Research On Energy Management Strategy For Hybrid Electric Vehicle Based On Adaptive Dynamic Programming Algorithm

With the rapid development of the automobile industry,the demand for oil resources is increasing which caused the problem of environmental polluted by automobile exhaust and other pollutants is worsening.Due to the development of battery technology,the ideal EV is difficult to be popularized quickly.The parallel hybrid electric vehicle has dual advantages of both traditional and pure electric vehicles because of the dual system structure of engine and power driven,and it has great promotion value in the short term.At present,the research on energy management strategy of HEV(Hybrid Electric Vehicle)has become a key issue in the development of HEV industry.In this paper,parallel hybrid electric vehicle is selected as the research object.The following works are mainly done.Firstly,the research background of HEV is introduced.The development of HEV at home and abroad and the research status of HEV's energy management strategy are briefly described.Secondly,in order to maximize the fuel economy and maintain the stable operation of the battery power in the efficient region,the ADP(Adaptive Dynamic Programming)method is proposed which based on optimization,and it is compared with the rule based fuzzy logic control strategy and the adaptive fuzzy neural network control method.Thirdly,because the study of HEV's energy management strategy is easily disturbed by complex and random uncertain factors,such as complex nonlinear system,actual operating condition and driver operation,it is difficult to establish an accurate system model.So the model part of ADP algorithm is approximated by BP neural network algorithm.Then,take the SOC(State of Charge)and the demand torque and speed of the engine as the state variables,take the torque of the motor as the system control variable,considering the instantaneous fuel consumption,design the cost function and select the performance index of the system,explain the energy management strategy under the ADP algorithm in detail.Finally,in order to achieve better control purposes,the initial weight of the action network in the ADP algorithm is approximated and preselected by using the data of the electric auxiliary control strategy in the ADVISOR platform.At the same time,the UDDS(Urban Dynamometer Driving Schedule)of the American urban dynamometer is selected,and these three control strategies are compared and analyzed through the ADVISOR software platform.The results of this paper show that the optimized ADP energy management control strategy can improve the fuel economy of HEV while maintaining stable SOC changes and working in high efficiency compared to the rule based fuzzy control and adaptive fuzzy neural network control strategy.