自适应稀疏贝叶斯滤波在轴承故障提取中的应用  

Application of Parameter-adaptive Spare Bayesian Step-filtering in Bearing Fault Detection

作　　者：杨娜[1,2] 刘晔 徐元博[3] 刘静超 YANG Na;LIU Ye;XU Yuanbo;LIU Jingchao(Department of Mechanical and Electrical Technology,Xijing University,Xi′an 710023,China;School of Electrical Engineering,Xi′an Jiaotong University,Xi′an 710049,China;School of Automation,Xi′an University of Posts and Telecommunications,Xi′an 710121,China)

机构地区：[1]西京学院机电技术系,西安710123 [2]西安交通大学电气工程学院,西安710049 [3]西安邮电大学自动化学院,西安710121

出　　处：《噪声与振动控制》2023年第3期132-138,201,共8页Noise and Vibration Control

基　　金：西京学院科研基金资助项目(XJ220206)。

摘　　要：稀疏贝叶斯滤波作为一种简单、新颖的滤波器,对噪声中的步进动态具有较好鲁棒性。同时,该滤波器引入一种L1正则化,其稀疏解可通过标准凸优化方法快速获得,因此它也具有较高的运算效率。但是在原始的稀疏贝叶斯滤波中,正则化参数必须提前设定,而该种参数的选择主要依靠人为经验,这就可能导致所选择的参数无法满足要求。针对现有不足,提出一种基于樽海鞘群优化算法的自适应稀疏贝叶斯滤波的轴承故障提取方法。该种自适应滤波方法采用轴承故障信号的包络谱峭度和负熵为目标函数选择最优的正则化参数,从而得到最优的滤波信号。最后通过包络分析得到轴承故障特征频率。通过模拟数据和真实数据证明该方法的有效性和优越性。The sparse Bayesian step-filtering(SBSF)technique is a simple and novel filter method that is robust to steplike dynamics in noise.This technique introduces an L1-regularized global filter whose sparse solution can be rapidly obtained by standard convex optimization methods.Thus,it also has a higher computation efficiency.However,in the original SBSF method,the regularization parameters have to be pre-defined,and the selection of such parameters mainly depends on human experience,which may lead to the difficulty to select satisfactory parameters.In order to overcome the shortcoming,a parameter-adaptive SBSF based on Salp Swarm Algorithm(SSA)is proposed in this paper.This parameteradaptive SBSF method uses the envelope spectrum kurtosis and negative entropy of bearing fault signal as the objective functions to select the optimal regularization parameter,so as to obtain the optimal filtering signal.Finally,the bearing fault characteristic frequency is detected by envelope analysis.The validity and superiority of this method are demonstrated by simulation and real data.

关 键 词：故障诊断 轴承故障检测 包络谱峭度 负熵 樽海鞘群算法 稀疏贝叶斯滤波 

分 类 号：TH133.33[机械工程—机械制造及自动化] TH165.3