Time-frequency ridge estimation: An effective tool for gear and bearing fault diagnosis at time-varying speeds. (15th April 2023)

• Record Type: Journal Article
• Authors: Li, Yifan; Zhang, Xin; Chen, Zaigang; Yang, Yaocheng; Geng, Changqing; Zuo, Ming J.
• Is Part Of: Mechanical systems and signal processing. Volume 189(2023)
• Publication Date: 2023-04-15
• Subjects: Instantaneous frequency -- Ridge estimation -- Tacho-less order tracking -- Fault detection -- Rotating machinery -- Instantaneous angular speed; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2023.110108
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: A novel time–frequency ridge estimation (TFRE) method is proposed. The TFRE can detect ridges with higher accuracy from more complicated TFRs. The TFRE executes automatically without parameter setting and adjustment. Abstract: For a rotary machine vibration signal collected under variable speed conditions, its time–frequency representation (TFR) contains abundant oscillatory components with time-varying amplitudes and frequencies. A single component with a sequence of peaks in the TFR is called a ridge. Accurate ridge detection from TFRs can boost rotary machine health condition assessment without rotation speed measurement. Nowadays, cost function ridge estimation and fast path optimization ridge estimation are the most widely utilized techniques. However, the unreasonable kernel function definitions and inappropriate search region selections significantly restrict the performance of instantaneous frequency estimation of target ridges. To address the deficiencies, this paper proposes a novel time–frequency ridge estimation (TFRE) method. The TFRE integrates a new cost kernel function and an adaptive search region detection principle. For the former, it comprehensively considers the trade-off between seeking peaks and ensuring the smoothness of a ridge. The latter varies the search bandwidth in real-time according to instantaneous signal signatures to effectively isolate interferences and neighboring ridges. A unique advantage of the proposed method is that it dispenses with the tuning of parameters. As a consequence, human intervention is minimized. Experimental gear and bearing vibration signals were analyzed to demonstrate the performance of the TFRE. Results indicated that the proposed TFRE is characterized by superior ridge estimation performance compared to the state-of-the-art methods.