Experimental study of design parameter influence on thermal and hydraulic performance of cold plates

An experimental setup was designed to test the influence of several design parameters on the thermal and hydraulic performance of a liquid cooled cold plate with serpentine channels, to improve the design and validate the modelling of liquid cold plates. Heaters are installed on the cold plate which are controlled to deliver heat dissipation rates of 200 W to 1200 W and corresponding heat flux densities of 4 kW/m² to 24 kW/m². A 50%–50% mixture of water and ethylene glycol is used as coolant. Several cold plates were tested, with variations in channel hydraulic diameter, cold plate thickness, number of bends, thermal contact interface of the heaters and the use of turbulators. The thermal resistance and pumping power are determined from the measurements and the different cases are compared to each other. Increasing the channel diameter with fixed channel length results in a design with decreasing overall performance. Increasing the cold plate thickness only has a very small influence on the thermal resistance, because the conductive thermal resistance of the aluminum cold plate is small compared to the other thermal resistances. Increasing the total length of the coolant channel for a fixed diameter increases the total heat transfer area for convection, thereby increasing the overall performance. The thermal contact resistance of the heaters to the cold plate is a critical parameter. When using no thermal interface material, the total thermal resistance increases up to 36% compared to using thermal paste. Finally turbulators can be used to increase the heat transfer rate at the cost of a higher pressure drop. The difference in performance when using turbulators is within the measurement error, so no significant increase nor decrease was perceived. The liquid heat transfer coefficient is also evaluated in function of the Reynolds number. This shows no abrupt transition from laminar to turbulent flow, which is consistent with other research on convection in serpentine channels, where the smooth transition is explained by the appearance of Dean vortices due to the bends.