混合动力汽车的动力最优控制策略研究
|
摘要
提出一种基于行驶特征预测和离线最优轨迹的最优控制策略,将离线最优轨迹运用到在线中。利用动态规划(DP)算法获得了离线全局最优轨迹。将行驶特征分为行驶工况和行驶模式。选取了11种标准行驶工况,利用欧几里得贴近度法实现了对行驶工况的预测;将行驶模式定义为5种类别,利用BP神经网络实现了对行驶模式的预测。采用神经网络对标准工况下离线最优轨迹及相应汽车状态进行学习,设计了基于神经网络和离线最优轨迹的能量分配模型,以及基于行驶特征预测和离线最优轨迹的最优控制策略并进行仿真验证。结果表明:与ADVISOR自带的电机助力策略相比,所提的最优控制策略使得燃油经济性提高了7.51%,同时工况适应性良好。
It presents an online control strategy based on the prediction of driving characteristics and the off-line optimal trajectory. The off-line optimal trajectory is applied to the on-line control. The off-line global optimal trajectory is obtained in dynamic programming(DP) algorithm. The driving characteristics are divided into driving conditions and driving modes. Eleven standard driving conditions are selected, and the Euclidean proximity method is used to realize the prediction of driving conditions. The driving modes are defined as five categories, and the BP neural network is used to realize the prediction of driving modes. The off-line optimal trajectory and the corresponding vehicle state are studied in the neural network. The energy allocation model based on the neural network and the off-line optimal trajectory is designed. Combined with the prediction of driving characteristics, a comprehensive on-line control strategy based on the prediction of driving characteristics and the off-line optimal trajectory is designed and verified by simulation. The results show that the fuel economy of the proposed optimal control strategy is improved by 7.51% compared with the motor-assisted strategy of ADVISOR, and the operation condition adaptability of the strategy is good.
It presents an online control strategy based on the prediction of driving characteristics and the off-line optimal trajectory. The off-line optimal trajectory is applied to the on-line control. The off-line global optimal trajectory is obtained in dynamic programming(DP) algorithm. The driving characteristics are divided into driving conditions and driving modes. Eleven standard driving conditions are selected, and the Euclidean proximity method is used to realize the prediction of driving conditions. The driving modes are defined as five categories, and the BP neural network is used to realize the prediction of driving modes. The off-line optimal trajectory and the corresponding vehicle state are studied in the neural network. The energy allocation model based on the neural network and the off-line optimal trajectory is designed. Combined with the prediction of driving characteristics, a comprehensive on-line control strategy based on the prediction of driving characteristics and the off-line optimal trajectory is designed and verified by simulation. The results show that the fuel economy of the proposed optimal control strategy is improved by 7.51% compared with the motor-assisted strategy of ADVISOR, and the operation condition adaptability of the strategy is good.
引文
[1] TAN R, TANG D, LIN B. Policy impact of new energy vehicles promotion on air quality in Chinese cities[J]. Energy Policy, 2018, 118: 33-40.
[2] 秦大同,赵新庆,苏玲,等.插电式混合动力汽车变参数能量管理策略[J].中国公路学报,2015,28(2): 112-118.
[3] 林歆悠,孙东野.基于工况识别的混联式混合动力客车控制策略研究[J].中国机械工程,2012,23(7): 869-874.
[4] 詹森,秦大同,曽育平.基于遗传优化K均值聚类算法工况识别的混合动力汽车能量管理策略[J].中国公路学报,2016,29(5):130-138.
[5] 连静,常静,李琳辉,等.基于模糊在线识别的并联混合动力客车自适应控制策略[J].北京理工大学学报, 2016,36(3):264-270.
[6] CASTLE N, ENGBERG J B, WAGNER L M, et al. Resident and facility factors associated with the incidence of urinary tract infections identified in the nursing home minimum data set[J]. Journal of Applied Gerontology, 2017, 36(2): 173-194.
[7] 楚峰.中国新能源汽车发展历程[J].运输经理世界, 2015(7):58-59.
[8] RUI W,SRDJAN K. Review of driving conditions prediction and driving style recognition based control algorithms for hybrid electric vehicles[C]//Proceedings of VPCC 2011 Vehicle Power and Propulsion Conference. New York:IEEE, 2011:1-7.
[9] MURPHEY Y L, PARK J, KILIARIS L, et al.Intelligent hybrid vehicle power control-Part Ⅱ: Online intelligent energy management[J]. Vehicular Technology, IEEE Transactions on, 2013,61(1):69-79.
[10] PARK J, CHEN Z H, KILIARIS L,et al.Intelligent vehicle power control based on machine learning ofoptimal control parameters and prediction of road type and traffic congestion[J].Vehicular Technology, IEEE Transactions on, 2009,58(9):4741-4756.
[11] WANG Z,XU G, LI W,et al.Driving Load Forecasting Using Cascade Neural Networks[M]. Berlin:Springer,2007.
[12] 田毅,张欣,张良,等.神经网络工况识别的混合动力电动汽车模糊控制策略[J].控制理论与应用,2011,28(3):363-369.
[13] 姜平.城市混合道路行驶工况的构建研究[D].合肥:合肥工业大学,2011.
[14] 张月琴,刘翔,孙先洋.一种改进的BP神经网络算法与应用[J].计算机技术与发展, 2012,22(8): 163-166.
[15] MURPHEY Y L, PARK J, KILIARIS L, et al.Intelligent hybrid vehicle power control-Part I:Machine learning of optimal vehicle power[J].Vehicular Technology, IEEE Transactions on, 2013,61(8): 3519-3530.
[16] JIE Xing, HAN Xuefeng,YE Hui,et al.Driving cycle recognition for hybrid electric vehicle [C]//IEEE Conference and Expo. Beijing:[s.n.],2014:1-6.
[17] 李翔晟,杨三英,周永军.混合动力汽车电池SOC 的仿真与修正[J].电源技术,2011,35(12):1589-1591.
[2] 秦大同,赵新庆,苏玲,等.插电式混合动力汽车变参数能量管理策略[J].中国公路学报,2015,28(2): 112-118.
[3] 林歆悠,孙东野.基于工况识别的混联式混合动力客车控制策略研究[J].中国机械工程,2012,23(7): 869-874.
[4] 詹森,秦大同,曽育平.基于遗传优化K均值聚类算法工况识别的混合动力汽车能量管理策略[J].中国公路学报,2016,29(5):130-138.
[5] 连静,常静,李琳辉,等.基于模糊在线识别的并联混合动力客车自适应控制策略[J].北京理工大学学报, 2016,36(3):264-270.
[6] CASTLE N, ENGBERG J B, WAGNER L M, et al. Resident and facility factors associated with the incidence of urinary tract infections identified in the nursing home minimum data set[J]. Journal of Applied Gerontology, 2017, 36(2): 173-194.
[7] 楚峰.中国新能源汽车发展历程[J].运输经理世界, 2015(7):58-59.
[8] RUI W,SRDJAN K. Review of driving conditions prediction and driving style recognition based control algorithms for hybrid electric vehicles[C]//Proceedings of VPCC 2011 Vehicle Power and Propulsion Conference. New York:IEEE, 2011:1-7.
[9] MURPHEY Y L, PARK J, KILIARIS L, et al.Intelligent hybrid vehicle power control-Part Ⅱ: Online intelligent energy management[J]. Vehicular Technology, IEEE Transactions on, 2013,61(1):69-79.
[10] PARK J, CHEN Z H, KILIARIS L,et al.Intelligent vehicle power control based on machine learning ofoptimal control parameters and prediction of road type and traffic congestion[J].Vehicular Technology, IEEE Transactions on, 2009,58(9):4741-4756.
[11] WANG Z,XU G, LI W,et al.Driving Load Forecasting Using Cascade Neural Networks[M]. Berlin:Springer,2007.
[12] 田毅,张欣,张良,等.神经网络工况识别的混合动力电动汽车模糊控制策略[J].控制理论与应用,2011,28(3):363-369.
[13] 姜平.城市混合道路行驶工况的构建研究[D].合肥:合肥工业大学,2011.
[14] 张月琴,刘翔,孙先洋.一种改进的BP神经网络算法与应用[J].计算机技术与发展, 2012,22(8): 163-166.
[15] MURPHEY Y L, PARK J, KILIARIS L, et al.Intelligent hybrid vehicle power control-Part I:Machine learning of optimal vehicle power[J].Vehicular Technology, IEEE Transactions on, 2013,61(8): 3519-3530.
[16] JIE Xing, HAN Xuefeng,YE Hui,et al.Driving cycle recognition for hybrid electric vehicle [C]//IEEE Conference and Expo. Beijing:[s.n.],2014:1-6.
[17] 李翔晟,杨三英,周永军.混合动力汽车电池SOC 的仿真与修正[J].电源技术,2011,35(12):1589-1591.