Impact Factor (2025): 6.9
DOI Prefix: 10.47001/IRJIET
The rapid improvement of engineering
modeling is supported by the improvement of parallel GPU and CPU computational
capacities. However, due to space limitations, the improvement of the
computational capacities of GPU and CPU imposes challenges in the cooling
process. The liquid cooling method has attracted more interest as an effective
heat dissipation method. In this work, a new channel configuration is
introduced using the Archimedean spiral curve to generate the Archimedean
spiral channel configuration. The conjugate heat sink model was created to have
four different domains: liquid coolant (water), Cold plate (copper), glue layer
(ethoxy), and CPU (alumina). The effect of turbulence was incorporated by
varying the flow rate at a constant water inlet temperature of 25oCto
cover a range of Reynolds numbers (Re) from 3000 to 15000. The Shear Stress
Transport (k-w
SST) was the used turbulent model for a better capturing of the viscous,
high-frequency flow fluctuation near-wall region. Input power of 450 W was
subjected to the bottom surface of the CPU. The results showed that the
Reynolds number has a decisive impact on controlling the CPU temperature. As
higher Re decreased the average temperature developed within the CPU and
increased the pressure drop at an exponential rate. Darcy-Weisbach equation
confirmed these findings for internal flow when the pressure drop depends on
the squared average velocity. The hydrothermal performance of the Archimedean
spiral channel configuration rapidly decreased with the Re. Similar to the
velocity profile, the turbulent kinetic energy is generated at higher rate next
to the channels’ outer wall compared to the inner wall.
Country : Iraq
IRJIET, Volume 6, Issue 8, August 2022 pp. 61-71