Vertical track irregularity identification based on a detrending numerical integration algorithm with vehicle acceleration measurements
•A novel algorithm embedding automatic detrending in numerical integration is developed.•Parameter-adaptive VMD is designed for nonstationary acceleration signal decomposition.•Integration errors and nonzero initial conditions are compensated through optimization.•Track irregularities are identified...
Saved in:
| Published in: | Mechanical systems and signal processing Vol. 235; p. 112958 |
|---|---|
| Main Authors: | , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Elsevier Ltd
15.07.2025
|
| Subjects: | |
| ISSN: | 0888-3270 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | •A novel algorithm embedding automatic detrending in numerical integration is developed.•Parameter-adaptive VMD is designed for nonstationary acceleration signal decomposition.•Integration errors and nonzero initial conditions are compensated through optimization.•Track irregularities are identified using axle box or bogie acceleration measurements.•Identified track irregularities provide multi-resolution adaptive representation.
A persistent challenge in acceleration-based methods for identifying track irregularities lies in the trend issue that arises during measurement and numerical integration, which has yet to be effectively and adaptively addressed. This paper develops a novel detrending numerical integration algorithm that embeds the detrending process directly into the integration procedure. The algorithm combines parameter-adaptive variational mode decomposition and progressive filtering to remove trends present in the original acceleration signals. Numerical integration formulae with four variable parameters are designed to compensate for integration errors and nonzero initial conditions, with parameter optimization achieved through a particle swarm algorithm. Furthermore, the paper proposes a method for identifying vertical track irregularities using axle box or bogie accelerations based on the detrending integration algorithm, with explicit formulae derived for time-domain estimation based on the inverse dynamic model. The proposed algorithm and method are validated through numerical cases and field experiments, demonstrating that the identified vertical track irregularities not only achieve high accuracy but also offer multi-resolution adaptive representation, with the potential to enhance the detection and localization of track faults. |
|---|---|
| ISSN: | 0888-3270 |
| DOI: | 10.1016/j.ymssp.2025.112958 |