English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Data assimilation in subsurface flow systems is challenging due to the largenumber of flow simulations often required, and by the need to preservegeological realism in the calibrated (posterior) models. In this work wepresent a deep-learning-based surrogate model for two-phase flow in 3Dsubsurface formations. This surrogate model, a 3D recurrent residual U-Net(referred to as recurrent R-U-Net), consists of 3D convolutional and recurrent(convLSTM) neural networks, designed to capture the spatial-temporalinformation associated with dynamic subsurface flow systems. A CNN-PCAprocedure (convolutional neural network post-processing of principal componentanalysis) for parameterizing complex 3D geomodels is also described. Thisapproach represents a simplified version of a recently developedsupervised-learning-based CNN-PCA framework. The recurrent R-U-Net is trainedon the simulated dynamic 3D saturation and pressure fields for a set of random`channelized' geomodels (generated using 3D CNN-PCA). Detailed flow predictionsdemonstrate that the recurrent R-U-Net surrogate model provides accurateresults for dynamic states and well responses for new geological realizations,along with accurate flow statistics for an ensemble of new geomodels. The 3Drecurrent R-U-Net and CNN-PCA procedures are then used in combination for achallenging data assimilation problem involving a channelized system. Twodifferent algorithms, namely rejection sampling and an ensemble-based method,are successfully applied. The overall methodology described in this paper mayenable the assessment and refinement of data assimilation procedures for arange of realistic and challenging subsurface flow problems.

Deep-learning-based surrogate flow modeling and geological parameterization for data assimilation in 3D subsurface flow