【作者】 吴迪；

【导师】 赵韩；

【作者基本信息】 合肥工业大学 ， 车辆工程， 2013， 博士

【摘要】 能源危机和环境污染是当今世界所面临的两大挑战,混合动力汽车具有低油耗、低排放的特点,并且不受里程限制,被认为具有很好的产业化前景。ISG混合动力汽车能量管理策略设计是整车设计的重要环节,也是实现ISG混合动力汽车节能减排的关键技术之一因此能量管理策略成为ISG混合动力汽车研究的重点。本文以“十一五”国家863节能与新能源汽车重大专项ISG混合动力汽车项目及其整车控制策略设计子课题为背景,建立了ISG混合动力汽车的前向仿真模型,分别基于动态规划算法、模型预测控制算法及随机模型预测控制算法对ISG混合动力汽车的能量优化管理策略进行研究,并基于dSPACE实时控制系统进行ISG混合动力汽车转矩分配的快速原型仿真试验。本文所做主要工作及结论如下：基于Matlab/Simulink平台,采用实验建模和理论建模相结合的方法建立混合动力汽车整车前向仿真模型。通过实验建模的方法,分别建立发动机、ISG电机以及动力电池的仿真模型,通过理论建模的方法,建立离合器、变速箱等传动系模型。根据前向仿真的能量传递方向,基于上述部件模型,搭建ISG混合动力汽车整车模型,为ISG混合动力汽车能量管理策略的验证提供了基础。对给定的循环工况离散化,将ISG混合动力汽车的能量优化管理转化为多阶段决策过程,结合改进后的动态规划算法,以油耗、换挡次数以及SOC平衡为优化目标建立每个阶段的代价函数和目标函数,选取档位迁移和电机扭矩作为控制变量、SOC和档位为状态变量,优化得到在该工况下的最优电机扭矩序列和最佳档位迁移序列。仿真结果表明该能量管理策略具有很好的燃油经济性。提出了利用指数函数预测车轮扭矩的方法,将模型预测控制与动态规划相结合,建立了基于模型预测控制且可以燃油消耗为目标函数进行滚动优化的控制策略,通过优化得到最优电机控制扭矩。分析预测时间长短对ISG混合动力汽车燃油经济性的影响,选取合适的预测时间。基于ISG混合动力汽车整车仿真模型在NEDC工况下对该能量管理策略进行仿真,仿真结果表明,该能量管理策略不仅具有很好的燃油经济性,还具有较好的实时性。提出了基于随机模型预测控制的ISG混合动力汽车能量优化管理策略。建立了基于马尔科夫链的驾驶员模型,通过该模型预测驾驶员需求功率,建立随机模型预测控制与动态规划相结合的优化模型,通过滚动优化得到最佳扭矩分配方案。基于ISG混合动力汽车整车仿真模型,对该能量优化管理策略进行仿真,并选取恒值预测方法和具有预先知识的预测方法进行对比,仿真结果表明随机模型预测方法是可行的,ISG混合动力汽车的随机模型预测控制能量优化管理策略具有很好的燃油经济性和实时性。搭建基于dSPACE实时系统的仿真台架,进行了ISG混合动力汽车随机模型预测控制能量优化管理策略的快速原型仿真试验,验证了随机模型预测控制优化管理策略的基本性能,使得该控制策略得到进一步的完善。更多还原

【Abstract】 Resource shortage and environment pollution are two big challenges in the world nowadays. Hybrid electric vehicle is considered as most promising kind of vehicle for industrialization because it has many advantages, such as low energy consumption, low emission and long drive distance. The design of energy management strategy of ISG HEV is one of important part for the vehicle development, and one of key techlonogy for the energy saving and emission reduction of ISG HEV. Therefore the energy management strategy becomes the research focus for ISG hybrid electric vehicle.This dissertation takes a ’national863program’ project as background and takes an ISG HEV as research object. The forward simulation model for ISG hybrid electric vehicle was founded. The research on energy optimal management strategy for ISG hybrid electric vehicle based on dynamic programming, model predictive control and stomastic model predictive control were studied separately. The rapid control prototype simulation test were implemented based on dSPACE real-time syetem. The main research work and conclusions were summarized as following:Based on Matlab/Simulink, the ISG HEV forward simulation model was developed to provide the simulation and verification foundation of energy management strategy. The engine model, ISG motor model and power battery model were founded by using the experimental data. The clutch model and gearbox model were founded based on physical logic of parameters.The energy optimal management was converted into multistage decision process by discretization of driving cycle. The cost function for each stage and the object function were founded by using the fuel consumption, the number of shift and SOC balance as optimal objection. The gear shift and motor torque were selected as control variable. The SOC and gear were selected as state variable. The optimal motro torque horizon and gear shift horizon were got by the optimization. The simulation results show that the energe management strategy has good fuel economy performance.A method of exponential function was proposed to predict torque of the wheel. The energy management strategy was established by the rolling optimization for the fuel consumption based on the combination of model predictive control algorithm and dynamic programming. The motor torque horizon was got by the rolling optimization. The influence of the prediction horizon on the optimization results was studied. The simulation for the energy management strategy was developed based on the ISG HEV model under the NEDC. The results show that it has good fuel economy and it is real-time implementable. An energy management strategy based on stochastic model predictive control (SMPC) was proposed. The driver model based on Markov chain was developed to predict the required power of driver. The optimization model was developed by combining the dynamic programming and model predictive control. The optimal toque distribution was got by rolling optimization. The simulation was developed and the results were compared to the results for two other predictive methods. The results show that the method of stochastic predictive is feasiable and the energy management strategy has good economy performance.The test bench based on dSPACE real-time system was build and the rapid prototype simulation test for the SMPC energy management stategy of ISG HEV was implemented. The energy management strategy was further verified by the test.更多还原