Research On A Real-time Multi-objective Prediction Energy Management Strategy For Range Extended Fuel Cell Vehicles

With rapid development of hydrogen technology,proton exchange membrane fuel cells(PEMFCs),as an environmentally friendly technology,convert the chemical energy of hydrogen fuel and oxygen to electrical energy for driving vehicle.And the only product of the reaction is water.The PEMFCs system has the advantages of high efficiency,no pollution,and low noise.They can constitute range extended fuel cell vehicles(FCVs)with high efficiency and zero greenhouse gas(GHG)emissions.For a range extended FCV,the battery can be charged from the grid or the fuel cell,and the battery and fuel cell can directly provide power to the driving motor.Therefore,the vehicle can work in different working modes.It is very complex when selecting the working mode and controlling energy distribution.An efficient energy management strategy(EMS)is imperative for the vehicle controller.To improve the real-time performance and energysaving effect of the range extended FCV EMS,a research about real-time multi-objective prediction EMS for range extended ECV is carried out in this paper,including the following contents:(1)By analyzing the powertrain configuration of the range extended FCV,the parameters of powertrain components are determined and the models are established.Based on the established models,this paper proposes a mathematical model for multiobjective energy management optimal control problem of the range-extended FCV.Then,simulation and analyzation of the dynamic programming(DP)-based multi-objective EMS proves the huge energy-saving potential of the multi-objective EMS.Based on the theory of model predictive control(MPC),a multi-objective EMS for FCV is proposed to achieve real-time control.Through the simulation and analyzation of MPC-based multi-objective EMS,the main research target of this paper is determined.(2)A novel energy management optimization algorithm based on direct collocation method and sequential quadratic programming is proposed.To avoid the shortcomings of the traditional indirect method for solving energy management problems,a direct method is proposed to solve the multi-objective energy management optimization problem in this paper.The global optimization problem in the prediction horizon is transformed into a nonlinear programming problem by direct configuration method,and the sequential quadratic programming is used to solve it.The proposed novel optimization algorithm is applied to the MPC-based multi-objective EMS,then the strategy is simulated under standard driving cycles.Compared with the simulation results of traditional DP,the proposed algorithm can significantly reduce the calculation time and improve the realtime performance of MPC-based multi-objective EMS.(3)A rapid planning method for the global state of charge(SOC)reference trajectory based on traffic information is proposed.A SOC change predictor based on long short term memory network is designed.Taking optimal SOC trajectories under various standard driving cycles as training samples,the artificial neural network with an ability of fitting the nonlinear system is used to learn the change law of the optimal SOC trajectories.Then,the proposed method is compared with a mileage-based global SOC reference trajectory planning method under standard driving cycles.The analysis shows that the proposed method is closer to the optimal SOC trajectory calculated by the DP algorithm than the comparison method,and has a shorter calculation time.It proves that the proposed method can improve the accuracy of global SOC reference trajectory and meet the application requirements of the MPC-based real-time multi-objective EMS.(4)Based on optimizing the rolling optimization algorithm and SOC reference trajectory planning method of the MPC based multi-objective EMS,this paper proposes a real-time multi-objective EMS based on traffic information.A simulation of the proposed strategy and the original strategy under comprehensive driving conditions is performed to verify the improvement of the real-time performance and energy saving performance of the proposed strategy.In addition,Vissim software is used to establish a traffic environment model based on real roads to simulate actual driving environment,and a real-time EMS co-simulation experiment is carried out in this paper.The results of the co-simulation experiment prove that the proposed real-time multi-objective EMS based on traffic information can achieve the multi-objective optimization target,which includes the improvement of both energy consumption economy and fuel cell life economy,and the proposed strategy has good real-time performance.