基于数据/机理混合模型的DCT换挡系统滚动时域优化控制及实现研究

Dual Clutch Transmission (DCT) has advantages of high efficiency, energy saving and environmental protection. However, the control degrees and complexity of dynamic coupling are increased with the actuator increased. The automatic shift schedule optimization and gearshift control are basically two core issues for developing the DCT control system. There are a few issues are remaining, particularly in some special driving conditions, such as frequent and unexpected gearshift, shaking during gearshift process at low speed, large amount of calibration during design process. The control problem of DCT gearshift system is a constrained multi-variable and multi-objective optimization problem, which is suitable to be solved under the framework of Model Predictive Control (MPC),which is a control method often used in real time with ability to explicitly solve problems with system constraints based on updating data. In this project, under the framework of MPC, we focus on developing hybrid data/mechanism model in gear shift optimization and clutch control; proposing an intelligent strategy for gear shift optimization and clutch control during gear-changing process of vehicles with DCT, in which factors of drivability like dynamic performance, comfort of the driver and driving habits are considered using road information obtained by GPS and GIS, such as slope, bend and levels; developing a fast algorithm and achieving its real-time implementation with lower cost, for engineering applications to the moving horizon optimal control problem. The urgent problems of DCT in engineering practice are expected to be solved through the implementation of the project, and the longitudinal dynamic performance of vehicles would be improved. Moreover, the project will promote the trend of interdisciplinary of control science and vehicle engineering, and will broaden the field of control applications, and thus has a very important academic and practical meaning.

双离合器自动变速器(DCT)具有效率高、节能环保的优势，但随着执行机构的增加，控制自由度和动力学耦合复杂度也随之增加。换挡规律优化和换挡过程控制是DCT换挡电控系统开发的核心，目前仍存在“频繁换挡和意外换挡、低速换挡顿挫感明显、控制器标定量大”等问题。这类控制问题属于有系统约束的多目标协调优化问题，适合在具有信息实时更新及显式处理约束条件能力的模型预测控制MPC框架下加以解决。因此，本项目重点研究：建立面向DCT换挡系统优化和控制任务的数据/机理混合模型；利用车载定位系统等提供的道路信息(坡道，弯道等)，综合考虑动力性、舒适性及驾驶习惯等因素，提出DCT滚动优化智能换挡策略及换挡过程滚动时域优化控制方法；研究面向工程应用的滚动时域优化方法高效低成本实现。通过本项实施目有望解决DCT系统实际应用的迫切需求问题，充分发挥控制理论与控制科学在换挡控制系统优化、控制和实现技术的基础支撑作用。

随着人们对汽车性能需求的提高, 执行机构的增加，使其控制自由度和动力学耦合复杂度也随之增加，给系统建模及控制带来了新的挑战。本项目充分利用汽车行驶过程中产生的数据，研究了数据机理混合建模及控制方法。首先，建立了连续动态模型、数据图表及输入输出数据关系混合的模块化模型，为控制任务提供了可靠的模型；其次，综合考虑经济性和排放性的平衡需求，研究了节气门开度、节气门开度的变化率、车速、整车加速度及道路坡度等一系列参数与挡位的关系，构建了带有约束的多目标协调实时优化问题，提出了一种利用高精地图信息的预测巡航控制系统；针对预测巡航控制系统的实车实现困难问题，提出了车辆巡航工况切换表和一种基于二分法的快速求解算法，实现了预测算法的快速求解；同时，将所提出的前馈-反馈结构的滚动优化控制方法应用到车辆稳定性控制，解决了系统性能需求之间的矛盾，提高了系统的响应速度及稳定性。然后，针对汽车换挡控制多目标优化问题(本质上响应速度快和平顺性好的矛盾需求，离合器滑摩功小的需求)，考虑系统约束(包含发动机、离合器、电机等多个动力源和执行器饱和及频响受限，发动机转速不能过低等)，提出了基于预测控制方法换挡控制器，并将控制器在FPGA上进行硬件电路实现；最后，进行了相关领域的拓展研究，主要包括汽车底盘控制、转向控制及能源优化等方面。此项目的实施及研究成果充分表明了控制理论与控制科学在车辆动力传动系统优化、控制和实现技术方面的基础支撑作用，拓宽了控制应用领域，对解决提升汽车性能实际应用研究的迫切需求，具有重要的理论意义和实际价值。