Affiliations: [a] School of Mechatronic Engineering & Jiangsu Key Laboratory of Mine Mechanical and Electrical Equipment, China University of Mining and Technology, Xuzhou, China
| [b] Department of Mechanical and Industrial Engineering, University of Massachusetts Amherst, Amherst, MA, USA
Correspondence:
[*]
Corresponding author. Hua Zhu, School of Mechatronic Engineering & Jiangsu Key Laboratory of Mine Mechanical and Electrical Equipment, China University of Mining and Technology, Xuzhou 221116, China. E-mail: huazhua.cumt@gmail.com.
Note: [1] This project is supported by National Science Foundation of China (no. 51775546) and PDPA.
Abstract: As the rolling element bearing continues to soar in industry, the requirement for accurate bearing degradation prognostics becomes more and more crucial. A single Fuzzy predictor may suffer from its model parameters optimization. To this end, this paper proposed a new ensemble Fuzzy predictor model for estimating the degradation of a bearing using tribological responses among the rollers and the bearing races. This new method employs the genetic algorithm (GA) to assign an optimal weight vector to a set of adaptive network-based Fuzzy inference system (ANFIS) models. The ensemble of the predicted values of the ANFIS models is used as the prediction of the bearing degradation. Experimental data acquired from the degradation test of five rolling element bearings was used to evaluate the prediction performance of the proposed method. The analysis result demonstrates that the ensemble ANFIS model enables to improve the prediction accuracy against a single ANFIS one. The contribution of this paper is that the ensemble of the ANFIS models is not addressed in existing research and should be optimized.
