method of multipoint optimal minimum entropy deconvolution adjusted(MOMEDA) was adopted to weaken the influences of interferences such as transmission path firstly, and the weak transient impact components were enhanced initially. Furthermore, the problems of quality factor determination in resonance-based signal sparse decomposition(RSSD) method and the strict periodicity of fault frequency components in the envelope spectrum were considered and the PSO was used to optimize the quality factors based on proposed novel index called envelope spectrum multi-point kurtosis(ESMK). In consequence, a new self-adaptive transient impact extraction method called PSO-RSSD was proposed, which might effectively eliminate the impacts of high amplitude interference and background noises. Bearing simulation and measured signal analysis results show that, compared with minimum entropy deconvolution(MED)-RSSD method, ESMK may effectively measure periodic transient impacts under strong impact interference, and PSO-RSSD may separate adaptively the optimal low quality resonance components, which verifies the effectiveness and superiority of the proposed method.