Leeb 경도와 머신러닝 기법을 통한 암석 일축압축강도 추정
Estimation of Rock Uniaxial Compressive Strength Using Leeb Hardness and Machine Learning Techniques
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
본 연구에서는 비파괴적 시험 방법을 통해 다양한 암석의 일축압축강도(UCS)를 예측하는 모델을 개발 하고, Leeb 경도(LH)와 UCS 간의 비선형 관계를 규명하였다. 직접 실험을 통해 수집한 94개와 문헌에 서 수집한 101개를 포함하여 총 195개의 데이터셋을 분석하여 LH, P파 속도(VP), 건조밀도(Density) 와 UCS 간의 상관관계를 평가하였다. LH와 UCS의 관계 분석 결과, 암석 유형에 따라 상관성의 차이가 나타났으며, 특히 화성암에서 가장 높은 상관관계(R2 = 0.74)를, 변성암에서 가장 낮은 상관관계(R2 = 0.12)를 보였다. 예측 방법을 비교한 결과, 머신러닝 모델이 선형회귀(R2 = 0.626)와 비선형회귀(R2 = 0.644) 방법보다 현저히 우수한 성능을 보였다. 다양한 머신러닝 모델 중 XGBoost가 가장 높은 예측 성능(R2 = 0.902, RMSE = 16.282)을 보였으며, 특히 여러 변수(LH, VP, Density, Rock type)를 결합했을 때 예측 정확도가 크게 향상되었다. 본 연구는 비파괴적 시험과 머신러닝을 결합하여 현장과 실험실에서 암석 강도를 신속하고 경제적으로 예측할 수 있는 방법론을 제시하며, 다양한 암석 물성의 통합을 통해 예측 정확도를 더욱 향상시킬 수 있는 가능성을 보여준다.
Abstract
This study developed a model to predict the uniaxial compressive strength (UCS) of various rocks using non-destructive testing methods, identifying a non-linear relationship between Leeb hardness (LH) and UCS. A total of 195 datasets, including 94 experimental data and 101 literature data, were analyzed to evaluate correlations among LH, P-wave velocity (VP), dry density (Density), and UCS. The analysis of the relationship between LH and UCS showed variable correlations by rock type, with igneous rocks showing the highest correlation (R2 = 0.74) and metamorphic rocks showing the lowest (R2 = 0.12). In comparing prediction methods, machine learning models significantly outperformed both linear (R2 = 0.626) and non-linear regression (R2 = 0.644) approaches. Among various machine learning models tested, XGBoost demonstrated the highest performance (R2 = 0.902, RMSE = 16.282), particularly when combining multiple variables (LH, VP, Density, Rock type). This study suggests that non-destructive testing, combined with machine learning, can offer an efficient and economical approach to predicting UCS in both field and laboratory settings, with potential for enhanced accuracy through the integration of multiple rock properties.
참고문헌 (23)
- Aldeeky, H., "Assessing the uniaxial compressive strength and tangent Young’s modulus of basalt rock using the Leeb rebound hardness test", Materiales de Construcción, 70(340), pp.230~ (2020)
- Altman, N. S., "An introduction to kernel and nearest-neighbor nonparametric regression", The American Statistician, 46(3), pp.175~185 (1992)
- Aoki, H., "Estimating the unconfined compressive strength of intact rocks from Equotip hardness", Bulletin of Engineering Geology and the Environment, 67, pp.23~29 (2008)
- Aydin, A., "The Schmidt hammer in rock material characterization", Engineering Geology, 81(1), pp.1~14 (2005)
- Breiman, L., "Random forests", Machine Learning, 45(1), pp.5~32 (2001)
- Çelik, S. B., "Comparative investigation of Shore, Schmidt, and Leeb hardness tests in the characterization of rock materials", Environmental Earth Sciences, 78, pp.1~16 (2019)
- Çelik, S.B., "Investigation of the effect of saturated conditions and number of measurements on the Leeb hardness test and improved correlations to estimate basic rock properties", Acta Geotechnica, 18(8), pp.4261~4278 (2023)
- Çelik, S.B., "Investigation of the use of Leeb hardness in the estimation of some physical and mechanical properties of rock materials", Pamukkale Univ. J. Eng. Sci, 26(8), pp.1385~1392 (2020)
- Çelik, S.B., "Investigation of the Leeb hardness test in rapid characterisation of rock cores with particular emphasis on the effect of length to diameter ratio", International Journal of Mining, Reclamation and Environment, 37(7), pp.524~543 (2023)
- Chen, T., "XGBoost: A scalable tree boosting system", Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp.785~794 (2016)
- Corkum, A. G., "The Leeb hardness test for rock : An updated methodology and UCS correlation", Rock Mechanics and Rock Engineering, 51(3), pp.665~675 (2017)
- Corkum, A. G., "The Leeb hardness test for rock : an updated methodology and UCS correlation", Rock Mechanics and Rock Engineering, 51, pp.665~675 (2018)
- Cortes, C., "Support-vector networks", Machine Learning, 20(3), pp.273~297 (1995)
- Dipova, N., "Nondestructive testing of stabilized soils and soft rocks via needle penetration", Periodica Polytechnica Civil Engineering, 62(2), pp.539~544 (2018)
- Garrido, M. E., "Predicting the uniaxial compressive strength of a limestone exposed to high temperatures by point load and Leeb rebound hardness testing", Rock Mechanics and Rock Engineering, 55(1), pp.1~17 (2022)
- Ghorbani, S., "Adoption of ASTM A956-06 Leeb hardness testing standard to rock engineering applications", Construction and Building Materials, 364, pp.129886~ (2022)
- Gomez-Heras, M., "Ultrasonic pulse velocity as a way of improving uniaxial compressive strength estimations from Leeb hardness measurements", Construction and Building Materials, 261, pp.119996~ (2020)
- Guan, S., "Comparison and combination of Leeb hardness and point load strength for indirect measuring tensile and compressive strength of rocks", Bulletin of Engineering Geology and the Environment, 83(4), pp.109~ (2024)
- İnce, İ., "An investigation on sample size in Leeb hardness test and prediction of some index properties of magmatic rocks", Arab Journal of Geosciences, 14(3), pp.182~ (2021)
- Ke, G., "LightGBM : A highly efficient gradient boosting decision tree", Advances in Neural Information Processing Systems, 30 (2017)
- Leeb, D., "Dynamic hardness testing of metallic materials", NDT International, 12(6), pp.274~278 (1979)
- Verwaal, W., "Estimating rock strength with the equotip hardness tester", International Journal of Rock Mechanics and Mining Sciences and Geomechanics Abstracts, 30(6), pp.659~662 (1993)