Real-Time Optimization for Large Scale Processes: Nonlinear Model Predictive Control of a High Purity Distillation Column
Real-Time Optimization for Large Scale Processes: Nonlinear Model Predictive Control of a High Purity Distillation Column
复制标题
大规模过程的实时优化：高纯度蒸馏塔的非线性模型预测控制
DOI：
10.1007/978-3-662-04331-8_20
复制
发表时间：
2001
期刊：
影响因子：
--
通讯作者：
E. Stein
中科院分区：
文献类型：
--
作者：
M. Diehl;I. Uslu;R. Findeisen;S. Schwarzkopf;F. Allgöwer;H. Bock;T. Bürner;E. Gilles;A. Kienle;J. Schlöder;E. Stein
本文的目的是一个实验的概念验证的应用NMPC的大规模系统使用专门的动态优化策略。为了这个目的，我们研究现代的，计算效率高的NMPC计划和实时优化技术的应用程序的一个非平凡的过程控制的例子，即控制的高纯度二元蒸馏塔。所有必要的步骤进行了讨论，从制定的DAE模型与164个国家的最后应用程序的实验装置。特别介绍了一种基于直接多次打靶法的高效实时优化方案。它的特点是一个初始值嵌入策略，允许立即响应干扰，实时迭代，吻合优化迭代与真实的过程开发。使用该方案，10秒的采样时间在标准PC上是可行的。这表明，一个有效的NMPC计划的基础上大规模DAE模型的实时控制的中试规模的精馏塔是可行的。
The purpose of this paper is an experimental proof-of-concept of the application of NMPC for large scale systems using specialized dynamic optimization strategies. For this aim we investigate the application of modern, computationally efficient NMPC schemes and realtime optimization techniques to a nontrivial process control example, namely the control of a high purity binary distillation column. All necessary steps are discussed, from formulation of a DAE model with 164 states up to the final application to the experimental apparatus. Especially an efficient real-time optimization scheme based on the direct multiple shooting method is introduced. It is characterized by an initial value embedding strategy, that allows to immediately respond to disturbances, and real-time iterations, that dovetail the optimization iterations with the real process development. Using this scheme, sampling times of 10 seconds are feasible on a standard PC. This shows that an efficient NMPC scheme based on large scale DAE models is feasible for the real-time control of a pilot scale distillation column.