PREDICTION OF FAULT ROCK PERMEABILITY WITH DEEP LEARNING: TRAINING DATA FROM TRANSFER SAMPLES OF FAULT CORES
Schmatz J, Klaver J, Wellmann E, Jiang M, Kraus W, Freitag S, Hoffmann A, Deckert H (2022)
Publication Type: Conference contribution
Publication year: 2022
Publisher: European Association of Geoscientists and Engineers, EAGE
Book Volume: 4
Pages Range: 2794-2798
Conference Proceedings Title: 83rd EAGE Conference and Exhibition 2022
Event location: Madrid, Virtual, ESP
ISBN: 9781713859314
Abstract
While the influence of clay smear on the sealing properties of fault zones in siliciclastic rocks can be predicted by validated concepts such as the SGR or CSP reliable models for predicting the structural and hydraulic properties of faults in layered limestone-marl sequences do not yet exist. The main goal of our study is to analyse the development of fault sealing as a result of marl smearing in interaction with mechanical mixing as well as fracturing and cementation processes in dependence of mechanically alternating bedding and fault geometry. Oriented transfer samples of fault cores from different normal fault systems destabilised by the fault process with adjacent damage zone were successfully extracted from outcrops with Jurassic limestone in a quarry Northern Bavaria, Germany. Microanalytical tools and multiscale (m-nm) analyses workflows were developed to provide ground truth for the training of machine learning algorithms for the efficient interpretation of 2D microstructural image data. The systematic macroscopic and microstructural examination of the transfer specimens has shown that the fault zones are built up by recurrent building blocks, whose variation and expression are strongly influenced by the presence and nature of interbedded marly layers.
Authors with CRIS profile
Involved external institutions
MaP – Microstructure and Pores
Germany (DE)
Institut für geothermisches Ressourcenmanagement (igem)
Germany (DE)
Rheinisch-Westfälische Technische Hochschule (RWTH) Aachen
Germany (DE)
Germany (DE)
Institut für geothermisches Ressourcenmanagement (igem)
Germany (DE)
Rheinisch-Westfälische Technische Hochschule (RWTH) Aachen
Germany (DE)
How to cite
APA:
Schmatz, J., Klaver, J., Wellmann, E., Jiang, M., Kraus, W., Freitag, S.,... Deckert, H. (2022). PREDICTION OF FAULT ROCK PERMEABILITY WITH DEEP LEARNING: TRAINING DATA FROM TRANSFER SAMPLES OF FAULT CORES. In 83rd EAGE Conference and Exhibition 2022 (pp. 2794-2798). Madrid, Virtual, ESP: European Association of Geoscientists and Engineers, EAGE.
MLA:
Schmatz, J., et al. "PREDICTION OF FAULT ROCK PERMEABILITY WITH DEEP LEARNING: TRAINING DATA FROM TRANSFER SAMPLES OF FAULT CORES." Proceedings of the 83rd EAGE Conference and Exhibition 2022, Madrid, Virtual, ESP European Association of Geoscientists and Engineers, EAGE, 2022. 2794-2798.
BibTeX: Download