A fuzzy diagnosis of multi-fault state based on information fusion from multiple sensors

作者：Xue, Hongtao^*; Li, Zhongxing; Li, Yong; Jiang, Hong; Chen, Peng

来源：Journal of Vibroengineering, 2016, 18(4): 2135-2148.

DOI：10.21595/jve.2016.16712

• fuzzy diagnosis
• multi-fault state
• support vector machine
• possibility theory
• information fusion
• synthetic symptom parameter

This paper presents a fuzzy diagnosis for detecting and distinguishing multi-fault state, the method is constructed on the basis of possibility theory and support vector machines (SVMs) with information fusion from multiple sensors. Non-dimensional symptom parameters (NSPs) are defined to reflect the characteristics of vibration information, and principal component analysis (PCA) is used to evaluate and select sensitive NSPs of each sensor. SVMs are employed to fuse vibration information from different sensors into an effective synthetic symptom parameter (SSP) for increasing diagnostic sensitivity, then the possibility function of the SSP is used to construct a fuzzy diagnosis for fault detection and fault-type identification by possibility theory. Practical examples of diagnosis for a roller bearing used in a test bench are given to show that multi-fault states of bearing can be identified precisely by the proposed method.

• 出版日期2016-6
• 单位江苏大学