Physics-informed generative adversarial networks for synthetic uniaxial compressive strength data generation in geotechnical applications

Abstract

Data-driven rock mechanics has long been constrained by a chronic deficit: coordinated laboratory measurements of uniaxial compressive strength (UCS) and its physically related properties are not only limited in number but also unevenly distributed across rock types. To address this constraint, the present study proposes a diversity-tuned physics-informed generative adversarial network—termed PI-GAN—designed to produce synthetic UCS, porosity, and bulk-density data that preserve the fundamental mechanical relationships inherent to rock materials. The framework integrates adversarial learning with an intact-rock strength-envelope constraint, category-specific UCS–porosity–density correlation losses, moment and covariance matching, feature-envelope penalties, and a pairwise-distance diversity term. The model was calibrated on 944 complete records drawn from the P³ petrophysical database, grouped into dense crystalline, volcanic, pyroclastic, and sedimentary rock categories. The trained PI-GAN generated 944 synthetic samples, increasing the usable dataset to 1,888 records. The synthetic data preserved the dominant correlation structure, with the overall UCS–porosity correlation shifting from − 0.598 to − 0.675 and the UCS–density correlation changing from 0.566 to 0.650. Category-specific Fisher z-tests showed non-significant differences for 11 of 12 correlation pairs, with an average absolute correlation difference of 0.036. Distributional validation yielded a mean Kolmogorov–Smirnov statistic of 0.181, while diversity diagnostics produced a mean pairwise-distance ratio of 0.703 and a near-duplicate rate of 0.0%. Strength-envelope sensitivity checks returned zero violations across tolerance levels from 10% to 50% of category mean UCS. The results indicate that the proposed PI-GAN can provide physics-consistent augmentation data for machine-learning development in data-scarce geotechnical applications, while serving as a complement to, rather than a replacement for, site-specific laboratory testing.

Keywords

Generative adversarial networks / Synthetic data generation / Uniaxial compressive strength / Physics-informed learning / Rock mechanics / Data augmentation

References

[1] Amiri H, Vasconcelos I, Jiao Y, Chen P-E, Plümper O. Quantifying microstructures of earth materials using higher-order spatial correlations and deep generative adversarial networks. Scientific Reports, 2023, 13 1805

[2] Bär K, Reinsch T, Bott J. The petrophysical property database (P3) - A global compilation of lab-measured rock properties. Earth System Science Data, 2020, 12: 2485-2515

[3] Bieniawski ZT. Engineering rock mass classifications, 1989John Wiley & Sons

[4] Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP. SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 2002, 16: 321-357

[5] Chen M, Kang X, Ma X. Deep learning-based enhancement of small sample liquefaction data. International Journal of Geomechanics, 2023, 23(9 04023140

[6] Ferreira I, Ochoa L, Koeshidayatullah A. On the generation of realistic synthetic petrographic datasets using a style-based GAN. Scientific Reports, 2022, 12 12845

[7] Ge Q, Li J, Lacasse S, Sun H, Liu Z. Data-augmented landslide displacement prediction using generative adversarial network. Journal of Rock Mechanics and Geotechnical Engineering, 2024, 16(104017-4033

[8] Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio Y. Generative adversarial nets. Advances in Neural Information Processing Systems, 2014, 27: 2672-2680

[9] Gretton A, Borgwardt KM, Rasch MJ, Schölkopf B, Smola A. A kernel two-sample test. Journal of Machine Learning Research, 2012, 13(25): 723-773

[10] Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V., & Courville, A. (2017). Improved training of Wasserstein GANs. Advances in Neural Information Processing Systems, 30.

[11] He H, Bai Y, Garcia EA, Li S. ADASYN: Adaptive synthetic sampling approach for imbalanced learning. IEEE International Joint Conference on Neural Networks, 2008, 2008: 1322-1328

[12] Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., & Hochreiter, S. (2017). GANs trained by a two time-scale update rule converge to a local Nash equilibrium. Advances in Neural Information Processing Systems, 30.

[13] Hoek, E., Carranza-Torres, C., & Corkum, B. (2002). Hoek-Brown failure criterion - 2002 edition. In Proceedings of the 5th North American Rock Mechanics Symposium 1 267–273).

[14] Hoek E, Brown ET. The Hoek-Brown failure criterion and GSI - 2018 edition. Journal of Rock Mechanics and Geotechnical Engineering, 2019, 11(3): 445-463

[15] Karniadakis GE, Kevrekidis IG, Lu L, Perdikaris P, Wang S, Yang L. Physics-informed machine learning. Nature Reviews Physics, 2021, 3(6): 422-440

[16] Klyuchnikov N, Ismailova L, Kovalev D, Safonov S, Koroteev D. Generative adversarial networks for synthetic wellbore data: Expert perception vs mathematical metrics. Journal of Petroleum Science and Engineering, 2022, 211 110106

[17] Kynkäänniemi, T., Karras, T., Laine, S., Lehtinen, J., & Aila, T. (2019). Improved precision and recall metric for assessing generative models. Advances in Neural Information Processing Systems, 32.

[18] Lyu B, Wang Y, Shi C. Multi-scale generative adversarial networks (GAN) for generation of three-dimensional subsurface geological models from limited boreholes and prior geological knowledge. Computers and Geotechnics, 2024, 170 106336

[19] Raissi M, Perdikaris P, Karniadakis GE. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 2019, 378: 686-707

[20] Shams R, Masihi M, Boozarjomehry RB, Blunt MJ. A hybrid of statistical and conditional generative adversarial neural network approaches for reconstruction of 3D porous media (ST-CGAN). Advances in Water Resources, 2021, 158 104064

[21] Xu, L., Skoularidou, M., Cuesta-Infante, A., & Veeramachaneni, K. (2019). Modeling tabular data using conditional GAN. Advances in Neural Information Processing Systems 32.

[22] Yaghoubi E, Yaghoubi E, Khamees A, Vakili AH. A systematic review and meta-analysis of artificial neural network, machine learning, deep learning, and ensemble learning approaches in field of geotechnical engineering. Neural Computing and Applications, 2024, 36(21): 12655-12699

[23] Yang L, Zhang D, Karniadakis GE. Physics-informed generative adversarial networks for stochastic differential equations. SIAM Journal on Scientific Computing, 2020, 42(1): A292-A317

[24] Zhao G, Cai Z, Wang X, Dang X. GAN data augmentation methods in rock classification. Applied Sciences, 2023, 13(9): 5316

[25] Zheng Q, Zhang D. RockGPT: Reconstructing three-dimensional digital rocks from single two-dimensional slice with deep learning. Computational Geosciences, 2022, 26: 677-696

Rights & permissions

The Author(s)