한국지반환경공학회 논문집 (Journal of the Korean GEO-environmental Society)

한국지반환경공학회 (Korean Geo-Environmental Society)
권호 표지
DOI QR코드

DOI QR Code

기계학습기법을 통한 액상화 발생가능 지수 예측에 관한 연구

Study of Prediction of Liquefaction Potential Index Based on Machine Learning Method

  • Junseo Jeon (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology (KICT)) ;
  • Jongkwan Kim (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology (KICT)) ;
  • Jintae Han (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology (KICT)) ;
  • Seunghwan Seo (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology (KICT)) ;
  • Byeonghan Jeon (Institute of Technology, LT SAMBO)
  • 투고 : 2024.09.30
  • 심사 : 2024.10.22
  • 발행 : 2024.11.01
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

실제 시추공 정보 및 지진파를 이용하여 액상화 발생가능 지수를 산정하고, 기계학습기법을 이용하여 액상화 발생가능 지수 예측모델을 학습하였다. 학습을 위해 지진파의 특징을 반영한 인자를 포함하여 총 10가지의 특징을 선택하였다. 일반적으로 이용되는 기계학습기법 중 사전학습을 통해 후보 모델을 선정하고, 후보 모델에 대해 베이지안 최적화를 적용하여 초매개변수를 최적화시켰다. 인공신경망, 가우시안 프로세스 회귀, 랜덤 포레스트 중 평균제곱근오차, 결정계수 및 과대적합 여부를 종합한 결과, 랜덤 포레스트가 액상화 발생가능 지수를 잘 예측하였다. 다만, 액상화 발생가능 지수가 5 이상에서는 액상화 발생가능 지수를 과소예측하는 경향을 보였다.

In this study, the liquefaction potential index was assessed using actual borehole data and seismic waves, and a predictive model was developed based on machine learning methods. A total of 10 features were selected including factors reflecting the characteristics of the seismic waves. To identify candidate methods, a preliminary test was conducted using commonly used machine learning methods for regression, followed by Bayesian optimization to optimize the hyperparameters for these candidate methods. Among artificial neural networks, Gaussian process regression, and random forest, it was found that the random forest effectively predicted the liquefaction potential index, as indicated by a low root mean square error, a high coefficient of determination, and considerations regarding overfitting. However, it was noted that the model tends to underestimate the liquefaction potential index when the index was 5 or higher.

키워드

과제정보

본 연구는 과학기술정보통신부 한국건설기술연구원 연구운영비지원(주요사업)사업으로 수행되었습니다(과제번호 20240104-001, AI기반 지진 시 지반 액상화 평가를 위한 데이터베이스 구축, 과제번호 20240133-001, 지반분야 재난재해 대응과 미래 건설산업 신성장을 위한 지반 기술 연구).

참고문헌

  1. Baek, W. and Choi, J. (2019), Correlations of earthquake accelerations and LPIs for liquefaction risk mapping in Seoul & Gyeonggi-do area based on artificial scenarios, Journal of the Korean Geo-Environmental Society, Vol. 20, No. 5, pp. 5~12 (In Korean).
  2. Baziar, M. H., Jafarian, Y., Shahnazari, H., Movahed, V. and Tutunchian, M. A. (2011), Prediction of strain energy-based liquefaction resistance of sand-slit mixtures: An evolutionary approach, Computers & Geosciences, Vol. 37, No. 11. pp. 1883~1893.
  3. Chern, S. G., Lee, C. Y. and Wang, C. C. (2008), CPT-based liquefaction assessment by using fuzzy-neural network, Journal of Marine Science and Technology, Vol. 16, No. 2. pp. 139~148.
  4. Fattah, E. A., Ali, H. E. A. and Ebid, A. M. (2002), Prediction of soil liquefaction using genetic programming, Proceedings of III Middle East Regional Conference on Civil Engineering Technology and III International Symposium on Environmental Hydrology, ASCE-EGS, Cairo, Egypt.
  5. Goh, A. T. C. (1994), Seismic liquefaction potential assessed by neural networks, Journal of Geotechnical Engineering, Vol. 120, No. 9, pp. 1467~1480.
  6. Goh, A. T. C. and Goh, S. H. (2007), Support vector machines: Their use in geotechnical engineering as illustrated using seismic liquefaction data, Computers and Geotechnics, Vol. 34, No. 5, pp. 410~421.
  7. Ghani, S., Sapkota, S. C., Singh, R. K., Bardhan, A. and Asteris, P. G. (2024), Modelling and validation of liquefaction potential index of fine-grained soils using ensemble learning paradigms, Soil Dynamics and Earthquake Engineering, Vol. 177, pp. 108399.
  8. Iwasaki, T., Tatsuoka, F., Tokida, K. -I., and Yasuda, S. (1978), A practical method for assessing soil liquefaction potential based on case studies at various sites in Japan, Proceedings of the 2nd International Conference on Microzonation for Safer Construction Research and Application, NSF, San Francisco, USA, pp. 885~896.
  9. Jas, K. and Dodagoudar, G. R. (2023), Explainable machine learning model for liquefaction potential assessment of soils using XGBoost-SHAP, Soil Dynamics and Earthquake Engineering, Vol. 165, pp. 107662.
  10. Juang, C. H., Yuan, H., Lee., D.H. and Lin., P.S. (2003), Simplified cone penetration test-based method for evaluating liquefaction resistance of soils, Journal of Geotechnical and Geoenvironmental Engineering, Vol. 129, No. 1. pp. 66~80.
  11. Kohestani, V. R., Hassanlourad, M. and Ardakani, A. (2015), Evaluation of liquefaction potential based on CPT data using random forest, Natural Hazards, Vol. 79, pp. 1079~1089.
  12. Ozsagir, M., Erden, C., Bol, E., Sert, S. and Ozocak, A. (2022), Machine learning approaches for prediction of fine-grained soils liquefaction, Computers and Geotechnics, Vol. 152, pp. 105014.
  13. Pal, M. (2006), Support vector machines-based modelling of seismic liquefaction potential, International Journal for Numerical and Analytical Methods in Geomechanics, Vol. 30, No. 10, pp. 983~996.
  14. Sabbar, A. S., Chegenizadeh, A. and Nikraz, H. (2019), Prediction of liquefaction susceptibility of clean sandy soils using artificial intelligence techniques, Indian Geotechnical Journal, Vol. 49. pp. 58~69.
  15. Samui, P. and Sitharam, T. G. (2011), Machine learning modelling for predicting soil liquefaction susceptibility, Natural Hazards and Earth System Sciences, Vol. 11, pp. 1~9.