Research On Operation Control Strategy Of Urban Rail Transit Train With The Consideration Of Multi-Objective Optimization

Urban rail transit system is an important facility to relief the traffic pressure of the city with superiorities in convenience,efficiency,punctuality,safety and high capacity.As one of the important research contents in urban rail transit system,train operation control strategy optimization has great significance in the improvement of passenger experience and energy saving of urban rail transit system,which also can be further improved with the development of relevant technologies.There are many research hotspots in operation control strategy optimization of urban rail transit train,and train speed profile optimization,train timetable optimization and train running tracking controller design are focused by many researchers.But relevant researches do not take the advantages of each other,and relevant optimization techniques are not utilized to the full.Thus,this dissertation aims to provide a systematic optimization method for train operation control strategy.This dissertation optimizes train operation control strategy with multi-objective consideration,and proposes a series of optimization methods from different aspects for train running,as well as the simulation system integration and application.Based on the review of relevant researches at home and abroad,dynamic mechanism of train operation is deeply analyzed,and multi-dimension optimization objectives and evaluation indicators of train operation control strategy are established;A kind of hybrid scheme of train running is proposed to enlarge the optimization space,and improved multi-objective optimization algorithm is used to obtain the train section speed profile set;The optimized speed profile sets are added to the train timetable optimization model,which can realize the energy synchronization optimization between successive trains for single train timetable;In order to ensure the tracking effect of speed profile,adaptive train running tracking controllers are designed with intelligent technologies.The main research contents and achievements of the dissertation are shown as following:(1)A train dynamic model and multi-dimension operation optimization objectives under different aspects are established.The train dynamic model considers time-varying characteristics of parameters,and model assumptions,parameters and force analysis are illustrated.Meanwhile,according to the characteristics of the research object,objectives and evaluation indicators of section speed profile optimization,train timetable optimization and train running tracking control are analyzed in detail with relevant calculation equations,which can determine the optimization direction.(2)A Pareto optimization method of section speed profiles based on the hybrid scheme of train running is proposed to realize the optimization in multiple dimensions.The hybrid scheme contains multiple train running mode and enlarges the parameter range,which can make the optimal solution search space larger,and the corresponding multi-objective values and speed profiles can be obtained through the constructed train performance simulation(TPS)calculation model.The research designs hybrid scheme multi-objective particle swarm optimization(HS-MOPSO)algorithm to obtain the Pareto set of section speed profile,which solving efficiency is obviously improved compared with the original one under the same conditions.Numerical example analysis validates the efficiency and rationality of the proposed optimization method.(3)A successive train energy synchronization optimization method for single train combined with section speed profile set is proposed.Based on the analysis and modeling of energy synchronization between successive trains,the single train timetable optimization model is constructed.Meanwhile,serialized time synchronization is applied on the train timetable and section speed profiles to get the route net energy consumption for single train timetable,and the optimization effect of both studies can be superimposed.The research uses particle swarm optimization(PSO)algorithm to solve the timetable optimization model,and outputs recommended speed profiles for each section.Numerical example shows that net energy consumption for train running can be effectively optimized for long downhill section through the proposed timetable optimization method.(4)Model-free adaptive train running tracking controllers based on proportionalintegral-derivative(PID)structure are designed.This dissertation uses model-free adaptive control method to deal with the rail transit train which is a complex nonlinear system,and the proposed relevant controllers have good structure inheritance for currently widely used PID structure and can avoid conflicts with the plant.The research mainly focuses on the design of a kind of neural network proportional-integralderivative(NNPID)controller which can deal with the problems of PID gain range and actuator saturation,and this controller performs the best in comprehensive performance compared with other kinds of controllers in Simulink simulation.(5)The train running optimization simulation system is constructed and the case is applied.The system combines the above research achievements and is systematically designed in interfaces,functions and framework.The system includes a sub-system for generating speed profile Pareto set,a sub-system of line timetable optimization and a sub-system of train running tracking control,and the application for a whole line validates the proposed multi-objective optimization method of train operation control strategy.