Research Initiation Award: Robust Control of Chemical Processes

化学处理系统的模型可能会受到不同程度的误差的影响。在许多物质中，可以定量地确定误差范围的合理范围。特别是，椭圆形置信域--确定性模型--可以作为对过程工厂建模的副产品而产生。该项目致力于设计能够满足正常对象以及所有属于邻近椭球不确定区域的对象的性能和稳定性指标的鲁棒控制器。本研究的目的是：(1)开发一种综合的方法来分析具有椭球不确定对象的系统的鲁棒性；(2)开发工具来综合这类对象的鲁棒控制器，并通过仿真研究来测试和评估所得到的控制策略。稳健性分析将通过扩展频域技术的结果来进行，这将导致一个简单的鲁棒稳定的充要条件。计划使用通用控制结构以及诸如极点配置等专门设计来规划新的综合方法。化学工程系统的计算机模拟研究将被用来批判性地评估使用与化学加工工业相关的系统的新的控制策略。

Models of chemical processing systems may be affected by varying degrees of error. In many substances reasonable bounds on the extent of the errors can be determined quantitatively. In particular, ellipsoidal confidence regions-- certainty models-- can be generated as a by product of efforts to model process plants. This project addresses the design of robust controllers capable of satisfying performance and stability specifications for a normal plant as well as for all plants belonging to a neighboring ellipsoidal uncertainty region. The objectives of this research are to (1) develop a comprehensive methodology for analyzing the robustness of systems with ellipsoidally uncertain plants, (2) produce tools for synthesizing robust controllers for such plants, and test and evaluate the resulting control strategies via simulation studies. The analysis of robustness will be carried out by extending the results of a frequency-domain technique which leads to a simple necessary and sufficient condition for robust stability. New synthesis methods are planned using generic control structures as well as specialized designs such as pole-placement. Computer simulation studies of chemical engineering systems will be used to critically evaluate the new control strategies using systems of relevance to the chemical processing industry.