Acoustic-vibration baseline for heavy-haul railways using DM-MFOIS fiber sensing technology
Year: 2026
Type: article
Abstract: Aiming at the monitoring needs of heavy-haul railway infrastructure, this study established an off-track, non-contact monitoring system based on Dual-measuring-arm Michelson Fiber Optic Interferometric Sensing (DM-MFOIS) technology, enabling real-time acquisition of train-induced acoustic-vibration signals and baseline construction. However, systematic baselines for different track sections—such as tangent sections, curves, and bridges—using such sensing remain limited. To address this gap, the STA/LTA method was employed for automatic signal segmentation and denoising, while train positioning was achieved by analyzing spectral centroid variations, with an error range of 0.4–0.8 seconds. By applying S-transform and multi-dimensional features, the response characteristics of tangent sections, small-radius curves, and bridge sections under both heavy and light train passages were systematically analyzed. The results indicate: (i) In curved sections, the locomotive position has a significant influence, with distinct impact signals observed at characteristic points. (ii) Bridge sections exhibit intense signal responses; the RMS surges approximately six-fold for light trains and twelve-fold for heavy trains when crossing bridges, with the most pronounced impact at the track-bridge transition zone. (iii) In the case of rail emergency clamp damage, features such as energy concentration at 471 Hz, a spectral centroid valley, and a 7.6% increase in RMS were identified at the damaged location, validating the baseline’s detection capability. The established acoustic-vibration baseline can provide a basis for anomaly warning and intelligent maintenance of railway infrastructure.
DOI: 10.1093/iti/liag004
Authors: Zheng Li, Jieling Xiao, Xi Yang, Jinyuan Cao, Xiyun Wang
Institution: Southwest Jiaotong University
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7140606432