On-Line Train Speed Profile Generation of High-Speed Railway With Energy-Saving: A Model Predictive Control Method

By considering dynamic operational conditions in high-speed railway, this paper focuses on the on-line generation problem of train speed profile with energy-saving. This problem is formulated via the model predictive control framework in a moving-horizon manner, such that the real-time running conditions (e.g., temporary speed restrictions) can be involved in the on-line scheduling process of the train. At each control step, a hybrid scheme combining the energy-efficient and time-optimal train control strategies is proposed to ensure the feasibility of the optimal train control problem within the prediction horizon. The optimal control problem in the horizon is transformed into a multi-phase optimal control problem, which is then solved efficiently on-line using the pseudospectral method. By repeatedly solving the train control problem at each step, the energy-efficient train speed trajectory for the whole trip involving dynamic operational conditions can be obtained on-line. In addition, a delay recovery process is designed to re-schedule the train operation if the delay time during the trip exceeds a given threshold value. Finally, numerical examples using data for a real high-speed railway line are given to demonstrate the effectiveness and robustness of the proposed approach.

Automated summary

This paper focuses on the online generation problem of energy-saving train speed profile in high-speed railway, taking into account dynamic operational conditions. Using the model predictive control framework, the problem is formulated in a moving-horizon manner, allowing for real-time running conditions to be involved in the online scheduling process of the train. A hybrid scheme combining energy-efficient and time-optimal train control strategies is proposed to ensure feasibility within the prediction horizon. The problem is transformed into a multi-phase optimal control problem and efficiently solved online using the pseudospectral method. Numerical examples with real high-speed railway line data demonstrate the effectiveness and robustness of the proposed approach.