Ridha Gharbi

A look at dispersion in porous media through computed tomography imaging

This paper presents a method for measuring the longitudinal dispersion coefficient and the adsorption properties of porous media by X-ray computed tomography (CT) imaging. The in situ solvent concentration profiles of tracer tests were imaged and fitted to a convection-dispersion mathematical model. In addition to providing the average dispersion coefficient traditionally measured with breakthrough curves, our method allows the dispersion phenomenon to be visualized and the contribution of heterogeneity to the dispersion coefficient to be estimated. The method is demonstrated by measuring the dispersion coefficients and adsorption properties of a sandpack and a Berea sandstone. Results show that heterogeneity increases the dispersion coefficient above that of a homogeneous porous medium. Adsorption was found to be higher in the Berea sandstone, a natural porous medium, than in the clean sandpack. CT imaging of the tracer tests has shed more light on the dispersion and adsorption phenomena in porous media.

On scaling immiscible displacements in permeable media

Abstract This study was aimed at scaling unstable immiscible laboratory corefloods in oder to generalize their results. The corefloods were imaged by CT to measure and visualize their fluid saturation distributions in time and space. The saturation data were scaled with a dimensionless self-similarity variable to derive the characteristics dimensionless response curves for the unstable corefloods. The results show that unstable immiscible displacements are self-similar processes. When presented as a function of the self-similar dimensionless variable, the spatial and temporal saturation data collapsed into one dimensionless response curve. The saturation data obtained by different experimenters on similar porous media that previously appeared to be random and unpredictable were found to fall on the same dimensionless response curve under the new scaling. We also found that the dimensionless response function for a water-wet medium was significantly different from that of an oil-wet medium. The response function for the water-wet medium had a well-defined front, whereas that for the oil-wet medium had a more diffuse and less well-developed front. These features reflect the higher displacement efficiency in the water-wet medium than in the oil-wet medium. Theser results show that unstable immiscible displacements are not so unpredictable as we once thought them to be. When properly scaled, unstable immiscible displacements are just as reproducible and predictable as stable displacements. The study has resulted in significant new insights into the performance of unstable immiscible displacements. The methodology developed in this study can be applied to other laboratory experiments, both to plan the experiments and to present their results meaningfully. Further, the methodology of this study when combined with numerical simulation and the principles of similitude, provides a way to scale the results of the laboratory experiments to other systems. This novel approach to performance prediction that combines laboratory imaging experiments with numerical modeling is the subject of a companion paper.

Numerical modeling of laboratory corefloods

Abstract This paper describes a procedure for scaling laboratory coreflood experiments in order to generalize their results to other systems. The procedure combines numerical modeling with laboratory imaging experiments. The corefloods were imaged by computed tomography (CT) to obtain their saturation distributions in time and space. The saturation data were transformed by a self-similar dimensionless variable to obtain the characteristic dimensionless response curves. The characteristics response curves were calculated numerically using fine-grid numerical simulation and compared with the experimental curves. The numerical simulation was repeated with adjusted parameters until the experimental and the computed response curves were in good agreement. After successfully matching the experimental response curves, the well-tuned numerical model was used to scale the results of the laboratory coreflood experiments to other systems by changing the appropriate dimensionless scaling groups in the model. This procedure was used to scale the results of laboratory corefloods of unstable immiscible displacements with excellent results. It was found that using appropriate relative permeability and capillary pressure curves, the dimensionless response curves for unstable immiscible displacements were accurately simulated numerically. It was also found that the simulation of the dimensionless response curves in the corefloods automically simulated the experimental recovery curves as well. The results show that the numerical simulation of the dimensionless response functions of corefloods is a powerful technique for scaling the results of laboratory corefloods to other systems.

Scaling coreflood experiments to heterogeneous reservoirs

Abstract This paper describes a procedure for scaling a laboratory waterflood experiment in a homogeneous sandpack to predict its expected performance in heterogeneous reservoirs. The dimensionless response function of the laboratory coreflood obtained by computed tomography (CT) imaging was history-matched with a numerical simulator. The well-tuned numerical simulator was then used to predict the expected waterflood performance in several two-dimensional heterogeneous reservoirs generated geostatistically. We found that if the heterogeneous medium is characterized by high variability and high correlation in its permeability distribution, then the waterflood performance in the medium will be significantly less than in the laboratory sandpack. In this case, the laboratory waterflood must be scaled in order to use it to predict the expected performance in the field. However, if the heterogeneous medium is characterized by low variability in the permeability distribution, then the waterflood response will be the same as in the laboratory sandpack regardless of the correlation structure of the permeability distribution. In this case, no scaling is required in using the laboratory waterflood to predict the expected field performance. Further, if the heterogeneous medium is characterized by low correlation in the permeability distribution, then the waterflood response will be the same as in the laboratory sandpack, regardless of the variability in its permeability distribution. In this case, no scaling of the laboratory coreflood is required to predict expected field performance. The results from this study clearly demonstrate that the combination of CT-imaging experiments with numerical modeling provides a novel way to scale laboratory corefloods in homogeneous media to forecast their expected performance in heterogeneous reservoirs.