Performance analysis of virtual platforms focused on multi-objective process optimization

Several systematic methodologies for designing and optimizing conceptual processes have been reported by combining different computer-aided tools, mainly based on process simulators and programming environments that can help to improve any processes significantly. Thus, the objective of this work involves the performance analysis for virtual platforms focused on multi-objective process optimization, using a framework based on identical objective functions (economic, exergetic, and environmental targets), similar algorithms and hyperparameters (NSGA-II), for the separation and purification of the benzene, toluene, m-xylene mixture (BTX). Two computational interfaces were developed: Aspen Plus-Matlab (A) and Aspen Plus-Python (B), both using direct communication based on COM. Results were compared in terms of computational time, objective functions, and state variables values. Finally, both computational interfaces have a similar performance in all terms: interface B showed better computational time, but the use of the utopia-tracking approach let to conclude that interface A has better objective function values.