Title: Machine learning based progressive crack fault monitoring on spur gear using vibration analysis

Authors: Manoj Kumar Gangwar; Neelesh Kumar Sahu; Rajendra Kumar Shukla; Chitresh Nayak

Addresses: Mechanical Engineering Department, Medi-Caps University, Indore, MP, India ' Mechanical Engineering Department, Medi-Caps University, Indore, MP, India; Mechanical Engineering Department, Gyan Ganga Institute of Technology and Science, Jabalpur, MP, India ' Mechanical Engineering Department, Medi-Caps University, Indore, MP, India ' Mechanical Engineering Department, Medi-Caps University, Indore, MP, India

Abstract: This paper describes ability to diagnose progressive crack fault on spur gear using signal processing and machine learning (ML) techniques. Experiments are performed on three different conditions of spur gear, healthy as well as crack at tooth root (50%, 90%). Time-domain and frequency-domain signal processing methods as well as machine learning techniques have been used to process and analyse the acquired signals. This study is motivated by the process of spur gear fault diagnosis by machine learning algorithms as J48 decision tree and support vector machine (SVM). Noise level is also considered during the meshing of gears. The results of this investigation revealed that J48 decision tree outperforms from SVM with 93.3% accuracy. It has been noted that both sides' signals and noise levels must be analysed in order to improve gear health condition monitoring. The proposed method can be used to diagnose the fault on different gears and other elements.

Keywords: spur gear; fault diagnosis; condition monitoring; signal processing and machine learning; support vector machine; SVM.

DOI: 10.1504/IJVNV.2024.138126

International Journal of Vehicle Noise and Vibration, 2024 Vol.20 No.1, pp.89 - 106

Received: 23 Aug 2023
Accepted: 28 Dec 2023
Published online: 29 Apr 2024 *

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