Prediction of Structural Displacement From Acceleration Based on Improved Long Short‐Term Memory Networks
Structural Control and Health Monitoring · 2025-01 · 2025 (1)
DOI: 10.1155/stc/2290381 ↗Kurz gefasst
Fachartikel in der Zeitschrift Structural Control and Health Monitoring, online erschienen am 24.11.2025. Verfasst von Junkai Shen und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Accurate estimation of structural displacement is crucial for assessing the damage state of structures. However, displacement sensors are rarely deployed on practical structures. Moreover, the traditional integration method based on acceleration data entails sophisticated processing steps, such as multiple filtering and baseline correction procedures. Therefore, an end‐to‐end displacement prediction method based on the long short‐term memory (LSTM) model integrated with an attention mechanism is presented herein to mitigate the limitations. The proposed model can predict structural displacement using noisy acceleration data acquired from accelerometer sensors mounted on structures. The overlapping window technique was employed to augment the dataset and enhance training efficiency. Furthermore, an attention mechanism was incorporated to improve prediction accuracy. When compared with the traditional integration method and the LSTM model without the attention mechanism, the proposed Attention‐LSTM model exhibited superior performance on the testing dataset. Additionally, the effects of key parameters were systematically investigated, such as the normalization method, cropping window size, number of LSTM layers, and number of units per layer. It is demonstrated that the proposed Attention‐LSTM model possesses robust generalization capability and antinoise performance, which contributes to the reliable estimation of structural displacement in engineering structures.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Shen, J., Zhang, L., & Zhu, B. (2025). Prediction of Structural Displacement From Acceleration Based on Improved Long Short‐Term Memory Networks. Structural Control and Health Monitoring, 2025(1). https://doi.org/10.1155/stc/2290381
Vancouver
Shen J, Zhang L, Zhu B. Prediction of Structural Displacement From Acceleration Based on Improved Long Short‐Term Memory Networks. Structural Control and Health Monitoring. 2025;2025(1). doi:10.1155/stc/2290381
BibTeX
@article{Shen2025,
author = {Shen, Junkai and Zhang, Lingxin and Zhu, Baijie},
title = {Prediction of Structural Displacement From Acceleration Based on Improved Long Short‐Term Memory Networks},
journal = {Structural Control and Health Monitoring},
year = {2025},
volume = {2025},
number = {1},
doi = {10.1155/stc/2290381}
}
EUVIMED