Tunnel and Underground Space (터널과지하공간)

Korean Society for Rock Mechanics and Rock Engineering (한국암반공학회)
권호 표지
DOI QR코드

DOI QR Code

A Study on the Applicability of Machine Learning Algorithms for Detecting Hydraulic Outliers in a Borehole

시추공 수리 이상점 탐지를 위한 기계학습 알고리즘의 적용성 연구

  • Received : 2023.12.07
  • Accepted : 2023.12.15
  • Published : 2023.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Korea Atomic Energy Research Institute (KAERI) constructed the KURT (KAERI Underground Research Tunnel) to analyze the hydrogeological/geochemical characteristics of deep rock mass. Numerous boreholes have been drilled to conduct various field tests. The selection of suitable investigation intervals within a borehole is of great importance. When objectives are centered around hydraulic flow and groundwater sampling, intervals with sufficient groundwater flow are the most suitable. This study defines such points as hydraulic outliers and aimed to detect them using borehole geophysical logging data (temperature and EC) from a 1 km depth borehole. For systematic and efficient outlier detection, machine learning algorithms, such as DBSCAN, OCSVM, kNN, and isolation forest, were applied and their applicability was assessed. Following data preprocessing and algorithm optimization, the four algorithms detected 55, 12, 52, and 68 outliers, respectively. Though this study confirms applicability of the machine learning algorithms, it is suggested that further verification and supplements are desirable since the input data were relatively limited.

한국원자력연구원은 심부 암반의 수리/지화학 특성 분석을 위해 KURT (KAERI Underground Research Tunnel)를 건설하였고, 다수의 조사용 시추공을 시추하여 각종 시험을 수행 중이다. 시추공 조사에서 목적에 적합한 조사 구간 선정은 매우 중요하며 수리 유동 파악 및 지하수 채수가 목적인 경우, 유량이 풍부한 구간이 조사 목적에 부합한다. 본 연구에서는 이러한 구간을 수리 이상점으로 정의했으며, 심도 1km 수준의 시추공 물리검층 자료(온도, 전기전도도)를 활용하여 이를 탐지하고자 하였다. 체계적이고 효율적인 이상점 탐지를 위해 기계학습 알고리즘 중 DBSCAN, OCSVM, kNN, isolation forest을 적용하고 그 적용성을 파악하였다. 데이터 전처리와 알고리즘 최적화를 수행했으며, 그 결과 네 가지 알고리즘은 각각 55, 12, 52, 68개의 수리 이상점을 탐지하였다. 본 논문을 통해 기계학습 알고리즘의 활용 가능성을 확인했으나, 학습에 활용된 입력자료가 제한적이었기 때문에, 향후 추가적인 검증과 보완이 바람직한 것으로 판단된다.

Keywords

Acknowledgement

본 연구는 2023년도 정부(과학기술정보통신부)의 재원으로 고준위폐기물관리차세대혁신기술개발사업의 지원(2021M2E3A2041312)과 사용후핵연료관리핵심기술개발사업단 및 한국연구재단의 지원(No.2021M2E1A1085200)을 받아 수행되었습니다.

References

  1. Ali, M., Zhu, P., Huolin, M., Pan, H., Abbas, K., Ashraf, U., Ullah, J., and Zhang, H., 2023, A novel machine learning appraoch for detecting outliers, rebuilding well logs, and enhancing reservoir characterization, Natural Resources Research, 32(3), 1047-1066.  https://doi.org/10.1007/s11053-023-10184-6
  2. Bressan, T.S., de Souza, M.K., Girelli, T.J., and Junior, F.C., 2020, Evaluation of machine learning methods for lithology classification using geophysical data, Computers and Geosciences, 139, 104475. 
  3. Dastjerdy, B., Saeidi, A., and Heidarzadeh, S., 2023, Review of applicable outlier detection methods to treat geomechanical data, Geotechnics, 3(2), 375-396. 
  4. Ester, M., Kriegel, H.P., Sander, J., and Xu, X., 1996, A density-based algorithm for discovering clusters in large spatial database with noise, Proceedings of the Second International Conference on Knowledge Discovery and Data Mining, 96, 226-231. 
  5. Hu, W., Liao, Y., and Vemuri, V.R., 2003, Robust anomaly detection using support vector machines, Proceedings of International Conference on Machine Learning, 282-289. 
  6. Jung, S., Choi, J.M., and Jeong, D., 2021, Lithological analysis of DB-2 borehole around KURT with depth, KAERI/TR-9012/2021, KAERI, Daejeon, Korea. 
  7. Kim, H.C. and Song, M.Y., 1999, A study on the effective utilization of temperature logging data for calculating geothermal gradient, Economic and Environmental Geology, 32(5), 503-517. 
  8. Lee, S.M., Kim, H.S., and Na, K.C., 1980, Explanatory text of the geological map of Daejeon. KIGAM, Daejeon, Korea. 
  9. Liu, F.T., Ting, K.M., and Zhou, Z.H., 2008, Isolation forest. Proceedings of the 8th IEEE International Conference on Data Mining, 413-422. 
  10. Park, H.I., Lee, J.D., and Cheong, J.G., 1977, Explanatory text of the geological map of Yuseong, KIGAM, Daejeon, Korea. 
  11. SolExperts, 2023, Multipacker systems, https://www.solexperts.com. 
  12. Taunk, K., De, S., Verma, S., and Swetapadma, A., 2019, A brief review of nearest neighbor algorithms for learning and classification, Proceedings of the International Conference on Intelligent Computing and Control System (ICCS), 1255-1260.