A new interpretable dynamic ensemble learning model with sample augmentation for predicting seismic responses of vehicle-track-bridge systems

Abstract

Ensuring the operational safety of high-speed trains during earthquakes is a core challenge for China’s extensive high-speed rail network. While machine learning (ML) -based seismic response assessment has become a mainstream approach, conventional ML methods suffer from limitations such as heavy training data demands, poor interpretability, and over-reliance on deterministic predictions. This study proposes an interpretable dynamic ensemble learning model integrated with sample augmentation to predict extreme seismic responses of vehicle-track-bridge (VTB) systems. The framework combines Generative Adversarial Networks (GAN) for data generation, the Kepler Optimization Algorithm (KOA)—chosen for its superior convergence speed and optimization performance over classical algorithms—for hyperparameter tuning, and a dynamically weighted ensemble of Long Short-Term Memory (LSTM)-Attention and Support Vector Machine (SVM). A 3D nonlinear VTB model under bidirectional seismic excitation serves as the physical basis, with GAN-based augmentation mitigating data imbalance. Comprehensive validation against traditional ML models confirms significant accuracy gains, marked by reduced Mean Absolute Error (MAE) and coefficient of determination (R²) values consistently exceeding 0.97. SHapley Additive exPlanation (SHAP) analysis identifies key input features affecting wheel-rail interaction parameters, and Gaussian probabilistic interval prediction quantifies predictive uncertainty with adaptive confidence bounds. The findings offer references for seismic prediction and safety risk assessment of high-speed railways.

Keywords

ensemble learning model / generative adversarial networks / seismic response prediction / SHAP analysis / probabilistic interval prediction / vehicle-track-bridge system

References

[1] Jiang L-z, Zhang Y-t, Feng Y-l, et al. . Simplified calculation modeling method of multi-span bridges on high-speed railways under earthquake condition [J]. Bulletin of Earthquake Engineering, 2020, 18(5): 2303-2328

[2] Feng Y-l, Jiang L-z, Zhou W-bao. Dynamic response of a three-beam system with intermediate elastic connections under a moving load/mass-spring [J]. Earthquake Engineering and Engineering Vibration, 2020, 19(2): 377-395

[3] Zhang Y-t, Jiang L-z, Zhou W-b, et al. . Study of bridge-subgrade longitudinal constraint range for high-speed railway simply-supported beam bridge with CRTSII ballastless track under earthquake excitation [J]. Construction and Building Materials, 2020, 241: 118026

[4] LAI Zhi-peng, KANG Xin, JIANG Li-zhong, et al. Earthquake influence on the rail irregularity on high-speed railway bridge [J]. Shock and Vibration, 2020: 4315304. DOI:https://doi.org/10.1155/2020/4315304.

[5] Guo W, He X-e, Jiang L-z, et al. . The effect of high-speed railway train on the seismic response of bridge under earthquakes [J]. Engineering Structures, 2024, 303: 117446

[6] Wei B, Xiao B-q, Hu Z-l, et al. . Damage control analysis of components in high-speed railway bridgetrack system based on combined seismic isolation design under earthquake [J]. Structures, 2023, 54: 515-526

[7] Zhou W-b, Ren Z-b, Liu S-h, et al. . Simplified and rapid prediction of earthquake-induced track dynamic irregularity of high-speed railway bridges under different site conditions [J]. Soil Dynamics and Earthquake Engineering, 2024, 179: 108563

[8] Kang X, Jiang L-z, Bai Y, et al. . Seismic damage evaluation of high-speed railway bridge components under different intensities of earthquake excitations [J]. Engineering Structures, 2017, 152: 116-128

[9] Connolly D P, Kouroussis G, Woodward P K, et al. . Field testing and analysis of high speed rail vibrations [J]. Soil Dynamics and Earthquake Engineering, 2014, 67: 102-118

[10] Panda J P. Machine learning for naval architecture, ocean and marine engineering [J]. Journal of Marine Science and Technology, 2023, 28(1): 914

[11] Tian Q-h, Li J-l, Ran G-t, et al. . Path planning based on improved deep Q-network algorithm for bionic robotic fish with ocean currents [J]. Neurocomputing, 2025, 653: 131173

[12] Chen Z-y, Zhang Y-p, Li J-b, et al. . Diagnosing tunnel collapse sections based on TBM tunneling big data and deep learning: A case study on the Yinsong Project, China [J]. Tunnelling and Underground Space Technology, 2021, 108: 103700

[13] Zhao C-y, Mahmoudi E, Luo M-m, et al. . Unfavorable geology recognition in front of shallow tunnel face using machine learning [J]. Computers and Geotechnics, 2023, 157: 105313

[14] Lazaridis P C, Kavvadias I E, Demertzis K, et al. . Structural damage prediction of a reinforced concrete frame under single and multiple seismic events using machine learning algorithms [J]. Applied Sciences, 2022, 12(8): 3845

[15] DEMIR A, DEMIR S, SAHIN E K. Machine learning based prediction of peak floor acceleration in low- to mid-rise RC buildings using ground motion intensity measures [J]. Iranian Journal of Science and Technology, Transactions of Civil Engineering, 2025. DOI:https://doi.org/10.1007/s40996-025-02006-x.

[16] Wang Z-w, Lu X-f, Zhang W-m, et al. . Deep learning-based reconstruction of missing long-term girder-end displacement data for suspension bridge health monitoring [J]. Computers & Structures, 2023, 284: 107070

[17] Yang J-x, Yang F, Zhang L-k, et al. . Bridge health anomaly detection using deep support vector data description [J]. Neurocomputing, 2021, 444: 170-178

[18] Ji Y-j, Huang Y-p, Yang M, et al. . Physics-informed deep learning for virtual rail train trajectory following control [J]. Reliability Engineering & System Safety, 2025, 261: 111092

[19] Han H-g, Zhang Y-b, Huang Y-ting. Collision-free motion-constrained path planning for multiple Unmanned Delivery Vehicles based on heuristic deep reinforcement learning [J]. Neurocomputing, 2025, 648: 130586

[20] Dupuis M, Schill C, Lee R, et al. . A deep-learning-based model for quality assessment of earthquake-induced ground-motion records [J]. Earthquake Spectra, 2023, 39(4): 2492-2517

[21] Wang Y-w, Zhao Q-x, Qian K, et al. . Cumulative absolute velocity prediction for earthquake early warning with deep learning [J]. Computer-Aided Civil and Infrastructure Engineering, 2024, 39(11): 1724-1740

[22] He X-h, Zhao Y-s, Shi K, et al. . A deep learning model for predicting the stochastic responses of vehicle-bridge system based on temporal convolutional network and Gaussian process [J]. Engineering Structures, 2025, 328: 119745

[23] Yang M-x, Yi R, Zhu S-yu. Reliability analysis of the vehicle-bridge system with uncertain parameters by SPEM integrating deep leaning method [J]. Structures, 2023, 58: 105535

[24] Liao Y-c, Lin R, Zhang R-y, et al. . Attention-based LSTM (AttLSTM) neural network for seismic response modeling of bridges [J]. Computers & Structures, 2023, 275: 106915

[25] Peng K, Jiang L-z, Zhou W-b, et al. . A seismic response prediction method based on a self-optimized Bayesian Bi-LSTM mixed network for high-speed railway track-bridge system [J]. Journal of Central South University, 2024, 31(3): 965-975

[26] Peng K, Zhou W-b, Jiang L-z, et al. . Multimodal fusion hybrid neural network approach for multi-class damage classification in high-speed rail track-bridge systems with multi-parameter [J]. Engineering Structures, 2025, 328: 119710

[27] Wu L-x, Zhou W-b, Zhong T-x, et al. . Uncertainty quantification in predicting seismic response of high-speed railway simply-supported bridge system based on bootstrap [J]. Reliability Engineering & System Safety, 2025, 260: 111006

[28] Liu H-y, Wang Z-y, Han Y, et al. . RIME-VMDBiLSTM: A surrogate model for seismic response prediction of nonlinear vehicle-track-bridge system [J]. Journal of Central South University, 2025, 32(10): 4073-4091

[29] Peng K, Lai Z-p, Jiang L-z, et al. . A high-precision surrogate model for seismic vehicle-track-bridge system based on hybrid deep learning for nonlinear structural restoring forces [J]. Computers & Structures, 2025, 316: 107870

[30] MATHWORKS. MATLAB [CP]. Natick: MathWorks.

[31] Nguyen V T, Ahn J H, Haldar A, et al. . Fragility-based seismic performance assessment of modular underground arch bridges [J]. Structures, 2022, 39: 1218-1230

[32] GB 50111-2006. Standard for seismic design of railway engineering [S].

[33] GB/T 50011-2010. Standard for seismic design of buildings [S].

[34] Chen Y-j, Jiang L-z, Lai Z-p, et al. . A novel evaluation of train running safety on the HSR bridges under earthquakes based on the observed track deformation rate [J]. Engineering Structures, 2023, 275: 115260

[35] Kalker J J. A fast algorithm for the simplified theory of rolling contact [J]. Vehicle System Dynamics, 1982, 11(1): 1-13

[36] Shen Z Y, Hedrick J K, Elkins J A. A comparison of alternative creep force models for rail vehicle dynamic analysis [J]. Vehicle System Dynamics, 1983, 12(1–3): 79-83

[37] Lai Z-p, Jiang L-z, Chen Y-j, et al. . An improved model for the nonlinear simulation of the highspeed vehicle-track-bridge coupling system under seismic shaking [J]. Earthquake Engineering & Structural Dynamics, 2024, 53(12): 3775-3797

[38] Goodfellow I J, Pouget-Abadie J, Mirza M, et al. . Generative adversarial nets [J]. NeurIPS, 2014, 27: 2672-2680 Arxiv:1406.2661

[39] Likas A, Vlassis N, J Verbeek J. The global k-means clustering algorithm [J]. Pattern Recognition, 2003, 36(2): 451-461

[40] Abdel-Basset M, Mohamed R, Abdel Azeem S A, et al. . Kepler optimization algorithm: A new metaheuristic algorithm inspired by Kepler’s laws of planetary motion [J]. Knowledge-Based Systems, 2023, 268: 110454

[41] Dong X-b, Yu Z-w, Cao W-m, et al. . A survey on ensemble learning [J]. Frontiers of Computer Science, 2020, 14(2): 241-258

[42] Hochreiter S, Schmidhuber J. Long short-term memory [J]. Neural Computation, 1997, 9(8): 1735-1780

[43] Chen Y-j, Sun Z-y, Zhang R-y, et al. . Attention mechanism based neural networks for structural post-earthquake damage state prediction and rapid fragility analysis [J]. Computers & Structures, 2023, 281: 107038

[44] SRIVASTAVA N, HINTON G E, KRIZHEVSKY A, et al. Dropout: A simple way to prevent neural networks from overfitting [J]. Journal of Machine Learning Research, 15(1): 1929–1958.

[45] Chandra M A, Bedi S S. Survey on SVM and their application in image classification [J]. International Journal of Information Technology, 2021, 13(5): 80

[46] Gautam D, Bhattarai A, Rupakhety R. Machine learning and soft voting ensemble classification for earthquake induced damage to bridges [J]. Engineering Structures, 2024, 303: 117534

[47] Yang Y-d, Li S-f, Li W-q, et al. . Power load probability density forecasting using Gaussian process quantile regression [J]. Applied Energy, 2018, 213: 499-509

RIGHTS & PERMISSIONS

Central South University

Just Accepted

This article has successfully passed peer review and final editorial review, and will soon enter typesetting, proofreading and other publishing processes. The currently displayed version is the accepted final manuscript. The officially published version will be updated with format, DOI and citation information upon launch. We recommend that you pay attention to subsequent journal notifications and preferentially cite the officially published version. Thank you for your support and cooperation.