铁道科学与工程学报2026，Vol.23Issue(7)：3110-3122,13.DOI:10.19713/j.cnki.43-1423/u.T20251365
面向虚拟编组的高速列车鲁棒模型预测控制
Robust model predictive control for high-speed trains in virtual formation
摘要
Abstract
To address the growing transport demands of high-speed railways and enhance the overall operational efficiency of the high-speed network,this study focused on the safe and efficient control of high-speed trains operating under a virtual formation mode.Taking the CRH380A high-speed train as the research subject,a virtual formation operation mode covering the entire phase from departure formation to disassembly upon arrival was proposed.To ensure adherence to this mode,a multi-train,multi-mass dynamics model was established,accounting for non-linear operational resistance during train movement,inter-carriage interactions,and external unknown disturbances.Based on this,an optimized operational objective function was designed incorporating constraints such as speed limitations,operational spacing restrictions,and control quantity limitations during high-speed train operations.Subsequently,a robust model predictive control method based on a hierarchical architecture was proposed.The solution layer employed a quadratic programming strategy to generate optimal reference trajectories for train group operations in real time,while the control layer calculates optimal control forced through a robust model predictive controller,enabling coordinated operation of train groups under the new virtual formation mode.Simulation results demonstrate that,compared to traditional model predictive control,the proposed method can exhibit superior performance in train speed tracking error,acceleration fluctuation range,and control force output when subjected to identical external disturbances.Speed tracking accuracy can improve by up to 47%,maximum operational acceleration decreased by up to 16.8%.The control force variation is reduced,confirming the method's robust disturbance rejection capability.Furthermore,when encountering sudden disturbances,the proposed method can reduce the time required for virtual formation trains to complete marshalling.It also exhibits superior stability during emergency braking scenarios.These findings can provide valuable insights for enhancing the operational safety of high-speed trains and improving the transport efficiency of high-speed rail networks.关键词
虚拟编组/分层控制/模型预测/鲁棒性/高速列车协同运行Key words
virtual formation/hierarchical control/model predictive/robustness/high-speed train coordinated operation分类
交通工程引用本文复制引用
付雅婷,周坤,李中奇..面向虚拟编组的高速列车鲁棒模型预测控制[J].铁道科学与工程学报,2026,23(7):3110-3122,13.基金项目
山区土木工程安全与韧性全国重点实验室自主课题(HJGZ2024209) （HJGZ2024209）
国家自然科学基金资助项目(62363011,52472342) （62363011,52472342）
江西省自然科学基金资助项目(20242BAB26022) （20242BAB26022）
中国博士后科学基金资助项目(2024M760904) （2024M760904）
