Control strategies applied to a chemical plant with interacting control loops

Interaction among control loops in multivariable systems has been the subje ct of much research in the last decades. Distillation columns that are wide ly used in the process industries are the typical example of these systems. In this work, Wood and Berry test-bed problem has been used due to the str ong interaction among the control variables. Therefore, different control s trategies have been applied to this binary distillation column with two int eracting control loops. A diagonal controller structure with proportional-i ntegral controllers using several tuning methods has been evaluated. Two In ternal Model Control structures have been implemented utilizing in one of t hem a neural network model of the plant. Model identification by neural net works has also been presented.