BEL1D (Bayesian Evidential Learning 1D imaging) has recently been introduced as a viable option for the stochastic imaging of the subsurface geophysical properties (Michel et al., 2020). This methodology has been applied to surface nuclear magnetic resonance and surface wave data in order to produce sets of probable models of the subsurface. Here, we improve the accuracy of this algorithm by the introduction of iterative prior resampling. We further validate results against a state-of-the-art McMC method.

Presentation Date: Wednesday, October 14, 2020

Session Start Time: 1:50 PM

Presentation Time: 3:05 PM

Location: 361A

Presentation Type: Oral

Keywords:
reservoir simulation, imaging, thomas herman, reservoir characterization, accuracy, bel1d accuracy, machine learning, upstream oil & gas, bel1d, posterior
Subjects:
Reservoir Characterization, Reservoir Simulation, Information Management and Systems, Seismic processing and interpretation
This content is only available via PDF.
2020. Society of Exploration Geophysicists
You can access this article if you purchase or spend a download.