Identifier

etd-04272012-114012

Degree

Master of Science in Chemical Engineering (MSChE)

Department

Chemical Engineering

Document Type

Thesis

Abstract

Micromodels are used to visualize and study pore-scale phenomena such as immiscible displacements in porous media, foam flow behavior, and CO2 flooding. The understanding gained from these experiments can be used to develop models to predict future behavior of the reservoir. Most micromodels are constructed using lithography techniques that are restricted to 2D patterns that require artificial generation or manipulation of images to develop connected micromodels. Characteristics innate to the original rock structure are often lost or skewed in developing micromodels that bear little resemblance to the original media. Alternative microfabrication techniques using a micromilling tool have allowed us to vary the floor height in a micromodel, thus giving some variation in the third dimension. We refer to these structures (with varying floor height and fixed ceiling, and which cannot have passages on top of one another) as 2.5D micromodels. Using a technique called depth averaging, in which we take a section of a 3D voxel image of porous media and project the solid voxels down while simultaneously pushing the void space above, we generate micromodels that may allow for more accurate representations of the pore structure in 3D rock. The design of the etched pattern requires the selection of a specific depth (or number of XMCT image slice) over which to average the image data. The 2.5-D pattern was obtained by optimizing a series of parameters to ensure the structure and flow patterns matched as closely as possible to the equivalent 3D structure and flow as can be accommodated given the restricted dimensionality. Parameters considered include flow-based parameters, common statistical correlations, and a host of topological parameters obtained by network model generation techniques. For a Boise sandstone core sample imaged at 5.07 µm/pixel, an optimized depth of 115 µm gave the most accurate measures across the range of parameters. However, due to constraints regarding the resolution of the micromilling process, a second series of flow simulations were conducted in the originally optimized region of interest (100-150 µm) for a lower resolution image that resulted in the selection of 130 µm to depth average. This design was then used to fabricate a brass mold insert. The process of developing the microchips for nanosensor experiments is currently in the stage of assembling the PMMA chips.

Date

2012

Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Thompson, Karsten E.

Share

COinS