复合功率分流混合动力系统全过程多目标变时尺非线性模型预测控制

Power-split hybrid powertain system is the industrialized scheme for the plug-in hybrid electric vehicle(PHEV), which has been vigorously promoted by the Toyota, GM and other automotive companies. In order to break through the technology monopoly, the applied project will be implemented basing on the development and industrialization of compound power-split PHEV which have independent intellectual property rights, the methods of its nonlinear dynamic modeling in the multiple energy domain and model-based torque estimation for the engine, brakes and clutch will be discussed, and the model predictive realtime optimization control algorithm characterized by multi-objective and variable time scale during the whole process will be finally developed and apllied for the compound power-split PHEV. Firstly, the steady operating mode and transient switching process are analyzed, and the different graphical dynamic models are established and analysis of system performance are conducted considering the elasticity of planetary gear ,the friction of tranmission shafts,the nonlinear characteristics of the torsional spring damper, brakes and clutch, the dynamic torque characteristics of the engine and two motors. Secondly, the engine output torque and the torque transmitted by the two brakes are observed accurately by using the nonlinear Kalman state estimation theory. especially steady energy management strategy and transient mode switching control algorithm are onlinely and realtimely optimized through the integration of dynamic programming and model predictive control theory.Finally, the provided powertrain's nolinear dynamic models and realtime optimal control methods for the compound power-split hybrid system are respectively tested and validated by using the hardware in-the-loop simulation platform ,the powertrain's dynamic performance test bench and the real protype car. For the power-split hybrid powertrain system, implementation of this project will expectly not only produce a theoretical methods of nonlinear dynamics modeling and real-time online optimization control, but also provide an engineering solution scheme for the synchronous improvement of fuel ecnomy and driviability performance.

功率分流混合动力系统已成为丰田、通用等汽车公司力推的产业化插电式混合动力方案，为打破技术垄断，项目结合自主品牌复合功率分流混合动力系统开发，探讨其多能域非线性动力学建模和基于模型的转矩估计方法，旨在开发全过程多目标变时间尺度的模型预测实时优化控制算法。首先，分析其稳态工作模式及瞬态切换过程，考虑行星齿轮弹性、轴系摩擦、扭转减振器和制动器非线性特性、发动机和电机动态转矩特性，进行图形化动力学建模和系统性能分析；其次，使用非线性卡尔曼状态估计理论，对发动机、制动器和离合器转矩进行观测，并综合运用动态规划和模型预测控制理论，对其稳态能量管理及瞬态模式切换过程进行在线优化；最后，利用硬件在环仿真台架、动力总成动态台架和样车对所开发动态模型和优化控制算法进行试验。项目实施将为认识和分析功率分流混合动力系统动力学本质，提供实时优化控制方法，充分挖掘其节能潜力并改善其驾乘性能提供理论依据和解决对策。

功率分流混合动力系统利用行星传动装置和电机实现发动机工作状态与车轮转速、需求转矩的解耦，是深度和插电式混合动力系统的重要技术方案。为充分挖掘功率分流混合动力系统的节能潜力并改善其驾驶平顺性，项目研究了其构型设计优化方法及稳态、瞬态全过程动态控制策略。主要研究工作及结论如下：. 1）实现了基于图论的双行星排功率分流混合动力系统构型设计及软件开发.提出了基于图论模型的数学描述方法，研究了基于设计实验抽样的分步构型筛选方法，建立了双行星排功率分流混合动力系统构型评价体系，开发了系统构型图形化自动设计及优化软件。. 2）搭建了复合功率分流混合动力系统非线性动态模型.通过机理分析并结合试验数据，搭建了发动机阻力矩和湿式制动器等关键零部件的非线性动态模型。通过集成各零部件模型并考虑传动系轴的等效刚度及阻尼，搭建了复合功率分流混合动力汽车数学模型，并利用实车进行模型验证。. 3）开发了功率分流混合动力系统稳态及瞬态全过程模型预测优化控制策略.针对全过程稳态工作模式，设计了发动机工作点的最优决策律和模型预测能量管理策略。针对瞬态切换过程，结合离线动态规划和在线模型预测控制，对模式切换过程各动力元件及执行元件的目标输出进行求解。. 4）实现了复合功率分流混合动力系统转矩估计及其补偿控制.设计了基于T-S模糊模型的未知输入观测器，实现了变速箱输入轴和动力输出端转矩估计。基于转矩估计结果，开发了电机转矩补偿控制策略以抑制传动系转矩波动。. 5）完成了转矩估计及系统全过程预测控制的试验验证.设计并搭建了功率分流混合动力系统动态性能测试台架，通过硬件在环测试验证了转矩估计和系统全过程预测优化控制策略的有效性，基于道路实车试验验证了系统瞬态预测控制方法。. 项目提出了基于图论的功率分流系统构型设计方法，通过综合利用转矩估计、动态规划和预测控制，开发了系统全过程动态优化控制策略，为全面提升复合功率分流混合动力系统的稳、动态性能提供了理论依据和解决方案。