Research On Energy Management Strategy Of Plug-in Hybrid Electric Vehicle Based On Online Cycle Recognition

New energy vehicle technology is an important direction to reduce vehicle exhaust emissions and alleviate global energy shortages.Plug-in Hybrid Electric Vehicle(PHEV),as a type of new energy vehicle,takes into account the advantages of pure fuel vehicles and pure electric vehicles.As the core technology of PHEV,energy management strategy determines the advantages and disadvantages of energy consumption economy,drivability and emission performance of the whole vehicle.The current mainstream energy management strategy is mainly designed for a specific working condition and does not consider uncertainty of actual driving conditions.Therefore,this paper takes a single-axle parallel PHEV as the research object,and proposes an energy management strategy based on online identification of operating conditions to adapt to different driving conditions,and ultimately achieve the purpose of improving vehicle energy consumption and economy.The main research contents are as follows:(1)Parameter matching and model building of the vehicle.The single-axle parallel PHEV is regarded as the target model.Firstly,its structural characteristics and working mode are analyzed.Then the parameters of the single-axle parallel PHEV are matched.Then,based on Matlab/Simulink software,a single-axis parallel PHEV vehicle backward simulation model and a CD-CS control strategy model are built.Finally,the correctness of the model is verified by simulation analysis.(2)Construction of comprehensive driving condition library.In order to construct the comprehensive driving condition database,firstly,the typical operating conditions are selected from the standard operating condition library as the initial data set for constructing the comprehensive driving condition database,and the kinematic segments are divided.Secondly,the characteristic parameters of each kinematic segment are selected,and principal component analysis is used to reduce its dimension.Then the k-means clustering algorithm is used to cluster the kinematic segments.Firstly,the silhouette function is used to determine the number of clusters,and the characteristics of the comprehensive driving conditions after clustering are analyzed.Finally,a comprehensive driving condition database is constructed.(3)The establishment of a global offline optimization strategy based on DP.Based on the comprehensive driving condition database constructed above,the offline global optimization of the target vehicle is carried out.Using the prepared dynamic programming to solve it,the optimal torque distribution method of the engine and motor under different working conditions is obtained,and the global optimal control rules are extracted.The optimal control rule base is constructed to provide a basis for designing control strategies based on online identification of driving conditions.(4)Design of energy management strategy based on online recognition of driving conditions.Firstly,the comprehensive driving condition database is divided into condition blocks by the compound equal division method,and its characteristic parameters are selected.Secondly,in order to design a condition identification strategy with a higher recognition accuracy,five condition identification strategies based on neural networks are designed in this paper.The identification accuracy and adaptability of them are compared,the optimal driving condition identification strategy is selected,and the driving condition identification period and prediction period are determined.Then,the optimal operating condition identification strategy is combined with the optimal control rules under different comprehensive driving conditions obtained under the global optimization strategy.An energy management strategy based on online identification of conditions is constructed.Finally,a random driving condition is selected for simulation verification,and it is compared with the CD-CS strategy and the global optimization strategy.The advantages of the energy management strategy based on online identification of operating conditions in terms of energy consumption economy and operating conditions adaptability of vehicles are verified.