09. Heat and Mass Transfer

Recently, the rapid development of Artificial Intelligence (AI) has led to the need for AI servers to compute vast amounts of data. During these computational processes, chips generate a significant amount of heat. One of the most common methods for dissipating CPU heat is tower coolers. Tower coolers consist of fins and fans, which must work collaboratively to dissipate the heat efficiently. This study adopts CFD simulations and the Taguchi method to identify the optimal fin design. Additionally, this study modifies the optimized fin shapes and analyzes their impact on-chip temperature. The study introduces the Figure of Merit (FOM), a metric that combines the cooler's temperature and fan energy consumption performance, to evaluate the relative performance of different fin shapes. The results indicate that modifying the Taguchi-optimized fins into a staggered configuration yields the highest FOM value, making it the optimal fin design in this study.