基于ITS和随机预演控制的HEV能量管理策略研究

The main orientation of the trends for new energy automobile is developing hybrid electric vehicles. As the brain and soul of hybrid electric vehicles, energy management system has direct effects on vehicle performance,such as ecomony, safety, reliability and comfort. So, it makes great sense to do researches on energy management problems of HEV. A hybrid electric vehicle always consists of many highly nonlinear and strong coupling components, and the operating circumstance for hybrid electric vehicles is very complicate. Then, the energy management problem of HEVs can be concluded as a class of stochastic global optimization problem with preview nature, whivh is too difficult to solve using traditional methods. So far, energy management problem is still the focus and difficulty in the research field of HEVs. In fact, modern intelligent transportation system can provide us with accurate trip preview information, which can be used to improve the efficiency and performance of energy management system for HEVs. However, there are not yet effective methods, which can make use of the trip preview information to further optimize energy management strategie. Therefore, it is very urgent to establish new theory to resolve it. At first, trip preview information from intelligent transportation systems will be analyzed in this project. Secondly, stochastic traffic model will be constructed and optimized based on preview information. Then, preview control based robust energy management strategies will be developed. Finally, the effectiveness of the new theory and approaches will be illustrated on experimental platform. This proposal is a frontier research involving nulti-disciplines of control theory and engineering, intelligent transportation, vehicle engineering and so on. This research not only makes great sense in paving the way to improve the overall performance of HEV, but also significantly promotes the development of theory and application of related subjects.

混合动力汽车(HEV)是当今新能源汽车发展的主流方向，而能量管理系统直接影响着车辆的经济性、安全性、可靠性和舒适性，是车辆的大脑和灵魂，对其研究意义重大。HEV由多个高度非线性、强耦合部件构成，运行环境复杂多变，其能量管理问题可归结为一类具有预演本质的随机全局优化问题，常规方法难以解决，历来是该领域研究热点和难点之一。事实上，现代智能交通系统(ITS)可为用户提供精确的行程预演信息，可藉此从根本上提升能量管理系统的效能。然而，目前仍缺乏充分利用预演信息实现能量管理最优化的有效方法，亟待寻求新理论予以突破。本项目拟首先定量分析基于ITS的行程预演信息，然后基于预演信息实现随机工况建模与优化，进而提出基于预演控制的鲁棒能量管理策略，并搭建平台予以验证。本项目属于控制理论、智能交通和车辆工程多学科交叉研究，有望为提高HEV整体性能提供一条崭新途径，同时对相关学科理论和应用研究也产生显著促进作用。

本课题组从电动汽车动力总成系统参数匹配优化入手，系统研究了动力总成系统参数匹配优化、转矩协调控制、主动悬架控制及能量管理等一系列问题。所取得的主要成果如下：（1）基于组合矩阵实验和多目标优化方法提出了一类电动汽车动力总成系统匹配优化策略，有效提高了整车动力性和经济性；（2）针对感应电机高阶非线性、强耦合等特点，提出了一种基于平衡点计算的感应电机哈密顿控制策略，顿控制器具有良好的转速响应速度、稳态精度和鲁棒性；（3）提出了一种新的结合模型预测与模型参考控制的转矩协调控制策略，与传统的基于驾驶经验的模式切换控制方法相比，有效降低了模式切换过程的噪声和冲击度，提高了模式切换的平顺性和舒适性；（4）利用径向基函数神经网络等数据驱动方法实现路况的精确预测，进而提出了基于路况信息的滚动优化能量管理策略和回馈制动策略，明显提升了整车的燃油经济性和驾驶性；（5）基于HEV硬件在环综合实验平台验证了所提出控制策略的有效性