Multi-objective optimisation of micromixer design using genetic algorithms and multi-criteria decision-making algorithms
Eduardo Henrique Taube Cunegatto, Flávia Schwarz Franceschini Zinani, Sandro José Rigo
Abstract
This work employed the constructal design method (CDM) to optimise a micromixer's shape. The micromixer had five degrees of freedom, optimised to maximise the mixing ratio and minimise the pressure drop across it, for Peclet numbers equal to 250, 500, and 1,000. Computational fluid dynamics (CFD) was used for simulations that generated second-order metamodels, employed within the NSGA-II algorithm for multi-objective optimisation. Upon defining the best set of solutions, multi-criteria decision-making algorithms aided in choosing solutions that would meet the objectives, namely LINMAP, TOPSIS, and VIKOR. Our analysis revealed that shapes with the highest mixing ratios also exhibited the highest pressure drops, with the VIKOR algorithm favouring this trade-off. Conversely, TOPSIS solutions tended to minimise pressure drop and mixing ratios, while LINMAP solutions fell between these extremes. This integrated approach provided a curated selection of optimal choices, a crucial advantage given the many potential solutions inherent in passive micromixer design.