A Physics-Informed Neural Network (PINN) Framework for Predicting Stress-Dependent Permeability Evolution and Pressure Depletion in Compacting Reservoirs

Accurately modeling the coupled hydro-geomechanical behavior of compacting hydrocarbon reservoirs is critical for reliable production forecasting and subsurface management. Traditional numerical simulators, such as finite element and finite difference methods, often suffer from massive computational overheads and complex grid-generation bottlenecks when handling stress-dependent permeability evolution. Conversely, purely data-driven machine learning models frequently fail to respect fundamental physical laws, leading to unphysical predictions and poor extrapolation capabilities. To address these limitations, this paper proposes a physics-informed neural network (PINN) framework explicitly designed to simulate transient pore pressure depletion and stress-dependent permeability degradation in compacting reservoirs. By embedding mass conservation governing equations, exponential permeability constitutive relations, and boundary conditions directly into the neural network's multi-objective loss function, the proposed architecture acts as a continuous, mesh-free surrogate model. Utilizing automatic differentiation to evaluate exact spatial and temporal derivatives, the framework eliminates spatial discretization errors while maintaining strict physical consistency. Furthermore, a staged hybrid optimization strategy combining Adam and L-BFGS algorithms is implemented to navigate the complex optimization landscape. Benchmarking results demonstrate that the proposed PINN framework successfully achieves near-instantaneous surrogate evaluations and significant computational speedups compared to conventional simulators, offering a robust and efficient pathway for real-time reservoir management and subsurface monitoring.

Physics-Informed Neural Networks; Stress-Dependent Permeability; Hydro-Poromechanical Coupling; Mesh-Free Surrogate Modeling; Reservoir Simulation.

1. Wang, H., Ji, B., Lv, C., Zhang, L., Li, X., Cui, C., Yu, H., & Nie, J. (2019). The stress sensitivity of permeability in tight oil reservoirs. Energy Exploration and Exploitation, 37(4), 1364–1376. [Google Scholar] [Crossref]

2. https://doi.org/10.1177/0144598719855819 [Google Scholar] [Crossref]

3. Nguyen, V. H., Gland, N., Dautriat, J., David, C., Wassermann, J., & Guélard, J. (2014). Compaction, permeability evolution and stress path effects in unconsolidated sand and weakly consolidated sandstone. International Journal of Rock Mechanics and Mining Sciences, 67, 226–239. [Google Scholar] [Crossref]

4. https://doi.org/10.1016/j.ijrmms.2013.07.001 [Google Scholar] [Crossref]

5. Zhang, H., Liu, H., Luan, G., He, S., Gu, D., Mo, S., & Lei, G. (2014). A novel quantitative petrophysical model for the stress sensitivity of tight sandstones. Journal of Petroleum Science and Engineering, 122, 657–666. https://doi.org/10.1016/j.petrol.2014.09.008 [Google Scholar] [Crossref]

6. Han, G., & Dusseault, M. B. (2003). Description of fluid flow around a wellbore with stress-dependent porosity and permeability. Journal of Petroleum Science and Engineering, 40(1–2), 1–16. [Google Scholar] [Crossref]

7. https://doi.org/10.1016/S0920-4105(03)00047-0 [Google Scholar] [Crossref]

8. Jiang, G. C., Sheng, K. M., He, Y. B., Yang, L. L., Dong, T. F., Sun, Z., & Jiang, K. L. (2024). Numerical simulation of the temporal and spatial evolution of sandstone pore type reservoir damage types and severity. Scientific Reports, 14(1). https://doi.org/10.1038/s41598-024-76383-0 [Google Scholar] [Crossref]

9. Akhyani, S., Shahbodagh, B., & Khalili, N. (2026). Fully coupled physics-informed neural networks for hydro-mechanical analysis of saturated poroelastic media. Computers and Geotechnics, 193. [Google Scholar] [Crossref]

10. https://doi.org/10.1016/j.compgeo.2026.107954 [Google Scholar] [Crossref]

11. Cai, E., Gao, Z., Xu, Z., Zuo, S., Wen, M., & Zhang, Y. (2026). A Physics-Informed Deep Learning Framework for Multi-Field Hydro-Mechanical Consolidation Analysis. International Journal for Numerical and Analytical Methods in Geomechanics, 50(7), 3000–3028. [Google Scholar] [Crossref]

12. https://doi.org/10.1002/nag.70260 [Google Scholar] [Crossref]

13. Han, J. X., Xue, L., Wei, Y. S., Qi, Y. D., Wang, J. L., Liu, Y. T., & Zhang, Y. Q. (2023). Physics-informed neural network-based petroleum reservoir simulation with sparse data using domain decomposition. Petroleum Science, 20(6), 3450–3460. https://doi.org/10.1016/j.petsci.2023.10.019 [Google Scholar] [Crossref]

14. Chen, H. Y., Xue, L., Liu, L., Zou, G. F., Han, J. X., Dong, Y. Bin, Cong, M. Z., Liu, Y. T., & Hosseini-Nasab, S. M. (2025). Physics-informed graph neural network for predicting fluid flow in porous media. Petroleum Science, 22(10), 4240–4253. https://doi.org/10.1016/j.petsci.2025.06.007 [Google Scholar] [Crossref]

15. Badawi, D., & Gildin, E. (2023). Physics-Informed Neural Network for the Transient Diffusivity Equation in Reservoir Engineering. http://arxiv.org/abs/2309.17345 [Google Scholar] [Crossref]

16. Zhang, Y., Chen, X., Li, K., & Zou, Y. (2026). A Physics-Informed Neural Network Model for Reservoir Seepage in Porous Media Based on Darcy’s Law. Processes, 14(10). https://doi.org/10.3390/pr14101578 [Google Scholar] [Crossref]

17. Almajid, M. M., & Abu-Al-Saud, M. O. (2022). Prediction of porous media fluid flow using physics informed neural networks. Journal of Petroleum Science and Engineering, 208. [Google Scholar] [Crossref]

18. https://doi.org/10.1016/j.petrol.2021.109205 [Google Scholar] [Crossref]

19. Millevoi, C., Spiezia, N., & Ferronato, M. (2024). On Physics-Informed Neural Networks training for coupled hydro-poromechanical problems. Journal of Computational Physics, 516. [Google Scholar] [Crossref]

20. https://doi.org/10.1016/j.jcp.2024.113299 [Google Scholar] [Crossref]

21. Ur Rashid, H., Pachalieva, A., & O’Malley, D. (2026). Differentiable multiphase flow model for physics-informed machine learning in reservoir pressure management. Scientific Reports, 16(1). [Google Scholar] [Crossref]

22. https://doi.org/10.1038/s41598-026-37063-3 [Google Scholar] [Crossref]

23. Haghighat, E., Amini, D., & Juanes, R. (2022). Physics-informed neural network simulation of multiphase poroelasticity using stress-split sequential training. Computer Methods in Applied Mechanics and Engineering, 397. https://doi.org/10.1016/j.cma.2022.115141 [Google Scholar] [Crossref]

24. Wahyudi, A., Ariadji, T., Marhaendrajana, T., Sidarto, K. A., & Syihab, Z. (2025). A Fully Implicit Reservoir Simulation Using Physics Informed Neural Networks. Scientific Contributions Oil and Gas, 48(3), 131–147. https://doi.org/10.29017/scog.v48i3.1858 [Google Scholar] [Crossref]

25. Pang, M., Zhang, T., Gao, L., & Qin, B. (2021). Investigating the effects of effective stress on pore-dependent permeability measurements of crushed coal. PLoS ONE, 16(12 December). [Google Scholar] [Crossref]

26. https://doi.org/10.1371/journal.pone.0261678 [Google Scholar] [Crossref]

27. An, C., Killough, J., & Xia, X. (2021). Investigating the effects of stress creep and effective stress coefficient on stress-dependent permeability measurements of shale rock. Journal of Petroleum Science and Engineering, 198. https://doi.org/10.1016/j.petrol.2020.108155 [Google Scholar] [Crossref]

28. Li, B., Zhou, F., Fan, W., Ren, D., Li, H., & Li, M. (2020). Experimental investigation and theoretical modeling of stress-dependent permeability in naturally fractured tight gas reservoir. Journal of Petroleum Science and Engineering, 188. https://doi.org/10.1016/j.petrol.2020.106949 [Google Scholar] [Crossref]

29. Dautriat, J., Gland, N., Youssef, S., Rosenberg, E., Bekri, S., & Vizika, O. (n.d.). Stress-Dependent Directional Permeabilities of Two Analog Reservoir Rocks: A Prospective Study on Contribution of m-Tomography and Pore Network Models. [Google Scholar] [Crossref]

30. Suri, P., Azeemuddin, M., Zaman, M., Kukreti, A. R., & Roegiers, J.-C. (1997). ENGINEERING Stress-dependent permeability measurement using the oscillating pulse technique. In Journal of Petroleum Science and Engineering (Vol. 17). [Google Scholar] [Crossref]

31. Lin, R., Ren, L., Zhao, J., Tan, X., Rasouli, V., Wang, X., Wu, J., Song, Y., & Shen, C. (2022). Stress and pressure dependent permeability of shale rock: Discrete element method (DEM) simulation on digital core. Journal of Petroleum Science and Engineering, 208. https://doi.org/10.1016/j.petrol.2021.109797 [Google Scholar] [Crossref]

32. Cao, N., Lei, G., Dong, P., Li, H., Wu, Z., & Li, Y. (2019). Stress-Dependent permeability of fractures in tight reservoirs. Energies, 12(1). https://doi.org/10.3390/en12010117 [Google Scholar] [Crossref]

33. Feng, Y., Tang, H., Tang, H., Leng, Y., Shi, X., Liu, J., Wang, Z., & Deng, C. (2023). Influence of geomechanics parameters on stress sensitivity in fractured reservoir. Frontiers in Earth Science, 11. https://doi.org/10.3389/feart.2023.1134260 [Google Scholar] [Crossref]

34. Ali, T. A., & Sheng, J. J. (2015). SPE-177299-MS Evaluation Of The Effect Of Stress-dependent Permeability On Production Performance in Shale Gas Reservoirs. [Google Scholar] [Crossref]

35. An, C., Killough, J., & Mi, L. (2019). Stress-dependent permeability of organic-rich shale reservoirs: Impacts of stress changes and matrix shrinkage. Journal of Petroleum Science and Engineering, 172, 1034–1047. https://doi.org/10.1016/j.petrol.2018.09.011 [Google Scholar] [Crossref]

36. Li, Y., Shen, X., Li, D., Liu, Y., Huang, W., Qiu, J., Yan, X., & Zhang, K. (2026). Two-phase flow simulation in low-permeability heterogeneous reservoirs with a fully implicit scheme-based enriched physics-informed neural network. Journal of Petroleum Exploration and Production Technology, 16(11). [Google Scholar] [Crossref]

37. https://doi.org/10.1007/s13202-026-02216-7 [Google Scholar] [Crossref]

38. Fuks, O., & Tchelepi, H. A. (2020). LIMITATIONS OF PHYSICS INFORMED MACHINE LEARNING FOR NONLINEAR TWO-PHASE TRANSPORT IN POROUS MEDIA. In Journal of Machine Learning for Modeling and Computing (Vol. 1, Number 1). www.begellhouse.com [Google Scholar] [Crossref]

39. Yuan, R., Huang, X. L., Fang, Y., Kasama, K., Cheng, Y., & He, Y. (2026). A physics-informed machine learning computational framework for solving Mohr-Coulomb plasticity in geomechanics. Advances in Engineering Software, 212. https://doi.org/10.1016/j.advengsoft.2025.104061 [Google Scholar] [Crossref]

40. Xing, H., Yao, G., Yuan, H., Zhao, J., & Wen, D. (2025). Physics-Informed Neural Networks for Solving Parameterized Dual-Domain Darcy–Brinkman Flows in Gradient Porous Mediums. International Journal for Numerical Methods in Engineering, 126(16). https://doi.org/10.1002/nme.70110 [Google Scholar] [Crossref]

41. Cuomo, S., Di Cola, V. S., Giampaolo, F., Rozza, G., Raissi, M., & Piccialli, F. (2022). Scientific Machine Learning Through Physics–Informed Neural Networks: Where we are and What’s Next. Journal of Scientific Computing, 92. https://doi.org/10.1007/s10915-022-01939-z [Google Scholar] [Crossref]

42. Meng, X., Liu, X., Zhang, H., Guo, X., Wang, Y., & Jeng, D. S. (2026). A physics-informed neural network framework for wave-induced dynamic response in a poroelastic seabed. Computers and Geotechnics, 198. https://doi.org/10.1016/j.compgeo.2026.108250 [Google Scholar] [Crossref]

43. Tartakovsky, A. M., Marrero, C. O., Perdikaris, P., Tartakovsky, G. D., & Barajas-Solano, D. (2020). Physics-Informed Deep Neural Networks for Learning Parameters and Constitutive Relationships in Subsurface Flow Problems. Water Resources Research, 56(5). https://doi.org/10.1029/2019WR026731 [Google Scholar] [Crossref]

44. Ji, Y., Peng, Y., Wang, X., Chen, Z., & Wu, K. (2026). Physics-Informed Neural Networks for Near-Wellbore Stress Field Prediction with Enhanced Generalization. Processes, 14(14). [Google Scholar] [Crossref]

45. https://doi.org/10.3390/pr14142254 [Google Scholar] [Crossref]

46. Wang, S., Zhang, S., Xiao, T., Ren, W., Zhang, L., & Tian, Y. (2026). Physics-informed neural networks for predicting the coupled hydro-mechanical behaviour of partially saturated soils. Georisk. https://doi.org/10.1080/17499518.2026.2643668 [Google Scholar] [Crossref]

47. Bodong, Z., Maojin, T., Guohua, Z., & Chenglin, L. (2026). An Integrating Physics-Informed Neural Networks and Attention Mechanism based Intelligent Permeability Prediction Method for Shale Reservoirs. In SPE Journal. [Google Scholar] [Crossref]

48. Xiong, S., Liu, S., Weng, D., Shen, R., Yu, J., Yan, X., He, Y., & Chu, S. (2022). A Fractional Step Method to Solve Productivity Model of Horizontal Wells Based on Heterogeneous Structure of Fracture Network. Energies, 15(11). https://doi.org/10.3390/en15113907 [Google Scholar] [Crossref]

49. Ishitsuka, K., & Lin, W. (2023). Physics-informed neural network for inverse modeling of natural-state geothermal systems. Applied Energy, 337. https://doi.org/10.1016/j.apenergy.2023.120855 [Google Scholar] [Crossref]

50. Zhang, Z. (2022). A physics-informed deep convolutional neural network for simulating and predicting transient Darcy flows in heterogeneous reservoirs without labeled data. Journal of Petroleum Science and Engineering, 211. https://doi.org/10.1016/j.petrol.2022.110179 [Google Scholar] [Crossref]

51. Aziz, Md. A., Strauss, T., Mohebujjaman, M., & Khan, T. (2025). Self-adaptive physics-informed neural network for forward and inverse problems in heterogeneous porous flow. http://arxiv.org/abs/2512.14610 [Google Scholar] [Crossref]

52. Kalule, R., Abderrahmane, H. A., Ahmed, S., Alameri, W., & Al-Shalabi, E. W. (2026). Relative permeability and capillary pressure estimation via physics-informed machine learning and reinforcement learning. Petroleum, 12(3), 430–443. https://doi.org/10.1016/j.petlm.2026.04.005 [Google Scholar] [Crossref]

53. Liu, Y., & Arson, C. (2026). A Physics-Informed Neural Network for Pressurized Cavities of Arbitrary Shape in Heterogeneous Rock. Rock Mechanics and Rock Engineering. https://doi.org/10.1007/s00603-026-05882-5 [Google Scholar] [Crossref]

54. Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2017). Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations. http://arxiv.org/abs/1711.10561 [Google Scholar] [Crossref]

55. Cai, M., Gu, H., Hong, P., & Li, J. (2023). A combination of physics-informed neural networks with the fixed-stress splitting iteration for solving Biot’s model. Frontiers in Applied Mathematics and Statistics, 9. https://doi.org/10.3389/fams.2023.1206500 [Google Scholar] [Crossref]

56. Samarkanov, O. L., Riazi, M., Irawan, S., & Kamrava, S. (2026). Applications of physics-informed neural networks in fluid flow in porous media - A review study. In Energy and AI (Vol. 24). Elsevier B.V. https://doi.org/10.1016/j.egyai.2026.100747 [Google Scholar] [Crossref]

57. Han, J., Xue, L., Jia, Y., Mwasamwasa, M. S., Nanguka, F., Sangweni, C., Liu, H., & Li, Q. (2024). Prediction of Porous Media Fluid Flow with Spatial Heterogeneity Using Criss-Cross Physics-Informed Convolutional Neural Networks. CMES - Computer Modeling in Engineering and Sciences, 138(2), 1323–1340. https://doi.org/10.32604/cmes.2023.031093 [Google Scholar] [Crossref]

58. Pachalieva, A., O’Malley, D., Harp, D. R., & Viswanathan, H. (2022). Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management. Scientific Reports, 12(1). https://doi.org/10.1038/s41598-022-22832-7 [Google Scholar] [Crossref]

• An Efficient Multi-Cluster Wireless Sensor Network Based on Optimal Cluster Head Selection Using Particle Swarm Optimization
• Enhancing Employee Psychological Well-Being of Tourism SME Workers through Cross-Cultural Knowledge Sharing in Ubud, Bali
• A Theoretical Framework for Supply Chain Resilience: Integrating Complexity and Network Theories to Understand Adaptive Capabilities