Design and Monitoring of Cooperative, Distributed Control Systems for Nonlinear Processes

1027553 christofides异常情况的优化操作和管理是过程工业面临的主要挑战，例如，仅在美国，异常情况每年至少造成100亿美元的收入损失。这一认识激发了过程控制领域的重要研究，以确保安全高效的过程运行。传统上，控制系统依赖于集中式控制架构，利用专用的有线链路连接到测量传感器和控制执行器，以将适当的过程变量调节到所需值。虽然这种过程控制的范例是成功的，但当化工厂的过程状态变量、被操纵的输入和测量的数量变得很大时（这在现代工厂中很常见），解决集中控制问题所需的计算时间可能会显著增加，并可能阻碍集中控制系统（特别是当使用模型预测控制- mpc等基于非线性约束优化的控制系统时）在过程动力学和操作条件设定的限制内进行实时计算的能力。克服这个问题的一个可行的替代方案是利用协作的分布式控制体系结构，在这种体系结构中，通过在单独的处理器中以协调的方式解决多个控制（优化）问题来计算被操纵的输入。然而，严格设计非线性过程的协作、分布式控制体系结构是一项具有挑战性的任务，这是处理集中控制系统设计的传统过程控制方法无法解决的。为了设计协作的分布式控制系统，需要解决的关键基本问题包括单个控制系统的设计及其通信策略，以便它们有效地合作实现闭环工厂目标，以及开发有效的故障检测，隔离和管理策略。基于上述考虑，本研究计划的目标是发展设计和监测大规模非线性过程的协作分布式控制系统所需的理论和方法，并展示其在工业重要性过程系统中的应用和有效性。严格的方法和架构将被开发用于设计协作，分布式控制系统，明确地考虑异步和延迟测量的影响，并将开发新的监测和可重构容错控制策略来处理执行器/传感器/控制器故障。具体而言，研究项目包括：1)基于lyapunov模型预测控制技术的非线性过程协同分布式控制系统设计；控制系统架构、模型不确定性和状态估计问题将被明确地解决，2)为合作、分布式控制系统设计故障检测和隔离系统，3)开发可重构的容错控制策略，明确考虑稳定性、性能和鲁棒性，以及4)应用于对化学和水工业重要的模拟和实验室规模的过程系统。大规模非线性过程协同、分布式控制系统设计和监测方法的发展有望显著改善化工过程的运行和性能，提高过程的安全性和可靠性，并最大限度地减少过程故障的负面经济影响，从而直接影响美国经济。将研究成果整合到过程控制和操作的高级课程中，并撰写一本关于？容错过程控制？将使该领域的学生和研究人员受益。软件开发、短期课程和讲习班以及持续参与异常情况管理（ASM）联盟将是将这项研究成果转移到工业部门的手段。此外，将通过加州大学洛杉矶分校工程教育与多样性中心（CEED）的参与，以及通过向高素质学生提供暑期实习机会，与波莫纳加州州立理工大学（California State Polytechnic University in Pomona）的联系，让不同群体的本科生和研究生参与研究。最后，这项研究将受益于加州大学洛杉矶分校校园信息技术领域的教育倡议和创新，并为其做出贡献。

1027553ChristofidesOptimal operation and management of abnormal situations are major challenges in the process industries since, for example, abnormal situations account for at least $10 billion in annual lost revenue in the US alone. This realization has motivated significant research in the area of process control to ensure safe and efficient process operation. Traditionally, control systems rely on centralized control architectures utilizing dedicated, wired links to measurement sensors and control actuators to regulate appropriate process variables at desired values. While this paradigm to process control has been successful, when the number of the process state variables, manipulated inputs and measurements in a chemical plant becomes large - a common occurrence in modern plants -, the computational time needed for the solution of the centralized control problem may increase significantly and may impede the ability of centralized control systems (particularly when nonlinear constrained optimization-based control systems like model predictive control-MPC are used), to carry out real-time calculations within the limits set by process dynamics and operating conditions. One feasible alternative to overcome this problem is to utilize cooperative, distributed control architectures in which the manipulated inputs are computed by solving more than one control (optimization) problem in separate processors in a coordinated fashion. However, the rigorous design of cooperative, distributed control architectures for nonlinear processes is a challenging task that cannot be addressed with traditional process control methods dealing with the design of centralized control systems. To design cooperative, distributed control systems, key fundamental issues that need to be addressed include the design of the individual control systems and of their communication strategy so that they efficiently cooperate in achieving the closed-loop plant objectives, as well as the development of efficient strategies for fault detection, isolation and management.Intellectual Merit Motivated by the above considerations, the objective of this research program is to develop the theory and methods needed for the design and monitoring of cooperative, distributed control systems for large-scale nonlinear processes and demonstrate their application and effectiveness in the context of process systems of industrial importance. Rigorous methods and architectures will be developed for the design of cooperative, distributed control systems accounting explicitly for the effect of asynchronous and delayed measurements, and novel monitoring and reconfigurable fault-tolerant control strategies will be developed to deal with actuator/sensor/controller failures. Specifically, the research projects include: 1) Design of cooperative, distributed control systems for nonlinear processes using Lyapunov-based model predictive control techniques; control system architecture, model uncertainty and state estimation issues will be explicitly addressed, 2) Design of fault-detection and isolation systems for cooperative, distributed control systems, 3) Development of reconfigurable fault-tolerant control strategies accounting explicitly for stability, performance and robustness considerations, and 4) Applications to simulated and lab-scale process systems of importance to chemical and water industries.Broader Impact The development of cooperative, distributed control system design and monitoring methods for large-scale nonlinear processes is expected to significantly improve the operation and performance of chemical processes, increase process safety and reliability, and minimize the negative economic impact of process failures, thereby impacting directly the US economy. The integration of the research results into advanced-level classes in process control and operations and the writing of a new book on ?Fault-Tolerant Process Control? will benefit students and researchers in the field. The development of software, short courses and workshops and the on-going participation in the Abnormal Situation Management (ASM) Consortium will be the means for transferring the results of this research into the industrial sector. Furthermore, the involvement of a diverse group of undergraduate and graduate students in the research through participation in the Center for Engineering Education and Diversity (CEED) at UCLA, and outreach to the California State Polytechnic University in Pomona by offering summer internships to highly-qualified students, will be pursued. Finally, the research will benefit from and contribute to educational initiatives and innovations on the UCLA campus in the area of information technology directed by the co-PI.