CS-5 Shellside Condensation of Mixtures and Pure Components in the Presence of Noncondensable Gases on Horizontal Tube Bundles

K. Ishihara

New horizontal shellside condensation data have been obtained for both plain and finned tubes since the completion of the inert gas recirculation loop in August, 1983, using mixtures and pure components with and without noncondensable gases. This is the interim report of the shellside condensation data and research work up to the end of 1984 including the data of mixtures of n-pentane/p-xylene, N2/steam, N2/p-xylene, N2/n-pentane/p-xylene and CO2/n-pentane/p-xylene. The test bundles used for this series of runs are baffled bundles of plain tubes, 19 fins/in. tubes and 30 fins/in. tubes. The data obtained during 1985 and 1986 will be included in the final report of the same subject at the end of the current five-year period.

Deterioration of condenser performance due to the presence of inert gases was found to be much more severe than expected for finned tubes bundles, especially at high inert gas concentration in the vapor. Effectiveness of external low0integral fins was severely compromised due to the presence of inert gases that blanket the fin surface and due to the condensate retention around the fins. As a result, the finned tube bundles are found to show almost no enhancement effect on heat transfer coefficients compared to the plain tube bundle and may give a penalty for heat transfer under some conditions. The only advantage of the finned tube bundle over the plain tube bundle is due to the increased surface area which compensates for the decrease of the heat transfer coefficient to give better overall performance on duty.

New correlations have been proposed for calculating the vapor-phase heat transfer coefficients for both plain and finned tube bundles based on the rigorous formulation of the Resistance Proration Method. The new correlation incorporates the correction factor for temperature profile distortion due to interface mass transfer, the correction factor for flooding which is a function of fin geometry and fluid property, and the correction factor for the effects of bundle geometry, vapor flow, and mass diffusion at the interface. The new correlation does not require phase-equilibrium properties at the interface. The prediction by the new correlation is consistently good from low to high concentration of noncondensable gas or a light component with a 30 percent error range for about 90 percent of the data points.