{"title":"Identification of Mechanical Fault of Induction Motor by Combining Lyapunov Exponent and Random Forest Algorithm","authors":"Yan Liu, Jincheng Geng, Yuchen Li, Yiming He","doi":"10.1115/imece2022-94758","DOIUrl":null,"url":null,"abstract":"\n In this paper, a scheme by combining Lyapunov exponent and random forest algorithm for mechanical fault identification of induction motor is proposed. During the implementation of the scheme, the severity of pedestal looseness is identified. First, operating states of motors are simulated in an experimental platform. Vibration signals are measured and the evolution of signal waveform are revealed. Subsequently, for vibration signals of different input frequencies or different number of pedestal loose bolts, Lyapunov exponential spectrums are calculated by BBA algorithm. Features extracted from Lyapunov exponential spectrums, such as largest Lyapunov exponent and Kolmogorov entropy, are utilized to demonstrate the chaotic characteristics of signals. At last, extracted features are fed into random forest model for the fault classification. The feasibility of this scheme is validated by accuracy of fault recognition greater than 90%. Several groups of experimental results indicate that proposed scheme has high effectiveness and generalization performance.","PeriodicalId":302047,"journal":{"name":"Volume 5: Dynamics, Vibration, and Control","volume":"9 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Volume 5: Dynamics, Vibration, and Control","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1115/imece2022-94758","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
In this paper, a scheme by combining Lyapunov exponent and random forest algorithm for mechanical fault identification of induction motor is proposed. During the implementation of the scheme, the severity of pedestal looseness is identified. First, operating states of motors are simulated in an experimental platform. Vibration signals are measured and the evolution of signal waveform are revealed. Subsequently, for vibration signals of different input frequencies or different number of pedestal loose bolts, Lyapunov exponential spectrums are calculated by BBA algorithm. Features extracted from Lyapunov exponential spectrums, such as largest Lyapunov exponent and Kolmogorov entropy, are utilized to demonstrate the chaotic characteristics of signals. At last, extracted features are fed into random forest model for the fault classification. The feasibility of this scheme is validated by accuracy of fault recognition greater than 90%. Several groups of experimental results indicate that proposed scheme has high effectiveness and generalization performance.