Optimal Design of Distillation Columns Using a Combined Optimisation Approach

Distillation is the most important separation process in the chemical industry despite being very energy-intensive, hence any attempt to optimise its design may potentially have a significant impact. In this work, a combined optimisation strategy based on Genetic Algorithm (GA, stochastic method) and an outer approximation method (OAERAP, deterministic method) is proposed and demonstrated to reliably and efficiently optimise different distillation processes by achieving significant energy and/or capital savings. Three different case studies are presented to compare the combined strategy to the stand-alone GA and OAERAP methods. The combined optimisation strategy shows excellent results, with the total annualised costs obtained from the combined strategy being similar or better than for the stand-alone methods. More importantly, the combined strategy requires much shorter CPU time, as well as significantly less manual effort, thus greatly increasing the time efficiency when compared to the OAERAP method.