CAJ | 학술논문

为了解决智能监测和故障诊断中故障样本缺乏的问题,提出一种支持向量数据描述(support vector data description,SVDD)和经验模态分解(empirical mode decomposition,EMD)相结合的单分类方法.该方法在只有正常状态数据样本而无需故障样本的情况下可以建立起单值分类器,从而区分出机器的运行状态.采用经验模态分解对数据进行预处理,提取信号在不同频带的能量特征作为SVDD的输入参数进行分类.将该方法应用于滚动轴承的故障诊断中,试验结果表明,该方法可以比传统的SVDD方法更有效地识别轴承的运行状态.
위료해결지능감측화고장진단중고장양본결핍적문제,제출일충지지향량수거묘술(support vector data description,SVDD)화경험모태분해(empirical mode decomposition,EMD)상결합적단분류방법.해방법재지유정상상태수거양본이무수고장양본적정황하가이건립기단치분류기,종이구분출궤기적운행상태.채용경험모태분해대수거진행예처리,제취신호재불동빈대적능량특정작위SVDD적수입삼수진행분류.장해방법응용우곤동축승적고장진단중,시험결과표명,해방법가이비전통적SVDD방법경유효지식별축승적운행상태.
In order to solve the problem of insufficient fault samples in intelligent monitoring and fault diagnosis, a one-class classification method combined support vector data description(SVDD) and empirical mode decomposition(EMD) is proposed. With this method, one-class classifier can be built when only the object of normal condition is available, and the abnormal condition can be distinguished from normal condition. Empirical mode decomposition (EMD) is used as data preprocessing to extract the energy varies in different frequencies bands, and the energy features extracted by EMD can be served as the input parameters of SVDD for classification. Applying this method to the rolling bearing fault diagnosis, the test result shows that this method is superior to traditional SVDD method and would identify rolling bearing fault patterns more effectively.