Interpretation of Miscible Displacement Experiments Using Solutions in Laplace Space and Deconvolution

Abstract

Abstract This paper presents a new methodology to interpret miscible displacement experiments. From the monitoring and analysis of effluent tracer concentration in laboratory cores it has been possible to determine the parameters which characterize the porous media and the displacement process. Miscible displacement models of tracer transport in both homogeneous and heterogeneous porous media have been considered. The three-parameter models for heterogeneous porous media have been solved by numerically inverting solutions obtained in Laplace domain. The numerical inversion has been carried by Crump's algorithm, which yields solutions virtually free of numerical dispersion. The methodology consists initially of the analysis of the effluent concentration data obtained from one-dimensional tracer injection experiments, by simple graphical methods. The results of the graphical interpretation are then used as initial estimates of the model parameters in an automated curve fitting process. Application of graphical analysis is only possible in continuous (step) injection experiments. For experiments consisting of injection of a tracer slug, the effluent composition data are deconvolved. The deconvolution process yields the equivalent effluent composition data that would have been obtained from a continuous injection experiment, thus allowing the application of graphical analysis. An experimental apparatus has been constructed and a series of experiments have been made. Experimental results have been analized with the new interpretation methodology. The combination of graphical and automated curve fitting analysis has proved to be fast and reliable, and can be employed in on-line experiments.