English
 
Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
Add to Basket
 Local TagsRelease HistoryDetailsSummary
  Ongoing Approaches for utilizing Neural Networks in DC Resistivity Inversion

Weit, S., Spitzer, K., Scheunert, M., Rheinbach, O. (2024): Ongoing Approaches for utilizing Neural Networks in DC Resistivity Inversion. - In: Börner, J., Yogeshwar, P. (Eds.), - Protokoll über das 30. Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung: St. Marienthal, 25. September - 29. September 2023, 30. Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung (St. Marienthal 2023), 31-31, 1 p.

Item is

 

show all hide all

Basic

show hide
Item Permalink: https://gfzpublic.gfz-potsdam.de/pubman/item/item_5026104 Version Permalink: https://gfzpublic.gfz-potsdam.de/pubman/item/item_5026104_3
Genre: Conference Paper

Files

show Files
hide Files
:
EMTF_2023_St_Marienthal_13_Weit.pdf (Publisher version), 2MB
View   Save
File Permalink:
https://gfzpublic.gfz-potsdam.de/pubman/item/item_5026104_2/component/file_5026106/EMTF_2023_St_Marienthal_13_Weit.pdf
Name:
EMTF_2023_St_Marienthal_13_Weit.pdf
Description:
Fulltext
Visibility:
Public
MIME-Type / Checksum:
application/pdf / [MD5]
Technical Metadata:
View
Copyright Date:
-
Copyright Info:
-
License:
-

Locators

show
hide
Locator:
https://gfzpublic.gfz-potsdam.de/pubman/item/item_5026039 (Supplementary material)
Description:
Conference Proceedings

Creators

show
hide
 Creators:
Weit, Sascha1, Author
Spitzer, Klaus1, Author
Scheunert, Mathias1, Author
Rheinbach, Oliver1, Author
Affiliations:
130. Kolloquium, 2023, Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung EMTF, External Organizations, ou_5026038              

Content

show
hide
Free keywords: -
 Abstract: As part of an investigation into applications of Neural Networks for EM problems, different approaches have been tested for DC resistivity modeling and inversion. The first approach consists of using Convolutional Neural Networks (CNN) for DC resistivity inversion. For this purpose, dipole-dipole pseudosections were simulated using the in-house FEMALY toolbox and used as input data for a CNN, which was trained to output underground resistivity. Training results showed qualitatively good match with the ground truth. However, the predictions are characterized by lack of extrapolation to unseen types of data (e.g. homogeneous half-spaces) and coarse grid enforced by the approach. The second approach laid in the use of Physics Informed Neural Networks (PINN). In this approach, the relevant partial differential equation is included as a regularization term in the loss function, leading to a network whose outputs are guided by physics. Derivatives for the PDE termare obtained via automatic differentiation, removing the need for discretization. This also necessitated a move to solving the forward problem. While this approach has the benefits of being mesh-free and incorporating physics into the training process, in practice it failed at even elementary modeling cases, particularly involving resistivity anomalies. A third approach aimed at combining the previous two, by creating a physics-informed Convolutional Neural Network. This was achieved by replacing the previous loss approaches by a convolution with a Laplace-operator Kernel. This approach produces results that look promising qualitatively for homogeneous half-spaces, however full Dirichlet boundary conditions are required and resistivity anomalies can again not be easily incorporated.

Details

show
hide
Language(s): eng - English
 Dates: 20242024
 Publication Status: Finally published
 Pages: 1
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: -
 Degree: -

Event

show
hide
Title: 30. Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung
Place of Event: St. Marienthal
Start-/End Date: 2023-09-25 - 2023-09-29

Legal Case

show

Project information

show

Source 1

show
hide
Title: Protokoll über das 30. Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung: St. Marienthal, 25. September - 29. September 2023
Source Genre: Proceedings
 Creator(s):
Börner, Jana1, 2, Editor
Yogeshwar, Pritam2, 3, Editor
Affiliations:
1 Technische Universität- Bergakademie Freiberg , ou_persistent22            
2 30. Kolloquium, 2023, Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung EMTF, External Organizations, ou_5026038            
3 Universität zu Köln, ou_persistent22            
Publ. Info: Potsdam : Deutsche Geophysikalische Gesellschaft e. V.
Pages: 146 Volume / Issue: 30 Sequence Number: 13 Start / End Page: 31 - 31 Identifier: ISSN: 2190-7021