TY - JOUR
T1 - Development of ensemble smoother–neural network and its application to history matching of channelized reservoirs
AU - Kim, Sungil
AU - Lee, Kyungbook
AU - Lim, Jungtek
AU - Jeong, Hoonyoung
AU - Min, Baehyun
N1 - Funding Information:
This research was supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry and Energy (MOTIE), Korea (No. 20172510102160 ) and also Korea Institute of Geoscience and Mineral Resources (KIGAM) ( GP2020-006 ). Baehyun Min was supported by the National Research Foundation of Korea (NRF) grants (No. 2018R1A6A1A08025520 and No. 2019R1C1C1002574 ). Kyungbook Lee was supported by Gas Hydrate R&D Organization (GHDO), MOTIE , and KIGAM ( GP2020-004 ).
Publisher Copyright:
© 2020 Elsevier B.V.
PY - 2020/8
Y1 - 2020/8
N2 - This study develops ensemble smoother–neural network (ES-NN) that combines an ensemble smoother (ES) with a convolutional autoencoder (CAE) to yield comparable performance at a lower computational cost to that of an ensemble smoother–multiple data assimilation (ES-MDA). The ES-NN updates reservoir facies models using CAE trained by importing initial and updated ensembles of ES as input and output of the CAE, respectively, which aims to learn the principle of assimilation of the ES. The trained CAE is recurrently applied in reservoir model calibration without additional forward simulation. The ES-NN yields satisfactory history matching results in terms of production profiles and facies distributions compared to ES and ES-MDA in two case studies. This comparison highlights the efficacy of ES-NN as a prospective data assimilation tool for history matching.
AB - This study develops ensemble smoother–neural network (ES-NN) that combines an ensemble smoother (ES) with a convolutional autoencoder (CAE) to yield comparable performance at a lower computational cost to that of an ensemble smoother–multiple data assimilation (ES-MDA). The ES-NN updates reservoir facies models using CAE trained by importing initial and updated ensembles of ES as input and output of the CAE, respectively, which aims to learn the principle of assimilation of the ES. The trained CAE is recurrently applied in reservoir model calibration without additional forward simulation. The ES-NN yields satisfactory history matching results in terms of production profiles and facies distributions compared to ES and ES-MDA in two case studies. This comparison highlights the efficacy of ES-NN as a prospective data assimilation tool for history matching.
KW - Convolutional autoencoder
KW - Ensemble smoother–multiple data assimilation
KW - Ensemble smoother–neural network
KW - History matching
UR - http://www.scopus.com/inward/record.url?scp=85081907280&partnerID=8YFLogxK
U2 - 10.1016/j.petrol.2020.107159
DO - 10.1016/j.petrol.2020.107159
M3 - Article
AN - SCOPUS:85081907280
SN - 0920-4105
VL - 191
JO - Journal of Petroleum Science and Engineering
JF - Journal of Petroleum Science and Engineering
M1 - 107159
ER -