Research On Multi-Objective Energy Management Strategy For Fuel Cell Vehicle

Fossil fuel shortages,the awakening of environmental awareness,and the guidance of Carbon Peaking and Carbon Neutrality strategic decisions and related policies have accelerated the development of clean energy-related industries represented by hydrogen energy.Hydrogen energy represents a promising avenue for clean,sustainable energy production,and hydrogen fuel cells are a critical application of this technology.As a clean secondary energy source,hydrogen energy is poised to play an increasingly vital role in driving future energy transformations.Today,proton exchange membrane fuel cell(PEMFC)and battery have been extensively studied for their ability to equip fuel cell vehicles(FCVs)with high efficiency,zero emissions,low noise,and long-range capabilities.Accordingly,achieving optimal power allocation between fuel cells and batteries constitutes a crucial avenue of investigation towards enhancing vehicle energy efficiency,prolonging the operational life of both power sources,and curtailing the overall lifespan expenses of the vehicle.Therefore,starting with a comprehensive fuel cell vehicle model,this paper places emphasis on the multi-objective optimization of energy management strategies that take into account hydrogen consumption,fuel cell life,and battery life.The main contents of this paper are as follows:(1)Research on fuel cell vehicle modeling.This paper has developed a fuel cell vehicle dynamics model and power system key components model on the Matlab/Simulink platform.The models include the fuel cell model,battery equivalent internal resistance model,DC/DC converter model,and drive motor model,among others.Moreover,this paper has established a comprehensive mechanism model of the fuel cell system,comprising five components,namely the fuel cell stack,humidifier,radiator,reservoir,and condenser.In addition,this paper has developed a temperature model of the fuel cell system.To evaluate the accuracy of the fuel cell and temperature model,simulations were conducted.The results obtained provide a solid foundation for developing an energy management strategy for the fuel cell vehicle in subsequent research.(2)The optimal control problem of energy management in fuel cell vehicles is converted into a multi-objective optimization problem based on the Pontryagin’s Minimal Value Principle(PMP).This principle is introduced briefly,and the evaluation index,objective function,state variables,and control variables are clarified.Additionally,corresponding life models are established to quantify the decline of fuel cell and battery life.Subsequently,an energy management strategy with multi-objective optimization is formulated to consider hydrogen consumption,fuel cell life,and battery life.The strategy is modeled and solved,and its results are compared with those of a single-objective dynamic planning energy management strategy that only considers optimal hydrogen consumption.The effectiveness of the algorithm is demonstrated through the comparison of the two strategies.(3)Deep reinforcement learning(DQN)has been employed to investigate the multiobjective optimal control of fuel cell vehicles(FCV).To address this issue,a novel optimal energy management control strategy based on DQN with multi-objective optimization(DQN-MOEMS)has been proposed.In designing the reward function,the four crucial factors,including hydrogen consumption,beginning and end SOC offsets,fuel cell life,and battery life,have been taken into account.Moreover,the empirical-first replay technique has been utilized to enhance the algorithm’s convergence speed.Ultimately,simulation comparison analysis has been conducted to validate the effectiveness of the DQN-MOEMS strategy in optimizing vehicle economy,fuel cell life,and battery life,as well as the algorithm’s adaptability to diverse operating conditions.The proposed DQN-MOEMS strategy offers a promising approach for addressing the challenges associated with the optimal control of FCV.