Multi-objective optimization of hydrothermal performance of a porous minichannel heat sink using RSM and NSGA-II algorithm

Abstract

The key considerations in developing heat sinks for the thermal management of electronic devices are enhancing heat transfer and minimizing entropy generation. This study aims to achieve significant cooling capability with reduced irreversibility by integrating nanofluids and porous media within a circular minichannel heat sink. This heat sink (40 × 40 × 10 mm) is composed of aluminum and is designed to operate within Reynolds numbers of 300–1300, featuring a bronze porous substrate and Fe₃O₄/water nanofluid coolant. ANSYS Fluent was utilized for computational fluid dynamics simulations, while the Non-dominated Sorting Genetic Algorithm-II was employed to optimize hydrothermal performance. The investigation examined key input parameters, including volume flow rate (Q), nanofluid concentration (φ), porosity (∊), and channel count (n) to optimize the responses: Nusselt number and entropy generation. Latin hypercube sampling is utilized to explore the design space, while Response Surface Methodology developed regression models linking objectives and design variables. The findings demonstrated that a porous channel increases the Nusselt number by approximately 515.12 % and reduces entropy generation by 83.65 % compared to a non-porous channel (∊=1). Moreover, the optimal trade-off between objective functions suggests that enhancing the volume flow rate while reducing the nanofluid concentration, porosity level, and channel count significantly improves the hydrothermal performance of heat sink. Additionally, the optimal Pareto front indicates superior cooling performance with design variables set to Q = 3.239 cm³/s, φ = 1 %, ∊=0.75, and n = 05. This configuration yields a 23.8 % increase in Nusselt number and a 25.2 % decrease in entropy generation compared to the reference study (Q = 0.841 cm³/s, φ = 1 %, ∊=0.75, and n = 05).