Decentralized Adaptive Control of Distillation Columns. A Case Study

摘要

The contribution presents a decentralized adaptive approach to control of multi input - multi output (MIMO) continuous-time dynamic systems. The controlled plant is represented by a binary distillation column with two inputs and two outputs. The external model of the controlled plant and the control design are formulated using continuous-time transfer functions in the matrix fraction description. The decentralization of a MIMO plant is performed using the mathematical decomposition and the controlled system is decomposed into subsystems. The designed control loop consists of local and global control loops. The local controllers arc adaptive ones obtained via the pole-placement method and they control isolated subsystems. The global control compensates influence of interactions between subsystems. The decentralized control algorithm is based on recursive identification of an explicit plant model and it is adaptive one.