History-Matching Improvement of Complex Geomodels
May 18, 2021
PONENCIAS DICT will be hosting on Friday, May 21 @ 10h (Mexico City, CT) the webinar “Deep-learning-based flow modeling and geological parameterization for 3D subsurface flow” by Dr. Louis J. Durlofsky, professor at Stanford University. Register now!
In this webinar, Dr. Durlofsky will introduce geological parameterization procedures based on the use of principal component analysis and convolutional neural networks (CNN-PCA procedure) to enable the concise and geologically realistic representation of complex geomodels. Furthermore, Dr. Durlofsky will,
-
present a recurrent residual U-Net method as a surrogate for the flow problem,
-
apply the 3D CNN-PCA parameterization and the recurrent residual U-Net flow surrogate in combination for a challenging history-matching problem.
Join us this Friday, May 21 @ 10 h (Mexico City, CT).
Save your seat @ t.ly/etyM
