IJMO 2024 Vol.14(1): 13-17
DOI: 10.7763/IJMO.2024.V14.843
A Novel Approach for Fault Diagnosis in Mechanical Systems Using Time-Frequency Analysis and Unsupervised Learning
Kuo-Hao Li1 and Yao-Chi Tang 2,*
1. Department of Engineering Science and Ocean Engineering, National Taiwan University, Taipei City 106319, Taiwan
2. Department of Systems Engineering and Naval Architecture, National Taiwan Ocean University, Keelung City 202301, Taiwan
Email: d10525008@ntu.edu.tw (K.-H.L.); tom@mail.ntou.edu.tw (Y-C.T.)
*Corresponding author
Manuscript received April 10, 2023; revised May 5, 2023; accepted August 23, 2023; published February 6, 2024
Abstract—The development of a new method for fault diagnosis in mechanical systems is a critical field of research due to the increasing demand for machine reliability and maintenance efficiency. In this study, a novel approach to fault diagnosis using time-frequency analysis and unsupervised learning techniques is proposed. Firstly, the proposed method converts the vibration signal into a time-frequency domain signal using the Short-Time Fourier Transform (STFT) and integrates it with respect to time to obtain the Marginal Frequency (MF). The Area Under the Frequency Curve (AUFC) is calculated and an unsupervised 1D K-means clustering algorithm is used to cluster the feature vectors. Each cluster is assigned a normal or failure state and the maximum value of the normal region is used as the threshold for failure detection. The method is tested on a set of vibration data from normal and failed bearings, and the results demonstrate the effectiveness and robustness of the proposed approach for fault detection in different bearings. The proposed approach represents a promising solution for fault diagnosis in mechanical systems, which can significantly improve the reliability and maintainability of machinery. The combination of time-frequency analysis and unsupervised learning techniques provides a powerful tool for fault detection in mechanical systems.
Keywords—fault diagnosis, K-means clustering algorithm, marginal frequency, time-frequency analysis
[PDF]
Cite: Kuo-Hao Li and Yao-Chi Tang, "A Novel Approach for Fault Diagnosis in Mechanical Systems Using Time-Frequency Analysis and Unsupervised Learning," International Journal of Modeling and Optimization, vol. 14, no. 1, pp. 13-17, 2024.
Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).
