Predictive Control Of Hybrid Electric Vehicle Energy Management Syste

Under the background of global resource shortage and increasingly serious environmental pollution,plug-in hybrid electric vehicle(PHEV)with energy-saving and emission-reduction characteristics has become the key development direction of the automotive industry.In this thesis,we study the energy management strategy of PHEVs based on predictive control for parallel plug-in hybrid vehicles.The research includes the following parts:(1)Aiming at the problem that the accuracy of calculating the state of charge is not high by using the static battery model,the first-order RC equivalent circuit model of the power battery pack is established to simulate the battery voltage by increasing the dynamic response of the RC loop,and then the relationship between the parameters and the state of charge SOC is given by using data fitting.Furthermore,considering a priori knowledge of the unknown actual operating conditions of the vehicle,a Markovian acceleration predictive model is established based on the data extracted from the cyclic road conditions,which is a prerequisite for the subsequent research on energy management strategy based on model predictive control.(2)Aiming at the problem that the global optimization algorithm cannot be applied online,an energy management strategy for plug-in hybrid vehicles is proposed based on DP-MPC.Firstly,through the analysis of energy management strategy based on dynamic programming algorithm,the energy management strategy of plug-in hybrid vehicle is constructed based on dynamic programming algorithm.Then,based on the above dynamic programming method,the energy management strategy of plug-in hybrid vehicle based on model predictive control is given,and the optimal torque distribution is solved based on the introduced penalty function optimization.Furthermore,the method for determining the SOC reachable area of the power battery is adopted in the solution process,which effectively improves the calculation efficiency.Finally,the numerical simulation results show that the fuel consumption of the energy management strategy based on DP-MPC is close to the global optimization algorithm and is significantly lower than that of the rulebased energy management strategy.(3)Aiming at the hybrid dynamic characteristics in the hybrid system,by combining the hybrid logic dynamic model with the model predictive control,an energy management strategy for plug-in hybrid vehicles is proposed based on HLD-MPC.Firstly,multi-stage linearization are respectively performed on the fuel consumption per unit time and SOC change rate of the engine with nonlinear characteristics.Then,the state and output expressions of the vehicle in the six working modes are derived,and the expressions in different working modes are unified into the general form of the mixed logic dynamic model by introducing the logical variables and auxiliary variables.Furthermore,the system constraints and auxiliary variables in different working modes can be converted into mixed integer linear inequalities to establish a plug-in hybrid vehicle hybrid dynamic model.Finally,based on the idea of moving horizon optimization and online feedback of model predictive control,as well as heuristic algorithms to reduce the scope of the solution,the HLD-MPC optimization problem can be transformed into a mixed integer linear programming problem.Experimental results verify that the HLD-MPC-based energy management strategy proposed has a good optimization effect.This thesis solves the problems of low accuracy of static battery model calculation state of charge in PHEV energy management,the problem that the global optimization algorithm cannot be applied online,and the problem of mixed dynamic characteristics in hybrid system,and proves that the proposed energy management strategy has significant advantages through simulation results.