Research On Energy Management Strategy Of P2.5 Plug-in Hybrid Electric Vehicle Based On Driving Pattern Recognition

Plug-in hybrid vehicles with dual power sources can effectively solve the range anxiety of pure electric vehicles and are one of the mainstream solutions for new energy vehicles today.The development of a reasonable energy management strategy can realise the complementary advantages of the engine and the electric motor,and improve energy saving and emission reduction capabilities.Most of the existing energy management strategies adopt rule-based control strategies,ignoring the influence of actual driving conditions on the control effect,and the economy needs to be further improved.In this paper,the global optimal energy management strategy is used as the basis,combined with real-time driving conditions,to develop an energy management strategy based on condition identification,in order to further improve the fuel economy of the vehicle.The main research elements of this paper are as follows:(1)Analysis of the structural characteristics of the P2.5 configuration plug-in hybrid electric vehicle system,elaboration of the power flow of each operating mode and establishment of numerical models of key components.The energy management strategy based on Charge Depleting-Charge Sustaining(CD-CS)is established to determine the operating modes and switching rules for each phase.(2)19 standard working conditions are selected to form a standard working condition library,and the characteristic parameters of each working condition are clustered and analysed to classify each standard working condition into three categories and determine the corresponding representative working conditions of the three categories.Based on the three typical working conditions,Support Vector Machine(SVM)is then used to identify the working conditions.The Whale Optimization Algorithm(WOA)is used to optimize the key parameters of the SVM model in order to improve the classification and recognition accuracy.(3)A global optimal energy management strategy based on Dynamic Programming(DP)algorithm is established to achieve optimal torque allocation.Compared with the CD-CS strategy,the economic performance of the DP-based strategy is significantly improved,which verifies the effectiveness of the proposed energy management strategy based on the DP algorithm and provides reference and methodological support for the energy management strategy in the later paper.(4)The energy management strategy based on working condition identification is proposed.To address the shortcomings of the DP strategy that requires known working conditions and for specific working conditions to be optimised,a neural network is used to learn the power allocation results of typical working conditions offline.In driving,the corresponding neural network is selected to allocate the power by the working condition recognition results to achieve online control.Finally,simulation validation is carried out to compare the simulation results of the neural network energy management strategy and the CD-CS strategy with and without working condition recognition.The simulation results show that the economy of the neural network energy management strategy with and without condition recognition is better than that of the CD-CS strategy;the economy of the energy management strategy based on condition recognition is improved by 1.56%compared with that of the strategy without condition recognition,thus verifying the effectiveness of the energy management strategy based on condition recognition proposed in this paper.