A Physics-Informed Neural Network Approach for Constitutive Modeling of Oil-Well Cement Slurries Under Cyclic Loading
Abstract
Deep learning (DL) based on Artificial Neural Networks (ANN) demonstrates robust performance, strong nonlinear mapping capabilities, and powerful self-learning capacities, enabling its widespread application in learning and predicting stress-strain constitutive relationships under quasi-static loading paths (e.g., uniaxial/triaxial compression tests). However, significant gaps remain in applying ANN methods to learn, characterize, or predict constitutive relationships for geotechnical materials under complex multi-cycle dynamic loading paths. This limitation primarily arises from the nonlinearity, variability, and complexity inherent in cyclic stress-strain hysteresis loops. This study systematically investigates oil-well cement sheaths subjected to high-intensity multi-frequency cyclic compression. Samples were set and cured at four downhole temperatures (25, 90, 115, and 140 °C), then tested under four constant-amplitude loading levels (30%, 50%, 70%, and 90%). A comprehensive analysis of 480 hysteresis loops from 16 sample groups was conducted, evaluating accumulated plastic strain, dissipation energy proportion, and the proportion of plastic damage energy. The analysis reveals that conventional fatigue life prediction models fail to effectively capture the complex evolutionary characteristics of hysteresis loops. Subsequently, the cyclic stress-strain constitutive relationship was treated as time-series data. Through preprocessing and cycle-by-cycle segmentation of mechanical test data, single-step rolling prediction of hysteresis loops was achieved. An LSTM (The Long Short-Term Memory) architecture was developed to address long-term dependencies and complex nonlinear features in different hysteresis loops. By constructing a Physics-Informed Neural Network (PINN) that integrates physical laws and data patterns, the physics-guided learning capability of the improved model was enhanced. Recursive prediction with additional physical constraints enabled full-process continuous rolling prediction, demonstrating superior predictive performance. The proposed methodology provides novel perspectives for advancing intelligent inversion methods in geotechnical engineering applications under dynamic loading conditions.
DOI:
https://doi.org/10.46690/gs.2026.01.04Keywords:
Oil-well cement, hysteresis loop evolution, cyclic constitutive modeling, physics-informed neural networks (PINN), cyclic loading pathReferences
Banimahd M, Yasrobi SS, Woodward PK. 2005. Artificial neural network for stress–strain behavior of sandy soils: Knowledge based verification. Computers and Geotechnics, 32:377–386. https://doi.org/10.1016/j.compgeo.2005.06.002.
Basheer IA. 2002. Stress-strain behavior of geomaterials in loading reversal simulated by time-delay neural networks. Journal of Materials in Civil Engineering, 14:270–273. https://doi.org/10.1061/(ASCE)0899-1561(2002)14:3(270).
Basheer IA. 2000. Selection of Methodology for Neural Network Modeling of Constitutive Hystereses Behavior of Soils. Computer-Aided Civil and Infrastructure Engineering, 15:445–463. https://doi.org/10.1111/0885-9507.00206.
Beaudoin JJ, Feldman RF. 1985. High-strength cement pastes—A critical appraisal. Cement and Concrete Research, 15:105–116. https://doi.org/10.1016/0008-8846(85)90015-8.
Carleo G, Cirac I, Cranmer K, et al. 2019. Machine learning and the physical sciences. Reviews of Modern Physics, 91:045002. https://doi.org/10.1103/RevModPhys.91.045002.
Chen BR, Feng XT, Ding WX, et al. 2004. Evolutionary neural network constitutive model for complete stress-strain relationship of rock under chemical corrosion. Journal of Northeastern University (Science), 25(7):695-698.
Chung J, Gulcehre C, Cho K, et al. 2014. Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555. https://doi.org/10.48550/arXiv.1412.3555.
Deng K, Yuan Y, Hao Y, et al. 2020. Experimental study on the integrity of casing-cement sheath in shale gas wells under pressure and temperature cycle loading. Journal of Petroleum Science and Engineering, 195:107548. https://doi.org/10.1016/j.petrol.2020.107548.
Diaz MB, Kim KY, Jung SG. 2020. Effect of frequency during cyclic hydraulic fracturing and the process of fracture development in laboratory experiments. International Journal of Rock Mechanics and Mining Sciences, 134:104474. https://doi.org/10.1016/j.ijrmms.2020.104474.
Eilers LH, Nelson EB, Moran LK. 1983. High-Temperature Cement Compositions - Pectolite, Scawtite, Truscottite, or Xonotlite: Which Do You Want? Journal of Petroleum Technology, 35:1373–1377. https://doi.org/10.2118/9286-PA.
Ellis G W, Yao C, Zhao R, Penumadu D. 1995. Stress strain modeling of sands using artificial neural networks. Journal of Geotechnical Engineering, 121:429–435. https://doi.org/10.1061/(ASCE)0733-9410(1995)121:5(429).
Evans DJ. 2007. An appraisal of Underground Gas Storage technologies and incidents, for the development of risk assessment methodology, British Geological Survey Open Report. British Geological Survey Open Report.
Ghaboussi J, Sidarta DE. 1998. New nested adaptive neural networks (NANN) for constitutive modeling. Computers and Geotechnics, 22:29–52. https://doi.org/10.1016/S0266-352X(97)00034-7.
Gholami R, Aadnoy B, Fakhari N. 2016. A thermo-poroelastic analytical approach to evaluate cement sheath integrity in deep vertical wells. Journal of Petroleum Science and Engineering, 147:536–546. https://doi.org/10.1016/j.petrol.2016.09.024.
Gorji M B, Mozaffar M, Heidenreich J N, et al. 2020. On the potential of recurrent neural networks for modeling path dependent plasticity. Journal of the Mechanics and Physics of Solids, 143:103972. https://doi.org/10.1016/j.jmps.2020.103972.
Grabowski E, Gillott JE. 1989. Effect of replacement of silica flour with silica fume on engineering properties of oilwell cements at normal and elevated temperatures and pressures. Cement and Concrete Research, 19:333–344. https://doi.org/10.1016/0008-8846(89)90023-9.
Habibagahi G, Bamdad A. 2003. A neural network framework for mechanical behavior of unsaturated soils. Canadian Geotechnical Journal, 40:684–693. https://doi.org/10.1139/t03-004.
Johari A, Javadi AA, Habibagahi G. 2011. Modelling the mechanical behaviour of unsaturated soils using a genetic algorithm-based neural network. Computers and Geotechnics, 38:2–13. https://doi.org/10.1016/j.compgeo.2010.08.011.
Karakosta E, Lagkaditi L, ElHardalo S, et al. 2015. Pore structure evolution and strength development of G-type elastic oil well cement. A combined 1H NMR and ultrasonic study. Cement and Concrete Research, 72:90–97. https://doi.org/10.1016/j.cemconres.2015.02.018.
Karniadakis GE, Kevrekidis IG, Lu L, et al. 2021. Physics-informed machine learning. Nature Reviews Physics, 3:422–440. https://doi.org/10.1038/s42254-021-00314-5.
Li K, Li D, Chen D, et al. 2021. A generalized model for effective thermal conductivity of soils considering porosity and mineral composition. Acta Geotechnica, 16:3455–3466. https://doi.org/10.1007/s11440-021-01282-x.
Li K, Yin Z, Zhang N, et al. 2023. A data-driven method to model stress-strain behaviour of frozen soil considering uncertainty. Cold Regions Science and Technology, 213:103906. https://doi.org/10.1016/j.coldregions.2023.103906.
Li X, Yao Z, Huang X, et al. 2021. Investigation of deformation and failure characteristics and energy evolution of sandstone under cyclic loading and unloading. Rock and Soil Mechanics, 42:1693–1704. https://doi.org/10.16285/j.rsm.2020.1463.
Li Z, Guo X, Han L, et al. 2007. Improvement of latex on mechanical deformation capability of cement sheath under triaxial loading condition. Acta Petrolei Sinica, 28:126–129. https://doi.org/10.7623/syxb200704027.
Liu X, Ning J, Tan Y, et al. 2016. Damage constitutive model based on energy dissipation for intact rock subjected to cyclic loading. International Journal of Rock Mechanics and Mining Sciences, 85:27–32. https://doi.org/10.1016/j.ijrmms.2016.03.003.
Luke K. 2004. Phase studies of pozzolanic stabilized calcium silicate hydrates at 180 °C. Cement and Concrete Research, 34:1725–1732. https://doi.org/10.1016/j.cemconres.2004.05.021.
Ma Q, Liu Z, Qin Y, et al. 2021. Rock plastic-damage constitutive model based on energy dissipation. Rock and Soil Mechanics, 42:1210–1220. https://doi.org/10.16285/j.rsm.2020.1091.
Ma Q, Qin Y, Zhou T, et al. 2019. Mechanical properties and constitutive model of porous rock under loading and unloading. Rock and Soil Mechanics, 40:2673–2685. https://doi.org/10.16285/j.rsm.2018.0513.
Mayergoyz ID. 1985. Hysteresis models from the mathematical and control theory points of view. Journal of Applied Physics, 57:3803–3805. https://doi.org/10.1063/1.334925.
Meng Q, Zhang M, Han L, et al. 2016. Effects of Acoustic Emission and Energy Evolution of Rock Specimens Under the Uniaxial Cyclic Loading and Unloading Compression. Rock Mechanics and Rock Engineering, 49:3873–3886. https://doi.org/10.1007/s00603-016-1077-y.
Mozaffar M, Bostanabad R, Chen W, et al. 2019. Deep learning predicts path-dependent plasticity. Proceedings of the National Academy of Sciences, 116:26414–26420. https://doi.org/10.1073/pnas.1911815116.
Nardin A, Schrefler B, Lefik M. 2003. Application of artificial neural network for identification of parameters of a constitutive law for soils. Developments in Applied Artificial Intelligence, 16:545–554.
Pang X, Qin J, Sun L, et al. 2021. Long-term strength retrogression of silica-enriched oil well cement: A comprehensive multi-approach analysis. Cement and Concrete Research, 144:106424. https://doi.org/10.1016/j.cemconres.2021.106424.
Peng X, Wang Z, Luo T, et al. 2008. An elasto-plastic constitutive model of moderate sandy clay based on BCRBFNN. Journal of Central South University of Technology, 15:47–50. https://doi.org/10.1007/s11771-008-0312-4.
Penumadu D, Zhao R. 1999. Triaxial compression behavior of sand and gravel using artificial neural networks (ANN). Computers and Geotechnics, 24:207–230. https://doi.org/10.1016/S0266-352X(99)00002-6.
Rafiai H, Jafari A. 2011. Artificial neural networks as a basis for new generation of rock failure criteria. International Journal of Rock Mechanics and Mining Sciences, 48:1153–1159. https://doi.org/10.1016/j.ijrmms.2011.06.001.
Raissi M, Perdikaris P, Karniadakis GE. 2019. Physicsinformed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378:686–707. https://doi.org/10.1016/j.jcp.2018.10.045.
Rashidian V, Hassanlourad M. 2014. Application of an Artificial Neural Network for Modeling the Mechanical Behavior of Carbonate Soils. International Journal of Geomechanics, 14:142–150. https://doi.org/10.1061/(ASCE)GM.1943-5622.0000299.
Read J S, Jia X, Willard J, et al. 2019. Process‐guided deep learning predictions of lake water temperature. Water Resources Research, 55:9173–9190. https://doi.org/10.1029/2019WR024922.
Romo MP, García SR, Mendoza MJ, et al. 2001. Recurrent and Constructive-Algorithm Networks For Sand Behavior Modeling. International Journal of Geomechanics, 1:371–387. https://doi.org/10.1080/15323640108500167.
Shadravan A, Kias E, Lew R, et al. 2015. Utilizing the Evolving Cement Mechanical Properties under Fatigue to Predict Cement Sheath Integrity. SPE Kuwait Oil & Gas Show and Conference. https://doi.org/10.2118/175231-MS.
Shaikh FUA, Supit SWM, Sarker PK. 2014. A study on the effect of nano silica on compressive strength of high volume fly ash mortars and concretes. Materials & Design, 60:433–442. https://doi.org/10.1016/j.matdes.2014.04.025.
Shi LL, Zhang J, Zhu QZ, et al. 2022. Prediction of mechanical behavior of rocks with strong strain-softening effects by a deep-learning approach. Computers and Geotechnics, 152:105040. https://doi.org/10.1016/j.compgeo.2022.105040.
Song Z, Frühwirt T, Konietzky H. 2018. Characteristics of dissipated energy of concrete subjected to cyclic loading. Construction and Building Materials, 168:47–60. https://doi.org/10.1016/j.conbuildmat.2018.02.076.
Tan Y, Wang C. 2001. A fast approaching model for rock constitutive equation by radial basis function neural network. Chinese Journal of Geotechnical Engineering, 23:14–17.
Qiu TM, Yuntian F, Mengqi W, et al. 2021. Constitutive relations of granular materials by integrating micromechanical knowledge with deep learning. Chinese Journal of Theoretical and Applied Mechanics, 53:2404–2415. https://doi.org/10.6052/0459-1879-21-221.
Wang M, Qu T, Guan S, et al. 2022. Datadriven strain–stress modelling of granular materials via temporal convolution neural network. Computers and Geotechnics, 152:105049. https://doi.org/10.1016/j.compgeo.2022.105049.
Wang Y, Wang W, Ma Z, et al. 2023. A deep learning approach based on physical constraints for predicting soil moisture in unsaturated zones. Water Resources Research, 59:e2023WR035194. https://doi.org/10.1029/2023WR035194.
Wu L, Ma D, Wang Z, et al. 2023. A deep CNN-based constitutive model for describing of statics characteristics of rock materials. Engineering Fracture Mechanics, 279:109054. https://doi.org/10.1016/j.engfracmech.2023.109054.
Wu Z, Song Z, Tan J, et al. 2020. The evolution law of rock energy under different graded cyclic loading and unloading modes. Journal of Mining Safety Engineering, 37:836–845. https://doi.org/10.13545/j.cnki.jmse.2020.04.23.
Xi Y, Li J, Tao Q, et al. 2020. Experimental and numerical investigations of accumulated plastic deformation in cement sheath during multistage fracturing in shale gas wells. Journal of Petroleum Science and Engineering, 187:106790. https://doi.org/10.1016/j.petrol.2019.106790.
Xiao JQ, Ding DX, Jiang FL, et al. 2010. Fatigue damage variable and evolution of rock subjected to cyclic loading. International Journal of Rock Mechanics and Mining Sciences, 47:461–468. https://doi.org/10.1016/j.ijrmms.2009.11.003.
Xiao JQ, Ding DX, Xu G, et al. 2009. Inverted S shaped model for nonlinear fatigue damage of rock. International Journal of Rock Mechanics and Mining Sciences, 46:643–648. https://doi.org/10.1016/j.ijrmms.2008.11.002.
Xie K, Liu P, Zhang J, et al. 2021. Physics-guided deep learning for rainfall-runoff modeling by considering extreme events and monotonic relationships. Journal of Hydrology, 603:127043. https://doi.org/10.1016/j.jhydrol.2021.127043.
Xie W, Kimura M, Takaki K, et al. 2022. Interpretable framework of physics‐guided neural network with attention mechanism: simulating paddy field water temperature variations. Water Resources Research, 58:e2021WR030493. https://doi.org/10.1029/2021WR030493.
Xiong W, Wang J, Wu M. 2023. Data-driven constitutive modelling of granular soils considering multiscale particle morphology. Computers and Geotechnics, 162:105699. https://doi.org/10.1016/j.compgeo.2023.105699.
Yang C, Kim Y, Ryu S, et al. 2020. Prediction of composite microstructure stress-strain curves using convolutional neural networks. Materials & Design, 189:108509. https://doi.org/10.1016/j.matdes.2020.108509.
Yang H, Wang L, Yang CH, et al. 2024. Mechanical performance of oil-well cement slurries cured and tested under high-temperatures and high-pressures for deep-well applications. Cement and Concrete Research, 175:107355. https://doi.org/10.1016/j.cemconres.2023.107355.
Yang HZ, Wang L, Huang G, et al. 2024. Effects of compressive cyclic loading on the fatigue properties of oil-well cement slurries serving in deep downhole environments. Construction and Building Materials, 428:136360. https://doi.org/10.1016/j.conbuildmat.2024.136360.
Yin H, Ma H, Shi X, et al. 2024. Leakage Risk Analysis of Underground Gas Storage Salt Caverns with Micro leakage Interlayer in Bedded Rock Salt of Jiangsu, China. Rock Mechanics and Rock Engineering, 58:2829–2845. https://doi.org/10.1007/s00603-024-04316-4.
Youssef MAH, Camilo M, Jamshid G, et al. 2006. Novel Approach to Integration of Numerical Modeling and Field Observations for Deep Excavations. Journal of Geotechnical and Geoenvironmental Engineering, 132:1019–1031. https://doi.org/10.1061/(ASCE)1090-0241(2006)132:8(1019).
Zhang P, Yin Z, Jin Y. 2021. State-of-the-art review of machine learning applications in constitutive modeling of soils. Archives of Computational Methods in Engineering, 28:3661–3686. https://doi.org/10.1007/s11831-020-09524-z.
Zheng Y, Liu P, Cheng L, et al. 2022. Extracting operation behaviors of cascade reservoirs using physics-guided long-short term memory networks. Journal of Hydrology: Regional Studies, 40:101034. https://doi.org/10.1016/j.ejrh.2022.101034.
Zhu JH, Zaman MM, Anderson SA. 1998. Modelling of shearing behaviour of a residual soil with Recurrent Neural Network. International Journal for Numerical and Analytical Methods in Geomechanics, 22:671–687. https://doi.org/10.1002/(SICI)1096-9853(199808)22:8<671::AID-NAG939>3.0.CO;2-Y
Downloads
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 GeoStorage
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.