Spatiotemporal Compression of Structural Vibration Data Using a Bayesian Reconstruction Framework with Active Sensor Placement
Civil Engineering Sciences · 2026
DOI: 10.34133/cesci.0034 ↗Kurz gefasst
Fachartikel in der Zeitschrift Civil Engineering Sciences, 2026. Verfasst von Linfeng Zuo und 4 weiteren Autor:innen. Die Arbeit nennt 59 Literaturangaben. Gefördert von: Key Technologies Research and Development Program.
Abstract (Originalsprache)
This study presents a Bayesian strategy for spatiotemporal compression of structural vibration data in health monitoring (SHM) systems. Conventional dense sensor arrays and high-frequency sampling produce massively redundant data in both space and time, imposing severe burdens on storage, transmission, and processing. To mitigate this redundancy, we propose an integrated framework comprising 3 main components. First, an active learning-based optimal sensor placement method is developed within a Bayesian full-field reconstruction model, which sequentially selects sensor locations by minimizing the reconstruction error, thereby achieving spatial compression. Second, temporal compression is realized by directly downsampling the measurements, while retaining high-rate data only at a few key locations. Finally, the target high-resolution response is recovered from the multirate observations via their joint Gaussian distribution, whose covariance matrix is constructed using a mode-weighted multiperiodic kernel function. Numerical simulations on a cantilever beam and experimental validation on a steel plate demonstrate that the proposed approach attains high-efficiency spatiotemporal compression without sacrificing reconstruction accuracy, offering a robust and cost-effective solution for SHM data reduction.
Förderung
- Key Technologies Research and Development Program
Zitieren
APA
Zuo, L., Lu, Z., Zhou, M., Deng, X., & Wang, L. (2026). Spatiotemporal Compression of Structural Vibration Data Using a Bayesian Reconstruction Framework with Active Sensor Placement. Civil Engineering Sciences. https://doi.org/10.34133/cesci.0034
Vancouver
Zuo L, Lu Z, Zhou M, Deng X, Wang L. Spatiotemporal Compression of Structural Vibration Data Using a Bayesian Reconstruction Framework with Active Sensor Placement. Civil Engineering Sciences. 2026. doi:10.34133/cesci.0034
BibTeX
@article{Zuo2026,
author = {Zuo, Linfeng and Lu, Zhong-Rong and Zhou, Minghao and Deng, Xiangbin and Wang, Li},
title = {Spatiotemporal Compression of Structural Vibration Data Using a Bayesian Reconstruction Framework with Active Sensor Placement},
journal = {Civil Engineering Sciences},
year = {2026},
doi = {10.34133/cesci.0034}
}
EUVIMED