改进融合指标的新型盲解卷积算法在轴承故障诊断中的应用  

A Novel Blind Deconvolution Algorithm with Improved Fusion Index for Bearing Fault Diagnosis

作　　者：田甜 唐贵基[1] 田寅初 王晓龙[1] TIAN Tian;TANG Guiji;TIAN Yinchu;WANG Xiaolong(Department of Mechanical Engineering,North China Electric Power University,Baoding 071000,Hebei,China)

机构地区：[1]华北电力大学机械工程系,河北保定071000

出　　处：《噪声与振动控制》2024年第1期162-167,共6页Noise and Vibration Control

基　　金：国家自然科学基金资助项目(52005180);河北省自然科学基金资助项目(E2022502003)。

摘　　要：为解决现有盲解卷积算法易受随机脉冲影响的问题,综合时域特征和频域特征,提出一个新的故障敏感指标,即包络谱峭度-包络基尼系数融合指标(Envelope Spectral Kurtosis-envelope Gini Index,ESKEG)。该指标对周期性脉冲更敏感,不易受随机脉冲的影响。基于该指标,提出一个新的解卷积算法,即基于最大ESKEG的盲解卷积,并采用粒子群算法(Particle Swarm Optimization,PSO)求解滤波器系数。通过仿真振动信号和实验仿真信号进行验证,结果表明相比于其他盲解卷积算法,所提出的PSO-ESKEG算法在故障先验知识未知的情况下,能更有效避免受到随机脉冲信号的影响。In order to solve the problem that the existing blind deconvolution algorithms are vulnerable to random pulses,a new approach is proposed.In this method,the time domain and frequency domain features are integrated into a composite index named Envelope spectral kurtosis-envelope Gini(ESKEG)index.This new index is more sensitive to periodic pulses and less susceptible to random pulses.On this basis,a new blind deconvolution algorithm based on the maximum ESKEG is proposed,which uses the particle swarm optimization(PSO)algorithm to solve the filter coefficients.By comparing its results with the simulated vibration signals and experimental simulation signals,the correctness and efficiency of this algorithm is verified.It is demonstrated that the proposed PSO-ESKEG algorithm outperforms other blind deconvolution algorithms in overcoming the influence of random pulse signals when the prior knowledge of the fault is unknown.

关 键 词：故障诊断 盲解卷积 包络谱峭度-包络基尼系数 粒子群优化 随机脉冲 

分 类 号：TH165[机械工程—机械制造及自动化] TH133[电子电信—通信与信息系统] TN911[电子电信—信息与通信工程]