Research On Energy Management Strategy Of Fuel Cell Vehicle Based On Multi-dimensional DP Algorithm

In the context of increasingly severe environmental pollution and depletion of petroleum resources,new energy vehicles like Fuel Cell Vehicles(FCV)are gradually being promoted around the world with the merit of long cruising range,fast hydrogen replenishment,little pollution and wide raw material sources,which have become an important direction for future automobile development and one of the main technical routes to solve environmental pollution and depletion of petroleum resources.FCV is typically powered by a combination of energy sources due to weak output of fuel cells,therefore,system parameter matching and Energy Management Strategy(EMS),as an essential part of FCV research and development,will affect the power performance,economical efficiency and system durability of the vehicle,and has emerged as a research hotspot by scholars at home and abroad.Firstly,based on the research and development needs of fuel cell vehicle,we analyzed the characteristics of various configurations of the fuel cell hybrid energy source system,and the semi-active fuel cell+power battery hybrid energy source configuration was selected to be the research configuration so as to do the analysis on the characteristics of the important parts in the configuration and construct the corresponding mathematical model.The main work includes rested on the analysis of the characteristics of battery equivalent circuit models,Thevenin model that can describe the battery polarization phenomenon was selected for mathematical modeling.To ensure the accuracy of the model description,the battery performance test and the parameter identification of Thevenin equivalent circuit model were completed in accordance with the comprehensive analysis of the battery operating characteristics and the corresponding battery performance test experimental program;on the basis of the comparative analysis of the characteristics of different types of fuel cells,it was determined to use the Proton Exchange Membrane Fuel Cell(PEMFC)as the research object and establish the steady-state output model of PEMFC considering the voltage loss;the DC/DC converter was introduced into the system as well as to analyze its efficiency characteristics and establish the efficiency model.Secondly,analysis was conducted on the working mode of the semi-active fuel cell+battery hybrid energy source,which contributes to establish the bus power balance model on the basis of the power relationship between the energy source and the bus.Rested on the FCV vehicle parameters and power performance indicators,the preliminary parameter matching of the important components of the fuel cell hybrid energy source system was completed,and the basic of parameters such as the power,speed,and torque of the motor were determined sequentially,based on which the normalized efficiency model of the drive motor was constructed by experimental modeling;the maximum output power of the fuel cell was matched up with the rated power and the number of series and parallel of the battery initially,which lays a foundation for the application of EMS solution and multi-objective optimization.Not only can Thevenin equivalent circuit model reflect the ability to follow the response of the battery to the load,but also describe the polarization of the battery.Therefore,in view of Thevenin model,BSOC and polarization voltage V_p were taken as the state variables and the equivalent hydrogen consumption of the system was selected as the cost function with the use of multi-dimensional DP algorithm to provide a solution for the fuel cell hybrid energy source system EMS.In addition,for purpose of improving the running speed of the algorithm,the state variables in each stage were divided into main and auxiliary parts.In the reverse solving process,the dimension of the three-dimensional correspondence between the state variable grid and the optimal cumulative function was reduced,along with reducing the solution dimension to the two-dimensional correspondence between the main state variable and the optimal cumulative function.The results of EMS solution were analyzed,and its correctness is explained.At last,the EMS which was based on multi-dimensional DP algorithm and the multi-objective optimization were applied to the research of parameter matching,the life model deprived from fuel cell performance degradation and the battery equivalent life model deprived from battery ampere-hour flow were introduced.Taking the maximum output power of fuel cell and battery capacity as parameter matching variables,a two-layer nested multi-objective parameter matching optimization process was proposed,which took the equivalent hydrogen consumption of the system as the economic index,as well as the fuel cell performance degradation rate and battery ampere hour flow as the system durability index,and applied the multi-dimensional DP algorithm to solve EMS.Based on the results of two-layer nested multi-objective parameter matching optimization and synthesized the power performance indicators of fuel cell vehicle and preliminary parameter matching,the economy and durability of the system were further analyzed at the same time.In the end,the optimal parameter design space is obtained by combining the manufacturing cost and weight of the hybrid energy source system with above analysis,thereby enabling designers to acquire the optimal combination of different system component parameters according to the different performance trends of the vehicle in the parameter matching.