Prediction model for indoor rock compression failure time based on ensemble learning and optimization algorithms

Abstract

The prediction of rock failure, a key fundamental research for addressing mining safety issues (such as mine slope stability and rockburst), faces challenges with traditional methods due to their complex generalization and computational processes that struggle to describe the entire failure process. Consequently, 12 prediction models integrating ensemble learning and optimization algorithms were established to predict rock peak stress and failure time using strain, elastic modulus, density, mass, and confining pressure as inputs. Five-fold cross-validation was used to optimize hyperparameters, significantly improving the model’s generalization ability, robustness, and stability. Dataset was established through rock mechanics experiments, with strain increments configured at 0.008‰, 0.01‰, and 0.012‰ in the test set. The cross-validation optimized particle swarm optimization eXtreme gradient boosting (CV-PSO-XGBoost) model performed best under a strain increment of 0.01‰, and its stress prediction achieved coefficient of determination R² = 0.904, mean absolute error (MAE) = 4.315, and root mean square error (RMSE) = 5.435; while the failure time prediction demonstrated R² = 0.811, mean absolute percentage error (MAPE) = 7.842%, and MAE = 30.343. Finally, SHapley Additive exPlanations (SHAP) analysis showed strain and stress significantly impact the model, with strain positively predicting failure time, aligning with traditional rock failure models, validating reliability. This study provides insights into the research on rock strata stability in mining.

Keywords

rock failure / strain increment / failure time / ensemble learning / SHapley Additive exPlanations interpretation

References

[1] Pan XH, Sun HY, Wu ZJ, Lü Q. Study of the failure mechanism and progressive failure process of intact rock patches of rock slope with weak surfaces. Rock Mech. Rock Eng., 2017, 504951

[2] Li SJ, Huang Z, Yang DX, Zeng W, Zhao K. Study of the acoustic characteristics and evolution of the failure mode of yellow sandstone under uniaxial compression. Rock Mech. Rock Eng., 2024, 5721059

[3] H.Z. Jiao, X.Y. Ge, Q. Wang, et al., Solidification/stabilization mechanisms of heavy metal ions in cemented paste backfill for green mine operations: A review, Int. J. Miner. Metall. Mater., 33(2026), No. 2, art. No. 382.

[4] J. Zhou, W.S. Xu, G.M. Zhao, et al., Energy evolution law and fractal characteristics of different rock specimen sizes on dynamic compression, Geofluids, 2022(2022), art. No. 5339603.

[5] Li XH, Niu ZJ, Yao QL, Wang WN, Yu LQ. Particle flow analysis of crack propagation characteristics in three trigeminal fractured sandstone with holes. J. China Coal Soc., 2020, 45113735

[6] Zhou ZW, Ma WB. Study on the damage mechanism of rock-like materials containing bicircular holes under uniaxial compression. Chin. J. Appl. Mech., 2025, 4261364

[7] Bai YL, Wang HY, Xia MF, Ke FJ. Statistical mesomechanics of solid, linking coupled multiple space and time scales. Appl. Mech. Rev., 2005, 586372

[8] Feng ZC, Shen YX. Study on precursor identification of rock instability based on surface displacement. J. Rock Mech. Eng., 2021, 40S23024

[9] M. Gao, Z.Z. Liang, S.P. Jia, Q.H. Zhang, and J.Q. Zou, Energy evolution analysis and related failure criterion for layered rocks, Bull. Eng. Geol. Environ., 82(2023), No. 12, art. No. 439.

[10] Xia MF, Han WS, Ke FJ, Bai YL. Statistical mesoscoplc damage mechanics and damage evolution induced catastrophe. Adv. Mech., 1995, 252145

[11] Yu HZ, Yin XC, Liang NG. Use damage rheology model research on the precursor of earthquakes. Prog. Geophys., 2004, 192404

[12] Hao SW, Rong F, Lu MF, et al. . Power-law singularity as a possible catastrophe warning observed in rock experiments. Int. J. Rock Mech. Min. Sci., 2013, 60253

[13] Wang J, Gong JW, Zhao ZY. Position, direction of strain localization of rock-like specimens under uniaxial compression and its application to early-warning. Rock Soil Mech., 2018, 39S2186

[14] Zhang HM, Gao PX, Jia HL, Wang T, Tang LY. Study on the damage localization law in frozen sandstone during uniaxial compression. J. Glaciol. Geocryol., 2024, 462565

[15] L.Q. Ma, H. Sun, I. Ngo, and J. Han, Infrared radiation quantification of rock damage and its constitutive modeling under loading, Infrared Phys. Technol., 121(2022), art. No. 104044.

[16] Liu W, Ma LQ, Gao QQ, et al. . Fracture precursor recognition and damage quantitative characterization of stressed rock using infrared radiation. Rock Mech. Rock Eng., 2023, 5685567

[17] T.X. Hao, M.Q. Yuan, F. Li, and G.Q. Wang, Quantitative analysis method for infrared characterization of coal damage under load based on image enhancement and pixel extraction, Geomat. Nat. Hazards Risk, 14(2023), No. 1, art. No. 2272574.

[18] C. Zhang, X.J. Wang, P.H. Han, T. Zhang, L. Zhang, and F.T. Wang, Acoustic emission and splitting surface roughness of sandstone in a Brazilian splitting test under the influence of water saturation, Eng. Geol., 329(2024), art. No. 107369.

[19] G.L. Feng, Q. Ma, Z. He, et al., Time-delayed failure process of granite and its energy evolution and acoustic emission characteristics, Eng. Fail. Anal., 157(2024), art. No. 107854.

[20] Z.K. Zhang, Z.P. Song, J.X. Lai, J.L. Qiu, Y. Cheng, and J. Zhang, Critical slowing down precursor information for the acoustic emission response characteristics of defective tuffs, Theor. Appl. Fract. Mech., 129(2024), art. No. 104220.

[21] X.X. Kong, H.F. Lu, C.Y. Liu, and B.L. Zhao, Experimental study on precursor characteristics of rock failure based on strain and temperature changes, Case Stud. Therm. Eng., 41(2023), art. No. 102632.

[22] Cao KW, Ma LQ, Wu Y, et al. . Cyclic fatigue characteristics of rock failure using infrared radiation as precursor to violent failure: Experimental insights from loading and unloading response. Fatigue Fract. Eng. Mater. Struct., 2021, 442584

[23] Liu G, Xiao FS, Qin T. Rock mechanics characteristics and acoustic emission rule under small-size effect. J. Rock Mech. Eng., 2018, 37S23905

[24] Liu G, Li YM, Xiao FS, Huang SJ, Zhang RF. Study on failure mechanics behavior and damage evolution law of yellow sandstone under uniaxial triaxial and pore water action. J. Rock Mech. Eng., 2019, 38S23532

[25] Zhao Y, Gu XS, Wang SH, Gao Y. Common precursor characteristics of uniaxial compression failure in rocks of different rock types. J. Northeast. Univ. Nat. Sci. Ed., 2023, 4491309

[26] Chen GQ, Zhang Y, Li Y, Pan YG, Jin CY. Thermal-acoustic precursor information chain of rock failure under true triaxial loading. J. Rock Mech. Eng., 2021, 4091764

[27] Luo Y. Damage Analysis and Failure Prediction of Jointed Rock Mass. 2019, Guangzhou, South China University34

[28] Liu XH, Zhang K, Li N, Qi FF, Ye JM. Quantitative identification of the failure behavior of the 3D printed rock-like specimen with one hole and two flaws. Rock. Soil Mech., 2021, 42113017

[29] Qi FF, Zhang K, Xie JB. Fracturing mechanism of rocklike specimens with different joint densities based on DIC technology. Rock. Soil Mech., 2021, 4261669

[30] Zhang QB, Zhao J. Determination of mechanical properties and full-field strain measurements of rock material under dynamic loads. Int. J. Rock Mech. Min. Sci., 2013, 60423

[31] Li YL, Guo SJ, Yang RS, Xie LY, Lu SH. Effects of gangue particle-size gradation on damage and failure behavior of cemented backfill under uniaxial compression. Int. J. Miner. Metall. Mater., 2025, 3271483

[32] Wang SS, Yang RS, Li YL, Yue ZW. Effects of cement content, polypropylene fiber length and dosage on fluidity and mechanical properties of fiber-toughened cemented aeolian sand backfill. Int. J. Miner. Metall. Mater, 2024, 31112404

[33] X. Zhu, J. Fan, C.L. He, and Y. Tang, Identification of crack initiation and damage thresholds in sandstone using 3D digital image correlation, Theor. Appl. Fract. Mech., 122(2022), art. No. 103653.

[34] P.K. Luo, L.C. Wang, D.Y. Li, J.S. Yang, and X.X. Lv, Deformation and failure mechanism of horizontal soft and hard interlayered rock under uniaxial compression based on digital image correlation method, Eng. Fail. Anal., 142(2022), art. No. 106823.

[35] Teng YC. Research on the Constitutive Model of Frozen Silty Clay Based on Machine Learning. 2023, Daqing, Northeast Petroleum University27

[36] Wang SF, Wu YM, Shi XL, Cai X, Zhou ZL. Strength prediction and cuttability identification of rock based on monitoring while cutting (MWC) using a conical pick. Int. J. Miner. Metall. Mater., 2025, 3251025

[37] X.L. Jin, R. Zhao, and Y.L. Ma, Application of a hybrid machine learning model for the prediction of compressive strength and elastic modulus of rocks, Minerals, 12(2022), No. 12, art. No. 1506.

[38] Lin YG, Wang HB, Wei LX, Xu JP, Ma H. Analysis on mechanical properties of fractured rock mass and intelligent prediction based on machine learning. Yangtze River, 2024, 557240

[39] Wu XR. Research on Early Warning Method of Coal-Rock Combination Damage Based on GA-BP Neural Network. 2021, Chongqing, Chongqing University51

[40] Wei MD, Meng WZ, Dai F, Wu W. Application of machine learning in predicting the rate-dependent compressive strength of rocks. J. Rock Mech. Geotech. Eng., 2022, 1451356

[41] El Fil H, Pyrak-Nolte LJ, Bobet A. The role of a rock discontinuity’s micro-structure and aperture on the detection of seismic precursors to shear failure. Rock Mech. Rock Eng., 2024, 5742469

[42] X.L. Liu, Z. Liu, X.B. Li, F.Q. Gong, and K. Du, Experimental study on the effect of strain rate on rock acoustic emission characteristics, Int. J. Rock Mech. Min. Sci., 133(2020), art. No. 104420.

[43] Sun XM, Shi FK, Ding JX, et al. . Triaxial mechanical properties and precursors of instability damage in rocks with different laminar dip angles. J. Mt. Sci., 2024, 212648

[44] Yan WC, Wu C, Xin L. Bridge fatigue failure and weight limit analysis based on machine learning. J. Railway Sci. Eng., 2025, 2273292

[45] J.L. Li, J.P. Qian, J.Y. Chen, et al., Recent advances of machine learning in the geographical origin traceability of food and agro-products: A review, Compr. Rev. Food Sci. Food Saf., 24(2025), No. 1, art. No. e70082.

[46] Grinsztajn L, Grinsztajn L, Oyallon E, Oyallon E, Varoquaux G, Varoquaux G. Why do tree-based models still outperform deep learning on typical tabular data?. Proceedings of the 36th International Conference on Neural Information Processing Systems. 2022507

[47] Ke GL, Meng Q, Thomas F, et al. Guyon I, Von Luxburg U, Bengio S, Wallach H, Fergus R, Vishwanathan S, Garnett R, et al. . Lightgbm: A highly efficient gradient boosting decision tree. Advances in Neural Information Processing Systems. 2017

[48] Breiman L. Random forests. Mach. Learn., 2001, 4515

[49] Chen TQ, Guestrin C. XGBoost: A scalable tree boosting system. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2016785

[50] J. Kennedy and R. Eberhart, Particle swarm optimization, [in] Proceedings of ICNN’95 - International Conference on Neural Networks, Perth, 2002, art. No. 1942.

[51] Yang XSGonzález JR, Pelta DA, Cruz C, Terrazas G, Krasnogor N. A new metaheuristic bat-inspired algorithm. Nature Inspired Cooperative Strategies for Optimization (NICSO 2010). 201065

[52] L.Y. Wu, D. Ma, Z.F. Wang, et al., A deep CNN-based constitutive model for describing of statics characteristics of rock materials, Eng. Fract. Mech., 279(2023), art. No. 109054.

[53] Jiang Z, Jin Y, Xia M, et al. . Multi scale and multi physics verification of power law singular precursors of disaster destruction. Abstract Collection of Papers from the 12th National Conference on Physical Mechanics. 201270

[54] Ma G. Elasticity: Power Law Singularity Analysis of Precursors to Catastrophic Failure in Statistical Brittle Models. 2013, Beijing, University of the Chinese Academy of Sciences

[55] Jiang HP, Jiang AN, Zhang FR, Yang XR. Study on mechanical properties and statistical damage constitutive model of red sandstone after heating and water-cooling cycles. Int. J. Damage Mech., 2022, 317975

[56] Z.N. Zhu, S.Q. Yang, P.G. Ranjith, H. Tian, G.S. Jiang, and B. Dou, A statistical thermal damage constitutive model for rock considering characteristics of the void compaction stage based on normal distribution, Bull. Eng. Geol. Environ., 81(2022), No. 8, art. No. 306.

[57] Feng ZC, Lv ZX, Zhao YS. Research progress on short-impending prediction of rock failure—Comments on the competition of short-impending prediction of rock failure. J. Rock Mech. Eng., 2022, 41122522

[58] Liu XF, Zhao YQ, Wang XR, Yu HC. Current status and prospects of research on fatigue damage and failure precursors of rocks. Igneous Rocks Gen., 2022, 4762190

[59] L.Y. Wu, S. Zhao, D. Ma, Y.H. Zhu, W. Cai, and W. Qiao, Study on water inrush mechanism of bed separation overburden based on AI-based mesoscopic particle parameter calibration model, Tunn. Undergr. Space Technol., 169(2026), art. No. 107258.

RIGHTS & PERMISSIONS

University of Science and Technology Beijing