Research On Energy-saving Optimization Of Train Speed Trajectories Based On Swarm Intelligence Algorithms

Railway transportation is an important part of transportation industry in our country, it shoulders the crucial task for national economic and social development. Railway transportation energy consumption presents an overall upward trend, with the rapid development of railway construction. Traction energy consumption accounts for 60%～70% of total railway transportation energy consumption. Therefore, reducing the locomotive traction energy consumption is significant for railway energy saving.The optimization of the speed curve is a principle way of improving the efficiency of train traction system at the technical level. This thesis presents studies based on application of optimization algorithms for the train operation target curve optimized that there is potential for improvements in the total system energy efficiency. With the mathematical model and energy-saving mechanism of trains, Particle Swarm Optimization, Improved Ant Colony Algorithm and their hybrid algorithm were applied to the energy-saving optimization of the train operation target curve in the simulation of Matlab. This paper can be divided into five parts, which are elaborated as follows:(1) Under certain operational, geographic and physical constraints, the objective function and constraint conditions of the train traction system are built. The transformation principles of train operation mode were analyzed by Hamiltonian function combined Karush-Kuhn-Tucker conditions with Lagrange operators. Key concepts of steep uphill and steep downhill in the complex line were introduced and optimization strategies of each stage under ideal condition were analyzed to derive the optimal control scheme of continuous ramp.(2) To solve the problem of feasibility in train target speed curve, Particle Swarm Optimization was studied to build the energy-saving optimization model including preprocessing module, query matrix module, ramp experience module, algorithm module and path information module and to construct speed control signals and fitness function which could balance energy consumption and running time. The analysis of simulation results indicated that train energy consumption decreased with the increasing of the planed travel time, which represented high computing precision and computing speed.(3) With the characteristics of fast computing speed, Ant Colony Optimization combined with train energy-saving experience improved pheromone adjustment scheme. Target speed curve optimization model was built based on ant colony algorithm and sub module to balance the energy consumption of train and running time was developed utilizing heuristic factor. The simulation results suggested that improved ant colony algorithm promoted both the computing precision and computing speed.(4) To solve the parameter self-adaption in the Particle Swarm Optimization, hybrid algorithm model was constructed according to the merits and demerits of Particle Swarm and Ant Colony Optimization. In addition, feedback mechanism between algorithms was built to save energy. Optimization results showed that the perfection improved the comprehensive performance of the algorithm.(5) Using algebraic convergence, convergence accuracy and convergence speed as evaluation index of algorithm performance to compare and analyze the Particle Swarm Optimization, Ant Colony Optimization, hybrid algorithm. In this study, we compared and analyzed optimization target speed curve simulation verification of these three algorithms using DYNAMIS software.This paper was dedicated to analyze the theory and practice of the energy-saving optimization of the speed curve, mainly highlighting the optimization algorithms and their application in order to lay a foundation for the development of outstanding online optimization guiding system. Hopefully, it can promote the efficiency of train traction system and reduce the energy consumption of railway transportation.