Title: Multi-well placement optimisation using sequential artificial neural networks and multi-level grid system

Authors: Ilsik Jang; Seeun Oh; Hyunjeong Kang; Juhwan Na; Baehyun Min

Addresses: Department of Energy and Resources Engineering, Chosun University, 309 Pilmun-daero Dong-gu, Gwangju 61452, South Korea ' Department of Energy and Resources Engineering, Chosun University, 309 Pilmun-daero Dong-gu, Gwangju 61452, South Korea ' Department of Energy and Resources Engineering, Chosun University, 309 Pilmun-daero Dong-gu, Gwangju 61452, South Korea ' Department of Energy and Resources Engineering, Chosun University, 309 Pilmun-daero Dong-gu, Gwangju 61452, South Korea ' Department of Climate and Energy Systems Engineering, Ewha Womans University, 52 Ewhayeodae-gil, Seodaemun-gu, 03760, South Korea

Abstract: This study suggests a sequential artificial neural network (ANN) method coupled with a multi-level grid system to optimise multi-well placement in petroleum reservoirs. As the number of scenarios for placing wells increases exponentially with the number of wells, the difficulty in finding the global optimum increases accordingly due to the intrinsic uncertainty of ANNs. The multi-level grid system can reduce the size of the search space by allocating only one well grid block per several grid blocks in the basic grid system. A higher level of grid system consists of finer grid blocks to gradually improve the resolution of the grid system. Repetitive implementation of the sequential ANN at each level of the grid system narrows the search space, and the global optimum is determined. The proposed algorithm is validated with applications to two- and three-infill-well problems in a coal-bed methane (CBM) reservoir. [Received: March 16, 2018; Accepted: September 19, 2018]

Keywords: sequential artificial neural network; multi-level grid system; multi-well placement; optimisation.

DOI: 10.1504/IJOGCT.2020.108047

International Journal of Oil, Gas and Coal Technology, 2020 Vol.24 No.4, pp.445 - 465

Received: 16 Mar 2018
Accepted: 19 Sep 2018
Published online: 02 Jul 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article