microfin; refrigerant; mixture; condensation; correlation
A new correlation is presented to predict the heat transfer coefficients (HTCs) of pure refrigerants and refrigerant mixtures undergoing convective condensation within horizontal microfin tubes. Preliminary results are reported here for a 318–point experimental database put together from 8 sources. The data includes CO2, R134a, R22, R407C, and R410A, 3.6–9.0 mm fin root diameter tubes, 14.3°C to 47.0°C saturation temperatures, vapor qualities from 4% to 99%, reduced pressures from 0.25 to 0.98, and heat and mass fluxes ranging from 2.9 to 39.9 kW/m2 and 98.5 to 800 kg/s.m2 respectively. Sixty unique dimensionless parameters pertinent to convective condensation in microfin tubes were first selected. Multi-variable regression analysis was then applied to identify the most significant variables influencing the condensing Nusselt number. The new correlation was evaluated and compared with two extant correlations (Yu and Koyama (1998 and 2008). Overall evaluation for the entire database shows that the new correlation is significantly better than these extant correlations. For this overall assessment, the new correlation predicts 82.6% of the data within ±30% error bands, with a mean absolute deviation (MAD) of 18.8%. For the 127 CO2 points and the remaining 191 points for halogenated refrigerants, the new correlation predicts 85.4% and 80.6% of the data respectively within ±30% error bands, with a MAD of 17.8% and 19.4% respectively. In comparison, the Yu and Koyama (1998) and (2008) correlations give the following performance respectively for the same database: (a) entire database: 45.4% and 25.7% of the data within ±30% error bands, and a MAD of 66.9% and 56.6%, (b) CO2: 18.1% and 15.0% of the data within ±30% error bands, and a MAD of 79.1% and 59.5%, (c) halogenated refrigerants: 66.0% and 33.0% of the data within ±30% error bands, and a MAD of 24.1% and 54.7%. The database will be expanded to include more refrigerants and operating conditions/tube geometries of practical interest. The performance of additional extant correlations such as those of Kedzierski and Goncalves (1999), Cavallini et al. (1995), and Cavallini et al. (2009) for the extended database will also be explored. The distribution of the entire database and the performance of the new correlation will be examined relative to refrigerant, and for various bins of fin root diameter, Dr, mass flux, G, all-liquid Reynolds number, Real, reduced pressure, PR, saturation temperature, Tsat, heat flux q, and vapor quality, x. Based on the bin analysis, parameter ranges in which one of the extant correlations gives better predictions than the new correlation will also be identified. Based on this analysis, overall recommendations for using the correlation most effectively will be provided.