THREE-DIMENSIONAL PRINTING FOR CARBONATE AQUIFER CHARACTERIZATION

Some strata in eogenetic karst carbonate aquifers, such as the Biscayne aquifer of Southeast Florida, are characterized by centimeter-scale, touching-vug megapores related to fabric-selective dissolution of ichnofabrics in a porous limestone matrix. Megaporosity values of up to 80 percent have been measured in decimeter-scale outcrop samples from the Biscanyne aquifer. Borehole imaging has revealed that these permeable layers can be correlated over broad areas in the Biscayne aquifer; hydrologic testing has demonstrated that these layers are the primary contributor to aquifer transmissivity. This megaporous limestone is fragile during normal drilling and coring operations, and may thus be incorrectly logged as large cavities or recovered as rubble that does not reflect the in situ properties of the rock. This may, in turn, lead to an undersampling bias in core-based studies of aquifer permeability. Computed tomography (CT) of carbonate rock samples obtained from outcrops has allowed digital trimming of the samples into convenient shapes for direct computation of permeability using megapore-scale Lattice Boltzmann modeling. Difficulties in cutting the real samples for laboratory permeability measurements required using a 3D printed epoxy core rendered as a subset of a CT-scanned outcrop sample.

The epoxy model of a ‘core’ with the long axis measuring 0.230 m in the z-direction and 0.1 m in both the x- and y-directions was printed with an epoxy sheath enclosing the epoxy rendering of the megaporous rock. The novel dual wax-epoxy prototyping method was used to replicate the intricate megapores in the core to an accuracy of +/- 0.00001 m/m in all three dimensions. This core has been used for nuclear magnetic resonance imaging (NMRI) of a flow field inside the touching-vug megapores and for specialized laboratory measurements of the permeability using glycerol as a viscous fluid. The permeability measurements agree very well with Lattice Boltzmann model-based estimates.