A general algorithmic approach to the optimal synthesis of energy-efficient distillation train designs

Abstract Significant energy savings in conventional distillation trains can be achieved by (i) altering the processing sequence, and (ii) performing intercolumn energy matches through an appropriate choice of the column pressures. Prior works have shown that the best thermally-coupled distillation systems usually feature near-minimum utility costs. By choosing the least utility cost as the design target, the synthesis of a heat-integrated distillation train is formulated as a mixed-integer linear programming problem to find simultaneously all the best separation strategy, column pressures and HX-network. Every feasible high-order intercolumn heat match is considered in the model. The search for the global optimum is made by applying a standard branch-and-bound technique for MILP problems. The proposed formulation has successfully been applied to a five-component mixture separation problem.