Optimization of Offshore Saline Aquifer CO2 Storage in Smeaheia Using Surrogate Reservoir Models

Aug 4, 2026

Upload by Asher

Machine learning-based surrogate reservoir models (SRMs) are used to replicate reservoir simulation outputs with far lower computational cost while keeping high accuracy relative to 3D numerical simulators. The study targets CO2 storage in the Smeaheia saline aquifer and builds two deep neural network SRMs to predict CO2 saturation and pressure for 50-year periods across all grid blocks. Training relies on 18 million samples with 31 static and dynamic input features, using 3–5 hidden layers. Results show about 300× runtime improvement and 99% accuracy, followed by genetic-algorithm optimization of injection rate and duration under operational safety constraints.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages27

Reading Time1 h 8 min

AccessFree to download

Introduction

Climate-change context and CCS motivation

CO2 geological storage risks and optimization goals

Methods (SRMs and optimization workflow)

Deep neural network surrogate reservoir models

Genetic algorithm for injection rate and duration

Results and implications

Runtime and accuracy comparison with numerical simulation

Long-term optimization outcomes for Smeaheia storage

POLITECNICO DI TORINO Repository ISTITUZIONALE Optimization of Offshore Saline Aquifer CO2 Storage in Smeaheia Using Surrogate Reservoir Models Original Optimization of Offshore Saline Aquifer CO2 Storage in Smeaheia Using Surrogate Reservoir Models / Amiri, Behzad; Jahanbani Ghahfarokhi, Ashkan; Rocca, Vera; Shang Wui Ng, Cuthbert. -In: ALGORITHMS. -ISSN 1999-4893. -ELETTRONICO. -17:10(2024) . [10 .3390/a17100452] Availability: This version is available at: 11583/2993378 since: 2024-10-14T08:02:09Z Publisher: MDPI Published DOI:10.3390/a17100452 Terms of use: This article is made available under terms and conditions as specified in the corresponding bibliographic description in the repository Publisher copyright (Article begins on next page) 17 February 2025 algorithms Article Optimization of Offshore Saline Aquifer CO 2 Storage in Smeaheia Using Surrogate Reservoir Models Behzad Amiri 1, Ashkan Jahanbani Ghahfarokhi 2, *, Vera Rocca 3 and Cuthbert Shang Wui Ng 2 Citation: Amiri, B.; Jahanbani Ghahfarokhi, A.; Rocca, V.; Ng, C.S.W. Optimization of Offshore Saline Aquifer CO2 Storage in Smeaheia Using Surrogate Reservoir Models. Algorithms 2024, 17, 452. [https://](https://)[ ](https://)[doi.org/10.3390/a17100452](doi.org/10.3390/a17100452) Academic Editors: Szymon Łukasik, Piotr A. Kowalski and Rohit Salgotra Received: 9 September 2024 Revised: 3 October 2024 Accepted: 9 October 2024 Published: 11 October 2024 Copyright: © 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 Department of Energy Resources, University of Stavanger, 4021 Stavanger, Norway; [behzad.amiri@uis.no](behzad.amiri@uis.no) 2 Department of Geosciences, Norwegian University of Science and Technology, 7031 Trondheim, Norway; [cuthbert.s.w.ng@ntnu.no](cuthbert.s.w.ng@ntnu.no) 3 Department of Environment, Land and Infrastructure Engineering, Politecnico di Torino, 10129 Torino, Italy; vera.rocca@polito.it * Correspondence: [ashkan.jahanbani@ntnu.no](ashkan.jahanbani@ntnu.no) Abstract: Machine learning-based Surrogate Reservoir Models (SRMs) can replace/augment multiphysics numerical simulations by replicating the reservoir simulation results with reduced computational effort while maintaining accuracy compared with numerical simulations. This research will demonstrate SRMs' potential in long-term simulations and optimization of geological carbon storage in a real-world geological setting and address challenges in big data curation and model training. The present study focuses on CO 2 storage in the Smeaheia saline aquifer. Two SRMs were created using Deep Neural Networks (DNNs) to predict CO 2 saturation and pressure over all grid blocks for 50 years. 18 million samples and 31 features, including reservoir static and dynamic properties, build the input data. Models comprise 3–5 hidden layers with 128–512 units apiece. SRMs showed a runtime improvement of 300 times and an accuracy of 99% compared to the 3D numerical simulator. The genetic algorithm was then employed to determine the optimal rate and duration of CO2 injection, which maximizes the volume of injected CO 2 while ensuring storage operations'safety through constraints. The optimization continued for the reproduction of 100 generations, each containing 100 individuals, without any hyperparameter tuning. Finally, the optimization results conﬁrm the signiﬁcant potential of Smeaheia for storing 170 Mt CO 2 . Keywords: geological carbon storage; surrogate reservoir model; artiﬁcial intelligence; deep learning; optimization 1. Introduction The main target of the Paris Agreement in December 2015 was to concentrate efforts on limiting temperature in

FAQs

What problem do surrogate reservoir models address in CO2 storage simulation?

They replace or augment multiphysics numerical simulations by reproducing reservoir simulation results with greatly reduced computational effort while maintaining accuracy.

How were the surrogate reservoir models built for the Smeaheia saline aquifer?

Two deep neural network SRMs were trained to predict CO2 saturation and pressure over all grid blocks for 50 years using 18 million samples and 31 input features.

What optimization method was used to determine injection strategy?

A genetic algorithm optimized injection rate and duration to maximize injected CO2 volume while enforcing constraints to ensure safe storage operations.

Aug 4, 2026

Upload by Asher