Research On Energy Management Strategy Of Parallel Hybrid Electric Vehicle Based On Dynamic Programming

With energy and environmental issues have become increasingly prominent,the research on hybrid electric vehicle(HEV)is of great significance.As a vehicle with different energy sources,energy management and torque distribution strategy is the key to improve fuel economy and reduce emissions.Energy management strategy of hybrid electric vehicle is to achieve a reasonable distribution of energy in the HEV,and improve vehicle fuel economy and emissions performance,according to the characteristics of hybrid powertrain and real-time driving cycle,on the promise of meeting the dynamic requirements.For the sake of improving fuel economy of a single axis parallel hybrid electric vehicle,the paper addresses optimal energy management and the control strategy of the vehicle,and the main research work is summarized as follows:(1)Based on the analysis of parallel hybrid electric vehicle's structural characteristics and experimental data,the mathematical models of vehicle and major components in HEV are constructed in Matlab environment,using the empirical modeling approach with the aid of theoretical modeling approach.It provides the essential simulation platform for the further research on the energy management control strategy.(2)Research on energy management strategy based on dynamic programming.In this paper,the battery SOC and vehicle driving requirement torque as the state variables,the engine output torque is the control variable and achieve the minimum fuel consumption of the engine for the control target,the energy management strategy based on dynamic programming is also designed.And the designed energy management strategy is simulated in the Matlab platform.(3)Optimization of real time energy management strategy for multi neural network model.Clustering the global optimal datasets obtained by dynamic programming using kernel-fuzzy C-means clustering algorithm,and design a multi neural network(MNN)model controller,the simulation results are presented that the multi neural network mode strategy has the excellent learning capability to dynamic programming(DP),and it effectively make up for the defects of DP.