Modeling, Identification and Validation of Uncertain Linear and Nonlinear Feedback Systems

拟议工作的目标是开发一套集成的工具，用于不确定系统的建模，以用于反馈控制设计目的。我们的建模框架基于鲁棒控制理论，并由几种复杂的控制设计方法支持。这些模型实际上是集合描述，其中集合是由模型中未知但有界的动态扰动生成的。这些模型允许设计者捕获物理系统将生成的各种不确定行为。虽然这是一个强大而灵活的建模框架，但可用于对物理系统建模的工具是有限的。我们在此提出的工作的基础上，对鲁棒控制模型的模型验证进行了先前的工作。这种方法导致了一些工具，允许设计者通过实验来量化物理系统和稳健控制模型集中最接近的成员之间的差异。目前的工作着眼于将这些方法扩展到替代的实验场景。一个方向将是使用相关的正弦扫描实验；这是一种在工业实践中经常使用的程序。第二个方向是考虑来自不同信号类别的噪声和扰动；例如随机噪声和增益有界扰动。这些模型更接近于典型的物理系统，并将使这些方法适用于更广泛的工业过程。我们还建议开发一个鲁棒控制建模框架，它将处理某些类型的非线性系统。该项目将考虑这些课程中的两个工业驱动的例子：喷气发动机压缩机中的旋转失速；以及燃烧室中的热声相互作用。这些方法将被应用于这些问题的简化模拟版本。非线性行为和动态扰动的结合开辟了非线性稳健性分析的一个领域，这一领域也将被讨论。有三个实验可用来支持这项研究：(1)金属有机化学气相沉积(MOCVD)反应器，用于蓝色LED和蓝色激光二极管的外延晶体生长。(2)磁悬浮轴承系统，作为人工心脏泵(左心室辅助装置)的原型开发。(3)快速热处理系统，模拟半导体和胶片加工行业中典型加工步骤的动态。这些实验中的每一个都表现出理论建模和辨识方法尚未很好地解决的共同实际困难，使它们成为所提出工作的理想试验台。

The objective of the proposed work is to develop an integrated suite of tools and approached for the modeling of uncertain systems for feedback control design purposes. Our modeling framework is based upon robust control theory, which is supported by several sophisticated control design methods. These models are in effect set descriptions, where the sets are generated by unknown but bounded dynamic perturbations within the model. The models allow the designer to capture the variety of uncertain behaviors that a physical system will generate. Although this a powerful and flexible modeling framework, the tools available for modeling physical systems are limited.Our prior work on model validation for robust control models from the basis for the work proposed here. This approach has led to tools which allow the designer to experimentally quantify the difference between a physical system and the closest member of a robust control model set. The current work looks at extending these approaches to alternative experimentally scenarios. One direction will be the use of correlated swept-sine experiments; a procedure frequently used in industrial practice. A second direction will be the consideration of noise and perturbations from different signal classes; for example stochastic noise and gain bounded perturbations. Such models agree more closely with typical physical systems, and will make these methods applicable to a wider range of industrial processes. We also propose to develop a robust control modeling framework which will handle certain classes of nonlinear systems. The project will consider two industrially motivated examples from these classes: rotating stall in jet engine compressors; and thermo-acoustic interactions in combustors. The methods will be applied to simplified simulation versions of these problems. The combination of nonlinear behaviors and dynamic perturbations opens up an area of nonlinear robustness analysis which will also be addressed. Three experiments are available and will be used in support of this research: (1) A metalorganic chemical vapor deposition (MOCVD) reactor, used for epitaxial crystal growth of blue LEDs and blue laser diodes. (2) A magnetically levitated bearing system, developed as a prototype for an artificial heart pump (left ventricular assist device). (3) A rapid thermal processing system, which models the dynamics of typical processing steps in the semiconductor and film processing industries. Each of these experiments exhibits common practical difficulties which are not yet well handled by theoretical modeling and identification methods, making them ideal testbeds for the work proposed.