An experimental and analytical investigation of forced convective condensation in smooth circular tubes is performed for R-22 and R32+R125 mixture (50% by mass). The refrigerant is condensed by water inside a 0.315 inch ID copper tube along a 20 ft. long tube-in-tube counterflow type test condenser. Instrumentation is provided at five locations along the condenser length for recording data under steady state conditions. Mass flux rates of 110,000 to 350,000 lbm/hr-ft2 are tested for a condensing temperature range of 75 to 115 ?F and oil concentrations of 2.6 and 5.37 percent for oil contamination tests with a maximum of 10 percent heat balance error. A FORTRAN program calculated the local and average heat transfer coefficients and total pressure drop. The local heat transfer coefficient is calculated by applying the one-dimensional heat diffusion equation in the radial direction. The data compared with Shah (1981) and Traviss et al. (1973) correlations to within 20 and 15 percent respectively. For the new refrigerant accurate pressure drops could not be obtained due to high condensing pressures. Oil reduced the heat transfer coefficient by 10 and 15 percent for 2.6 and 5.37 percent oil concentrations respectively. An annular two-phase flow model is developed to predict the local heat transfer coefficient using the Prandtl mixing length theory, Van- Driest's approach and Reynolds analogy. This model requires only the global parameters and no local experimental data are needed, is valid for oil-refrigerant mixtures, and an equation to calculate the viscosity of the oil-refrigerant mixture is recommended. Population balance model (surface renewal theory) based on homogeneous flow assumption is applied to the current experimental conditions. Two correlations found, one (empirical) from the experimental data and the other (semi-empirical) from the population balance model correlated the data to within 15 percent. Flow pattern studies using Taitel and Dukler (1976) map indicated predominantly annular flow and possible slug/plug flow for low qualities and/or mass flux rates. Film thickness is found to affect the heat transfer coefficient most.
