CAREER: Multi-Rate Model-Based Control of Nonlinear Processes

摘要 CTS-9703278 通用驱动器可更稳定地实现高产品质量、更有效地使用能源以及更严格的安全和环境法规，这对控制系统提出了比单独线性控制更严格的要求。 化学/石化和生物化学工业中非线性过程的丰富性、非线性过程的线性控制和状态估计的不足性、过程在其最佳所在的最大容量下操作的市场压力、以及许多关键过程测量的多速率/延迟性质，是控制器和状态估计器，可以有效地处理这些过程特性的发展的动机。 该项目的研究内容涉及多速率非线性控制和状态估计，并实时应用于工业和中试规模的过程。 具体的研究目标是：（1）通过输入-输出线性化的模型预测控制公式，直接解决离散时间输入-输出线性化控制中的约束处理和积分饱和等基本问题，并推导出存在主动输入约束时闭环稳定的条件;（ii）研究了多速率、延迟测量的非线性状态估计的基本问题，保证了估计器误差动态的渐近稳定;（iii）在IO线性化的框架内开发多速率非线性控制的通用和实用方法;以及（iv）在工业聚合反应器和中试规模生化反应器上实时实施所开发的多速率控制和状态估计方法。 聚合反应器的实时实施将与E。I.杜邦公司 教学活动将包括实验室的发展，监督高中和本科生参加研究，并在德雷克塞尔招聘和专业意识的推广活动。 此外，PI计划专注于两项活动，制定一套开放式工业过程控制问题（每个问题将在本科或研究生过程控制课程中作为项目分配），以及为过程控制和过程建模课程开发交互式，媒体丰富的计算机辅助教学模块。

Abstract CTS-9703278 The universal drive for more consistent attainment of high product quality, more efficient use of energy, and tighter safety and environmental regulations, has imposed far stricter demands on the control system than can be met by linear control alone. The abundance of nonlinear processes in the chemical/petrochemical and biochemical industries, the inadequacy of linear control and state estimation for nonlinear processes, the market pressure for operation of the processes at their maximum capacity where optimum lies, and the multi-rate/delayed nature of many key process measurements, are the motivations for the development of controller and state estimators that can effectively deal with these process characteristics. The research component of this project concerns multi-rate nonlinear control and state estimation with real-time application to industrial and pilot-scale processes. The specific research objectives are: (i) to solve directly the fundamental problems of constraint handling and integral windup in discrete-time input-output (IO) linearizing control, via a model predictive control formulation of IO linearization, and to derive conditions for closed-loop stability in the presence of active input constraints; (ii) to study the fundamental problem of nonlinear state estimation from multi-rate, delayed measurements with guaranteed asymptotically stable estimator error dynamics; (iii) to develop a general and practical method of multi-rate nonlinear control within the framework of IO linearization; and (iv) to implement in real-time, the developed methods of multi-rate control and state estimation on an industrial polymerization reactor and a pilot-scale biochemical reactor. The real-time implementation on the polymerization reactor will be carried out in collaboration with the E. I. duPont deNemours Company. The instructional activities will include laboratory development, supervising high-school and undergraduate students to participate in the resear ch, and outreach activities for recruitment and professional awareness at Drexel. In addition, the PI plans to focus on two activities, formulation of sets of open-ended industrial process control problems (each to be assigned in undergraduate or graduate process control courses as projects), and the development of interactive, media-rich computer-aided instructional modules for the Process Control and Process Modeling courses.