3D reservoir model history matching based on machine learning technology

Egor Illarionov, Pavel Temirchev, Dmitry Voloskov, Anna Gubanova, Dmitry Koroteev, Maxim Simonov, Alexey Akhmetov, Andrey Margarit

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

In adaptation of reservoir models a direct gradient backpropagation through the forward model is oftenintractable or requires enormous computational costs. Thus one have to construct separate models thatsimulate them implicitly, e.g. via stochastic sampling or solving of adjoint systems. We demonstrate that ifthe forward model is a neural network, gradient backpropagation becomes naturally involved both in modeltraining and adaptation. In our research we compare 3 adaptation strategies: variation of reservoir modelvariables, neural network adaptation and latent space adaptation and discuss to what extent they preserve thegeological content. We exploit a real-world reservoir model to investigate the problem in practical case. Thenumerical experiments demonstrate that the latent space adaptation provides the most stable and accurateresults.

Original languageEnglish
Title of host publicationSociety of Petroleum Engineers - SPE Russian Petroleum Technology Conference 2020, RPTC 2020
PublisherSociety of Petroleum Engineers
ISBN (Electronic)9781613997451
Publication statusPublished - 2020
EventSPE Russian Petroleum Technology Conference 2020, RPTC 2020 - Virtual, Online
Duration: 26 Oct 202029 Oct 2020

Publication series

NameSociety of Petroleum Engineers - SPE Russian Petroleum Technology Conference 2020, RPTC 2020

Conference

ConferenceSPE Russian Petroleum Technology Conference 2020, RPTC 2020
CityVirtual, Online
Period26/10/2029/10/20

Fingerprint

Dive into the research topics of '3D reservoir model history matching based on machine learning technology'. Together they form a unique fingerprint.

Cite this