Optimization of manufacturing process automation in the presence of time-varying operating conditions

自动控制系统对今天的制造业至关重要，因为它们允许稳定的过程操作，一致的产品质量，减少浪费和控制排放。然而，目前大多数系统的设计都不能最优地处理由时变因素引入的时变操作条件，如原料变化、产品等级变化、需求变化、过程非线性引起的多模式甚至操作员位移变化。无法精细控制过程可能导致产品质量降低，额外的浪费，并增加对环境有害物质的排放。因此，传统的自动化控制很难同时满足日益提高的产品质量规格和更严格的环境法规——这是当今工业的现实。我们建议开发一套工具来改进具有时变特性的系统的过程控制。我们将采用的方法是基于系统识别和优化一组控制律，包括一类时变过程的时不变和时变控制律。我们的目标是制定策略，以确保产品质量规范，同时最大限度地减少控制输入能量。时变系统辨识和过程控制，特别是混合或切换系统的时变系统辨识和过程控制，是目前研究的热点。我们的解决方案旨在利用自动化系统基础设施，从现有到可预见的10年，以便它可以在工业分布式控制系统（DCS）中实施，而无需额外的资本成本。这将使该技术迅速应用于工业成为可能。改进的控制减少了可变性，允许过程在不违反约束的情况下更接近设定值。这将导致更好的产品质量和减少环境足迹。优化时变操作的控制将使行业在全球市场上更有效率和更具竞争力。

Automatic control systems are critical to manufacturing today because they allow for steady process operations, consistent product quality, less waste, and control of emissions. However, most current systems are not designed to optimally handle time-varying operating conditions introduced by time-varying factors such as feedstock changes, product grade changes, demand changes, multi-modes due to process nonlinearities and even operator shift changes. The inability to finely control processes can lead to lower product quality, additional waste, and increased emissions of environmentally undesirable substances. As a result, it is difficult to meet both increased product quality specifications and more stringent environmental regulations - facts of life for industry today - with conventional automation control. We propose to develop a set of tools for improved process control of systems with time-varying characteristics. The approach we will take is based on system identification and optimization of a set of control laws including time invariant as well as time variant control laws for a class of time-varying processes. Our objective is to develop strategies to ensure product quality specifications while minimizing control input energy. Time-varying system identification and process control, particularly for the hybrid or switching systems, are areas of active research. Our solution will be designed to leverage automation system infrastructure from the existing to a foreseeable 10-year horizon so that it can be implemented in the industrial distributed control systems (DCS) without additional capital cost. This will make it possible to quickly apply the technology to industry. Improved control reduces variability, allowing processes to be operated closer to their setpoints without violating constraints. This will lead to better product quality and a reduced environmental footprint. Optimizing the control of time-varying operations will make industry more efficient and more competitive in the global marketplace.