Hybrid CFD and Monte Carlo-Driven Optimization Approach for Heat Sink Design

This study introduces a hybrid topology optimization methodology aimed at improving heat sink efficiency through a data-driven approach. The method integrates CFD simulations in Ansys Fluent with a Monte Carlo-driven optimization algorithm, modeling the design of a heat sink domain as a porous mediu...

Descripción completa

Detalles Bibliográficos
Autores: Busqué, Raquel, Bossio, Matias, Fabregat, Raimon, Bonada, Francesc, Maicas, Héctor, Pijuan, Jordi, Brigido, Albert
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2025
País:España
Institución:Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya)
Repositorio:Recercat. Dipósit de la Recerca de Catalunya
OAI Identifier:oai:recercat.cat:10459.1/468309
Acceso en línea:https://doi.org/10.3390/en18112801
https://hdl.handle.net/10459.1/468309
Access Level:acceso abierto
Palabra clave:Heat sink
Topology optimization
Data-driven optimization
Descripción
Sumario:This study introduces a hybrid topology optimization methodology aimed at improving heat sink efficiency through a data-driven approach. The method integrates CFD simulations in Ansys Fluent with a Monte Carlo-driven optimization algorithm, modeling the design of a heat sink domain as a porous medium. Porosity is used as a design variable, iteratively adjusted in a binary manner to optimize fluid-solid distribution. Three design variants were evaluated, with the selected optimized configuration reaching a maximum temperature of 57.11 °C, compared to 46.15 °C for a baseline serpentine channel. Despite slightly higher peak temperature, the optimized design achieved a substantial reduction in pressure drop, up to 91.57%, translating into significantly lower pumping power requirements and thus lower energy consumption. Experimental validation, using physical prototypes of both the reference and optimized channels, confirmed strong agreement with simulation results, with average surface temperatures of 29.27 °C and 30.03 °C, respectively. These findings validate the accuracy of the simulation-based approach and highlight the potential of data-driven optimization in thermal management system designs.