Research On Energy Management Strategy Of Electric Vehicle With Multiple Energy Sources And Battery SOC Estimation

Guided by the increasing environmental problems and the development strategy of the double carbon target,electric vehicles have good prospects for development in the future.Traditional electric vehicles generally use lithium batteries with high energy density as the energy source,however,lithium batteries have problems such as low power density and short cycle life.Fuel cells have the advantages of high efficiency and no pollution,while supercapacitors have high power density,fast charging and discharging speed and long cycle life,and there is the possibility of complementary advantages between different characteristic energy sources.Coupling other energy sources is therefore an important means of promoting the development of battery vehicles.However,the joint action of multiple energy sources greatly increases the difficulty of power distribution,and designing effective energy management strategy to achieve a reasonable distribution of power and to take into account the advantages of energy sources with different characteristics is the focus of the research.The main work of this paper includes.(1)The modelling of an electric vehicle with a mixture of three energy sources is completed.Based on the theoretical analysis,a simulation model of the electric vehicle and each energy source is established,and the characteristics of each energy source are analysed.The topology of the hybrid energy storage system consisting of lithium batteries and supercapacitors is outlined,and the connection method between each energy source is determined.(2)A layered energy management strategy is designed for the situation where the characteristics of different energy sources differ.In the upper layer,an evaluation factor is introduced and an adaptive sliding average filter is designed to supply the low frequency component of the fuel cell by means of frequency splitting and decoupling,and then the fuel cell output power is optimised by means of a power following control strategy that combines the fuel cell efficient operating zone and the variation of the lithium battery state of charge(SOC);in the lower layer,a novel algorithm-In the lower layer,a new algorithm-deep reinforcement learning-is used to manage the energy of the Li-ion battery and supercapacitor,and various spatial variables are analytically designed to achieve a reasonable distribution of power while satisfying the optimization objectives in the lower layer.(3)Simulations are conducted to validate the hierarchical energy management strategy designed in this paper.The results show that the strategy proposed in this paper,when applied to an electric vehicle with a mixture of three energy sources,can take advantage of the different energy sources,improving the efficiency of the fuel cell in the upper layer,and making full use of the supercapacitor in the lower layer to achieve the "power distribution".This not only maintains the stability of the SOC of the Li-ion battery and the supercapacitor,but also avoids the frequent charging and discharging of the Li-ion battery,which affects its service life.(4)An improved CNN-LSTM neural network is designed to estimate the SOC of the battery for the importance of SOC application in energy management strategy,and the weight is assigned to the dimensionality and temporal characteristics of the input data by the improved attention mechanism,and the results show that the proposed method has a better SOC estimation performance,and the error value is generally below the traditional CNN-LSTM,and the SOC estimation following performance is better.The effectiveness of the proposed neural network in estimating the battery SOC is verified.