Control Theory for Nonlinear, Distributed, Mechatronic Systems with Applications to Overhead Crane Manufacturing

本计画开发一套适用于精密运输及组装作业之桥式起重机控制系统。该起重机系统为非线性、分布式、机电一体化系统的先进控制理论的发展提供了一个很好的试验平台。起重机模型包括非线性分布式电缆和典型的电驱动马达。 该理论可以应用于许多类似的系统，如智能结构，由磁轴承支撑的柔性转子，以及高速材料运输机械。 桥式起重机具有更高的负载能力，更大的工作空间，更简单的位置监控，以及比叉车或自动引导车更容易的避障。桥式起重机的有效载荷位置和方向控制较差，限制了其在精密运输和装配操作中的应用。由于柔性电缆，高架起重机的有效载荷在运输期间摆动，使得手动远程定位困难并且可能造成损坏或伤害。 在制造环境中，需要高速响应、精确跟踪和良好的稳定性裕度。本研究所提出的集中式与分散式模型控制器是以李雅普诺夫理论为基础。本研究针对全非线性分布模型，研究集总式与分布式李雅普诺夫方法之融合。此外，“低通滤波器状”电动态特性倾向于滤除控制器的部分高频分量。 虽然高增益电流反馈往往会部分抵消这种滤波效果，本研究的目标是设计的backstepping型控制器，精确补偿的电动力学的影响。具体的研究目标是：i）为电驱动的非线性刚性电缆模型开发饱和控制器; ii）为线性分布模型研究可实现的边界控制器; iii）精确地建模非线性柔性电缆系统; iv）发展非线性分布式机电系统的控制理论;以及v）在机器人和机电一体化实验室的小规模实验和全尺寸高架起重机上实现控制器（DEMAG 5吨）在克莱姆森大学的新工程创新大楼。

This project develops an overhead crane control system for precision transport and assembly operations. The crane system provides an excellent testbed for the development of advanced control theory for nonlinear, distributed, mechatronic systems. The crane model includes a nonlinear, distributed cable and typically an electrical drive motor. The theory can then be applied to many similar systems, such as smart structures, flexible rotors supported by magnetic bearings, and high speed material transport machinery. Overhead cranes have higher load capacity, larger workspace, simpler position monitoring, and easier obstacle avoidance than forklifts or Automatic Guided Vehicle. The poor payload position and orientation control of overhead cranes has limited their application in precision transport and assembly operations. Due to the flexible cable, the payload of an overhead crane swings during transit, making manual remote positioning difficult and damage or injury possible. In manufacturing environments, high speed response, accurate tracking, and good stability margins are desired. The lumped and distributed model controllers advanced in this research are based on Lyapunov theory. This research studies the fusion of the lumped and distributed Lyapunov techniques for the full nonlinear, distributed model. Additionally, the `low-pass filter-like` electrical dynamics tend to filter out part of the high-frequency component of the controller. While high-gain current feedback tends to partially null out this filtering effect, this research targets the design of backstepping-type controllers which exactly compensate for the effects of the electrical dynamics. The specific research objectives are to:i) develop saturation controllers for the electrically-driven, nonlinear rigid cable model; ii) investigate implementable boundary controllers for the linear distributed model; iii) accurately model the nonlinear, flexible cable system; iv) develop control theory for nonlinear, distributed, mechatronic systems; and v) implement the controllers on a small-scale experiment in the Robotics and Mechatronics Laboratory and a fullscale overhead crane (DEMAG 5 ton) in the new Engineering Innovation Building at Clemson University.