Phase Change Materials (PCMs) are widely used in latent thermal energy storage systems owing to their high energy storage density. However, their low thermal conductivity limits the heat transfer efficiency. To address this challenge, this study proposes a surrogate-based multi-objective optimization framework to improve the thermal performance of a PCM-based finned heat sink while reducing its geometric footprint. Transient Computational Fluid Dynamics (CFD) simulations were performed using the enthalpy–porosity method to model the melting behavior of a high-Prandtl-number PCM (lauric acid). The flow was assumed to be incompressible, Newtonian, laminar, and transient, with buoyancy effects and temperature-dependent thermophysical properties included to ensure physical fidelity. Design of Experiments (DoE) was constructed by systematically varying the fin lengths of an aluminum heat sink, generating a high-fidelity CFD database. This dataset was used to train a Response Surface Model (RSM) capable of accurately representing the nonlinear relationships between the fin geometry and key performance indicators, including the Nusselt number, liquid fraction, and volume fraction. The trained surrogate model was subsequently coupled with a Multi-Objective Genetic Algorithm (MOGA) to simultaneously maximize the thermal performance and minimize the geometric constraints. Pareto-front analysis revealed trade-offs between heat transfer enhancement and compactness, confirming the absence of a single global optimum point. The optimization converged toward a nonuniform fin configuration that improved the overall performance evaluation criterion by approximately 0.15% and increased the Nusselt number by approximately 0.23%, while reducing the heat sink volume by approximately 3.2%. Although the optimized design exhibited a reduction of approximately 4.6% in the liquid fraction at the evaluation time, the corresponding decrease in total thermal storage was limited to approximately 2% because of enhanced convective heat redistribution and higher sensible heat contributions. The results demonstrate that surrogate-based multi-objective optimization is an effective strategy for designing compact and thermally efficient PCM heat sinks, providing valuable guidance for latent thermal energy storage applications in which both performance and size are critical constraints. • Surrogate-based optimization applied to PCM finned heat sinks. • CFD–DoE dataset used to train an ANN response surface model. • MOGA revealed trade-offs between heat transfer and compactness. • Asymmetric fins increased Nu while reducing heat-sink volume by 3.25%. • Thermal storage decreased only 1.7% despite reduced melting.
L.G. et al. (Wed,) studied this question.