Because of the importance of diffusion in oil production, many studies have been carried out, ranging from theoretical modeling, experimental measurement, investigation of the role of diffusion, and improvement of simulation tools.
There is a decent knowledge about diffusion but we are still far from accurately determining its values and quantifying its impact through simulation. For instance, it is realized that multicomponent diffusion should be modelled by the Maxwell-Stefan diffusion model but Fick’s law is exclusively used in commercial simulators as an industrial standard; it is known that diffusion coefficients should be temperature, pressure and concentration dependent but they are treated as constants in practice.
Furthermore, those approximations are usually employed in the experimental measurement of diffusion coefficients which is heavily dependent on processing the experimental data using a presumed model. This means that the measured effective diffusion coefficients can be distorted by the inappropriate models or assumptions used in the data processing. The inaccurate effective diffusion coefficients can further affect the evaluation of the role of diffusion in the subsequent simulation studies.
It is proposed here to use a holistic approach to improve the quantification of the diffusion effect in oil and gas production. In the proposed study, it is aimed to study how to obtain adequate diffusion coefficients from a typical experimental measurement of multicomponent diffusion coefficients, and how to use the diffusion coefficients in the evaluation of the diffusion effect in the fractured reservoir under different scenarios.
The study will involve development of a compositional simulator with a diffusion model based on the Maxwell-Stefan scheme rather than the Fickian scheme. In addition, some experimental work may be carried out to validate the approach to determining diffusion coefficients. In the evaluation of the diffusion effect in fracture-matrix systems, comparisons with commercial simulators will be made.
Main supervisor: Wei Yan
Co- supervisor: Erling H. Stenby, Alexander A. Shapiro