HR: 0800h
AN: H21C-0705 [Abstracts]
TI: Monitoring of Saltwater Intrusion in a sediment model using Electrical Resistivity Tomography
in a High Performance Computing Environment
AU: * Tubbs, K R
EM: ktubbs2@lsu.edu
AF: Louisiana State University, Donald W. Clayton Program in Engineering Science, 3418G
Patrick F. Taylor Hall, Baton Rouge, LA 70803,
AU: Tsai, F T
EM: ftsai@lsu.edu
AF: Louisiana State University, Department of Civil and Environmental Engineering, 3418G
Patrick F. Taylor Hall, Baton Rouge, LA 70803,
AU: White, C
EM: cdwhite@lsu.edu
AF: Louisiana State University, Department of Petroleum Engineering, Baton Rouge, LA 70803,
AU: Allen, G
EM: gallen@cct.lsu.edu
AF: Louisiana State University, Department of Computer Science, Baton Rouge, LA 70803,
AU: Tohline, J
EM: tohline@physics.lsu.edu
AF: Louisiana State University, Department of Physics and Astronomy, Baton Rouge, LA 70803,
AB:
Numerical and experimental studies have been conducted using electrical resistivity tomography (ERT) to better
understand subsurface heterogeneity and saltwater intrusion mechanisms in a sediment model. ERT is a
geophysical method which calculates the electrical resistivity distribution in the subsurface environment from a
large number of electrical potential measurements made from electrodes. In this study, ERT is used to first
characterize the physical sediment model to estimate the resistivity distribution due to heterogeneity. ERT is then
used to image the evolution of the saltwater intrusion at various stages in the sediment model. The ERT is
formulated as a regularized least-squares (RLS) problem. The inversion of electrical resistivity is conducted
through an adjoint-state method. Both numerical and experimental studies are based on a 3D flow flume with
dimensions 1m by 1m by 0.06m. The electrode configuration consists of 48 pairs of electrodes with 48
measurements per dipole-dipole pattern. Due to the large number of measurements and unknown parameters,
ERT image reconstruction is computationally expensive for 3D inversion. The proposed ERT image
reconstruction is implemented in a high performance computing (HPC) environment using PetSc and Cactus
Framework to reduce the time for ERT inversion. We demonstrate the applicability of the parallel ERT inversion in
both synthetic and laboratory three-dimensional ERT problems.
DE: 1816 Estimation and forecasting
DE: 1846 Model calibration (3333)
DE: 3260 Inverse theory
SC: Hydrology [H]
MN: 2007 Fall Meeting