Adaptive parameter blind source separation technique for wheel condition monitoring. (June 2017)

• Record Type: Journal Article
• Authors: Zhang, Jie; Gao, Hongli; Liu, Qiyue; Farzadpour, F.; Grebe, C.; Tian, Ying
• Is Part Of: Mechanical systems and signal processing. Volume 90(2017)
• Page Start: 208
• Page End: 221
• Publication Date: 2017-06
• Subjects: Condition monitoring -- Blind source separation -- Wheel defect -- Non-stationary signal -- Adaptive time-frequency distributions; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2016.12.021
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: An adaptive parameter BSS for non-stationary signal processing is proposed. Simulations of frequency-varying and time-varying non-stationary illustrate its accuracy. Condition monitoring for railway wheels illustrate its effectiveness under varying fault complexity. Abstract: Wheel condition monitoring has played a key role in the safe operation of railway vehicles. Blind source separation (BSS) is an attractive tool due to its excellent performance in separating source signals from their mixtures when no detailed knowledge of defective sources and the mixing process is assumed. In this paper, we propose an adaptive parameter BSS approach based on the adaptive time-frequency distributions theory in order to deal with the non-stationary blind separation problem and apply it to wheel defect monitoring. Some classical time-frequency signal analysis and BSS methods are applied in comparison with the proposed approach through frequency-varying non-stationary and time-varying non-stationary simulations. Experiments of single and multi-fault wheels have been conducted using the wheel/rail simulation facility to illustrate the effectiveness of the proposed method in processing the non-stationary signals with varying fault complexity.