基于双启发式动态规划的PHEV能量管理策略

PHEV energy management strategy based on dual heuristic dynamic programming

为提高插电式混合动力汽车(PHEV)燃油经济性和排放性能,提出一种基于双启发式动态规划(DHP)的能量管理控制策略。采用3层BP神经网络分别设计了DHP能量管理控制器的执行网络、模型网络和评价网络。利用Matlab/Simulink与ADVISOR平台,在CYC_NDEC循环工况下对DHP能量管理控制策略进行联合仿真验证,并与执行依赖启发式动态规划(ADHDP)控制算法设计的能量管理控制策略进行对比分析。结果表明,CYC_NEDC循环工况经过1个、2个与3个循环周期后,DHP能量管理控制策略的蓄电池组荷电状态SOC维持在0.63以上,燃油消耗量分别降低了6.9%、4.1%与2.7%,尾气有害物排放量均有一定程度的降低。该方法可有效提升整车的燃油经济性和环保性。

In order to improve the fuel economy and emission performance of plug-in hybrid electric vehicle(PHEV),an energy management control strategy based on double heuristic dynamic programming(DHP)was proposed.Three-layer BP neural network is used to design the action network,model network and critic network of DHP energy management controller.Using Matlab/Simulink and ADVISOR platform,the energy management control strategy of DHP is jointly simulated and verified under CYC_NDEC cycle condition,and compared with the energy management control strategy designed by the action dependent heuristic dynamic programming(ADHDP)control algorithm.The results show that the final SOC value of the battery pack using DHP energy management control strategy is maintained above 0.63,the fuel consumption is reduced by 6.9%,4.1%and 2.7%respectively,and the exhaust harmful substances emissions are reduced to some extent,after one,two and three cycles under CYC_NEDC cycle.Under the condition of ensuring good driving performance,this method can effectively improve the fuel economy and environmental protection of the whole vehicle.