Title: A new rotating machinery fault diagnosis method based on data driven and expert knowledge

Authors: Zhenghui Li; Na Zhang; Feiya Lv; Jingya Yang; Rui Wang

Addresses: Henan Engineering Research Center of Rail Transit Intelligent Security, 461200, Zhengzhou, Henan, China ' Henan Engineering Research Center of Rail Transit Intelligent Security, 461200, Zhengzhou, Henan, China ' School of Software, Anyang Normal University, 455002, Anyang, China ' Henan Engineering Research Center of Rail Transit Intelligent Security, 461200, Zhengzhou, Henan, China ' Henan Engineering Research Center of Rail Transit Intelligent Security, 461200, Zhengzhou, Henan, China

Abstract: This paper presents data driven and expert knowledge method for diagnosing rotating machinery fault. The belief rule-based (BRB) inference method is used to model the complex nonlinear relationship between the abnormal vibration features of rotating machinery and its fault type. If the data of all features are used for diagnosis, then computation burden will be too large to realise real-time diagnosis. First, the inputs of BRB model are reduced and weighted through neural network based on data driven algorithm. The outputs of BRB model are fault types of rotating machinery. The belief rules activated by the inputs are combined by the evidential reasoning (ER) algorithm so as to obtain the fused belief structure about the fault, and then, the accurate diagnose result can be calculated from the fused result. The diagnosis results cannot only judge the fault type, but also give the probability of potential fault. The model parameters are open and interpretable. Finally, in the experiment of fault diagnosis of motor rotor, the effectiveness of the proposed method is illustrated.

Keywords: fault diagnosis; rotating machinery; data driven; expert knowledge.

DOI: 10.1504/IJQET.2024.140142

International Journal of Quality Engineering and Technology, 2024 Vol.10 No.1, pp.1 - 18

Received: 10 Jun 2022
Accepted: 29 Apr 2023
Published online: 25 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article