GDP-Distil is an interface for synthesizing a single distillation column.

Given a multicomponent feed with known flow and composition and the desired prodcuts specifications; the problem then consists in selecting the number of feedtrays, feed location, and condenser and reboiler duties and areas of a destilation column so as to minimize the total anualized investment and and operating cost.

The column represenatations modeled involve a rigorous GDP (General Disjunctive Programming) folmulation where the logic is represented through disjunctions and propositions. The formulation of this optimization problem includes a preprocessing phase where a good initial solution is generated for the economic problem.

The preprocessing phase includes two initialization Non Linear Programming (NLP) problems and two rigorous tray-by-tray NLP formulations.

The economic problem is modeled as a Generalized Disjunctive Programming (GDP) problem. The solution algorithm iterates between two subproblems: "master" which is a Mixed-Integer Linear Programming (MILP) problem that optimizes the number of trays of the column and "econ" an NLP subproblem that optimizes the operating conditions for the column with fixed number of trays.

GDP-DISTILL is based on the paper by Mariana Barttfeld, Pío Aguirre and Ignacio E. Grossmann, "Alternative Representations and Formulations for the Economic Optimization of Multicomponent Destilation Columns", Computers and Chemical Engineering" , 27, 363-383 (2003).

GDP-DISTILL has been developed by Gabriela Garcia-Ayala under the supervision of Ignacio E.Grossmann.