| Hybrid electric vehicle(HEV)is an effective way to solve the problem of excessive energy consumption and air pollution nowadays.Compared with other HEV powertrain structures,dual-mode power-split HEV can meet the special requirements of heavy-duty non-road vehicles.It can provide wide speed range,large driving power,auxiliary system power and other electrical power.But for a dual-mode power-split HEV,the control of energy distribution and components coordination is more complex.So it raises higher requirements for the research on the control strategy.In this paper,based on the research of online vehicle velocity prediction and optimization control,a real-time predictive energy management strategy is proposed for a dual-mode power-split HEV.Firstly,the structure and characteristics of the dual-mode power-split HEV which is researched in this paper are analyzed.The mathematical model of the system is established by the combination of experimental modeling method and theoretical modeling method.Based on Matlab/Simulink software,a comprehensive control strategy including mode switching control,rule-based energy management strategy and PID-based dynamic coordinated control is developed.The feasibility of this comprehensive control strategy is verified by software simulation and it gives the foundation of further research on energy management strategy.Then,an energy management strategy based on linear model predictive control is proposed.A prediction model for describing the future dynamics of the system is established.The system model is linearized and discretized at each sampling time,and the quadratic programming problem is solved online for optimal control of power distribution.The effectiveness of the proposed strategy is verified by simulation analysis.And the fuel economy under this strategy is greatly improved compared with the rule-based strategy.However there is still room for improvement.In order to make the energy management strategy have a better forward-looking,a comprehensive prediction method of the future vehicle velocity is proposed.The Kmeans clustering algorithm is used in the offline phase to classify the vehicle working conditions into stationary and changing conditions.And the current operating condition is judged in the online phase.A method of vehicle speed prediction based on adaptive Markov chains is proposed for the stationary condition.The adaptive updating coefficient is introduced to update the probability transfer matrix online.A vehicle speed prediction method based on adaptive radial basis function neural network(RBFNN)is proposed for the changing conditions.The neural network is used to predict the future vehicle speed from the historical vehicle speed and the current driver's pedal position.The validity of the comprehensive method is demonstrated by simulation and comparison with other methods.In order to further improve the optimization effect,an energy management strategy based on nonlinear model predictive control is proposed.An improved dynamic programming algorithm is proposed to solve the nonlinear optimization problem online.The range of dynamic programming search area is reduced after analyzing the system characteristics,which greatly reduces the computational complexity and makes it have practical application ability.Combining with the method of comprehensive future vehicle velocity prediction,the effectiveness of the proposed strategy is verified by software simulation,and the fuel economy of the vehicle is improved compared with other methods.Finally,a set of development environment including software simulation,hardware-in-the-loop simulation and bench test is developed.The bench test platform is built and the integrated control system is developed.The effectiveness of the energy management strategy proposed in this paper is verified by the bench test.The controller developed in this study has realized the design goal,and has practical value to improve the fuel economy of the HEV. |
PDF Full Text Request