Research On FCV Energy Management Strategy Based On Speed Prediction

Because of its advantages of no emission,high efficiency,fuel cell hybrid electric vehicles provide an important method to solve environmental and energy problems.Now fuel cell hybrid electric vehicles are known as research hotspots in the field of energy-saving and new energy automobiles.In this paper,an energy management strategy based on the vehicle speed prediction model is proposed for a new type of hybrid vehicle,Fuel Cell Vehicle(FCV),with the structure of fuel cell + battery.Firstly,the working principle and types of fuel cell were analyzed.The structure of proton exchange membrane fuel cell(PEMFC)+ battery was selected based on the working conditions of fuel cell electric vehicle.According to the required dynamic performance indicators,the problem of optimizing the matching of fuel cell electric vehicle power system parameters is proposed and optimized using genetic algorithms.Then,based on the optimized fuel cell and battery parameters,combined with the fuel cell's working principle and electrochemical theory,the simulation model of fuel cell hybrid power system was established using Matlab/Simulink and Advisor.Using long-short term memory neural network(LSTM)to predict vehicle speed is one of the focuses of this paper.After comparing and analyzing the advantages of LSTM and general recurrent neural network(RNN),a vehicle speed prediction model was built using a neural network with LSTM layers and historical vehicle speed data.It also analyzes the time consumption of the single-step prediction in GPU,and compares that under the CPU,which verifies that the model can improve the calculation efficiency under the GPU.Finally,an energy management strategy based on model prediction is constructed.Then combine the model prediction problem with dynamic programming algorithm(DP)to construct predict control model using fuel consumption minimal and SOC steady change.The results of optimization using the DP algorithm under NEDC,OCC,and WLTP-C3 standard conditions and the model prediction energy management strategy were calculated.The result of simulation verified that the constructed predictive control strategy in this paper has effectiveness in improving the fuel economy of FCV.