Nonlinear Feedback Control of Hybrid Process Systems

本研究的目的是为多变量混合非线性过程(即具有连续动态和离散事件的过程)建立一个基于非线性模型的反馈控制的综合框架。集中和空间分布的混合系统都将被研究。Lyapunov理论将被用来产生新的解析非线性控制器设计，明确地处理控制执行器的约束和模型的不确定性，并在闭环系统中强制执行期望的稳定性、性能和鲁棒性指标。这项研究的动机来自：a)连续过程动力学与离散事件的共同耦合，b)化学过程模型中丰富的非线性和不确定性，加上对控制执行器能力的硬约束的普遍发生，c)缺乏能够明确和同时处理非线性、不确定性和约束的实用的混合化学过程的非线性控制方法，以及d)日益需要改善化学过程的操作，以减少产品的变异性和异常生产，提高能源效率和减少对环境的影响。为了实现预期的目标，研究将集中于以下项目：a)具有输入约束的多变量混合非线性过程的非线性和鲁棒控制；b)基于非线性状态估值器的非线性和鲁棒控制器的输出反馈实现；c)空间分布混合过程的非线性和鲁棒控制；d)将非线性控制算法应用于具有不确定性和执行器饱和的模拟集中和空间分布混合过程；以及e)开发和实验应用于等离子体增强化学气相沉积(PECVD)反应堆的实时综合测量/混合反馈控制系统。这项研究将对非线性、不确定性、约束和离散事件的存在对我们修改化学过程动态的能力施加的限制提供基本的见解，提供易于在实践中实施的具体控制算法，说明控制方法的应用并得出控制器实施的调整指南，并基于PI在该领域的前人和当前的研究成果出版一本关于“非线性过程控制”的研究专著。新的控制算法有望显著改善具有连续/离散动态、非线性、不确定性和约束的化工过程系统的运行和性能。PECVD工艺的集成测量/控制系统的开发有望减少沉积的空间不均匀性，特别是随着晶片尺寸的不断增加和这些工艺在半导体制造中的广泛使用。

The objective of this research program is to develop a comprehensive framework for nonlinear model-based feedback control of multivariable hybrid nonlinear processes (i.e., processes with combined continuous dynamics and discrete events). Both lumped and spatially-distributed hybrid systems will be studies. Lyapunov theory will be employed to produce novel analytical nonlinear controller designs that deal explicitly with control actuator constraints and model uncertainty and enforce the desired stability, performance and robustness specifications in the closed-loop system. The motivation for the research is provided by: a) the common coupling of continuous process dynamics with discrete events, b) the abundance of nonlinearities and uncertainties in chemical process models coupled with the common occurrence of hard constraints on the capacity of control actuators, c) the lack of practical nonlinear control methods for hybrid chemical processes that can deal explicitly and simultaneously with nonlinearities, uncertainty and constraints and d) the increasing need to improve chemical process operation to reduce product variability and off-spec production, improve energy efficiency and reduce environmental impact. To realize the desired objective, the research will focus on the following projects: a) nonlinear and robust control of multivariable hybrid nonlinear processes with input constraints, b) output feedback implementation of the nonlinear and robust controllers using nonlinear state estimators, c) nonlinear and robust control of spatially-distributed hybrid processes, d) application of the nonlinear control algorithms to simulated lumped and spatially-distributed hybrid processes with uncertainty and actuator saturation, and e) development and experimental application of a real-time integrated measurement/hybrid feedback control system to a plasma-enhanced chemical vapor deposition (PECVD) reactor. The research will provide fundamental insights into the limitations imposed by the presence of nonlinearities, uncertainty, constraints and discrete events on our ability to modify the dynamics of a chemical process, provide concrete control algorithms that can be readily implemented in practice, illustrate the application of the control methods and derive tuning guidelines for the implementation of the controllers, and produce a research monograph on "nonlinear process control" based on previous and current research of the PI in this area. The new control algorithms are expected to lead to significant improvement in the operation and performance of chemical process systems with combined continuous/discrete dynamics, nonlinearities, uncertainty and constraints. The development of integrated measurement/control systems for PECVD processes is expected to reduce spatial non-uniformity of the deposition, especially as wafer dimensions continue to increase and broaden the use of these processes in semiconductor manufacturing.