Non-local investigation of miscible displacement in scale dependent porous media

Steven Paul Kirtley Sternberg, Purdue University

Abstract

Experimental and numerical investigation of dispersion between miscible fluids in a heterogeneous porous medium composed of homogeneous layers. This basic research examined the effect of large scale variations of the medium property of permeability on dispersion. Specific questions included if mixing can be modeled from Fick's law in scale dependent porous media, when and where does mixing behave asymptotic, and how does mixing behavior change at transitions in the porous medium properties. Experiments were performed in porous medium models constructed of layers of spherical glass beads packed into cylindrical Lucite columns. Flow was in a direction perpendicular to the layers. The concentration profiles of chloride tracer was measured at many locations along the column length. Each media was built at three lengths and two orders of the layers, giving 6 models. Each model was run at three velocities. All runs were repeated a minimum of two times. In each case the scale of measurement was finer than the scale of the heterogeneity. The experiments show that the mixing between miscible fluids was affected by transitions in the system parameters, before these transitions are reached. This phenomena has been interpreted as a non-local effect, and it is not predicted by theories based on Fick's law. The results also showed that the rate at which the mixing stabilizes after a transition is flow dependent. This work also included a numerical simulation of the experiments. The simulation was based on a random walk in a velocity field. A comparison was made between several ways of determining the dispersion coefficient for the random walk and the experimental results. This part of the work showed that a memory method or a non local method is required to predict the observed behavior of these systems.

Degree

Ph.D.

Advisors

Cushman, Purdue University.

Subject Area

Chemical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS