A novel approach to modeling breakdown pressure dynamics using machine learning

Abstract

This study presents a novel machine learning (ML)-based approach for predicting breakdown pressure (BP) in hydraulic fracturing using experimental data. Unlike traditional analytical models that rely on simplified assumptions, ML models can capture complex nonlinear relationships between BP and its influencing factors. However, a key limitation in BP prediction stems from dataset constraints, particularly the scale differences between experimental setups and real-world formations. To mitigate these limitations, this research utilizes a unique dataset of 144 BP data points, incorporating various rock mechanical properties, injection parameters, and fluid properties. Additionally, a separate analysis of pressurization rate, based on 32 additional experimental data points, was conducted to better understand its effect on fracture initiation — an aspect often overlooked in ML-based studies. The dataset includes critical parameters such as injection rate, confining pressure, tensile strength, Young's modulus, permeability, unconfined compressive strength, Poisson's ratio, porosity, wellbore radius, and fracture geometry ratio. Five ML models — LightGBM, CatBoost, XGBoost, Kolmogorov-Arnold Network (KAN), and TabNet — were trained and evaluated. TabNet achieved the highest predictive performance (R² = 0.94) due to its attention-based feature selection and deep-learning-based representation learning. Model performance was assessed using mean absolute error (MAE) and mean squared error (MSE) to ensure robustness. To further enhance model interpretability, SHapley Additive exPlanations (SHAP) and TabNet's attention mechanism were used to explicitly assess feature importance, providing insights into the relative influence of different parameters on BP predictions. Additionally, advanced feature-handling techniques were employed to address categorical variables automatically, ensuring minimal preprocessing bias. The findings demonstrate the scalability of ML models for BP prediction using experimental data, reducing reliance on costly and time-consuming laboratory testing. By incorporating advanced interpretability techniques, systematic pressurization rate analysis, and robust ML architectures, this research provides a more accurate, data-driven approach for optimizing hydraulic fracturing designs.

Keywords

AI / Machine learning / Hydraulic fracturing / Breakdown pressure / Geomechanics

CRediT authorship contribution statement

Subhan Aliyev: Writing - Original Draft, Methodology, Visualization, Software, Formal analysis, Data curation. Talal Al Shafloot: Supervision, Conceptualization, Writing - Review & Editing, Methodology. Murtada Saleh Aljawad: Supervision, Writing - Review & Editing, Validation. Abdulazeez Abdulraheem: Conceptualization, Methodology, Validation. Salaheldin Elkatatny: Supervision, Methodology, Writing - Review & Editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

The authors would like to acknowledge the support of the College of Petroleum Engineering & Geosciences at King Fahd University of Petroleum & Minerals.

References

[1] J.C. Glorioso, A. Rattia, Unconventional reservoirs: basic petrophysical concepts for shale gas,in:SPE/EAGE European Unconventional Resources Conference and Exhibition, Spe, 2012. SPE-153004.

[2] Q. Li, H. Xing, J. Liu, A review on hydraulic fracturing of unconventional reservoir, Petroleum 1 (1) (2015) 8-15.

[3] J. Guo, Y. Liu, A comprehensive model for simulating fracturing fluid leakoff in natural fractures, J. Nat. Gas Sci. Eng. 21 (2014) 977-985.

[4] G.A. Al-Muntasheri, A critical review of hydraulic-fracturing fluids for moderate-to ultralow-permeability formations over the last decade, SPE Prod. Oper. 29 (4) (2014) 243-260.

[5] X. Zhao, Mechanical behavior of rocks under complex stresses and high temperatures: effects on breakdown pressure, Int. J. Rock Mech. Min. Sci. 86 (2016) 90-102.

[6] E.K. Watkins, C.L. Wendorff, B.R. Ainley,A new crosslinked foamed fracturing fluid, in: SPE Annual Technical Conference and Exhibition?, SPE, 1983. SPE-12027.

[7] S. Al-Hajri, B.M. Negash, M.M. Rahman, M. Haroun, T.M. Al-Shami, Perspective Review of polymers as additives in water-based fracturing fluids, ACS Omega 7 (9) (2022) 7431-7443.

[8] M.J. Almarri, M.J. AlTammar, K.M. Alruwaili, S. Zheng, Numerical feasibility of near-wellbore cooling as a novel method for reducing breakdown pressure in hydraulic fracturing, J. Nat. Gas Sci. Eng. 102 (2022) 104549.

[9] B. S, B.M. Aadnoy, Elasto-thermo-poroelastic model for stress, pore pressure and temperature around a wellbore, J. Petrol. Sci. Eng. 60 (1) (2008) 1-13.

[10] I. Berchenko, E. Detournay, Deviation of hydraulic fractures through poroelastic stress changes induced by fluid injection and pumping, Int. J. Rock Mech. Min. Sci. 34 (6) (1997) 1009-1019.

[11] H. Yang, et al., Breakdown pressure prediction of tight sandstone horizontal wells based on the mechanism model and multiple linear regression model, Energies 15 (19) (2022) 6944.

[12] J. Zhang, Pore pressure prediction from well logs: methods, modifications, and new approaches, Earth Sci. Rev. 108 (1-2) (2011) 50-63.

[13] W. Zhu, Z. Chen, X. He, Z. Tian, M. Wang, Numerical investigation of influential factors in hydraulic fracturing processes using coupled discrete element-lattice Boltzmann method, J. Geophys. Res. Solid Earth 128 (9) (2023) e2023JB027292.

[14] L. Huang, X. Liao, M. Fan, S. Wu, P. Tan, L. Yang, Experimental and numerical simulation technique for hydraulic fracturing of shale formations, Adv. Geo-Energy Res. 13 (2) (2024) 83-88.

[15] R. Carbonell, E. Detournay, Modeling fracture initiation and propagation from a pressurized hole: a dislocation-based approach,in:ARMA US Rock Mechanics/Geomechanics Symposium, ARMA, 1995. ARMA-95.

[16] F. Guo, N.R. Morgenstern, J.D. Scott, Interpretation of hydraulic fracturing breakdown pressure, in: International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, Elsevier, 1993, pp. 617-626.

[17] N. Morita, A.D. Black, G.-F. Fuh, Borehole breakdown pressure with drilling fluids — I. Empirical results, in: International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, Elsevier, 1996, pp. 39-51.

[18] Y. He, et al., Investigation of low water recovery based on gas-water two-phase low-velocity Non-Darcy flow model for hydraulically fractured horizontal wells in shale, Petroleum 9 (3) (2023) 364-372.

[19] H. Peng, et al., Influence of supercritical CO₂ on the physical property of tight sandstone, Petroleum 10 (3) (2024) 520-526.

[20] T. Ishida, et al., Acoustic emission monitoring of hydraulic fracturing laboratory experiment with supercritical and liquid CO₂, Geophys. Res. Lett. 39 (16) (2012).

[21] Y. Gong, M. Mehana, I. El-Monier, F. Xu, F. Xiong,Machine learning for estimating rock mechanical properties beyond traditional considerations, in: Unconventional Resources Technology Conference, Unconventional Resources Technology Conference (URTeC); Society of, Denver, Colorado, 2019, pp. 466-480, 22-24 July 2019.

[22] T. Almani, K. Khan,Predicting Formation breakdown pressures for arbitrary wellbore orientation using machine learning techniques, in: ARMA/DGS/SEG International Geomechanics Symposium, ARMA, 2023 (ARMA-IGS).

[23] X. Xuesong, Y. Junliang, L.I. Zhonghui, S.U.N. Chong, Z. Yi, Determination of formation fracture pressure under high temperature and high pressure in deep water of the South China Sea, Petrol. Drill. Tech. 51 (6) (2023) 18-24.

[24] M.K. Hubbert, D.G. Willis, J. Petrol. Mechanics of hydraulic fracturing, Technol. 9 (6) (1957) 153-168, 1957.

[25] B. Haimson, C. Fairhurst, Initiation and extension of hydraulic fractures in rocks, Soc. Petrol. Eng. J. 7 (3) (1967) 310-318.

[26] X. Jin, S.N. Shah, J.-C. Roegiers, B. Hou,Breakdown pressure determination-a fracture mechanics approach, in: SPE Annual Technical Conference and Exhibition?, SPE, 2013 D011S006R006.

[27] N. Li, H. Xie, Z. Gao, C. Li, Study on the hydraulic fracturing failure behaviour of granite and its comparison with gas fracturing, Sustainability 14 (21) (2022) 14593.

[28] D. Lockner, J.D. Byerlee, Hydrofracture in Weber sandstone at high confining pressure and differential stress, J. Geophys. Res. 82 (14) (1977) 2018-2026.

[29] Y. Wu, J. Tao, J. Wang, Y. Zhang, S. Peng, Experimental investigation of shale breakdown pressure under liquid nitrogen pre-conditioning before nitrogen fracturing, Int. J. Min. Sci. Technol. 31 (4) (2021) 611-620.

[30] E. Fjar, R.M. Holt, A.M. Raaen, R. Risnes, P. Horsrud, Petroleum Related Rock Mechanics, Elsevier, 2008.

[31] C.A.I. Zhenzhong, et al., Geomechanics modeling of ultra-deep fault-controlled carbonate reservoirs and its application in development, Petrol. Geol. Exp. 46 (4) (2024) 868-879.

[32] D. Zhou, P. He,Major factors affecting simultaneous frac results, in: SPE Oklahoma City Oil and Gas Symposium/Production and Operations Symposium, SPE, 2015. SPE-173633.

[33] T. Almani, Y. Alquhaidan, K. Khan,Sensitivity analysis of the factors affecting formation breakdown pressure using a computational-based approach, in: ARMA/DGS/SEG International Geomechanics Symposium, ARMA, 2023 (ARMA-IGS).

[34] K.E. Gray, C.M. Kim, B.J. Hughes, Stress state, porosity,permeability and breakdown pressure around A borehole during fluid injection, in: ARMA US Rock Mechanics/Geomechanics Symposium, ARMA, 1981. ARMA-81.

[35] T. Ito, Effect of pore pressure gradient on fracture initiation in fluid saturated porous media: rock, Eng. Fract. Mech. 75 (7) (2008) 1753-1762.

[36] A.P. Bunger, A. Lakirouhani, E. Detournay,Modelling the effect of injection system compressibility and viscous fluid flow on hydraulic fracture breakdown pressure, in: ISRM International Symposium on In-Situ Rock Stress, ISRM, 2010 (ISRM-ISRS).

[37] M.J. AlTammar, D.P. Gala, M.M. Sharma,Application of different fluid injection methods to reduce breakdown pressure, in: International Petroleum Technology Conference, IPTC, 2020 D031S077R002.

[38] M. Kong, Z. Zhang, C. Zhao, H. Chen, X. Ma, Y. Zou, Rock mechanical properties and breakdown pressure of high-temperature and high-pressure reservoirs in the Southern Margin of Junggar Basin, Geofluids 2021 (1) (2021) 1116136.

[39] M. Bai, Why are brittleness and fracability not equivalent in designing hydraulic fracturing in tight shale gas reservoirs, Petroleum 2 (1) (2016) 1-19.

[40] Y. Zhang, J. Zhang, B. Yuan, S. Yin, In-situ stresses controlling hydraulic fracture propagation and fracture breakdown pressure, J. Pet. Sci. Eng. 164 (2018) 164-173.

[41] T. Ito, K. Hayashi, Physical background to the breakdown pressure in hydraulic fracturing tectonic stress measurements, in: International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, Elsevier, 1991, pp. 285-293.

[42] S. Wang, Y. Han, W. Hu, X. Zhao, L. Zhang, T. Wang, Modeling interactions between hydraulic fracture and pre-existing microcracks in crystalline rocks using hydro-grain-texture model, Geoenergy Sci. Eng. 244 (2025) 213459, https://doi.org/10.1016/j.geoen.2024.213459.

[43] C. Jiang, Y. Lei, B. Deng, X. Fan, M. Li, M. Wu, Effect of mesoscopic heterogeneity on hydraulic fracturing characteristics of granite under high temperature, Geoenergy Sci. Eng. 246 (2025) 213626, https://doi.org/10.1016/j.geoen.2024.213626.

[44] Wei Haifeng, Research progress on fracture propagation patterns of hydraulic fracturing in heterogeneous shale, Petrol. Geol. Rec. Eff. 30 (4) (2023) 156-166.

[45] F. Maojun, D.U. Xulin, B.A.I. Yuhu, L.I. Hao, Z. Hao, Z.H.U. Haiyan, Three-dimensional physical simulation experiments on large-scale hydraulic fracturing in multi-thin interbedded tight sandstone reservoirs, Petrol. Geol. Exp. 46 (4) (2024) 786-798.

[46] R.M. Tuzingila, L. Kong, R.K. Kasongo, A review on experimental techniques and their applications in the effects of mineral content on geomechanical properties of reservoir shale rock, Rock Mech. Bull. (2024) 100110.

[47] J. Wang, D. Elsworth, Y. Wu, J. Liu, W. Zhu, Y. Liu, The influence of fracturing fluids on fracturing processes: a comparison between water, oil and SC-CO₂, Rock Mech. Rock Eng. 51 (2018) 299-313.

[48] Z. Tariq, et al., Machine learning-based accelerated approaches to infer breakdown pressure of several unconventional rock types, ACS Omega 7 (45) (2022) 41314-41330.

[49] R. Roelofs, et al., A meta-analysis of overfitting in machine learning, Adv. Neural Inf. Process. Syst. 32 (2019).

[50] K. Ghosh, C. Bellinger, R. Corizzo, P. Branco, B. Krawczyk, N. Japkowicz, The class imbalance problem in deep learning, Mach. Learn. 113 (7) (2024) 4845-4901.

[51] N.U. Niaz, K.M.N. Shahariar, M.J.A. Patwary, Class imbalance problems in machine learning: a review of methods and future challenges,in: Proceedings of the 2nd International Conference on Computing Advancements, 2022, pp. 485-490.

[52] A. Muqtadir, S.M. Elkatatny, M.A. Mahmoud, A. Abdulraheem, A. Gomaa,Effect of saturating fluid on the geomechanical properties of low permeability scioto sandstone rocks, in: ARMA US Rock Mechanics/Geomechanics Symposium, ARMA, 2018. ARMA-2018.

[53] A. Farid Ibrahim, H. Nasr-El-Din, Evaluation of the breakdown pressure to initiate hydraulic fractures of tight sandstone and shale formations, in: SPE Trinidad and Tobago Section Energy Resources Conference?, SPE, 2018 D011S020R002.

[54] A. Mathur, et al., Crude steady-state permeability measurements on carbonate source rocks and effect of water block from fracturing fluid on productivity,in: International Petroleum Technology Conference, IPTC, 2024. IPTC-24536.

[55] Y. Zou, N. Li, X. Ma, S. Zhang, S. Li, Experimental study on the growth behavior of supercritical CO₂-induced fractures in a layered tight sandstone formation, J. Nat. Gas Sci. Eng. 49 (2018) 145-156.

[56] D.T. Nicholson, Pore properties as indicators of breakdown mechanisms in experimentally weathered limestones, Earth Surf. Process. Landf.: J. British Geomorphol. Res. Group 26 (8) (2001) 819-838.

[57] V. Marcon, et al., Experimental insights into geochemical changes in hydraulically fractured Marcellus Shale, Appl. Geochem. 76 (2017) 36-50.

[58] A.F. Siggins, Velocity-effective stress response of CO₂-saturated sandstones, Explor. Geophys. 37 (1) (2006) 60-66.

[59] S. Chatterji, P. Christensen, A mechanism of breakdown of limestone nodules in a freeze-thaw environment, Cement Concr. Res. 9 (6) (1979) 741-746.

[60] B. Haimson, C. Fairhurst, Hydraulic fracturing in porous-permeable materials, J. Petrol. Technol. 21 (7) (1969) 811-817.

[61] I. Song, M. Suh, K.S. Won, B. Haimson, A laboratory study of hydraulic fracturing breakdown pressure in tablerock sandstone, Geosci. J. 5 (2001) 263-271.

[62] F. Wu, X. Fan, J. Liu, X. Li, Analytical interpretation of hydraulic fracturing initiation pressure and breakdown pressure, J. Nat. Gas Sci. Eng. 76 (2020) 103185.

[63] S. Mishra, A. Datta-Gupta, Applied Statistical Modeling and Data Analytics: A Practical Guide for the Petroleum Geosciences, Elsevier, 2017.

[64] G. Ke, et al., Lightgbm: a highly efficient gradient boosting decision tree, Adv. Neural Inf. Process. Syst. 30 (2017).

[65] T. Yang, et al., Quantitative classification and prediction of pore structure in low porosity and low permeability sandstone: a machine learning approach, Geoenergy Sci. Eng. (2025) 213708, https://doi.org/10.1016/j.geoen.2025.213708.

[66] A. Anghel, N. Papandreou, T. Parnell, A. De Palma, H. Pozidis, Benchmarking and Optimization of Gradient Boosting Decision Tree Algorithms, 2018 arXiv preprint arXiv: 1809.04559.

[67] A. Choudhury, A. Mondal, S. Sarkar, Searches for the BSM scenarios at the LHC using decision tree-based machine learning algorithms: a comparative study and review of random forest, AdaBoost, XGBoost and LightGBM frameworks, Eur. Phys. J. Spec. Top. (2024) 1-39.

[68] L. Prokhorenkova, G. Gusev, A. Vorobev, A.V. Dorogush, A. Gulin, CatBoost: unbiased boosting with categorical features, Adv. Neural Inf. Process. Syst. 31 (2018).

[69] T. Chen, C. Guestrin, Xgboost: a scalable tree boosting system,in: Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, 2016, pp. 785-794.

[70] Y. Meng, N. Yang, Z. Qian, G. Zhang, What makes an online review more helpful: an interpretation framework using XGBoost and SHAP values, J. Theoret. Appl. Elect. Com. Res. 16 (3) (2020) 466-490.

[71] S. Ö O. Arik, T. Pfister, Tabnet: attentive interpretable tabular learning,in: Proceedings of the AAAI Conference on Artificial Intelligence, 2021, pp. 6679-6687.

[72] Z. Liu, et al., Kan: Kolmogorov-Arnold Networks, 2024 arXiv preprint arXiv:2404.19756.

[73] Y. Hou, D. Zhang, A Comprehensive Survey on Kolmogorov Arnold Networks (Kan), 2024 arXiv preprint arXiv:2407.11075.

[74] X. Wu, et al., Experimental study on cyclic hydraulic fracturing of tight sandstone under in-situ stress, Processes 11 (3) (2023) 875.

[75] S. Jung, M.B. Diaz, K.Y. Kim, H. Hofmann, G. Zimmermann, Fatigue behavior of granite subjected to cyclic hydraulic fracturing and observations on pressure for fracture growth, Rock Mech. Rock Eng. (2021) 1-14.

[76] Z. Guo, S. Tian, Q. Liu, L. Ma, Y. Yong, R. Yang, Experimental investigation on the breakdown pressure and fracture propagation of radial borehole fracturing, J. Pet. Sci. Eng. 208 (2022) 109169.

[77] R.D. Barree, J.L. Miskimins,Calculation and implications of breakdown pressures in directional wellbore stimulation, in: SPE Hydraulic Fracturing Technology Conference and Exhibition, SPE, 2015 D011S001R007.

[78] D. Garagash, E. Detournay, An analysis of the influence of the pressurization rate on the borehole breakdown pressure, Int. J. Solid Struct. 34 (24) (1997) 3099-3118.

[79] B. Lecampion, Hydraulic fracture initiation from an open-hole: wellbore size, pressurization rate and fluid-solid coupling effects, in: ARMA US Rock Mechanics/Geomechanics Symposium, ARMA, 2012. ARMA-2012.