Open Access

ARTICLE

Fault Diagnosis Method of Rolling Bearing Based on ESGMD-CC and AFSA-ELM

by Jiajie He^1,2, Fuzheng Liu³, Xiangyi Geng³, Xifeng Liang¹, Faye Zhang^3,*, Mingshun Jiang³

1 School of Traffic & Transportation Engineering, Central South University, Changsha, 410075, China
2 Project Management Department, CRRC Advanced Composites Co., Ltd., Qingdao, 266108, China
3 School of Control Science and Engineering, Shandong University, Jinan, 250061, China

* Corresponding Author: Faye Zhang. Email:

Structural Durability & Health Monitoring 2024, 18(1), 37-54. https://doi.org/10.32604/sdhm.2023.029428

Received 18 February 2023; Accepted 09 August 2023; Issue published 11 January 2024

Abstract

Incomplete fault signal characteristics and ease of noise contamination are issues with the current rolling bearing early fault diagnostic methods, making it challenging to ensure the fault diagnosis accuracy and reliability. A novel approach integrating enhanced Symplectic geometry mode decomposition with cosine difference limitation and calculus operator (ESGMD-CC) and artificial fish swarm algorithm (AFSA) optimized extreme learning machine (ELM) is proposed in this paper to enhance the extraction capability of fault features and thus improve the accuracy of fault diagnosis. Firstly, SGMD decomposes the raw vibration signal into multiple Symplectic geometry components (SGCs). Secondly, the iterations are reset by the cosine difference limitation to effectively separate the redundant components from the representative components. Additionally, the calculus operator is performed to strengthen weak fault features and make them easier to extract, and the singular value decomposition (SVD) weighted by power spectrum entropy (PSE) can be utilized as the sample feature representation. Finally, AFSA iteratively optimized ELM is adopted as the optimized classifier for fault identification. The superior performance of the proposed method has been validated by various experiments.

---

Keywords

---