Reduction of train and net energy consumption using genetic algorithms for Trajectory Optimisation

It is known that for a single DC powered train, energy savings can be obtained by a combination of motoring, braking and coasting during a journey. However, this does not necessarily yield all of the net energy savings that are possible if other trains are running within the same electrical section. Further savings may be available during motoring by using energy regenerated by other trains while they are braking. This paper first presents a single train trajectory optimisation to obtain minimum energy consumption with maximum regenerated energy, and then considers net energy reduction between two adjacent DC substations when the optimised trajectories are used. Each trajectory is optimised individually using a genetic algorithm to search for the best possible compromise between energy consumption and journey time requirements. A weighted combination of these two is used as the objective function and the rates of train acceleration, braking and coasting form a set of variables that define a driving strategy. In order to estimate the benefits and effects of optimised trajectories on net energy consumption, multi-train simulation was then performed for both fastest and optimised journeys. Both qualitatively and quantitatively, the results suggest that further considerable reductions of net energy consumption may be achieved by the adjustment of schedules for both the up and down direction so as to increase the receptivity of those trains within each subsection, or by the recalculation of single train trajectories with different optimisation criteria. Finally, consideration is given to the possible application of the technique on a real railway traction system. Although demonstrated here on a DC system, the method could equally be applied to AC electrified railways.