MODIFYING ``CONSTANT-RATE'' MASS TRANSFER MODEL FOR CFD ANALYSIS OF DIRECT CONTACT CONDENSATION OF STEAM JET IN WATER FLOW

Direct contact condensation is used in industrial applications, such as steam jet pumps, direct-contact heat exchangers and nuclear reactor cooling systems, due to its highly ecient heat and mass transfer. When steam is injected into subcooled water, a steam plume is generated at the exit of the steam nozzle, surrounded by an interface around the steam plume. The direct and quick transfer of heat, mass and momentum across steamwater interface makes the physics of direct contact condensation very complex. To eciently design the above mentioned equipment of direct contact condensation, a proper understanding of heat and mass transfer in this phenomenon is required. Several experimental and theoretical works have been performed on steam jet condensation in water. However, there have been a few numerical investigations of this phenomenon. Dierent heat and mass transfer models have been used by researchers to develop a suitable numerical tool to simulate steam condensation process. Constant-rate mass transfer model is one of the models used in numerical simulation of direct contact condensation in steam injection into water. This model needs a specic empirical coecientfor eachsimulation. In this study, considering available experimental results of previous investigations, a numerical simulation was performed in ANSYS Fluent software by using constant-rate mass transfer model to study the eects of dierent steam and water ow parameters, including water subcooling degree, Reynolds number of water ow, and steam mass ux, on dimensionless steam plume length. Moreover, constant-rate mass transfer model was revised by using the results obtained from the simulation and a new correlation for calculation of the constant coecient of the mass transfer model was proposed as a function of ow parameters that in uence the phenomenon to increase the accuracy of future simulations. The proposed correlation agrees well with numerical data and most of the data lie in the range of20% of the correlation.