Eventos Anais de eventos
ENCIT 2020
18th Brazilian Congress of Thermal Sciences and Engineering
Heat exchanger multi-objective optimization: A study of uniform division methods for the spherical pruning diversification mechanism
Submission Author:
Lucas Camilotti , PR
Co-Authors:
Lucas Camilotti, Cassiana Fagundes da Silva, Roberto Zanetti Freire
Presenter: Lucas Camilotti
doi://10.26678/ABCM.ENCIT2020.CIT20-0398
Abstract
Heat exchangers play an important role in several fields of engineering and science, and a considerable amount of research is being conducted since the last decades on how to further increase its performance. Through analytical modeling, it is possible to replicate the behavior of the system, and coupled with optimization, it can perform a search for the most adequate design parameters. Metaheuristics algorithms are well accepted for this task, since the design of a heat exchanger is complex and needs to take multiple criteria into consideration, giving the user a set of solutions with different compromises. The Multi-objective Differential Evolution with Spherical Pruning (sp-MODE) is a multi-objective variant of the classical Differential Evolution (DE), that uses a hyper-spherical grid-based diversity-ensuring mechanism for Pareto solutions. However, this mechanism shows a bias towards certain regions of the grid, because of non-uniform cell divisions. This work addresses this issue by adapting two uniform grid division methods and testing them over five benchmark functions and a Shell and Tube Heat Exchanger (STHE) design optimization problem. Three Pareto performance metrics are used in the comparison, and results show an increase in solution diversity of up to 43% for the functions and 17% for the heat exchanger case.
Keywords
Differential Evolution, Heat Exchanger Design, metaheuristics, Multi-Objective Optimization, Pareto diversity
DOWNLOAD PDF VIEW PRESENTATION
