Research On Adaptive Energy Management Strategy For Phev Based On Working Condition Recognition

The rapid development of the modern automobile industry and the dramatic growth of car ownership have brought about an energy crisis and environmental pollution problems.Scholars and engineers around the world have made unremitting efforts to improve the fuel economy and emission performance of automobiles,and new energy vehicles have emerged as a result.Among them,plug-in hybrid vehicles are the representatives of new energy vehicles,which can further optimize their fuel economy by carrying larger capacity batteries,leading to the energy management strategy of hybrid vehicles.This paper relies on the key science and technology research project in Jilin Province,and combines intelligent identification of driving conditions,clustering algorithms,and intelligent optimization algorithms to optimize the fuel economy of plug-in hybrid vehicles,and the specific work is as follows.(1)To meet the research requirements,a coaxial parallel plug-in hybrid vehicle simulation model is built in MATLAB/Simulink environment,including the driver model,powertrain components model and the whole vehicle dynamics model.The rule-based energy management strategy model is developed on the basis of the vehicle model,and simulation experiments are conducted under different driving cycle conditions to verify the control effect of the rule-based energy management strategy through comparative analysis.(2)To avoid the problem of dependence of the global optimal energy management strategy on the known global working conditions,this paper adopts the equivalent fuel consumption minimization strategy,analyzes the importance of the equivalence factor in the equivalent fuel consumption minimization strategy,and analyzes the equivalence between it and the equivalent fuel consumption minimization strategy from the perspective of whether the synergistic state in the Pontryagin minima is variable.On the basis of the above vehicle model,a simulation model of the equivalent fuel consumption minimization strategy is built,and the sensitivity of the equivalent factor to the type of working condition,driving distance and initial battery charge state is investigated through simulation experiments,based on which an improved particle swarm optimization algorithm is proposed to optimize the equivalent factor under different types of working conditions,different driving distances and different initial battery charge states.The simulation results show that the optimized equivalent factor can improve the fuel economy of the vehicle,which indicates that the algorithm has good optimization effect and can solve the optimization problem of the equivalent factor.(3)For the relationship between the equivalence factors and the working conditions,the working conditions data of the representative urban road sections were collected by driving a real vehicle,and the collected working conditions data were further processed to cluster the urban measured working conditions by K-mean clustering algorithm,and the equivalence factors under various working conditions were optimized by using the improved particle swarm optimization algorithm proposed above.Based on the clustering results,a fully connected neural network is used to build a model framework for working condition recognition,and a part of the clustered samples is used as the training set to train the fully connected neural network model,and another part is used as the test set to test the trained neural network recognition model.(4)An adaptive equivalent fuel consumption minimization strategy based on a fully connected neural network condition identification is proposed by combining the equivalent fuel consumption minimization strategy and the fully connected neural network condition identification algorithm.The strategy adopts the fully connected neural network for working condition identification,calls the optimized equivalent factor under the corresponding working condition,and uses the linear reference battery charge state generation method to adjust the equivalent factor in real time according to the deviation between the actual battery charge state and the reference battery charge state by using the logarithmic function method.The simulation results show that the strategy further improves the fuel economy of the vehicle,especially under urban real-world conditions.