Artificial intelligence – numerical study of a 3D model of latent heat thermal energy storage with sine-shaped fins

As the demand for cleaner and more efficient energy solutions grows, latent heat thermal energy storage systems (LHTESSs) utilizing phase change materials (PCMs) have gained increasing attention. This study presents a numerical investigation into how different fin configurations affect the thermal performance of PCMs during melting and solidification processes in a vertical shell-and-tube LHTESS. Using a pressure-based finite-volume method and the Boussinesq approximation, three fin designs—ring, axial, and spiral—were analyzed across multiple cases. In each case, the number of fins was varied while keeping the total fin volume constant. This approach ensured a fair comparison focused purely on geometry. The artificial intelligence in the form of a deep neural network was used to provide a generalized map of the system's behavior. Among all tested configurations, the ring fin design with six fins showed the best performance, reaching a melting volume fraction (MVF) of 95.5 % at 60 min, more than double that of the no-fin baseline (45 %) and achieving 99.9 % MVF and a 28.4 % increase in stored energy rate after 100 min. The spiral fins also performed well, with a 27.8 % energy gain, while the axial fins reached a comparable 95 % MVF but delivered a slightly lower energy improvement.