三峡大学学报（自然科学版）2025，Vol.47Issue(4)：96-103,8.DOI:10.13393/j.cnki.issn.1672-948X.2025.04.013
基于GASF-BMKELM的滚动轴承故障诊断方法
Method of Fault Diagnosis on Rolling Bearing Based on GASF-BMKELM
摘要
Abstract
Aiming at the problems that traditional methods of fault diagnosis are difficult to fully extract the fault information and the neural networks rely on the selection of initial parameter,a novel method of fault diagnosis based on Gramian angular summation field(GASF)and Bayesian optimization multi-kernel extreme learning machine(BMKELM)is proposed in this paper.Firstly,the original vibration signal is transformed into the logarithmic energy of the wavelet packet by using the logarithmic energy of the node and the Gram angle summation field(GASF).Secondly,the multi-kernel extreme learning machine(MKELM)is constructed by the weighted combination of multinomial kernel function and radial basis kernel function.At the same time,the Bayesian optimization algorithm is adopted to optimize the parameters of the multi-kernel extreme learning machine to improve the fault identification capability of the diagnostic model.Finally,the model of BMKELM is utilized to identify and classify the fault features with the input of wavelet packet logarithmic energy graph.Through verifying the proposed methed on two datasets,the experimental results show that it has an accuracy of 99.39%and 98.89%,respectively,and the fault recognition rate and stability is higher.关键词
滚动轴承/格拉姆角和场/小波包对数能量图/多核极限学习机/贝叶斯优化算法/故障诊断Key words
rolling bearing/Gramian angular summation field(GASF)/wavelet packet logarithmic-energy map/multiple-kernel extreme learning machine(MKELM)/Bayesian optimization algorithm/fault diagnosis分类
机械工程引用本文复制引用
杨锡发,王林军,邹腾枭,吴振雄,李响,陈保家..基于GASF-BMKELM的滚动轴承故障诊断方法[J].三峡大学学报（自然科学版）,2025,47(4):96-103,8.基金项目
国家自然科学基金项目(51975324) （51975324）
