Sub-Optimal Non-Linear Optimization of Trajectory Planning for the DLR Next Generation Train (NGT)

Abstract

Distance domain-based Single Train Performance Simulator (STPS) tools using the Euler’s method and the combination of forward- and backward-facing driving calculations have been utilized for the development of railway vehicle trajectories as conventional Trajectory Planning Tools (TPT). Industry, operators and academic groups have used STPS tools for the development of test cases for design and real-time operation purposes of railway vehicles in great extent. Similar methods have been used by the German Aerospace Center (DLR) for the DLR vehicle concepts Next Generation Train (NGT) HST, LINK and CARGO. In particular, the 230 km/h NGT LINK concept has been analysed under -all-out- or otherwise called -driving as fast as possible- vehicle trajectories as well as under simplified trajectories with time reserves according to timetables for 140 km/h and 240 km/h railway tracks. The main disadvantage of the STPS-based approach utilized so far is the long times required for trajectory-oriented sensitivity analysis of the NGT LINK’s energy usage. In addition, TPT approaches make the energy-oriented optimization of the train’s driving behaviour difficult when a certain schedule is to be met and other constraints are applied, such as when different or same prices are used for required and returned energy to the catenary. In this paper, a sub-optimal non-linear optimization methodology is adapted and utilized to solve the problem of trajectory generation for the vehicle concept NGT LINK for three journey time-related test cases on the same track, with prices for required and returned energy prices being different. The sub-optimal optimization methodology is described and the results of the analysis on a pre-defined route are presented and compared to the simulation results of a conventional TPT tool for three different journey time requirements.