1 Key Laboratory of Earthquake Engineering and Engineering Vibration, Institute of Engineering Mechanics, China Earthquake Administration, Harbin 150080, China
2 Laboratory of Earthquake Disaster Mitigation, Ministry of Emergency Management, Harbin 150080, China
3 College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, China
In order to achieve rapid prediction of structural seismic responses for emergency decision-making while addressing the high cost of acquiring target-specific data and the complexity of cross-scale prediction, a cross-scale transfer learning framework based on scaled-model data was used to investigate full-scale seismic response prediction for a 3D-printed concrete residential structure, and a hybrid GAE-Transformer-CNN model integrating graph attention mechanisms and positional encoding was developed, in which sensor spatial priors and structural topology were explicitly incorporated through a graph-aware module, long-range dependencies were captured by a Transformer, and local features were extracted by a CNN. The results show that after pre-training on the scaled model, fine-tuning was performed using only a small number of full-scale samples, and an average R2 of 0.929 was achieved on an independent test set, with the maximum interstory drift ratio at critical locations being accurately predicted from ground-acceleration inputs and sparse sensor observations. It is concluded that scaled experimental data can be effectively leveraged to support data-efficient and accurate cross-scale seismic response prediction, thereby providing a practical and scalable methodology for structural response assessment.
ZHANG Xin-yuan, WANG Zi-fa, SUN Xiao-yan, ZHAO Deng-ke, SHI Yi-qing, LI Zhao-yan.
Prediction of Structural Seismic Response Based on GAE-Transformer-CNN Hybrid Model and Transfer Learning[J].
Science Technology and Engineering, 2026, 26(21): 9002-9013 DOI:10.12404/j.issn.1671-1815.2507017
研究对象为河北省衡水市故城县一栋3D打印混凝土民居,原型建筑尺寸为21.3 m × 6.3 m × 4.6 m。按照1∶4.6的几何比例建立缩尺模型,并基于相似理论推导主要物理量关系(表1),其结果符合动力学相似规律。图4展示了足尺与缩尺模型的现场实景,两者在建模假设与构件参数上保持一致,为后续研究提供了可靠基础。
FanJiansheng, WangChen, SongLinghan. Research progress and applications of intelligent computational analysis in civil engineering[J]. Journal of Building Structures, 2022, 43(9): 1-22.
[3]
BaoY, ChenZ, WeiS, et al. The state of the art of data science and engineering in structural health monitoring[J]. Engineering, 2019, 5(2): 234-242.
[4]
ChaY J, AliR, LewisJ, et al. Deep learning-based structural health monitoring[J]. Automation in Construction, 2024, 161: 105328.
[5]
MorfidisK, KostinakisK. Approaches to the rapid seismic damage prediction of r/c buildings using artificial neural networks[J]. Engineering Structures, 2018, 165: 120-141.
ZhangChenyu. Rapid resilience assessment of rc frame structures based on deep learning under incomplete information[D]. Harbin: Harbin Institute of Technology, 2024.
[10]
CaoY C, ZhengJ F, XiongC. Prediction of inter-story drift ratio limits for reinforced concrete frame structures based on XGBoost[C]// Proceedings of the 2024 Eurasian OpenSees Days. Cham: Springer Nature Switzerland, 2025: 28-37.
[11]
LiaoQ, BaoY, HuH, et al. Mechanics-informed transformer-GCN for structural dynamic response prediction[J]. Engineering Structures, 2025, 325: 119470.
[12]
LiuC, LiangZ, LiuH, et al. Seismic performance of 3D printed reinforced concrete walls: experimental study and numerical simulation[J]. Engineering Structures, 2025, 333: 120143.
KangHao, GuoZixiong, LiuYang, et al. A CNN-based filtering method considering the influence of time-frequency characteristics of seismic waves[J]. Engineering Mechanics, 2025, 42(11): 103-114.
[15]
LiaoW, ChenX, LuX, et al. Deep transfer learning and time-frequency characteristics-based identification method for structural seismic response[J]. Frontiers in Built Environment, 2021, 7: 627058.
[16]
HuY, GuoW, XuZ, et al. Physics knowledge-based transfer learning between buildings for seismic response prediction[J]. Soil Dynamics and Earthquake Engineering, 2024, 177: 108420.
[17]
HaoS. Advanced machine learning approaches for force reconstruction and state estimation in structural dynamics[D]. Hong Kong: Hong Kong Polytechnic University, 2025.
[18]
ShenY, MaG, HwangH J, et al. Prediction of seismic response of building structures using a CNN-LSTM-ATT network with transfer learning[J]. Advances in Structural Engineering, 2025, 28(14): 2710-2725.
[19]
VeličkovićP, CucurullG, CasanovaA, et al. Graphattention networksC// TheProceedings of the 6th International Conference on Learning Representations ( ICLR). Vancouver: ICLR, 2018: 1-12.
[20]
LiP, WangY, WangH, et al. Distance encoding: design provably more powerful neural networks for graph representation learning[C]// The Proceedings of the 34th International Conference on Neural Information Processing Systems. Vancouver: NeurIPS, 2020: 4465-4478.
[21]
PokH, GermainS P. A real-time structural seismic response prediction framework based on transfer learning and unsupervised learning[J]. Engineering Structures, 2025, 323: 139237.
[22]
ZhangW, YanB, YiW. Research on multi-model prediction of skeleton curves of prefabricated concrete columns based on residual fusion long short-term memory-Transformer[J]. Journal of Building Engineering, 2023, 79: 107821.
ZhouFeng, WangShipu, ZhangKunpeng. Short-term traffic flow prediction model for urban road network based on graph transformer network[J]. Science Technology and Engineering, 2024, 24(10): 4307-4316.
FuHan, FanWei, LiuHaiwen, et al. Aircraft turbulence prediction method based on CNN-Transformer-BiLSTM[J]. Science Technology and Engineering, 2025, 25(33): 14363-14370.
YeWanrui, LiuZhe, SuYang, et al. Leakage fault diagnosis of heating pipe networks based on convolutional neural network and Transformer[J]. Science Technology and Engineering, 2025, 25(13): 5589-5597.
[29]
SuM, WanJ, ZhouQ, et al. Utilizing pretrained convolutional neural networks for crack detection and geometric feature recognition in concrete surface images[J]. Journal of Building Engineering, 2024, 98: 111386.
[30]
FengD C, YiX, DegerZ T, et al. Rapid post-earthquake damage assessment of building portfolios through deep learning-based component-level image recognition[J]. Journal of Building Engineering, 2024, 98: 111380.
ZhangXuanxuan, HanXiaolei, ShiYuanyuan, et al. Crack detection using a deeply supervised network integrated with multi-scale and multi-channel features[J]. Science Technology and Engineering, 2024, 24(13): 5595-5603.
ZhouZewen, ZhongZiting, ZhaiMusai, et al. Application of novelty detection technology based on convolutional neural network in structural damage identification[J]. Science Technology and Engineering, 2024, 24(21): 9069-9076.
[35]
MehtabS, SenJ. Analysis and forecasting of financial time series using CNN and LSTM-based deep learning models[C]// Advances in Distributed Computing and Machine Learning. Singapore: Springer, 2022: 405-423.
[36]
LinC, WengK, LinY, et al. Time series prediction of dam deformation using a hybrid STL-CNN-GRU model based on sparrow search algorithm optimization[J]. Applied Sciences, 2022, 12(23): 11951.
[37]
HeY, YanH, YangW, et al. Time-series analysis and prediction of surface deformation in the Jinchuan mining area, Gansu Province, by using InSAR and CNN-PhLSTM network[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022, 15: 6732-6751.
[38]
PanS J, YangQ. A survey on transfer learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2010, 22(10): 1345-1359.
WangErcheng, XiaoJunwei, LiJiahao, et al. Structural damage identification based on Inception-BiLSTM and transfer learning[J]. Science Technology and Engineering, 2024, 24(18): 7776-7784.
Ministry of Housing and Urban-Rural Development of the People’s Republic of China. Code for seismic design of buildings: partial revised provisions (2024 Edition) of GB 50011— 2010 (2016 Edition)[S]. Beijing: Standards Press of China, 2024.
[43]
DaiH, ZhouY, LiuH, et al. XGBoost-based prediction of on-site acceleration response spectra with multi-feature inputs from P-wave arrivals[J]. Soil Dynamics and Earthquake Engineering, 2024, 178: 108503.
YazdanpanahO, ParkM, ChangM, et al. Mastering seismic time series response predictions using an attention-Mamba transformer model for bridge bearings and piers across varied testing conditions[J]. Scientific Reports, 2024, 14(1): 29751.
[46]
YanL, LiuP, YangF, et al. Enhanced convolutional neural network-transformer framework for accurate prediction of the flexural capacity of ultra-high-performance concrete beams[J]. Buildings, 2025, 15(17): 3138.
[47]
XieY. Deep learning in earthquake engineering: a comprehensive review[J]. ASCE OPEN: Multidisciplinary Journal of Civil Engineering, 2025, 3(1): 03125001.