This study reviews 254 papers on artificial intelligence (AI) applications in the geoenergy sector, categorized into conventional and future-oriented technologies. Conventional geoenergy includes reservoir, production, and drilling, while future-oriented technologies cover geological CO₂ storage (GCS), gas hydrates (GH), and underground hydrogen storage (UHS). The 254 papers were analyzed systematically based on authorship, publication year, key findings, input-output data relationships, applied AI methods, and sample sizes for AI training. Results highlight the extensive use of machine learning (ML) and deep learning (DL) for tasks such as proxy modeling, dimensionality reduction, data generation, and optimization. Proxy modeling was the most commonly applied method, effectively predicting key economic and engineering parameters. Future-oriented applications require integrating multiple factors for safety and efficiency, emphasizing the need for advanced AI techniques and larger datasets. Conventional studies used larger datasets (average 49,794 samples) compared to future-oriented applications (average 13,779 samples), reflecting their longer research history. The average input-output data sizes (log scale) for ML and DL were 2.01 and 0.34 vs. 4.12 and 2.00, respectively, showing DL’s versatility in handling larger datasets. As geoenergy systems grow more complex, advanced AI methods must address interconnected geo-scales from pore to field levels. Future research should focus on developing automated, user-independent AI systems to enhance model selection and broaden AI applications across diverse datasets, driving greater efficiency and innovation in the geoenergy sector. Graphical abstract

This paper develops a reliable deep-learning framework to extract latent features from spatial properties and investigates adaptive surrogate estimation to sequester CO₂ into heterogeneous deep saline aquifers. Our deep-learning architecture includes a deep convolutional autoencoder (DCAE) and a fully-convolutional network to not only reduce computational costs but also to extract dimensionality-reduced features to conserve spatial characteristics. The workflow integrates two different spatial properties within a single convolutional system, and it also achieves accurate reconstruction performance. This approach significantly reduces the number of parameters to 4.3% of the original number required, e.g., the number of three-dimensional spatial properties needed decreases from 44,460 to 1920. The successful dimensionality reduction is accomplished by the DCAE system regarding all inputs as image channels from the initial stage of learning using the fully-convolutional network instead of fully-connected layers. The DCAE reconstructs spatial parameters such as permeability and porosity while conserving their statistical values, i.e., their mean and standard deviation, achieving R-squared values of over 0.972 with a mean absolute percentage error of their mean values of less than 1.79%. The adaptive surrogate model using the latent features extracted by DCAE, well operations, and modeling parameters is able to accurately estimate CO₂ sequestration performances. The model shows R-squared values of over 0.892 for testing data not used in training and validation. The DCAE-based surrogate estimation exploits the reliable integration of various spatial data within the fully-convolutional network and allows us to evaluate flow behavior occurring in a subsurface domain.

This paper presents a Pareto-based multi-objective optimization for operating CO₂ sequestration with a multi-well system under geological uncertainty; the optimal well allocation, i.e., the optimal allocation of CO₂ rates at injection wells, is obtained when there is minimum operation pressure as well as maximum sequestration efficiency. The distance-based generalized sensitivity analysis evaluates the influence of geological uncertainty on the amount of CO₂ sequestration through four injection wells at 3D heterogeneous saline aquifers. The spatial properties significantly influencing the trapping volume, in descending order of influence, are mean sandstone porosity, mean sandstone permeability, shale volume ratio, and the Dykstra–Parsons coefficient of permeability. This confirms the importance of storable capacity and heterogeneity in quantitatively analyzing the trapping mechanisms. Multi-objective optimization involves the use of two aquifer models relevant to heterogeneity; one is highly heterogeneous and the other is less so. The optimal well allocations converge to non-dominated solutions and result in a large injection through one specific well, which generates the wide spread of a highly mobile CO₂ plume. As the aquifer becomes heterogeneous with a large shale volume and a high Dykstra–Parsons coefficient, the trapping performances of the combined structural and residual sequestration plateau relatively early. The results discuss the effects of spatial heterogeneity on achieving CO₂ geological storage, and they provide an operation strategy including multi-objective optimization.

This study presents a front‐end engineering design (FEED) methodology for an integrated CO ₂ transport–injection–storage system, utilizing multiobjective optimization (MOO) and nodal analysis. The methodology’s performance is validated through a carbon capture and storage (CCS) demonstration project in the Gunsan Basin (GB), South Korea. This approach employs the dynamic inflow performance relationship (IPR)−outflow performance relationship (OPR) technique, applying it to the FEED of the CO ₂ transport–injection–storage system to enable CO ₂ injection into a saline aquifer via a single injection well connected through an onshore hub terminal and a subsea pipeline. By adjusting decision variables (CO ₂ discharge pressure at the onshore hub terminal, pipeline diameter, tubing diameter, and CO ₂ temperature at the wellhead), three objectives (CO ₂ storage capacity, safety, and economic benefit) are optimized through MOO, identifying the Pareto‐optimal front (POF) among objective functions. These trade‐off solutions provide reliable ranges for the four decision variables used in the nodal analysis, which considers real‐time pressure and temperature variations in the system during CO ₂ injection, along with the associated facility qualifications and operating conditions. This analysis determines the IPR−OPR at the bottom of the injection well and the corresponding pressure–flowrate, defining the practical FEED scope for the integrated CO ₂ transport–injection–storage system. By integrating optimal solutions from both MOO and nodal analysis, the study identifies the final nondominated solutions for efficient and stable CO ₂ geological storage. The proposed methodology offers decision‐makers robust scenarios for facility qualifications and operating conditions, considering CO ₂ storage capacity, safety, and economic efficiency at the FEED stage of a CCS demonstration project.