터널과지하공간 (Tunnel and Underground Space)

  • 제34권5호
  • /
  • Pages.527-542
  • /
  • 2024
  • /
  • 1225-1275(pISSN)
  • /
  • 2287-1748(eISSN)
한국암반공학회 (Korean Society for Rock Mechanics and Rock Engineering)
권호 표지
DOI QR코드

DOI QR Code

쉴드 TBM의 자동 방향제어를 위한 머신러닝과 반복계산법에 의한 중절잭 추진 거리 예측

Prediction of Articulation Jack Strokes for Automatic Steering Control of a Shield TBM Using Machine Learning and Iterative Calculation

  • 장수호 (한국건설기술연구원 지반연구본부) ;
  • 이철호 (한국건설기술연구원 지반연구본부) ;
  • 강태호 (한국건설기술연구원 지반연구본부) ;
  • 최순욱 (한국건설기술연구원 지반연구본부)
  • Soo-Ho Chang (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology) ;
  • Chulho Lee (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology) ;
  • Tae-Ho Kang (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology) ;
  • Soon-Wook Choi (Department of Geotechnical Engineering Research, Korea Institute of Civil Engineering and Building Technology)
  • 투고 : 2024.10.07
  • 심사 : 2024.10.14
  • 발행 : 2024.10.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

본 연구에서는 쉴드 TBM의 자율 운전을 위해 필요한 자동 방향제어에 대한 기초적인 연구로서, 쉴드 TBM에 사용되는 중절잭들의 추진 거리를 예측하고 이를 통해 쉴드 TBM의 3차원 경로 좌표를 계산할 수 있는 이론과 알고리즘을 정리하고 제시하였다. 중절잭들의 추진 거리를 예측하기 위하여 랜덤 포레스트 모델 기반의 머신러닝 모델과 설정된 허용 오차를 만족할 때까지 반복계산을 실시하는 방법을 적용하였다. 반복계산의 경우에는 계산시간을 단축하기 위한 최적화 방법들을 적용하였다. 허용오차를 제한하는 반복계산에 의한 상대오차의 평균과 분산이 머신러닝 모델의 예측결과보다 상대적으로 작게 나타났다. 그러나 반복계산의 경우에는 최적화 방법을 적용하더라도 계산시간 측면에서 적용할 수 있는 허용오차에 한계가 있는 것으로 나타났다. 따라서 실시간 계산속도가 중요한 경우에는 머신러닝 모델을 적용하는 것이 바람직하며, 사전에 계산된 결과를 활용할 수 있을 경우에는 반복계산법을 적용하여 정확도를 보다 높일 수 있을 것으로 판단된다.

A fundamental study was carried out on automatic steering control necessary for autonomous operation of shield TBMs in the future. It outlines and proposes theories and algorithms for predicting the strokes of articulation jacks used in a shield TBM and calculating the three-dimensional path coordinates of the shield TBM based on these predictions. To predict the strokes of articulation jacks, two methods were applied: a machine learning model based on the random forest regressor, and an iterative calculation method to satisfy a preset allowable error. For the iterative calculation, optimization methods were applied to reduce computation time. The mean and variance of the relative errors from the iterative calculation with allowable error were found to be relatively smaller than the predictions of the machine learning model. However, even with optimization methods applied, the iterative calculation method showed limitations in the allowable error that could be applied in terms of computation time. Therefore, it would be better to apply the machine learning model when real-time calculation speed is crucial. On the other hand, when pre-calculated results can be used during construction, the iterative calculation can be applied to achieve higher accuracy.

키워드

과제정보

본 연구는 한국건설기술연구원의 Pre-WTCL (World Top-Class Laboratory) 주요사업인 "곡선구간의 TBM 무인 방향제어를 위한 머신러닝 기반 직경 7~8 m급TBM 시뮬레이터 개발(과제번호: 20240069-001)"의 일환으로 수행되었습니다.

참고문헌

  1. Gokceoglu, C., Bal, C., and Aladag, C.H., 2023, Modeling of Tunnel Boring Machine Performance Employing Random Forest Algorithm, Geotechnical and Geological Engineering, 41(7), 4205-4231
  2. Ho, J.C.J., 2020, In Pursuit of the Autonomous TBM, Tunnelling Journal, February/March 2020, 18-22.
  3. ITA (International Tunnelling and Undergrround Space Association), 2019, Mechanized Tunnelling: Guidelines for the Design of Segmental Tunnel Linings, ITA Working Group 2, Paris, France.
  4. Mao, S., Shen, X., and Lu, M., 2015, Virtual Laser Target Board for Alignment Control and Machine Guidance in Tunnel-Boring Operations, Journal of Intelligent & Robotic Systems, 79, 385-400.
  5. Mao, S., Shen, X., Lu, M., and Wu, X., 2013, Real-time tablet-based virtual reality implementation to facilitate tunnel boring machine steering control in tunnel construction, Proceedings of the 13th International Conference on Construction Applications of Virtual Reality, October 30-31, London, UK, 219-229.
  6. McClellan, A., 2020, Big data, Writ large: Technology to advance tunnel boring, TBM: Tunnel Business Magazine, April 2020, 24-27.
  7. Ninic, J., and Meschke, G., 2015, Model update and real-time steering of tunnel boring machines using simulation-based meta models, Tunnelling and Underground Space Technology, 45, 138-152.
  8. Park, B., Choi, S.-W., Lee, C., Kang, T.-H., Do, S., Lee, W.-Y., and Chang, S.-H., 2021, An Operating Model for and EPB Shield TBM simulator by the Correlation Analysis of Operational Actions and Mechanical Responses, Applied Sciences, 11(23), 11443.
  9. Park, B., Choi, S.-W., Lee, C., Kang, T.-H., Do, S., Lee, W.-Y., and Chang, S.-H., 2023, A Geometric Model for a Shield TBM Steering Simulator, Applied Sciences, 13(18), 10087.
  10. Samadi, H., Mahmoodzadeh, A., Mohammed, A.H., Alenizi, F.A., Ibrahim, H.H., Nematollahi, M., and Elhag, A.B., 2024, Application of Several Fuzzy-Based Techniques for Estimating Tunnel Boring Machine Performance in Metamorphic Rocks, Rock Mechanics and Rock Engineering, 57(2), 1471-1494.
  11. Wang, L., Zhu, F., Li, J., and Sun, W., 2023, A data-driven approach for modeling and predicting the thrust force of a tunnel boring machine, Journal of Zhejiang University-SCIENCE A, 24(9), 801-816.
  12. Yu, Z., Li, C., and Zhou, J., 2023, Tunnel Boring Machine Performance Prediction Using Supervised Learning Method and Swarm Intelligence Algorithm, Mathematics, 11(20), 4237.
  13. Zhang, Y.-K., Gong, G.-F., Yang, H.-Y., Chen, Y.-X., and Chen, G.-L., 2022, Towards autonomous and optimal excavation of shield machine: a deep reinforcement learning-based approach, Journal of Zhejiang University-SCIENCE A, 23(6), 458-478.
  14. Zheng, Z., Luo, K., Tan, X., Jia, L., Xie, M., Xie, H., Jiang, L., Gong, G., Yang, H., and Han, D., 2024, Autonomous steering control for tunnel boring machines, Automation in Construction, 159, 105259.