A single fault detection method of gearbox based on random forest hybrid classifier and improved Dempster-Shafer information fusion. (June 2021)

• Record Type: Journal Article
• Authors: Tang, Xianghong; Gu, Xin; Rao, Lei; Lu, Jianguang
• Is Part Of: Computers & electrical engineering. Volume 92(2021)
• Publication Date: 2021-06
• Subjects: Gearbox compound faults -- Single fault detection -- Random forest -- Hybrid classifier -- Dempster-Shafer information fusion; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2021.107101
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: A framework based on random forest hybrid classifier (RFHC) is proposed for single fault detection, which not only identifies various fault types, but also separates the single fault from compound faults. An improved Dempster–Shafer (IDS) information fusion method is developed to fuse the result obtained by the hybrid classifier. Validity of the proposed method proposed is verified by QPZZ-II experimental platform datasets. Abstract: Gearbox fault diagnosis plays an irreplaceable role in ensuring the safe operation of rotating machinery equipment. However, many researches have only diagnosed single faults, and have not detected single faults from compound faults of gearbox. Therefore, in this paper, a framework based on random forest hybrid classifier (RFHC) is proposed for single fault detection, which not only identifies various fault types, but also separates the single fault from compound faults. Meanwhile, an improved Dempster–Shafer (IDS) information fusion method is developed to fuse the result obtained by the hybrid classifier. Extensive evaluations of the proposed methods on the QPZZ-II experimental platform datasets showed that the proposed framework detects the single faults from the compound faults effectively, which reduces the categorization complexity of a single classifier and improves the overall performance of the detection framework. Moreover, compared with the diagnosis result of a single sensor, IDS can achieve higher average fusion precision and improve the reliability of gearbox single fault detection. Graphical abstract: Image, graphical abstract