Most of the current air-to-refrigerant heat exchanger models use the classic ε-NTU approach. These models do not account for 2D longitudinal heat conduction in the tube and the longitudinal heat conduction in the fin. These models, also do not account for the heat conduction between different tubes, which is a consequence of the widely employed adiabaticfin- tip assumption. This paper presents a more fundamental numerical approach to heat exchanger modelling which takes into account the 2D longitudinal heat conduction in any element, does not apply the fin theory with its adiabatic-fin-tip assumption, and captures a more detailed representation of air properties. The new model uses a segment-by-segment approach and applies a 2D discretization for each segment. The goal of the present work is to evaluate the impact of all the assumptions used widely in the models based in the ε-NTU methodology. The paper includes a presentation of the numerical scheme, model validation, and a parametric study which tests the impact of the traditional heat exchanger model assumptions applied for a microchannel gas cooler with CO2 as working fluid. The study revealed significant differences in capacity predictions depending on the ε-NTU relationship adopted. Longitudinal heat conduction turned out to be not negligible only when the air-side heat transfer is high, with errors about 2.5%. Large errors in capacity prediction of individual tubes occurred due to the adiabatic-fin-tip assumption when the neighbouring tubes were of different temperature.
Citation: International Journal of Refrigeration
Pub Type: Journals
heat exchanger, modelling, simulation, microchannel, gas cooler