摘要
为了有效提高支持向量机(SVM)在轴承故障模型诊断识别中的故障识别率,提出了一种基于改进自适应白噪声的完备集成经验模态分解方法(ICEEMDAN)和支持向量机(SVM)相互结合的故障诊断识别方法。用ICEEMDAN对原始振动信号分解成若干IMFS(本征模态函数);利用相关系数方法选取有效性的IMFS进行信号重构,再把重构信号的特征向量输入SVM进行故障识别;利用美国凯斯西储大学的轴承实验数据验证了该模型的诊断有效性。结果表明:该方法能够有效提高滚动轴承故障诊断准确率。
In order to effectively improve the fault recognition rate of support vector machine(SVM)in bearing fault model diagnosis and recognition,a fault diagnosis and recognition method based on improved adaptive white noise(ICEEMDAN)and support vector machine(SVM)was proposed.Firstly,the original vibration signals were decomposed into IMFS(intrinsic modal functions)using ICEEMDAN.Secondly,the correlation coefficient method is used to select effective IMFS for signal reconstruction,and then the feature vector of the reconstructed signal is input into SVM for fault recognition.Finally,the validity of the model is verified by using experimental data of bearings from Case Western Reserve University.The results show that this method can effectively improve the accuracy of rolling bearing fault diagnosis.
作者
张平格
李业
刘文豪
ZHANG Pingge;LI Ye;LIU Wenhao(Hebei University of Engineering School of Mechanical and Equipment Engineering,Handan 056038,China)
出处
《技术与市场》
2022年第7期103-105,共3页
Technology and Market