Research Of Urban Rail Transit Operation Multi-objective Optimization And Intelligent ATO Control Algorithm

In recent years,urban railway transit has been booming.The huge scale of rail transit has caused a huge amount of energy consumption.In 2018,Beijing's subway consumed about 1.9 billion k Wh.Reducing subway energy consumption has become a popular research direction.The advent of communication-based train control systems allowing trains to quickly perceive about environment and react to the ground.This also improves the automation degree of train operation,so that the on-line operation optimization algorithm can be implemented in reality.This article proposes optimization methods from multiple levels to reduce the traction energy consumption of trains and improve punctuality and comfort.The main research contents are as follows.(1)The single train is taken as the research object,and its running speed curve is optimized for multiple objectives,and the running time of each section of the whole line is optimized and allocated.In the beginning,the kinematics equations of the train are established,and the necessary conditions for energy-saving optimal working stage are derived using the principle of maximum value.Then,the evaluation function of energy efficiency,punctuality and comfort and the mathematical model of speed curve optimization are established.An improved multi-object particle swarm optimal(MOPSO)algorithm is used to search for the optimal solution.Considering the line speed limit and gradient conditions,a two-stage optimization method is proposed.Using the data of the optimization results of the train speed curve,a function model of relationship between the energy consumption of the whole line and the running time of the inter-station was established.The steepest descent method was used to optimize the allocation of the running time for each inter-station in the whole line.Finally,the line data of Yizhuang Line of Beijing Metro was used for simulation,and the effectiveness of the algorithm was tested.(2)An on-line intelligent train control algorithm based on xgboost algorithm is proposed,which can learn rules from excellent train driving data and control train running.The algorithm takes the train real-time status information and line data provided by the ATO system as input features,and the train acceleration value as the control output.At the same time,an interval time allocation algorithm and a safety protection algorithm are designed to ensure that the train can run safely.Simulation results show that xgboost algorithm has certain advantages in energy saving and punctuality compared with PID algorithm.(3)From the perspective of improving the utilization rate of regenerative braking energy,the collaborative optimization of multiple trains is studied.First,the principle and utilization conditions of regenerative braking energy are introduced.The regenerative braking scenario under two trains is considered,and the time discrete-approximation calculation model of regenerative energy utilization is established in the case of multiple trains,and the factors that affect the effective regenerative energy are analyzed.The PSO algorithm is used to optimize the departure time interval and stop time during peak and trough periods of passenger flow.The results show that the optimization method proposed in this paper can effectively improve the utilization efficiency of regenerative braking energy,and the algorithm is more effective in the trough period.Figures 58,Tables 15,References 56.