All Issue

2025 Vol.41, Issue 3 Preview Page
Development of an MS–ANN Model for Predicting Ground Subsidence and Supporting Reinforcement Decision-Making in Abandoned Mine Areas

MS-ANN을 활용한 국내 폐광산 지반침하 예측 및 보강 의사결정 모델 개발

Min-Koan Kim1*
김 민관1*
1Member, Post Doctor, Korea Construction Standards Center, Korea Institute of Civil Engineering and Building Technology
1정회원, 한국건설기술연구원, 국가건설기준센터, 박사후연구원
*Corresponding Author
30 June 2025. pp. 25-33
PDF XML Citation
Abstract

Ground subsidence in abandoned mines is a critical geohazard that threatens infrastructure stability in Korea, necessitating accurate prediction and efficient reinforcement strategies. This study developed the Mine Subsidence– Artificial Neural Network (MS–ANN) model to predict ground subsidence and recommend reinforcement methods for abandoned mines in Korea. Using 247 subsidence cases from the Korean Mine Reclamation Corporation and 30 influencing factors, including geological, mining, and land-use data, the MS–ANN was designed with a feedforward neural network architecture optimized using the resilient backpropagation algorithm. The model comprises 21 input nodes, 2 hidden layers (with 21 and 28 nodes), and 9 output nodes to predict subsidence occurrence, extent, and reinforcement methods. Training on 229 cases yielded correlation coefficients of 0.968–0.9946 and an average inference error rate (AIER) of 4.5%–7.0%. However, validation on 18 test cases recorded correlation coefficients of 0.9855–0.9961 and an AIER of 2.5%–3.5%. The MS–ANN demonstrates stable prediction performance across various mine types, aiding subsidence risk assessment, reinforcement method recommendations, and sustainable land use. It is expected to enhance the cost-efficiency of mine reclamation by mitigating risks and costs associated with geohazards in Korea.

Keywords
Abandoned mine
Artificial neural network
Geohazard management
Ground subsidence
Mine Subsidence-Artificial Neural Network
Reinforcement method
RPROP algorithm
Stability assessment

국내 폐광산의 지반침하는 건축물 및 인프라 안정성을 위협하는 주요 지질재해로, 정확한 예측과 효율적인 보강 전략이 요구된다. 본 연구는 국내 폐광산의 지반침하 예측과 보강 공법 결정을 위해 Mine Subsidence-Artificial Neural Network(MS-ANN) 모델을 개발하였다. 한국광해광업공단의 247개 침하 사례와 30개 영향인자(지질, 채굴, 토지이용 데이터 등)를 활용하여 피드포워드 신경망(FNN) 구조를 기반으로 모델을 설계하였다. MS-ANN은 RPROP(Resilient Backpropagation) 알고리즘으로 최적화되었다. 이 모델은 입력층 21개 노드, 두 개의 은닉층(각각 21개 및 28개 노드), 출력층 9개 노드로 구성되어 지반침하 발생, 범위, 보강 공법을 예측한다. 229개 학습 데이터로 훈련한 결과, 예측 정확도(R2) 0.968~0.9946, 평균 추론 오차율(AIER) 4.5~7.0%를 달성하였으며, 시험 데이터(18개) 검증에서 R2 0.9855~0.9961, AIER 2.5~3.5%를 기록하였다. MS-ANN은 다양한 광산 유형에서 안정적인 예측 성능을 보이며, 침하 위험 평가, 보강 공법 추천, 그리고 지속가능한 토지 이용을 지원한다. 본 모델은 폐광산 복구 사업의 비용 효과성을 높이고, 지질재해로 인한 위험을 최소화하는 데 기여할 것으로 기대된다.

References
1

Bui, D. T., Tuan, T. A., Klempe, H., Pradhan, B., and Revhaug, I. (2016), "Spatial prediction models for shallow landslide hazards: a comparative assessment of the efficacy of support vector machines, artificial neural networks, kernel logistic regression, and logistic model tree", Landslides, Vol.13, No.2, pp.361-378, https://doi.org/10.1007/s10346-015-0557-6.

10.1007/s10346-015-0557-6
2

Choi, S. O., Kim, J. D., and Choi, G. S. (2009), "Application of Fuzzy Reasoning Method for Prediction of Subsidence Occurrences in Abandoned Mine Area", Tunnel and Underground Space, Vol.19, No.5, pp.463-472.

3

Korean Mine Reclamation Corporation. (2011), Guide book: Mine Rehabilitation technology in Korea, Korean Mine Reclamation Corporation.

4

Korean Mine Reclamation Corporation. (2016), Development of On-Site Monitoring Techniques for Ground Subsidence, Technical Report, pp.52-127.

5

Kriegeskorte, N. (2015), "Deep Neural Networks: A New Framework for Modeling", Annual Review of Vision Science, 1, pp.417-446, https://doi.org/10.1146/annurev-vision-082114-035447.

10.1146/annurev-vision-082114-03544728532370
6

Lee, S., Kim, Y. S., and Oh, H. J. (2012), "Spatial Prediction of Ground Subsidence Susceptibility Using an Artificial Neural Network", Environmental Management, Vol.49, No.2, pp.347-358, https://doi.org/10.1007/s00267-011-9766-5.

10.1007/s00267-011-9766-522005969
7

Park, H. B., Moon, S. W., Ju, S. J., Lee, J. E., and Seo, Y. S. (2024), "A Comparative Analysis of the Evaluation Methods for Ground Subsidence in Korea", The Journal of Engineering Geology, Vol.34, No.3, pp.381-401, https://doi.org/10.9720/KSEG.2024.3.381.

8

Riedmiller, M. and Braun, H. (1994), "A Direct Adaptive Method for Faster Backpropagation Learning: The RPROP Algorithm", Proceedings of the IEEE International Conference on Neural Networks, pp.586-591, http://dx.doi.org/10.1109/ICNN.1993.298623.

10.1109/ICNN.1993.298623
9

Ryu, D. W., Kim, T. H., and Heo, J. H. (2007), "A Study on the Evaluation Method of Subsidence Hazard by a Diffusion Equation and its Application", Tunnel and Underground Space, Vol.17, No.5, pp.116-126.

10

Shin, H. S. (2001), Neural network based constitutive models for finite element analysis, Ph.D. thesis, C/Ph/250/01, University of Wales Swansea, UK.

11

Siddique, T. and Mahmud, M. S. (2022), Ensemble deep learning models for prediction and uncertainty quantification of ground magnetic perturbation, Frontiers in Astronomy and Space Sciences, 9:1031407, https://doi.org/10.3389/fspas.2022.1031407.

10.3389/fspas.2022.1031407
12

Waltham, T., Bell, F., and Culshaw, M. (2007), "Sinkholes and subsidence: Karst and cavernous rocks in engineering and construction", Springer, https://doi.org/10.1007/978-1-84628-784-7.

13

Yang, I. J. and Lee, S. A. (2017), "A Study on the Status and Major Achievements on Mine Subsidence Prevention Technology", Tunnel and Underground Space, Vol.27, No.6, pp.357-365, https://doi.org/10.7474/TUS.2017.27.6.357.

14

Zhang, P., Yin, Z. Y., and Jin, Y. F. (2022), "Bayesian neural network-based uncertainty modelling: application to soil compressibility and undrained shear strength prediction", Canadian Geotechnical Journal, Vol.59, No.6, pp.463-472, https://doi.org/10.1139/cgj-2020-0751.

10.1139/cgj-2020-0751
Information
  • Publisher :The Korean Geotechnical Society
  • Publisher(Ko) :한국지반공학회
  • Journal Title :Journal of the Korean Geotechnical Society
  • Journal Title(Ko) :한국지반공학회 논문집
  • Volume : 41
  • No :3
  • Pages :25-33
  • Received Date : 2025-04-28
  • Revised Date : 2025-06-08
  • Accepted Date : 2025-06-12
  • DOI :https://doi.org/10.7843/kgs.2025.41.3.25
Journal Informaiton Journal of the Korean Geotechnical Society Journal of the Korean Geotechnical Society
Archives
: Vol.23~28 (2007~2012)
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close