Research On Energy Management Strategy Of Fuel Cell Vehicle Based On Condition Prediction

Nowadays,most vehicles on the road still rely on internal combustion engines（ICE）to convert the chemical energy in fossil fuels into mechanical energy.Fossil fuel is not only a limited resource,but also produces harmful gases such as carbon monoxide（CO）,carbon dioxide（CO₂）and nitrogen oxides（NOx）that have adverse effects on the environment and humans when burned.In response to energy and environmental issues,the world’s major automakers have invested in research and development of a large number of hybrid electric vehicles（HEV）,plug-in hybrid electric vehicles（PHEV）and pure electric vehicles（BEV）to meet the needs of different customers.Although the sales of these vehicles continue to climb,automakers have gradually realized that fuel cell vehicles（FCV）may be the ultimate solution for the electrification of personal transportation.Compared with traditional internal combustion engine vehicles or hybrid vehicles,fuel cell hybrid vehicles（FCV）have zero greenhouse gas emissions and are more environmentally friendly.Compared with pure electric vehicles,fuel cell vehicles have the advantages of fast feeding and long driving range.At present,significant progress has been made in the development of commercial fuel cell vehicle technology.However,there are still some challenges in pushing fuel cell technology to mass production.In addition to the high cost,how to effectively improve the economy of fuel cell vehicles and develop vehicle energy management strategies with higher hydrogen saving potential are also the focus and difficulty of current research.Therefore,starting from the establishment of a complete fuel cell vehicle model,this paper focuses on the short-term vehicle speed prediction algorithm based on BP neural network,and on this basis,develops the development of a fuel cell vehicle vehicle equivalent hydrogen consumption minimum strategy that integrates vehicle speed prediction.The main research contents of this paper are:1)First,carry out the modeling research of fuel cell hybrid electric vehicle.Through the analysis of the driving force of the fuel cell vehicle on the road,the vehicle dynamics model is obtained.On this basis,relying on Matlab/Simulink platform to develop and establish fuel cell vehicle power system component models.At the same time,reasonable simplification of the nonlinear fuel cell system model was carried out,focusing on the modeling of the air supply side,and a fourth-order simplified model of the fuel cell system was obtained.Taking into account the polarization phenomenon of the fuel cell during operation,the fuel cell output voltage model is derived.2)Research on short-term vehicle speed prediction algorithm based on BP neural network.The definition of neural network and the basic structure of neuron model are introduced,and the training process of BP neural network is explained in detail through formula derivation.Based on this,a vehicle speed prediction model based on BP neural network is proposed.After that,the parameters of the vehicle speed prediction model are reasonably selected,the characteristic parameters of the working conditions,and the input and output neurons are selected to finally determine the structure of the built vehicle speed prediction model,so as to complete the training and prediction of the vehicle speed prediction model.The comparative analysis of different speed prediction algorithms under selected CLTC-P conditions further verifies the accuracy and rationality of the BP neural network vehicle speed prediction model.3)Development of a strategy for minimizing equivalent hydrogen consumption that incorporates short-term vehicle speed prediction.For fuel cell hybrid electric vehicles,based on the short-term vehicle speed prediction model established in the previous stage,an energy management strategy combining vehicle speed prediction and equivalent hydrogen consumption minimum algorithm（ECMS）is designed,and the role of equivalent factors in ECMS is analyzed in detail,and An adaptive equivalent factor adjustment model based on vehicle speed prediction is established.Based on this,the energy management strategy with vehicle speed prediction（AECMS-pre）and the energy management strategy without vehicle speed prediction（AECMS-nopre）developed in this paper are compared and analyzed to verify the rationality of the adaptive equivalent factor adjustment scheme proposed in this paper.In order to explore the hydrogen saving potential of the developed energy management strategy,it is compared with dynamic programming strategy（DP）,offline PMP and rule-based strategy.It is confirmed that the proposed strategy has the ability of approaching the global optimal allocation.Finally,the hardware in the loop test of the control strategy is completed,and the results show that the developed vehicle energy management strategy has good real-time performance.