Analysis of Slope Failure Factors Based on Data and Prediction of Countermeasure Methods Using Artificial Intelligence

Sun-Kyum Kim; Seung-Hyun Kim

doi:10.9720/kseg.2024.4.563

All Issue

2024 Vol.34, Issue 4 Preview Page Next Page

Research Article

Analysis of Slope Failure Factors Based on Data and Prediction of Countermeasure Methods Using Artificial Intelligence 데이터 기반의 비탈면 붕괴 요인 분석 및 인공지능을 활용한 대책 공법 예측 연구

31 December 2024. pp. 563-575

PDF XML

Abstract

This study aims to derive meaningful insights by analyzing the correlation of collapse factors and predicting appropriate countermeasure methods using artificial intelligence. We utilize Cramér’s V correlation coefficient, a well-established method for correlation analysis, to explore the relationship between key factors contributing to slope collapse, such as slope type, surrounding topography, groundwater, weathering degree, discontinuity directionality, slope shape, lateral shape, Failure history, floating rock, rockfall, rock mass shape, and discontinuity within a basic investigation of the CSMS (cut slope management system). In addition, prediction models are designed for seven types of countermeasures—maintaining the status quo, reducing active load, suppressing active load, controlling rockfall, protecting the surface, controlling hydraulics, and other methods—using ensemble-based algorithms (Random Forest, XGBoost, LightGBM, and CatBoost) and deep learning-based algorithms (DNN (deep neural network) and TabNet). The analysis confirmed that slope type and weathering degree, collapse history and collapse type, and slope type and rock type showed the strongest correlations among the factors related to slope collapse. The ensemble-based algorithms demonstrated overall high prediction performance. This research is expected to contribute to the development of a high-quality database, enable the provision of more accurate countermeasure methods based on similar slope classifications, and enhance the usability of the CSMS.

Keywords

slope failure

Cramér’s V

correlation analysis

ensemble

deep learning

prediction algorithm

본 연구는 붕괴 요인 간의 상관 분석과 인공지능을 활용한 대책공법 예측을 통해 의미 있는 결과를 도출한다. 대표적인 상관분석 방법인 Cramér’s V 상관계수를 활용하여 도로비탈면관리시스템(cut slope management system, CSMS)의 기초 조사 데이터 중 비탈면 붕괴를 유발하는 주요 요인들(비탈면 종류, 주변 지형, 지하수, 풍화도, 불연속면 방향성, 비탈면 형상, 측면 현상, 붕괴 이력, 뜬돌, 낙석, 암반 형태, 불연속면) 간의 상관 관계를 살펴본다. 또한, Random Forest, XGBoost, LightGBM, CatBoost의 앙상블 기반 알고리즘과 DNN(deep neural network), TabNet의 딥러닝 기반 알고리즘을 활용하여 대책공법인 현상태 유지, 활동 하중 경감 공법, 활동 억제 공법, 낙석 제어 공법, 표면 보호 공법, 수리 제어 공법, 기타 공법의 총 7가지 공법에 대한 예측 모델을 설계한다. 그 결과, 비탈면 붕괴 요인 중 비탈면 종류와 풍화도, 붕괴 이력과 붕괴 유형, 비탈면 종류와 암반 형태가 가장 큰 관계를 가진 것으로 확인되며, 예측에 대해서도 앙상블 기반 알고리즘이 전반적으로 높은 예측 성능을 보였다. 이를 통해 고품질의 데이터베이스를 구축하고, 유사 비탈면 분류와 함께 더 정밀한 대책공법을 제공할 수 있으며, CSMS의 활용도를 높일 수 있을 것으로 기대된다.

키워드

비탈면 붕괴

Cramér’s V

상관분석

앙상블

딥러닝

예측 알고리즘

References

Akoglu, H., 2018, User's guide to correlation coefficients, Turkish Journal of Emergency Medicine, 18(3), 91-93.

10.1016/j.tjem.2018.08.00130191186PMC6107969

Arik, S.Ö., Pfister, T., 2021, TabNet: Attentive interpretable tabular learning, Proceedings of the AAAI Conference on Artificial Intelligence, 35(8) 6679-6687.

10.1609/aaai.v35i8.16826

Breiman, L., 2001, Random forests, Machine Learning, 45, 5-32.

10.1023/A:1010933404324

Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., Liu, T.Y., 2017, LightGBM: A highly efficient gradient boosting decision tree, Proceedings of the 31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, 3146-3154.

KICT (Korea Institute of Civil Engineering and Building Technology), 2019, Operation of road cut slope management system in 2018, Ministry of Land, Infrastructure and Transport, 355p (in Korean).

Kim, S.H., Koo, H.B., Hwang, J.H., Son, M., 2011, Case study on the cause of failure and characteristics of soil at a collapsed cut-slope at the ◯◯ Detour, Jeonranam-do, The Journal of Engineering Geology, 21(4), 313-322 (in Korean with English abstract).

10.9720/kseg.2011.21.4.313

Lee, K.B., Shin, H.S., Kim, S.H., Ha, D.M., Choi, I., 2019a, A study on automatic classification of characterized ground regions on slopes by a deep learning based image segmentation, Tunnel & Underground Space, 29(6), 508-522 (in Korean with English abstract).

10.7474/TUS.2019.29.6.508

Lee, Y., Kim, H., Lee, D., Lee, C., Lee, D., 2019b, Validation of forecasting performance of two-stage probabilistic solar irradiation and solar power forecasting algorithm using XGBoost, The Transactions of the Korean Institute of Electrical Engineers, 68(12), 1704-1710 (in Korean with English abstract).

10.5370/KIEE.2019.68.12.1704

Mittal, S., 2020, A survey on modeling and improving reliability of DNN algorithms and accelerators, Journal of Systems Architecture, 104, 101689.

10.1016/j.sysarc.2019.101689

MLTMA (Ministry of Land, Transport and Maritime Affairs), 2011, Practical manual for roadside slope management, 443p (in Korean).

Park, D.G., Kim, T.H., Park, J.H., 2006, Current status of slope disaster damage and countermeasures in Korea, Geotechnical Engineering, 22(6), 6-18 (in Korean).

Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A.V., Gulin, A., 2018, CatBoost: Unbiased boosting with categorical features, Proceedings of the 32nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montréal, 6639-6649.

Saqlain, M., Jargalsaikhan, B., Lee, J.Y., 2019, A voting ensemble classifier for wafer map defect patterns identification in semiconductor manufacturing, IEEE Transactions on Semiconductor Manufacturing, 32(2), 171-182.

10.1109/TSM.2019.2904306

Information

Publisher :Korean Society of Engineering Geology
Publisher(Ko) :대한지질공학회
Journal Title :The Journal of Engineering Geology
Journal Title(Ko) :지질공학
Volume : 34
No :4
Pages :563-575
Received Date : 2024-11-06
Revised Date : 2024-12-10
Accepted Date : 2024-12-13
DOI :https://doi.org/10.9720/kseg.2024.4.563

The Journal of Engineering Geology ISSN:1226-5268(Print) 2287-7169(Online) 지질공학

All Issue