Research On MPC-based Hybrid Energy Storage Control System For New Energy Urban Rail Vehicles

With the continuous development of China’s national economy and the acceleration of urbanization process,people’s demand for urban traffic is also growing.As urban rail transit has the characteristics of large traffic volume,friendly environment,punctuality and comfort,vigorous development of urban rail transit will relieve urban traffic pressure and improve urban traffic circulation efficiency.In recent years,on-board energy storage system has become the main research hotspot as the power source of rail transit vehicles.Compared with the traditional traction network power supply mode,the advantages of this mode are:reducing the visual pollution of traction network power supply lines,and making urban rail vehicles be able to work in no-network areas;absorbing braking energy by energy storage components can improve energy utilization ratio;avoiding the possible accidents of traction network,etc.According to the operation condition of rail transit,it is necessary to use power and energy storage components complementarily to match the power requirements of power systems in order to give full play to the power and energy characteristics of different energy storage elements.Therefore,in this paper,supercapacitors and lithium-ion batteries applying to rail transit power system are used as energy storage components.Considering the actual operation conditions and power parameters of rail transit,this paper study the modeling of energy storage components,system capacity allocation,energy management strategy and converter control strategy mainly,focusing on the application of model predictive control for hybrid energy storage of new energy rail transit vehicles.The specific research content and work are as follows.(1)Model analysis of hybrid energy storage system.According to the characteristics of rail transit vehicles and energy storage components,the advantages and disadvantages of different hybrid energy storage systems are compared.The state space model and efficiency model of lithium ion battery,supercapacitor equivalent circuit model and DC/DC converter are established to provide theoretical support for the follow-up control algorithm research.(2)Optimal capacity allocation of hybrid energy storage system.The calculation method of annual operation cost of hybrid energy storage system for rail transit vehicles is studied,including initial investment cost,operation cost and replacement cost.A capacity allocation estimation method based on energy matching is proposed for the characteristics of power demand and energy storage components At the same time,considering the boundary conditions such as charge-discharge ratio,weight and SOC interval,and aiming at the lowest economic cost of vehicles,the estimation method is precisely optimized by particle swarm optimization algorithm to obtain the capacity allocation parameters of the vehicle hybrid energy storage system under all the life cycle.In the condition of ensuring the accuracy of the optimization results,the optimization calculation time is greatly reduced.(3)Research on energy management strategy of hybrid energy storage system.Aiming at the efficiency model of energy storage system,an energy management strategy based on model predictive control is proposed,which aims at minimizing energy consumption in traction operation stage and takes the lower limit of supercapacitor voltage as the boundary condition.The single-step prediction and multi-step prediction methods of energy management in traction stage are studied,and the control effects and energy consumption of different strategies are compared and analyzed.Finally,the effectiveness of the proposed energy management strategy is verified by simulation results and the experiment results obtained by the experimental platform based on dSPACE.(4)Research on MPC based on DC/DC converters.To overcome the shortcomings of traditional PI control for DC/DC converters,a continuous control set control method based on model predictive control is studied.At the same time,aiming at the model mismatch problem of traditional model predictive system,a system model observation method based on Rumberger observer is proposed,which greatly improves the robustness of the control system.Finally,the effectiveness of the proposed method is verified by simulation and experiment.