Fault diagnosis of automaton based on local characteristic-scale decomposition and individual feature selection

3rd International Conference on Advances in Energy, Environment and Chemical Engineering

能源学;生态环境科学;化学工业

Du, Wei^1 ; Fang, Liqing^1

Department One, Mechanical Engineering College, Shijiazhuang

050003, China^1

关键词: Correlation dimensions;    Fault classification;    Feature matrices;    Feature selection methods;    Individual features;    Local characteristic-scale decompositions;    Multi-class support vector machines;    Vibration signal;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/69/1/012076/pdf
DOI  :  10.1088/1755-1315/69/1/012076

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

In order to accurately use the vibration signal to fault diagnosis, a method of automaton fault diagnosis based on local characteristic-scale decomposition (LCD) and individual feature selection (IFS) was proposed. Firstly, the vibration signal of automaton was decomposed with local characteristic-scale decomposition to obtain intrinsic scale components (ISC) in different scales, and the correlation dimension of these components were calculated as characteristic parameters. Then, the individual optimal feature subset for each pair of fault states was selected using an improved Fisher feature selection method. Finally, the feature matrix was input to multi-class support vector machine for fault classification and recognition. The results of automaton fault test demonstrate that the proposed method is of high accuracy in fault diagnosis.