Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention Mechanism

With the exponential growth of engineering monitoring data, data-driven neural networks have gained widespread application in predicting retaining structure deformation in foundation pit engineering. However, existing models often overlook the spatial deflection correlations among monitoring points....

Full description

Saved in:
Bibliographic Details
Main Authors: Yanyong Gao, Zhaoyun Xiao, Zhiqun Gong, Shanjing Huang, Haojie Zhu
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Buildings
Subjects:
Online Access:https://www.mdpi.com/2075-5309/15/14/2537
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849246472992718848
author Yanyong Gao
Zhaoyun Xiao
Zhiqun Gong
Shanjing Huang
Haojie Zhu
author_facet Yanyong Gao
Zhaoyun Xiao
Zhiqun Gong
Shanjing Huang
Haojie Zhu
author_sort Yanyong Gao
collection DOAJ
description With the exponential growth of engineering monitoring data, data-driven neural networks have gained widespread application in predicting retaining structure deformation in foundation pit engineering. However, existing models often overlook the spatial deflection correlations among monitoring points. Therefore, this study proposes a novel deep learning framework, CGCA (Convolutional Gated Recurrent Unit with Cross-Attention), which integrates ConvGRU and cross-attention mechanisms. The model achieves spatio-temporal feature extraction and deformation prediction via an encoder–decoder architecture. Specifically, the convolutional structure captures spatial dependencies between monitoring points, while the recurrent unit extracts time-series characteristics of deformation. A cross-attention mechanism is introduced to dynamically weight the interactions between spatial and temporal data. Additionally, the model incorporates multi-dimensional inputs, including full-depth inclinometer measurements, construction parameters, and tube burial depths. The optimization strategy combines AdamW and Lookahead to enhance training stability and generalization capability in geotechnical engineering scenarios. Case studies of foundation pit engineering demonstrate that the CGCA model exhibits superior performance and robust generalization capabilities. When validated against standard section (CX1) and complex working condition (CX2) datasets involving adjacent bridge structures, the model achieves determination coefficients (R<sup>2</sup>) of 0.996 and 0.994, respectively. The root mean square error (RMSE) remains below 0.44 mm, while the mean absolute error (MAE) is less than 0.36 mm. Comparative experiments confirm the effectiveness of the proposed model architecture and the optimization strategy. This framework offers an efficient and reliable technical solution for deformation early warning and dynamic decision-making in foundation pit engineering.
format Article
id doaj-art-2460841efad34ee98d975e8f87799cc2
institution Kabale University
issn 2075-5309
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Buildings
spelling doaj-art-2460841efad34ee98d975e8f87799cc22025-08-20T03:58:30ZengMDPI AGBuildings2075-53092025-07-011514253710.3390/buildings15142537Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention MechanismYanyong Gao0Zhaoyun Xiao1Zhiqun Gong2Shanjing Huang3Haojie Zhu4College of Civil Engineering, Huaqiao University, Quanzhou 361021, ChinaCollege of Civil Engineering, Huaqiao University, Quanzhou 361021, ChinaChina Construction Infrastructure Co., Ltd., Beijing 100044, ChinaChina Civil Engineering (Xiamen) Technology Co., Ltd., Xiamen 361000, ChinaChina Civil Engineering (Xiamen) Technology Co., Ltd., Xiamen 361000, ChinaWith the exponential growth of engineering monitoring data, data-driven neural networks have gained widespread application in predicting retaining structure deformation in foundation pit engineering. However, existing models often overlook the spatial deflection correlations among monitoring points. Therefore, this study proposes a novel deep learning framework, CGCA (Convolutional Gated Recurrent Unit with Cross-Attention), which integrates ConvGRU and cross-attention mechanisms. The model achieves spatio-temporal feature extraction and deformation prediction via an encoder–decoder architecture. Specifically, the convolutional structure captures spatial dependencies between monitoring points, while the recurrent unit extracts time-series characteristics of deformation. A cross-attention mechanism is introduced to dynamically weight the interactions between spatial and temporal data. Additionally, the model incorporates multi-dimensional inputs, including full-depth inclinometer measurements, construction parameters, and tube burial depths. The optimization strategy combines AdamW and Lookahead to enhance training stability and generalization capability in geotechnical engineering scenarios. Case studies of foundation pit engineering demonstrate that the CGCA model exhibits superior performance and robust generalization capabilities. When validated against standard section (CX1) and complex working condition (CX2) datasets involving adjacent bridge structures, the model achieves determination coefficients (R<sup>2</sup>) of 0.996 and 0.994, respectively. The root mean square error (RMSE) remains below 0.44 mm, while the mean absolute error (MAE) is less than 0.36 mm. Comparative experiments confirm the effectiveness of the proposed model architecture and the optimization strategy. This framework offers an efficient and reliable technical solution for deformation early warning and dynamic decision-making in foundation pit engineering.https://www.mdpi.com/2075-5309/15/14/2537deep foundation pitdeformation predictiondeep learningattention mechanismconvolutiongated recursive unit
spellingShingle Yanyong Gao
Zhaoyun Xiao
Zhiqun Gong
Shanjing Huang
Haojie Zhu
Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention Mechanism
Buildings
deep foundation pit
deformation prediction
deep learning
attention mechanism
convolution
gated recursive unit
title Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention Mechanism
title_full Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention Mechanism
title_fullStr Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention Mechanism
title_full_unstemmed Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention Mechanism
title_short Spatiotemporal Deformation Prediction Model for Retaining Structures Integrating ConvGRU and Cross-Attention Mechanism
title_sort spatiotemporal deformation prediction model for retaining structures integrating convgru and cross attention mechanism
topic deep foundation pit
deformation prediction
deep learning
attention mechanism
convolution
gated recursive unit
url https://www.mdpi.com/2075-5309/15/14/2537
work_keys_str_mv AT yanyonggao spatiotemporaldeformationpredictionmodelforretainingstructuresintegratingconvgruandcrossattentionmechanism
AT zhaoyunxiao spatiotemporaldeformationpredictionmodelforretainingstructuresintegratingconvgruandcrossattentionmechanism
AT zhiqungong spatiotemporaldeformationpredictionmodelforretainingstructuresintegratingconvgruandcrossattentionmechanism
AT shanjinghuang spatiotemporaldeformationpredictionmodelforretainingstructuresintegratingconvgruandcrossattentionmechanism
AT haojiezhu spatiotemporaldeformationpredictionmodelforretainingstructuresintegratingconvgruandcrossattentionmechanism