一种增强型的滚动轴承故障诊断  

An Enhanced Fault Diagnosis of Rolling Bearing

作　　者：袁东辉[1] 朱愉洁[1] 齐咏生[2] 王研凯[1] YUAN Dong-hui;ZHU Yu-jie;QI Yong-sheng;WANG Yan-kai(Inner Mongolia Electric Power Research Institute,Hohhot Inner Mongolia 010000,China;Institute of Electrical Power,Inner Mongolia University of Technology,Hohhot Inner Mongolia 010080,China)

机构地区：[1]内蒙古电力科学研究院,内蒙古呼和浩特010000 [2]内蒙古工业大学电力学院,内蒙古呼和浩特010080

出　　处：《计算机仿真》2022年第10期526-532,共7页Computer Simulation

基　　金：国家自然科学基金(61763037);内蒙古自然科学基金(2019LH06007)。

摘　　要：针对强背景噪声环境下,滚动轴承早期微弱故障特征难以准确提取的问题,提出一种参数优化的变分模态分解(VMD)和支持向量机(SVM)的故障诊断方法。先对轴承振动信号进行VMD分解,并以局部极大包络谱峰值因子作为适应度函数,利用粒子群寻优算法(PSO)对VMD的影响参数(惩罚因子α及分解个数K)进行自适应选择,获取包含故障特征的最佳模态分量;计算各个分量的指标特征,根据递归特征消除(RFE)方法筛选出能表征轴承运行状态的5个关键特征,构建故障特征向量组;将特征向量作为SVM的输入,轴承运行状态为输出,建立SVM轴承状态分类识别模型。通过西储大学平台轴承数据对算法进行验证,结果表明上述方法能够实现滚动轴承不同故障的准确识别。Aiming at the problem that it is difficult to accurately extract the early weak fault features of rolling bearings in the environment of strong background noise, a parameter optimized fault diagnosis method based on variational modal decomposition(VMD) and support vector machine(SVM) is proposed. First, the bearing vibration signal was decomposed by VMD,and the local maximum envelope spectrum crest factor was used as the fitness function. The particle swarm optimization algorithm(PSO) was used to self-decompose the impact parameters of VMD(penalty factor and decomposition number),so as to obtain the best modal component containing fault features;The index features of each component were calculated, and five key features that can characterize the running state of the bearing were selected according to the recursive feature elimination(RFE) method, and the fault feature vector group was constructed;The feature vector was taken as the input of SVM,and the running state of bearing was taken as the output. The SVM bearing state classification and recognition model was established. The algorithm was verified by the bearing data of the platform of Western Reserve University, and the results show that the above method can accurately identify different faults of rolling bearings.

关 键 词：故障诊断 变分模态分解 粒子群优化 特征提取 支持向量机 

分 类 号：TH17[机械工程—机械制造及自动化]