Research On Energy Management Strategy Of Parallel Phev Based On Trip Information Prediction

At present,the traditional internal combustion engine vehicle(ICE-V)has caused serious environmental pollution and energy shortage due to the shortcomings of its high energy consumption and high emissions.In order to achieve the sustainable development of the automotive industry,new energy vehicle schemes regarding the electric vehicles(EVs)as the development goal have been proposed in succession.However,the EV still has many technical bottlenecks,which make it difficult to completely replace the ICE-Vs in a short time.Under this background,parallel plug-in hybrid electric vehicles(PHEVs)with the excellent energy saving and emission reduction ability,are regarded as the ideal transition scheme from traditional ICE-Vs to pure electric vehicles.The energy of the parallel PHEV comes from fuel and electric energy.In order to realize the on-line optimal management of the vehicle energy and further improve its energy saving and emission reduction performance,this paper focuses on the energy consumption problem of parallel PHEV deeply.In this case,this paper respectively adopts the optimal control methods of the model predictive control(MPC)and heuristic dynamic programming(HDP),to design three energy management strategies(EMSs)based on trip information prediction.I.The online EMS based on hybrid trip modelConsidering that the vehicle driving state is affected not only by its own motion state,but also by road traffic state under the actual traffic environment,this paper designs the online EMS based on hybrid trip model through the analytical approach.First of all,according to the driving characteristics of the vehicle in actual road environment,this paper respectively describes the continuous motion process of the vehicle and the change process of the traffic flow by means of the designed vehicle kinematics model and cell transmission model(CTM)of the road.Then,the hybrid automata(HA)theory is employed to fuse the continuous vehicle motion process with its discrete state switching process,to construct the hybrid trip model for the online speed trajectory prediction.Based on this,this paper adopts the MPC to design the energy management controller based on hybrid trip model,and realize its EMS online via a control algorithm.II.The online EMS based on trip condition prediction modelFurther,considering the random change of the vehicle driving conditions in the actual traffic environment,this paper designs the online EMS based on trip condition prediction model by using the data-driven approach.For this goal,this paper firstly utilizes the back propagation neural network(BPNN)to construct the trip conditions prediction model of the vehicle according to the changing characteristics of the driving conditions data of the vehicle.Then the trip conditions prediction model is improved by means of the environmental modal refinement,genetic algorithm(GA)and particle swarm optimization algorithm(PSOA).And then,the GA/PSOA-based BPNN trip conditions prediction model is established for the online prediction of driving conditions.Based on this,this paper continually employs the MPC to design the energy management controller based on trip condition prediction model,and realize its EMS online via a control algorithm.III.The HDP-based online EMSFurther more,considering that the optimization effect of the MPC depends on the prediction quantity and accuracy of future information,this paper employs the idea of the adaptive dynamic programming(ADP)to design the HDP-based online EMS.First of all,considering that the motion process of the vehicle has strong uncertainty and high nonlinearity in the actual traffic environment,this paper uses the BPNN to establish the state space model of the parallel PHEV.Meanwhile,considering that HDP does not depend on the prediction quantity of future information during solving optimization problems,but uses Bellman optimum principle and reinforcement learning(RL)idea to achieve the infinite approximation of the optimal solution of the problem,thus the energy management controller based on HDP is designed with BPNN,and its EMS is realized online through a control algorithm.In order to verify the validity of the three EMSs designed above,this paper simulates and studies the EMSs by means of the actual road traffic data from the road network in Beijing.The experimental results show that,firstly,on the premise of guaranteeing real-time performance,the designed three EMSs can further improve the energy consumption of the parallel PHEV when compared with the existing online EMSs.Specifically,the fuel consumption of the parallel PHEV can be reduced by 3.46% to 60.27%.Secondly,although the fuel consumption and emissions of the designed three EMSs is more than the offline global optimization EMS by different degrees,their energy consumption optimization effect is quite close to that of the off-line global optimization EMS.In particular,the minimum difference between the energy consumption optimization effect of the HDP-based online EMS and that of the off-line global optimization EMS is only 2.13% and the maximum is only 4.5%.Besides,the designed three EMSs can be used for the real-time optimization of the energy consumption,while the off-line global optimization EMS can only be used for the off-line performance analysis of other EMSs.It indicates that the EMSs designed in this paper can further improve the overall energy saving performance of the parallel PHEV when compared with the existing EMSs,and provides a more effective solution for the on-line optimal management of its energy.In addition,the used optimization ideas and designed methods during implementing the above EMSs in this paper,which can also be used in EVs with multiple motor assemblies or energy storage systems.It provides an effective theoretical basis and technical method for EVs research on improving power utilization,increasing endurance mileage and prolonging the life of batteries.