A New Methodology for Cold Plate and Heat Sink Optimization

A new methodology of cold plate and heat sink optimization is presented which can reduce the computational time by more than two orders of magnitude when compared to CFD. This methodology is intended for high Prandtl number fluids or fully developed velocity fields. When compared to CFD, this model can predict the heat transfer to within 8% of the CFD solution for the intended operating ranges. The paper outlines the manipulation of established empirical and analytical correlations for developing thermal fields with constant wall temperatures using a scaling analysis. Since the real fin temperature is not constant, the proposed technique utilizes superposition to recreate the real fin temperature profile and enable the use of established analytical models. Finally, the fins are not individually modeled, which is computationally intensive, but captured using an assumption of anisotropic conduction, where the convective heat transfer is converted to a volumetric sink term.