A New Randomized Binary Prior Model for Hydraulic Tomography in Fractured Aquifers

Area of Science:

• Hydrogeology
• Geophysics
• Environmental Science

Background:

• Fractured aquifers exhibit complex subsurface flow paths, making property estimation challenging.
• Traditional methods struggle to accurately delineate preferential flow paths and hydraulic properties in heterogeneous fractured rock.

Purpose of the Study:

• To develop and validate a novel pilot-point-based hydraulic tomography (HT) inversion procedure using a randomized binary prior model.
• To improve the delineation of preferential flow paths and the estimation of hydraulic properties in fractured aquifers.

Main Methods:

• A randomized algorithm was used to create a binary prior model for guiding pilot point placement and constraining parameters.
• The methodology was tested on a 2D fractured granite block under controlled laboratory conditions.
• Pumping test data were preprocessed and analyzed using the proposed HT inversion procedure coupled with the HydroGeoSphere (HGS) model.

Main Results:

• The randomized binary prior model effectively guided pilot-point-assisted HT inversion.
• Parameter distributions (K and Ss tomograms) visually correlated well with known fracture topology.
• Model predictions showed reasonable accuracy in estimating drawdowns at validation ports.

Conclusions:

• The proposed HT inversion procedure with a randomized binary prior can accurately recover fracture connectivity in fractured aquifers.
• This approach offers a reliable method for predicting drawdowns compared to conventional pilot-point inversion schemes.