Energy-efficient driving strategies for rail vehicles

Abstract

The majority of railways internationally are striving to improve their financial performance while meeting competition from other modes of transport. A key factor in achieving this is by reducing energy consumption, which accounts for a significant proportion of all operating costs. The research undertaken for this Thesis addresses this challenge by applying optimization approaches to develop energy-efficient train operation strategies. It does this by developing a hybrid optimization approach, which combines global optimization techniques, for their "global" optimality properties, with local ones, for their faster convergence rate. Due to the number of control constraints and the number of decision stages involved for the control of a typical running train, a ruled-based quasi-global optimal control strategy is developed. This means that instead of first optimizing the control strategy for each particular running scenario, the Thesis shows how to develop optimized parameterized train operational control policies from empirical experience. The second step to develop the control sequence/strategy is using the control strategy generated from the optimized train operational control policies as initial searching point(s), then necessary optimality conditions are applied to locate the sub-optimal strategy for the vehicle in the particular running scenario. The proposed hybrid optimization method has been assessed and validated with the use of examples. The method shows good potential for significantly improving the fuel economy of running trains. The method has also shown significant numerical advantages over other conventional optimization methods in solving the optimization problem of the optimal/sub-optimal operation of a general running train with a long control horizon.