한국구조물진단유지관리공학회 논문집 (Journal of the Korea institute for structural maintenance and inspection)

  • 제27권3호
  • /
  • Pages.71-79
  • /
  • 2023
  • /
  • 2234-6937(pISSN)
  • /
  • 2287-6979(eISSN)
한국구조물진단유지관리공학회 (Korea Institute for Structural Maintenance Inspection)
권호 표지
DOI QR코드

DOI QR Code

임의 위치 구조물의 손상예측을 위한 인공지능 기반 지진가속도 추정방법

Method of Earthquake Acceleration Estimation for Predicting Damage to Arbitrary Location Structures based on Artificial Intelligence

  • 이경석 (부산대학교 지진방재연구센터) ;
  • 서영득 (부산대학교 지진방재연구센터) ;
  • 백은림 (부산대학교 지진방재연구센터)
  • 투고 : 2023.04.18
  • 심사 : 2023.05.02
  • 발행 : 2023.06.30
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

지진이 발생한 후 구조물의 안전성을 평가하기 위해 모든 교량 및 건축물에 지진가속도 및 변위를 계측하는 유지관리시스템을 구축하기는 효율적이지 않아, 이를 유지관리하기 위해서는 현장조사가 시행되며 조사범위가 넓은 경우 많은 시간이 소요된다. 그로 인해 2차 피해가 발생할 우려가 있으므로 신속한 개별 구조물의 안전성을 추정할 필요가 있다. 구조물의 지진 손상은 구조물에 인가된 지진력 정보와 구조해석모델을 이용하여 위험도평가 해석을 통해 예측할 수 있다. 이를 위해 지진 발생 시 임의위치에서 발생한 지진력을 추정할 필요가 있다. 본 연구에서는 국내 지진계측 기록과 선형추정방법 및 인공신경망 학습 방법을 활용한 임의위치의 지반 응답스펙트럼 및 가속도시간이력을 추정하는 방법들을 제안하고 적용성을 평가하였다. 선형추정방법의 경우 추정에 사용되는 인근 관측소의 위치가 가까울 경우 오차가 적었지만 멀어질 경우 오차가 크게 증가하였다. 인공신경망 학습 방법의 경우 동일한 조건에서 더 낮은 수준의 오차로 추정할 수 있었다.

It is not efficient to install a maintenance system that measures seismic acceleration and displacement on all bridges and buildings to evaluate the safety of structures after an earthquake occurs. In order to maintain this, an on-site investigation is conducted. Therefore, it takes a lot of time when the scope of the investigation is wide. As a result, secondary damage may occur, so it is necessary to predict the safety of individual structures quickly. The method of estimating earthquake damage of a structure includes a finite element analysis method using approved seismic information and a structural analysis model. Therefore, it is necessary to predict the seismic information generated at arbitrary location in order to quickly determine structure damage. In this study, methods to predict the ground response spectrum and acceleration time history at arbitrary location using linear estimation methods, and artificial neural network learning methods based on seismic observation data were proposed and their applicability was evaluated. In the case of the linear estimation method, the error was small when the locations of nearby observatories were gathered, but the error increased significantly when it was spread. In the case of the artificial neural network learning method, it could be estimated with a lower level of error under the same conditions.

키워드

과제정보

본 연구는 2022년도 행정안전부 재난 안전분야 연구개발사업(지역 맞춤형 재난안전 문제해결 기술개발지원)의 연구비 지원에 의해 수행되었습니다.

참고문헌

  1. Jeong, S. H., and Jang, W. S. (2017), Development of earthquake safety assessment software for cable-stayed bridges using seismic acceleration measurement data, Journal of the Korea Institute for Structural Maintenance and Inspection, KISMI, 21(1), 337-340. (in Korean) 
  2. Ahn, J. H., Jeong, J. W., Hong, Y. C., Park, J. B., and Choi, H. S. (2019), Proposal and Evaluation of Ground Response Spectrum Estimation Algorithm based on Seismic Observation Data, Journal of The Korea Institute for Structural Maintenance and Inspection, 23(5), 13-22. (in Korean)  https://doi.org/10.11112/JKSMI.2019.23.5.13
  3. Lee, K. S., Ahn, J. H., Park, H. Y., Seo, Y. D., and Kim, S. C. (2023), Seismic Acceleration Estimation Method at Arbitrary Position Using Observations and Machine Learning, KSCE Journal of Civil Engineering, 27(2), 712-726. (in Korean)  https://doi.org/10.1007/s12205-022-1235-6
  4. Abrahamson, N. A., and Somerville, P. G. (1996), Effects of the hanging wall and footwall on ground motions recorded during the Northridge earthquake, Bulletin of the Seismological Society of America, 86(1B), S93-S99. DOI: 10.1785/BSSA08601B0S93 
  5. Shiuly, A., Roy, N., and Sahu, R. B. (2020), Prediction of peak ground acceleration for Himalayan region using artificial neural network and genetic algorithm, Arabian Journal of Geosciences, 13(215). DOI: 10.1007/s12517-020-5211-5 
  6. Khosravikia, F., and Clayton, P. (2021) Machine learning in ground motion prediction, Computers and Geosciences, 148, 104700. DOI: 10.1016/j.cageo.2021.104700 
  7. Irvine, T. (2002), An introduction to the shock response spectrum, Vibrationdata. 
  8. Simard, P. Y., SteinKraus, D., and Platt, J. C. (2003), Best practices for convolutional neural networks applied to visual document analysis, Seventh International Conference on Document Analysis and Recognition, Edinburgh, UK, 958-963. DOI: 10.1109/ICDAR.2003.1227801. 
  9. Wei, X., Xie, C. W., Wu, J., and Shen, C. (2018), Mask-CNN: Localizing parts and selecting descriptors for fine-grained bird species categorization, Pattern Recognit, 76, 704-714. DOI: 10.1016/j.patcog. 2017.10.002. 
  10. Kolasa, M., Jozwicki, W., Wojtyna, R., and Jarzemski, P. (2007), Optimization of hidden layer in a neural network used to predict bladder-cancer patient-survival, SPA Conference, Poland. DOI: 10.1109/SPA.2007.5903302 
  11. Liu, Y., Starzyk, J. A., and Zhu, Z. (2007), Optimizing number of hidden neurons in neural networks, IASTED International Conference on Artificial Intelligence and Applications, Austria 1(1), 6. 
  12. Wagarachchi, N. M., and Karunananda, A. S. (2013), Optimization of multi-layer artificial neural networks using delta values of hidden layers, IEEE Symposium on Computational Intelligence, Cognitive Algorithms, Mind, and Brain (CCMB), Singapore. DOI: 10.1109/CCMB.2013.6609169 
  13. Zurada, J. M. (1992), Introduction to artificial neural networks systems, West Publishing Company.