摘要
针对滚动轴承的故障识别率低的问题,提出了一种改进的自适应噪声完备集合经验模态分解(ICEEMDAN)和排列熵结合的故障诊断方法。首先将振动信号通过ICEEMDAN分解成一系列的IMF并计算他们的排列熵值,再提取前8阶IMF排列熵值,最后结合GWO-SVM模型来进行故障分类诊断。该诊断方法在某轴承故障数据集上准确率为100%。为了进一步说明ICEEMDAN-GWO-SVM模型的优越性,另外设置了泛化实验,该实验采用了某大学的轴承故障数据,在三个工况下的故障诊断准确率也明显优于其他故障诊断方法。
Aiming at the problem of low fault recognition rate of rolling bearings,an improved fault diagnosis method combining adaptive noise complete set empirical mode decomposition(ICEEMDAN)and permutation entropy is proposed.Firstly,the vibration signal is decomposed into a series of IMF through iceemdan and their permutation entropy is calculated,then the first eight order IMF permutation entropy is extracted,and finally the fault classification and diagnosis is carried out combined with gwo-svm model.The accuracy of the diagnosis method is 100%on a bearing fault data set.In order to further illustrate the superiority of ICEEMDAN-GWO-SVM model,a generalization experiment is set up.The experiment adopts the bearing fault data of a university,and the accuracy of fault diagnosis under three working conditions is also significantly better than other fault diagnosis methods.
作者
顾云青
苏玉香
沈晓群
罗健锋
王婷
GU Yun-qing;SU Yu-xiang;SHEN Xiao-qun;LUO Jian-feng;WANG Ting(School of Marine Engineering Equipment,Zhejiang Ocean University,Zhoushan 316022,China;School of Electrical Engineering,Southwest Jiaotong University,Chengdu 611756,China)
出处
《组合机床与自动化加工技术》
北大核心
2022年第8期62-66,共5页
Modular Machine Tool & Automatic Manufacturing Technique
基金
浙江省自然科学基金青年基金(LQ18E070004)
舟山科技计划项目(2022C41009)
2021年浙江省教育厅一般科研项目(Y202148203)
2021浙江省大学生科技创新活动计划暨新苗人才计划项目(2021R411018)。
关键词
改进型CEEMDAN
排列熵
灰狼算法
支持向量机
故障诊断
improved CEEMDAN
permutation entropy
gray wolf optimizer
support vector machine
fault diagnosis
作者简介
顾云青(1997—),男,硕士研究生,研究方向为机械故障诊断,(E-mail)q651809269@qq.com;通信作者:苏玉香(1982—),女,副教授,博士,研究方向为海洋新能源发电、船舶电力系统故障诊断,(E-mail)suyuxiang82@163.com。
