MODEL DEVELOPMENT FOR MICRO–CHANNEL COOLING TECHNOLOGY

Abstract

The micro–channel research literature presents a clear need to provide accurate models that can predict pressure drop and heat transfer coefficients for a greater number of experimental data sets while capturing the physical phenomena of various flow patterns associated with the onset of nucleate boiling in the two–phase region. The model approach developed have been evaluated for the purpose of facilitating an efficient design instrument for micro–channels to predict the pressure drop related to heat input for the single phase through to the boiling [two–phase] region as well as heat transfer coefficient calculations for a single micro–channel. The simplified homogeneous model provides a lower bound for pressure drop estimates and the weighted annular–homogeneous model produces an upper bound value. Input parameters include the micro–channel dimensions, fluid flow rate, inlet temperature, thermo–physical properties of the respected fluid, and outlet pressure. Polynomial correlations for water are obtained from curve–fitting data available from the 1997 Ashrae Handbook over the temperature range of 0.01 to 200 ºC for the thermodynamic properties that include liquid density, viscosity, thermal conductivity, specific heat capacity and change of enthalpy [latent heat of vapourization].

The model results are evaluated over four independent experimental data sets that are available in the literature to demonstrate the sufficient accuracy for channel dimensions ranging from 50 µm to 713 µm. The independent data sets were numerically reproduced from figures presented in the research literature. The boiling front, pressure drop, heat transfer coefficient, wall temperature profile and vapour quality characteristics are evaluated. Heat transfer coefficient calculations were made via the Kandlikar correlation and the Kandlikar enhancement factor method that is corrected for vapour quality. The model produced pressure drop results that were within about a 30% error to the experimental data sets evaluated for heat fluxes in the range of 50 W/cm² to values exceeding 150 W/cm². Heat transfer coefficient values calculated between the two correlations of Kandlikar were within an estimated error of 30% to experimental measurements and demonstrated results in the two–phase region that exceeded 110 000 W/cm² K.