MS Thesis Presentation by Scott Wilson
Friday, April 8, 2005
(Dr. Yogendra Joshi, Chair)
"Investigation of Porous Foam Coldplates as a High Heat Flux Electronics Cooling Solution"
Compact heat exchangers such as porous foam coldplates have great potential as a high heat flux cooling solution for electronics due to their large surface area to volume ratio and tortuous coolant path. The focus of this work was the development of unit cell modeling techniques for predicting the performance of coldplates with porous foam in the coolant path.
Multiple computational fluid dynamics (CFD) models which predict porous foam coldplate pressure drop and heat transfer performance were constructed and compared to gain insight into how to best translate the foam microstructure into unit cell model geometry. Unit cell modeling in this study was realized by applying periodic boundary conditions to the coolant entrance and exit faces of a representative unit cell. A parametric study was also undertaken which evaluated dissimilar geometry translation recommendations from the literature. The use of an effective thermal conductivity for a representative orthogonal lattice of rectangular ligaments was compared to a porosity-matching technique of a similar lattice. Model accuracy was evaluated using experimental test data collected from a porous copper foam coldplate using deionized water as coolant. The compact heat exchanger testing facility which was designed and constructed for this investigation was shown to be capable of performing tests with coolant flow rates in excess of 300 mL/min and heat fluxes greater than 275 W/cm 2 . The greatest technical challenge of the testing facility design proved to be the method of applying the heat flux across a 1 cm 2 contact area. Based on the computational modeling results and experimental test data, porous foam modeling recommendations and porous foam coldplate design suggestions were generated.