Numerical modeling of fluid displacements in porous media assisted by computed tomography imaging

The displacement of one fluid by another in a heterogeneous porous medium is the basis of many industrial processes such as enhanced oil recovery and the remediation of contaminated aquifers. The interactions of heterogeneity with the several competing forces, namely, viscous, capillary, gravitational, and dispersive forces, can conspire to make the displacements unstable and difficult to model and to predict. The objective of this study was to develop a systematic procedure for modeling unstable fluid displacements in heterogeneous porous media.; A systematic procedure that combines fine-mesh numerical simulation with laboratory imaging experiments was developed to model unstable displacements. The procedure consists of the following steps: (1) the insitu saturation distributions of laboratory corefloods obtained by X-ray computed tomography imaging were subjected to a similarity transformation to make them easier to history-match with numerical models, (2) the transformed saturation data were history-matched with a numerical model, and (3) the well-tuned numerical model was then used to scale the experiments to other systems by changing the appropriate dimensionless scaling groups in the model.; The results show that, in the absence of significant gravity segregation, unstable miscible and immiscible displacements and their numerical models are self-similar processes, a fact that greatly enhances the ability to model and scale them from one system to another. When presented as a function of the self-similarity variable, the spatial and temporal saturation data collapsed into one dimensionless response curve that was easier to model than the original data. Further, it was found that laboratory corefloods in homogeneous porous media need to be scaled in order to use them to predict the expected performance in heterogeneous porous media with high permeability variation and high correlation. However, laboratory corefloods in homogeneous porous media need not be scaled in order to use them to predict the expected performance in heterogeneous porous media with low permeability variation regardless of their correlation structure.; The combination of imaging experiments with numerical modeling has resulted in new insights into the performance of unstable displacements in heterogeneous porous media. The methodology developed in this study will be of interest to those involved in forecasting the performance of enhanced oil recovery processes and the spreading of contaminants in heterogeneous aquifers.