The condensation of hydrocarbons in a vertical reflux condenser tube
A new test facility, with a vertical reflux condenser of 1500mm overall length and 45mm internal diameter, has been commissioned and tested and methods developed for measuring key process parameters. An experimental study of reflux condensation in a single tube using n-pentane and iso-octane and binary mixtures of these single component hydrocarbons has been undertaken. Using water as the cooling medium, a correlation was developed for determining the coolant-side heat transfer coefficient in the reflux condenser based on the Wilson plot method. The composition of binary liquid mixture samples from the test facility was determined using an empirical correlation developed using density measurements from a vibrating u-tube densitometer. The single components were condensed in the range 32.0-48.4°C and 0.106-1.515bara by adjusting the test condenser heat load for fixed conditions on the coolant side to investigate how the condensate-film heat transfer coefficient varied with the condensate film Reynolds number. The results show good agreement with the method recommended by HTFS for correcting the Nusselt theory for the effects of waves. A further small correction was made to improve the fit to the data. The binary hydrocarbon mixtures were condensed across the range 65.9-90.1°C and 0.729-1.531bara by conducting similar experiments where the feed vapour contained 50% and 70% n-pentane. Composition measurements of the condensate and vapour leaving the test condenser were made to examine the separation of components during partial reflux condensation. The results suggest that this separation is influenced by heat flux and that it would be improved if the test condenser were operated at a lower heat flux. Further experimental work is needed to verify this, and to investigate how this influences the number of thermodynamic stages, which was found to be less than one with the conditions reported here. Analysis of the heat transfer resistances on the vapour side showed that the standard procedure of using a dry-gas heat transfer coefficient, with or without a mass transfer correction term based on the film theory, poorly predicted the experimental values. These predictions were improved by the use of an enhancement factor, which may be more relevant in counter-current than co-current condensing situations. The results indicate that use of a dry-gas heat transfer coefficient with the film theory correction factor, over-predicts the mass transfer resistance. Comparison was made between the data and predictions based on the integral condensation curve, as might be used in Silver's method for condenser thermal design. It was shown that this method poorly predicted the surface area and the separation achieved in the test condenser. The results indicate that the heat and mass transfer coefficients obtained in a plain tube are significantly higher than those based on using a dry-gas heat transfer coefficient corrected by film theory. Implications for the design of reflux condensers have been presented.