Synthesis of Optimal Plantwide Control Systems for Integrated Chemical Processes

全厂控制是指对由许多相互关联的单元操作组成的整个（连续化工）工厂的控制。在20世纪50年代之前，人们普遍认为全厂控制的目标与伺服机构的目标相同（即良好的控制意味着快速的控制）。由于各种回路之间的相互作用，这种迫使变量尽可能快地达到稳态值的策略可能在全厂控制中失败。另一种策略是“设定值环境控制”，其基本理念是，如果工厂中的流量、温度、压力等保持不变，产品质量也将保持不变。这些要求可能是不一致的，例如，不可能保持回流流量恒定，同时操纵它来补偿干扰。“动态过程控制”的概念通过区分物料平衡控制器和产品质量控制器来解决这些不一致。物料平衡控制器充当低通滤波器（即调节），产品质量控制器充当高通滤波器（即伺服系统）。浪涌容量的选择可以使物料平衡控制系统的中断频率比质量控制系统的中断频率低一个数量级，以避免两个控制系统之间的显著相互作用。因此，可以通过在过程中安装足够大的浪涌容量来克服不良的质量控制。对于具有大浪涌容量和过度设计的单元操作的传统过程，这种方法足以设计有效的全厂控制系统。全球经济竞争迫使加工工业提高资本生产率和原材料利用率。本研究的总体目标是开发一种系统的方法来合成集成连续化学过程的最佳全厂控制系统。定义了可控性指标，量化了与动态可控性相关的成本。介绍了最优负荷分配原则（和最小动态成本原则），该原则允许在整个过程中对扰动传播和不确定性传播进行最佳处理。这两个概念允许系统地设计工厂范围内的控制系统，从而使总成本最小化。该理论还可用于定量比较基于稳态经济性和动态可控性的备选流程。案例研究将包括几个工业过程，例如，甲苯的加氢脱烷基反应以生产苯、苯乙烯和异丙烯，以及丁烷烷基化反应。结果将被纳入本科过程控制课程，以便学生可以为过程设计课程中设计的过程设计一个工厂范围的控制系统。

Abstract - Zheng - 9713599 Plantwide control refers to control of an entire (continuous chemical) plant consisting of many interconnected unit operations. Before the 1950's, it was commonly assumed that the objectives of plantwide control were the same as those of servomechanisms (i.e. good control means fast control). This strategy of forcing variables to their steady-state values as fast as possible may fail in plantwide control because of the interaction among various loops. An alternative strategy, "setpoint environment control" was developed, whose underlying philosophy was that if flows, temperatures, pressures, etc., in the plant were held constant, the product quality would also be constant. These requirements can be inconsistent, for example, it is impossible to hold reflux flow constant and at the same time manipulate it to compensate for a disturbance. The concept of "dynamic process control," resolves these inconsistencies by distinguishing between material balance controllers and product quality controllers. Material balance controllers act as low-pass filters (i.e. regulation) and product quality controllers act as high-pass filters (i.e. servo systems). Surge capacities can be chosen such that the break frequency of the material balance control system is an order of magnitude lower than that of the quality control system to avoid significant interaction between the two control systems. Thus, poor quality control can be overcome by installing sufficiently large surge capacities in the process. For traditional processes with large surge capacities and over-designed unit operations, this approach is adequate for designing an effective plantwide control system. Global economic competition has forced the process industries to improve their capital productivity and raw material utilization. The overall goal of this research is to develop a systematic method for synthesizing optimal plantwide control systems for integrated continuous chemical processes. A controllabil ity index, which quantifies the cost associated with dynamic controllability, is defined. The Optimal Load Distribution Principle (and the Minimal Dynamic Cost Principle), which allow optimal handling of disturbance propagation and uncertainty propagation throughout the entire process, are introduced. These two concepts permit a systematic design of a plantwide control system which minimizes the total cost. The theory can also be used to quantitatively compare alternate flowsheets based on both steady-state economics and dynamic controllability. Case studies will include several industrial processes, e.g. a hydrodealkylation of toluene to produce benzene, styrene, and cumene, and butane alkylation. The results will be incorporated into the undergraduate process control course so that students can design a plantwide control system for the process that is designed in the process design course.