Fault Diagnosis Based on Primary Vibration Feature Extraction of Dual-channel Data for Shaft Systems of Hydropower Units
Compared with feature-level and decision-level fusion,data-level fusion can preserve more raw information for further fault diagnosis so as to lay the foundation for subsequent fault diagnosis.In this paper,a fault diagnosis method based on primary vibration feature extraction of dual-channel data is presented for shaft systems of hydropower units.Firstly,the principal component analysis(PCA)model is used to integrate the dual-channel time-domain vibration signals collected from shaft systems into the composite vibration signals in the direction of the strongest vibration(i.e.the primary vibration).Secondly,time-domain features are extracted from the primary vibration.Finally,a classifier based on support vector ma-chine(SVM)is adopted for fault diagnosis.Experiments were carried out to validate the proposed method.The results show that compared with the methods based on single channel or on dual-channel feature-level fusion,the proposed method can eliminate the influence of changes in primary vibration directions,and improve the effects of intra-class clustering and inter-class separating thereby.As a result,it can improve the diagnostic accuracy.
万方数据
维普
