Prediction of Molecular Diffusion At Reservoir Conditions. Part 1- Measurement And Prediction of Binary Dense Gas Diffusion Coefficients

Abstract

Abstract Diffusion coefficients in binary dense, fluid systems are measured and used along with data available in the literature to obtain a generalised correlation for predicting binary molecular diffusion coefficients in dense gases over a broad range of conditions. Introduction Attempts to quantify mixing phenomena in porous media are often encountered in design studies of processes related to petroleum recovery. Dispersion, an important part of mixing phenomena, can be related to the distribution in travel times that results when a fluid passes through a porous media(1). On a microscopic scale, the two most important mechanisms contributing to dispersion are variations in velocity between fluid elements flowing in neighbouring pores or groups of pores (often referred to as convective mixing) and molecular diffusion(1, 2, 3). Although there is some controversy over the magnitude of convective effects, it has been thought that in vertical gravity stabilized miscible floods, such as those carried out in the carbonate pinnacle reefs of Alberta, molecular diffusion played an important role in the mixing of the solvent bank with drive gas and oil(3, 4, 5). In other processes, such as gas, solvent or CO2 horizontal miscible floods, the effects of molecular diffusion are difficult to quantify, but are thought to be important in the recovery of non-flowing oil by mass transfer(6, 7, 8). The estimation of molecular diffusion coefficients in low-pressure gases using the Chapman Enskog Theory(11) is adequate for a wide variety of single component and binary gases(10, 12), In liquids, the success of theoretical models has been more limited, perhaps because of the complex nature of molecular interactions in the liquid state, Nevertheless, except in the critical region, liquid diffusion coefficients can at present be predicted by empirical correlations with sufficient accuracy for many engineering purposes(9, 13, 14). In the critical or dense fluid state, there are few experimental studies of diffusion coefficients in binary mixtures. This has made it difficult to assess predictive techniques, which have so far been tested largely on the basis of self-diffusion data and are generally considered unreliable(9, 10). The present two-part paper describes Petroleum Recovery Institute work that was directed toward developing improved predictive methods for molecular diffusion coefficients in the high-pressure dense gas state which is commonly encountered under reservoir conditions. The first part presents previously unobtained data on binary diffusion coefficients for fluids in the dense gas state, and a general correlation for binary molecular diffusion coefficients at conditions ranging from ambient to those likely to exist in Alberta reservoirs. The second part presents a method for using the binary con-elation to estimate diffusion in multicomponent systems and shows how the effects of convection and diffusion can be combined to calculate concentration profiles for fluid mixtures flowing in porous media.