Research on intermittent fault diagnosis of rolling bearing based on interval-valued evidence construction and possibility

Abstract

Intermittent fault diagnosis plays a key role in reducing the maintenance cost of the machine and improving the safety of the equipment. It is very urgent to improve the detection rate and diagnostic accuracy of intermittent failure. This paper proposes a novel intermittent fault diagnosis method. The method firstly adopts the output results of traditional diagnostic methods and the power exponential to construct interval-valued evidence. Then the method adopts the possibility to process each single-element subset of the evidence. Next, a classification method of intermittent fault is proposed based on similarity measurement. The experimental results show that the accuracy is 0.98 in the scene with mild interference and 0.94 in the scene with heavy interference. The proposed method can identify intermittent fault quickly and it has strong robustness to interference. It can be better applied in intermittent fault diagnosis for rolling bearing than the existing models.