Energy Management Strategy For Hybrid Energy Storage System Of Electric Vehicle Based On Reinforcement Learning

Pure electric vehicle(EV)has a short driving range,and the capacity degradation rate of the battery increases under the condition of high-rate charging and discharging,which leads to the decrease of cycle life and limits the application and promotion of EV.To solve these problems,a reinforcement learning-based energy management strategy for hybrid energy storage system(HESS)of electric vehicle is proposed by studying the regenerative braking control strategy,reinforcement learning algorithm and HESS energy management strategy.The strategy consists of batteries and supercapacitors in a hybrid energy storage system,which makes full use of the energy in the regenerative braking process to improve the energy utilization efficiency,the impact of high current on the battery is reduced and the total energy loss of the system is reduced.And then the driving range of electric vehicle is extended.The main research contents of this paper are as follows:(1)The components of the HESS are modelled,and the parameters are matched.Firstly,the topology of the HESS is determined.Secondly,the model of battery,supercapacitor and DC/DC converter are modelled.Finally,the parameters of the battery and supercapacitor are matched according to the range of drive,average power,and peak power.(2)A regenerative braking control strategy is proposed.Firstly,the characteristics of the motor,the battery and the supercapacitor are analyzed,which play a restraining role in braking energy recovery.Secondly,the braking energy recovery process is analyzed.Finally,a regenerative braking control strategy is developed.(3)A HESS energy management control strategy is proposed.Firstly,the dynamic programming algorithm is introduced,and the corresponding energy management strategy is established.Secondly,the working principles of Q-learning algorithm,DQN algorithm,DDPG algorithm and TD3 algorithm are introduced.Finally,a HESS energy management strategy based on TD3 algorithm is proposed.(4)Experiment is validated.The regenerative braking control strategy and the HESS energy management control strategy proposed in this paper are simulated and the results are analyzed to verify the effectiveness of the proposed strategy under three typical operating conditions of UDDS,NEDC and CLTC.The innovative points of this thesis are as follows:(1)When recovering braking energy,the influence of motor,battery and supercapacitor on braking energy recovery is considered.In order to give full play to the advantage of high charging and discharging efficiency of supercapacitor,priority is given to supercapacitor to recover braking energy.(2)In order to extend the driving range and protect the battery,a HESS energy management strategy based on TD3 algorithm is proposed for the problem of overestimation of Q value by DDPG algorithm,which extends the driving range of electric vehicles by distributing the battery and supercapacitor power reasonably.