Three-dimensional modelling of controlled source electro-magnetic surveys using non-conforming finite element methods

SUMMARY The controlled source electro-magnetic (CSEM) method is increasingly used for in-land and off-shore subsurface characterization. Given its complex dependence between data and the parameters of interest, there is a crucial need for performant numerical algorithms that can simulate the CSEM re...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Geophysical journal international 2022-01, Vol.229 (2), p.1133-1151
Hauptverfasser: Elías, Matías W, Zyserman, Fabio I, Rosas-Carbajal, Marina, Manassero, María Constanza
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:SUMMARY The controlled source electro-magnetic (CSEM) method is increasingly used for in-land and off-shore subsurface characterization. Given its complex dependence between data and the parameters of interest, there is a crucial need for performant numerical algorithms that can simulate the CSEM response of 3-D geological structures. Here, we present two finite element (FE) algorithms for simulating the CSEM response in 3-D media with isotropic conductivity. A primary/secondary field approach is used to avoid the singularity introduced by the source. The primary field is computed semi-analytically for a horizontally layered model and different sources. The secondary field is obtained by discretizing the diffusive frequency-domain Maxwell’s equations with non-conforming FE. The two numerical algorithms are specifically designed to work on distributed-memory computers: (1) an iterative procedure with domain decomposition and (2) a direct and global algorithm. We evaluate their performance by computing their speed up on parallel processors, and solving problems with realistic conductivity structures. We also compare the accuracy of the solutions with published results on canonical models. The results shown here demonstrate the functionality of the two methodologies presented for specific cases when computing 3-D CSEM solutions.
ISSN:0956-540X
1365-246X
DOI:10.1093/gji/ggab524