Research On Energy Management Strategy Of Plug-in Hybrid Electric Vehicles Based On Prediction Of Working Conditions

With the advantages of superior fuel economy and environmental performance,plug-in hybrid electric vehicles(PHEVs)have become an effective tool to solve the problems regarding the lack of fossil energy and environmental pollution.Focusing on power-split PHEVs,this dissertation proposes a fast energy management strategy considering engine ON/ OFF control,and a drive-cycle-prediction based hierarchical energy management strategy,whose studies include three parts: the fast planning of state-of-charge(SOC)trajectory,the short-term velocity prediction and the real-time rolling predictive control.The main contributions are illustrated in the following.Introduced the power system of power-split PHEVs.The power-split PHEV's topology structure and operation modes are demonstrated profoundly,and meanwhile,a series of key vehicle units as well as the relationship of powertrain energy flows are depicted in detail.Then,by the software Autonomie,the Simulink model of the powersplit PHEV is also illustrated.In light of the engine work characteristic,the engine optimal operation line(OOL)is introduced,analyzed and extracted.Introduced an optimization based charge depleting-charge sustaining(CD/CS)strategy and dynamics programing(DP)based globally optimal energy management strategy.After constructing the vehicle energy balance model,the optimization based CD/CS strategy is put forward according to the equivalent consumption minimum theory.Additionally,by adopting the OOL,the optimal energy management strategy based on DP is achieved after introducing the related fundamentals.Simulation shows that this method can improve the fuel economy effectively.Proposed a fast energy management strategy considering the engine ON/ OFF control.It can conduct the SOC trajectory planning rapidly and efficiently.After analyzing the powertrain structure in detail,the fuel rate is fitted versus the battery power by a series of quadratic equations,and the related approximation accuracy is verified.On the basis of Pontryagin's minimum principle(PMP),the corresponding deductions are conducted so that the optimal engine ON/ OFF status and battery power can be determined.After that,estimation distribution algorithm(EDA)is applied to search the optimal equivalent factor with the objective cost of the final SOC constraint.Numerical simulations prove that the proposed strategy can achieve the near-optimal fuel performance and SOC trajectory planning which are similar to the result by DP.The calculation time of this approach is reduced for 96 % compared to that by DP,and the average calculation time for a driving cycle of 3600 s is deduced to be 8.6 s,illustrating the applicably computational cost.Moreover,the SOC curve optimized by the addressed method can be adopted as the control reference for online strategies.Constructed and achieved a hierarchical energy management strategy based on velocity prediction.The missions include the fast planning of battery SOC curve,velocity prediction and online energy management.In the upper control module,the SOC trajectory is planned rapidly by the forgoing optimization method,which is also adopted as the control reference in the low-level control.A velocity-prediction method based on wavelet transformation(WT)and radial basis function neural network(RBFNN)is introduced to realize accurate vehicle speed prediction.To balance the conflict between searching performance and constraints,a model prediction controller(MPC)with relaxation factor is established in the low-level control,as to realize the fast and local energy management based on the predicted velocity and the planned SOC trajectory.Simulation result demonstrates that,the proposed velocity predictive method can improve the predictive accuracy of 9.26 % and the fuel economy of 0.3 % to 0.78 %.Additionally,this hierarchical strategy can achieve the fuel economy similar to that by the fast energy management strategy,with the single-step calculation time of 2.44 ms to 3.3 ms.Therefore,in other words,the addressed strategy can increase the accuracy of the velocity prediction and can improve the fuel economy with a fast calculation speed.