Numerical and machine learning framework for strategic fin design to accelerate phase change material melting in advanced enclosures

The melting behavior of phase change materials (PCMs) is numerically investigated in rectangular, trapezoidal, and heptahedral enclosures equipped with identical V-shaped fins, under coupled parametric variations. The investigated parameters include PCM type (RT-25 to RT-38), fin material (aluminum, copper, steel), fin inclination angle (30°–90°), and heat flux (500–2000 W/m 2 ). The transient melting process is simulated using the enthalpy-based method, emphasizing the conduction-to-convection transition during early-stage melting dynamics. RT-25 exhibits the fastest melting (complete melting within 1500 s), while RT-27 shows the slowest behavior (liquid fraction of 0.172 after 1500 s). Aluminum provides a 27.6% higher average heat transfer coefficient compared to steel. The heptahedral enclosure achieves a 34.9% higher liquid fraction and 20.4% reduction in melting time compared to the rectangular configuration. Fin inclination angles between 70° and 90° yield the highest melting rates. Increasing heat flux from 500 to 2000 W/m 2 reduces the Fourier number for complete melting by 77.56%. A Random Forest regression model predicts the Nusselt number with R 2 values of 0.9896 (training) and 0.9729 (testing), with MAPE of 2.74% and 3.68%. SHAP analysis reveals that extreme heat flux conditions (500 and 2000 W/m 2 ) exert the greatest influence on predictions. The novelty lies in the systematic investigation of coupled geometric, material, and operational parameters, providing the first comparative assessment of these three enclosures with identical V-shaped fins. The findings establish practical design guidelines for selecting PCM-fin configurations in latent heat thermal energy storage systems, demonstrating that the combination of heptahedral enclosure, aluminum fins, and near-vertical inclination (70°–90°) provides the most effective thermal performance.