Preselection of a sorption model based on a column test: the algorithm and an example of its application

Abstract

In order to describe the contamination of saturated porous media, it is necessary to find an appropriate mathematical model that includes processes occurring in aquifers, such as advection, dispersion, diffusion, and various kinds of sorption. The identification of parameters of those processes is possible through laboratory column experiments, which result in records of breakthrough curves for a conservative tracer and a reactive tracer. An algorithm leading to the preliminary selection of the mathematical model that best describes transport processes of the reactive tracer in the experimental column is proposed in this article. A study published previously presented a sensitivity analysis for an arbitrarily adopted variability of the transport parameters. The analysis involved examining changes in the shape of breakthrough curves caused by the alteration of each parameter value. Specially defined indicators called descriptors were proposed to quantitatively describe the breakthrough curves. Then, formulas were proposed to determine the percentage deviations of descriptors of the breakthrough curve obtained for the reactive tracer in relation to the descriptors of the breakthrough curve of the conservative tracer. In the work described in this article, the deviations are analyzed and an algorithm is proposed that allows the preselection of the most suitable sorption model out of the five discussed simple (one-site) and six hybrid (two-site) models. The algorithm can facilitate and accelerate the interpretation of column experiments of contaminant transport in a porous medium. An example is provided to illustrate the usability of the proposed algorithm.