Flow boiling heat transfer of HFO1234yf and HFC32 refrigerant mixtures in a smooth horizontal tube: Part II. Prediction method

A prediction model has been developed using the results of prior experimental investigations into the flow boiling heat transfer of pure HFO1234yf, HFC32, and their refrigerant mixtures at two mass fractions (80/20 and 50/50 by mass%) in a smooth horizontal tube having an inner diameter of 2 mm. In the prior experiment, the saturation temperature was 15 degrees C, the mass fluxes ranged from 100 to 400 kg/m(2) s, and the heat fluxes ranged from 6 to 24 kW/m(2). In this paper, we propose a new semi-empirical correlation for pure refrigerants based on the superposition of the contributions from nucleate boiling and convection. We introduced two new factors in the correlation to account for the effects of the convection on two-phase flow and nucleate boiling. Comparisons between the experimental results and the predictions of seven correlations showed that the proposed correlation had reasonable agreement with the magnitudes and trends of the variations in the measured heat transfer coefficient of pure HFO1234yf, HFC32, and the 50/50 mass% mixture (with a mean absolute error of approximately 20%) when the average properties of the mixtures were used. However, this simple approach was found to overestimate the experimental results for the 80/20 mass% mixture. The inhomogeneity of the concentration in the mixture significantly affects its heat transfer with a high temperature glide; the predicted deviation was large when only the average properties of the mixture are considered. After taking into consideration the effects of mass diffusion on the flow boiling heat transfer of refrigerant mixtures, we introduced suppression factors for the mixture into both the existing and proposed correlations. Superior prediction accuracies were obtained using these modified correlations. When predicting the heat transfer coefficients of the refrigerant mixtures of HFO1234yf and HFC32 (80/20 and 50/50 by mass%), the average deviation of the proposed correlation with suppression factors was less than 20%. (C) 2013 Elsevier Ltd. All rights reserved.