An attention-guided graph neural network and U-Net++-based reservoir porosity prediction system

Abstract

Accurate prediction of reservoir porosity is fundamental for hydrocarbon resource evaluation and development planning, yet traditional methods struggle with spatial heterogeneity and complex geological structures. This study proposes a hybrid deep learning framework that integrates U-Net++ with an attention-guided graph neural network to simultaneously capture multiscale well logging data features and non-Euclidean spatial dependencies. The model incorporates dense skip connections, deep supervision, and dual-channel attention mechanisms to enhance both local feature extraction and global topological modeling. Experiments on a real-world continental sedimentary basin dataset (26 wells, ~40 km2) demonstrated that the proposed method achieved a mean squared error (MSE) of 4.62, mean absolute error of 1.24, coefficient of determination (R2) of 0.912, and structural similarity index measure of 0.831, representing a 14.9-38.7% reduction in prediction errors relative to widely used deep learning and graph-based baselines. Statistical tests (p<0.05) confirmed the significance of the improvements. The model was particularly robust in extreme porosity ranges (>16% or <8%), reducing errors by 23.1-42.6% compared to U-Net++. Ablation studies highlighted the contribution of graph structure (19.0% MSE reduction), attention mechanism (15.0%), and deep supervision (12.5%). Beyond predictive accuracy, attention-weight analysis revealed strong alignment with geologically meaningful features, such as faults and sedimentary facies boundaries, thereby enhancing interpretability. The proposed framework offers a scalable and interpretable solution for reservoir characterization, with broad potential applications in heterogeneous and faulted reservoirs.

Keywords

Reservoir porosity prediction / Graph neural network / U-Net++ / Attention mechanism / Spatial heterogeneity

Funding

This research was financially supported by Mahasarakham University; Scientific Research Fund of Institute of Seismology, China Earthquake Administration and National Institute of Natural Hazards, MEM, (No. IS202226322); 2025 Doctoral Special Support Program Project of Chengdu Jincheng College (NO.2025JCKY(B)0018); the Key Research Base of Humanities and Social Sciences of the Education Department of Sichuan Province, Panzhihua University, Resource based City Development Research Center Project (NO.ZYZX-YB-2404); and the Open Fund of Sichuan Oil and Gas Development Research Center (NO.2024SY017).

References

[1] Wei Y, Jia A, Xu Y, Fang J. Progress on the different methods of reserves calculation in the whole life cycle of gas reservoir development. J Nat Gas Geosci. 2021; 6(1):55-63. doi: 10.1016/j.jnggs.2021.04.001

[2] Zhang L, Wang B, Hu M, Shi X, Yang L, Zhou F. Research progress on optimization methods of platform well fracturing in unconventional reservoirs. Processes. 2025; 13(6):1887.doi: 10.3390/pr13061887

[3] Cao X, Liu Z, Hu C, Song X, Quaye JA, Lu N. Three-dimensional geological modelling in earth science research: An in-depth review and perspective analysis. Minerals. 2024; 14(7):686. doi: 10.3390/min14070686

[4] Saikia P, Baruah RD, Singh SK, Chaudhuri PK. Artificial Neural Networks in the domain of reservoir characterization: A review from shallow to deep models. Comput Geosci. 2020; 135:104357. doi: 10.1016/j.cageo.2019.104357

[5] Lu F, Fu C, Zhang G, Shi J. Adaptive multi-scale feature fusion based U-net for fracture segmentation in coal rock images. J Intell Fuzzy Syst. 2022; 42(4):3761-3774. doi: 10.21203/rs.2.23959/v2

[6] Khalili Y, Ahmadi M. Reservoir modeling & simulation: Advancements, challenges, and future perspectives. J Chem Pet Eng. 2023; 57(2):343-364. doi: 10.22059/jchpe.2023.363392.1447

[7] Ba J, Zhang L, Wang D, et al. Experimental analysis on P-wave attenuation in carbonate rocks and reservoir identification. J Seism Explor. 2018; 27(4):371-402.

[8] Shahin A, Myers M, Hathon L. Carbonates’ dual-physics modeling aimed at seismic reservoir characterization. J Seism Explor. 2017; 26(4):331-349.

[9] Soleimani F, Hosseini E, Hajivand F. Estimation of reservoir porosity using analysis of seismic attributes in an Iranian oil field. J Pet Explor Prod Technol. 2020; 10(4):1289-1316. doi: 10.1007/s13202-020-00833-4

[10] GhojehBeyglou M. Geostatistical modeling of porosity and evaluating the local and global distribution. J Pet Explor Prod Technol. 2021; 11(12):4227-4241. doi: 10.1007/s13202-021-01308-w

[11] Alimoradi I, Moradzadeh A, Bakhtiari MR. Reservoir porosity determination from 3D seismic data - Application of two machine learning techniques. J Seism Explor. 2012; 21(4):323-345.

[12] Santos JE, Xu D, Jo H, Landry CJ, Prodanović M, Pyrcz MJ. PoreFlow-Net: A 3D convolutional neural network to predict fluid flow through porous media. Adv Water Resour. 2020; 138:103539. doi: 10.1016/j.advwatres.2020.103539

[13] Azad R, Aghdam EK, Rauland A, et al. Medical image segmentation review: The success of u-net. IEEE Trans Pattern Anal Mach Intell. 2024; 46:10076-10095. doi: 10.1109/TPAMI.2024.3435571

[14] Liu W. Review of artificial intelligence for oil and gas exploration: Convolutional neural network approaches and the U-Net 3D model. Open J Geol. 2024; 14(4):578-593. doi: 10.4236/ojg.2024.144024

[15] Waikhom L, Patgiri R. A survey of graph neural networks in various learning paradigms: Methods, applications, and challenges. Artif Intell Rev. 2023; 56(7):6295-6364. doi: 10.1007/s10462-022-10321-2

[16] Vrahatis AG, Lazaros K, Kotsiantis S. Graph attention networks: A comprehensive review of methods and applications. Future Internet. 2024; 16(9):318. doi: 10.3390/fi16090318

[17] Hu J, Cao L, Li T, Dong S, Li P. GAT-LI: A graph attention network based learning and interpreting method for functional brain network classification. BMC Bioinformatics. 2021; 22(1):379. doi: 10.1186/s12859-021-04295-1

[18] Zangari L, Interdonato R, Calió A, Tagarelli A. Graph convolutional and attention models for entity classification in multilayer networks. Appl Netw Sci. 2021; 6(1):87. doi: 10.1007/s41109-021-00420-4

[19] Zhang S, Tong H, Xu J, Maciejewski R. Graph convolutional networks: A comprehensive review. Comput Soc Netw. 2019; 6(1):1-23. doi: 10.1186/s40649-019-0069-y

[20] Jiang Z. Spatial structured prediction models: Applications, challenges, and techniques. IEEE Access. 2020; 8: 38714-38727. doi: 10.1109/ACCESS.2020.2975584

[21] Zhao X, Zhong Y, Li P. RTG-GNN: A novel rock topology-guided approach for permeability prediction using graph neural networks. Geoenergy Sci Eng. 2024; 243:213358. doi: 10.1016/j.geoen.2024.213358

[22] Han S, Zhang Y, Wang J, Tong D, Lyu M. Graph neural network-based topological relationships automatic identification of geological boundaries. Comput Geosci. 2024; 188:105621. doi: 10.1016/j.cageo.2024.105621

[23] Tang M, He Y, Aslam M, Akpokodje E, Jilani SF. Enhanced U-Net++ for improved semantic segmentation in landslide detection. Sensors. 2025; 25(9):2670. doi: 10.3390/s25092670

[24] Ma Z, Xiong J, Gong H, Wang X. Adaptive depth graph neural network-based dynamic task allocation for UAV-UGVs under complex environments. IEEE Trans Intell Veh. 2024; 10:3573-3586. doi: 10.1109/TIV.2024.3457493

[25] Knobelreiter P, Reinbacher C, Shekhovtsov A, Pock T. End-to-end training of hybrid CNN-CRF models for stereo. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition; 2017. p. 2339-2348. doi: 10.48550/arXiv.1611.10229

[26] Ibrahim HA, Ewida HF, Senosy AH, Ebraheem MO. 3D subsurface modeling for studying structural features and identifying potential hydrocarbon zones using well logging and seismic reflection data in the Al Baraka Oil Field, Komombo Basin, Upper Egypt. Pure Appl Geophys. 2022; 179(12):4465-4487. doi: 10.1007/s00024-022-03175-w

[27] Zhao T, Chen G, Suraphee S, Phoophiwfa T, Busababodhin P. A hybrid TCN-XGBoost model for agricultural product market price forecasting. PLoS One. 2025; 20(5):e0322496. doi: 10.1371/journal.pone.0322496

[28] Zhao T, Chen G, Pang C, Busababodhin P. Application and performance optimization of SLHS-TCN-XGBoost model in power demand forecasting. CMES Comput Model Eng Sci. 2025; 143(3):2883-2917. doi: 10.32604/cmes.2025.066442

[29] Zhao T, Chen G, Gatewongsa T, Busababodhin P. Forecasting agricultural trade based on TCN-LightGBM models: A data-driven decision. Res World Agric Econ. 2025; 6(1):207-221. doi: 10.36956/rwae.v6i1.1429

[30] Manzoor U, Ehsan M, Hussain M, Bashir Y. Improved reservoir characterization of thin beds by advanced deep learning approach. Appl Comput Geosci. 2024; 23:100188. doi: 10.1016/j.acags.2024.100188

[31] Bashir Y, Akdeniz DN, Balci D, et al. 3D geo-seismic data enhancement leveraging geophysical attributes for hydrocarbon prospect and geological illumination. Phys Chem Earth Parts ABC. 2025; 138:103854. doi: 10.1016/j.pce.2025.103854

[32] Bashir Y, Kemerli BD, Yılmaz T, Saral M, Göknar EC, Korkmaz E. Reconstruction of subsurface potential hydrocarbon reservoirs through 3D seismic automatic interpretation and attribute analysis. Phys Chem Earth Parts ABC. 2024; 136:103751. doi: 10.2139/ssrn.4932237