Self-Optimizing Control Structures for Active Constraint Regions of a Sequence of Distillation Columns
MetadataVis full innførsel
Investigating and mapping active constraint regions for processes, and subsequently finding control structures for each region, is vital for their optimal operation. In this work, active constraint regions of three different case studies for the distillation process have been investigated: • A single distillation column with constant product prices. • A single distillation column with purity dependent prices. • Two distillation columns in sequence with constant prices. The active constraint regions for each case study have been identified and mapped with respect to the disturbances; energy price and feed flow rate. Selected stage temperatures and combinations of stage temperatures have been proposed as self – optimizing variables for the unconstrained degrees of freedom of each region. These were found by the use of the Minimum Singular Value Rule and the Exact Local Method. The methods requires the optimal sensitivities of measurements with respect to disturbances, which was obtained by using the software package Ipopt/sIpopt. To demonstrate applicability, a selection of the control structures for the different regions of each case study have been implemented and compared on the dynamic nonlinear models using Simulink. It has been shown that the first case study, a single distillation column with constant product prices, has 3 active constraint regions while the next, a single distillation column with purity dependent prices, has 5 active constraint regions. The last case study, two distillation columns in sequence with constant prices, has 8 active constraint regions.