Dry gas injected into wells will vaporize water from near the wellbore. The vaporization starts from the well and proceeds outward. Gas flowing to producers is in equilibrium with the reservoir brine, but water will be vaporized because the pressure drop that occurs toward the wellbore increases the ability of the gas to contain water. Thus, there are different mechanisms for injection and production. For both gas injection and gas production, vaporization concentrates solids in the brine that will precipitate into the formation when sufficiently concentrated. This paper reports on a combined experimental and theoretical analysis on the vaporization portion of this problem for dry gas injection. Experiments have been performed previously to determine the rate of water vaporization from Berea core samples at uniform initial water saturation (Zuluaga and Monsalve 2003). These experiments were performed by injecting dry methane into core samples that contained immobile water to represent water vaporization in a gas injector. Effluent water concentration curves showed two vaporization periods: a constant rate period and a falling rate period. The existence of a constant rate period means that the mass transfer within the core is occurring at conditions of local equilibrium. We interpret the failing rate period as the result of a moving capillary transition zone in which the amount of water vaporized decreases slowly because of capillary pressure effects. The falling rate period is the consequence of capillary imbibition of a wetting phase at very small saturation. We interpret the vaporization results with two traveling wave solutions. The first, which can be solved analytically, assumes that the capillary diffusion coefficient, D, and the volume fraction of water in the gaseous phase, C-wg, are constant. For this case, the results of the traveling wave solution are matched to the results of laboratory experiments by adjusting D. The second traveling-wave solution must be solved through numerical integration. In this case, the relative permeability scaling exponent is adjusted to match the laboratory experiments. The fitting provides insights into the nature of wetting phase flow at small saturation. Lastly, the experimental and mathematical procedure discussed in this paper is certainly a new method to obtain relative permeability exponents for the wetting phase at very low values of wetting-phase saturation (down to theoretically zero values).
