Diffusivity of Gas Into Bitumen: Part I—Analysis of Pressure-Decay Data With Swellingkyle2020-01-07T10:22:46-07:00
Diffusivity of Gas Into Bitumen: Part I—Analysis of Pressure-Decay Data With Swelling
Richardson, W.D.L., Schoeggl, F.F., Maini, B., Kantzas, A., Taylor, S.D, and Yarranton, H.W.
SPE Journal, 24, August 2019.
The oil-production rate of in-situ heavy-oil-recovery processes involving the injection of gaseous hydrocarbons partly depends on the diffusivity of the gas in the bitumen. The gas diffusivities required to model these processes are determined indirectly from models of mass-transfer experiments. However, the data in the literature are scattered partly because different methods and model assumptions are used in each case. In this work, the pressure-decay method is examined with a focus on accounting for swelling and the dependence of the diffusivity on the solvent content. To assess these issues, the diffusion of gaseous propane into bitumen is measured at conditions where significant swelling occurs. A numerical model is developed for the pressure-decay experiment that accounts for swelling (including excess volumes of mixing) and variable diffusivity. For gases, such as propane, with a relatively high solubility in bitumen, the error in the calculated diffusivity reached 25% when swelling was not included in the model. The error in the height of the gas/oil interface reached 15%. Nonideal mixing had no effect on the calculated diffusivity and only a small effect on the height of the interface. It was found that the diffusion data from a single experiment could be modeled equally well with a constant or a solvent-concentration-dependent (or viscosity-dependent) diffusivity. However, the apparent constant diffusivities for each experiment were different, confirming their dependence on the solvent content. The constant diffusivity approximately correlated to the viscosity of the oil. A larger data set is required to fully develop and test a correlation, and this work will be presented in Part II of this study (Richardson et al. 2019).
A full version of this paper is available on OnePetroOnline.