Research On The Predictive Energy Management Strategy Of Range Extended Electric Vehicles

In the face of the dual pressure of fuel consumption and emissions in the field of transportation and the development strategy of the "2030 carbon peak,2060 carbon neutrality"goals,the automotive power system accelerates the transformation to electrification,but the pure electric vehicles still face the realistic anxiety problems such as limited range,inconvenient charging,battery safety,etc.,and the range-extended electric vehicles with fuel to electricity and full-time electric drive emerge at the right moment.Energy management strategy is the core control algorithm of range-extended electric vehicle.Research on energy management strategy of range-extended electric vehicles,especially predictive energy management strategy(PEMS)based on instantaneous optimization algorithm,is of great significance for performance improvement and optimization of extended range electric vehicles.In order to solve the problem of predictive energy management strategy for range-extended electric vehicles,an advanced predictive energy management strategy(A-PEMS)framework was designed,which includes energy management strategy and dynamic coordination control strategy of range-extender.Based on MPC algorithm framework,a real-time speed prediction method based on adaptive generalized regression neural network(A-GRNN)and a real-time energy distribution method of prediction horizon based on dynamic programming(DP)were designed,and a dynamic coordination control strategy of range-extender based on feedforward added with feedback,and speed-torque-decoupling was designed.A forward co-simulation platform of rangeextended electric vehicle including driver,control strategy and vehicle controlled object was established,and A-PEMS was verified and optimized.Focusing on the PEMS of range-extended electric vehicles,the main research work and conclusions of this paper are as follows:(1)The real-time speed prediction algorithm was realized based on GRNN,and the influence rule of different input node number and radial basis function width on GRNN real-time speed prediction algorithm was explored.The adaptive adjustment of GRNN parameters was realized based on Akaike information criterion(AIC)grading criterion,and an adaptive A-GRNN realtime speed prediction algorithm was designed.The simulation results showed that the algorithm can effectively improve the accuracy of speed prediction,which provides a basis for the research of predictive energy management strategy.(2)The forward simulation platform of extended range electric vehicle was established,including controller strategy model,vehicle controlled object model and driver model.The power components of extended range electric vehicle were modeled,and the vehicle controlled object model was verified based on the road test data.The results showed that the controlled object model of the range-extended electric vehicle can reflect the running state of the actual driving well,and provides a simulation platform for the optimization and verification of A-PEMS strategy.(3)The global energy allocation based on DP for the prediction horizon of PEMS strategy was realized,and the problem of determining the target SOC at the end of the prediction horizon during the rolling time domain process was mainly solved.Combined with the realtime velocity prediction algorithm based on A-GRNN,the energy management strategy based on MPC was realized.The control principle of the range-extender was analyzed,and a dynamic coordinated control strategy based on feedforward added with feedback and speed-torque-decoupling was designed.A-PEMS strategy including MPC-based energy management and dynamic coordination control of range extender is designed and implemented.(4)Based on the bench test,the dynamic coordination control strategy of the range-extender was verified.The results show that the dynamic coordination control strategy of the rangeextender can control the range extender to respond to the target power quickly and accurately.The A-PEMS strategy was verified by the forward simulation platform of the range extender electric vehicle.The influences of the range of SOC discrete state space,the discrete degree of SOC state space grid,the deviation weight of SOC at the end of prediction horizon,and the length of prediction horizon on A-PEMS strategy were investigated,and each parameter was optimized.(5)The simulation results of A-PEMS strategy and other four energy management strategies under different driving cycles and different initial SOC values were compared.The results show that for the same driving cycle and the same initial SOC value,A-PEMS achieves better fuel economy performance than the other four energy management strategies.At the same time,APEMS will effectively reduce battery charge and discharge and prolong battery life.