한국산업정보학회논문지 (Journal of Korea Society of Industrial Information Systems)

한국산업정보학회 (Korea Society of Industrial Information Systems)
권호 표지
DOI QR코드

DOI QR Code

가속도 데이터 기반 교량 안전 판단을 위한 Edge AI 모델

Bridge Safety Determination Edge AI Model Based on Acceleration Data

  • 박진효 (동의대학교 인공지능학과) ;
  • 홍용근 (대전대학교 AI융합학과) ;
  • 윤주상 (동의대학교 인공지능학과)
  • Jinhyo Park ;
  • Yong-Geun Hong ;
  • Joosang Youn
  • 투고 : 2024.06.24
  • 심사 : 2024.08.02
  • 발행 : 2024.08.30
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

교량은 노후화와 지진, 유지보수 미비, 기상환경 등의 외부 요인에 의해 균열과 손상이 발생한다. 노후화 교량이 늘어나고 있는 상황에서 유지보수 작업을 진행하지 않으면 안전성이 저하되어 구조적 결함과 붕괴 문제가 발생할 수 있다. 이러한 문제를 예방하고 유지보수 비용을 절감하기 위해 교량의 상태를 모니터링하고 신속하게 대응할 수 있는 시스템이 필요하다. 이를 위해 기존의 연구에서 센서 데이터를 이용해 균열 위치와 정도를 파악하는 인공지능 모델이 제안되었다. 하지만 기존 연구에서 모델의 성능을 파악할 때 실제 교량의 데이터를 사용하지 않고 시뮬레이션을 통해서 교량의 형상을 제작하여 데이터를 획득하여 학습에 사용하였기 때문에, 실제 교량의 환경을 반영하지 못하고 있다. 본 논문에서는 실제 현장에서 발생하는 교량의 가속도 데이터를 활용하여 인공지능 기반 교량의 이상을 감지하는 '교량 안전 판단 Edge AI 모델'을 제안한다. 이를 위해 가속도 데이터에서 유효 데이터를 추출하기 위한 필터링 규칙을 새롭게 정의하고 이를 적용하는 모델을 구성하였다. 또한 현장에서 수집된 데이터를 기반의 제안된 교량 안전 판단 Edge AI 모델의 성능을 평가하였다. 그 결과 F1-Score가 최대 0.9565로 실제 교량의 데이터를 이용해 안전성을 판단할 수 있음을 확인할 수 있었고, 실제 충격 데이터를 유사한 데이터 패턴을 생성하는 규칙일수록 좋은 성능의 결과가 나왔다는 것을 확인하였다.

Bridges crack and become damaged due to age and external factors such as earthquakes, lack of maintenance, and weather conditions. With the number of aging bridge on the rise, lack of maintenance can lead to a decrease in safety, resulting in structural defects and collapse. To prevent these problems and reduce maintenance costs, a system that can monitor the condition of bridge and respond quickly is needed. To this end, existing research has proposed artificial intelligence model that use sensor data to identify the location and extent of cracks. However, existing research does not use data from actual bridge to determine the performance of the model, but rather creates the shape of the bridge through simulation to acquire data and use it for training, which does not reflect the actual bridge environment. In this paper, we propose a bridge safety determination edge AI model that detects bridge abnormalities based on artificial intelligence by utilizing acceleration data from bridge occurring in the field. To this end, we newly defined filtering rules for extracting valid data from acceleration data and constructed a model to apply them. We also evaluated the performance of the proposed bridge safety determination edge AI model based on data collected in the field. The results showed that the F1-Score was up to 0.9565, confirming that it is possible to determine safety using data from real bridge, and that rules that generate similar data patterns to real impact data perform better.

키워드

과제정보

이 논문은 과학기술정보통신부 및 정보통신기획평가원의 정보통신방송표준개발지원사업(RS-2024-00397768)과 부산광역시 및 (재)부산테크노파크의 BB21plus 사업으로 지원된 연구결과임.

참고문헌

  1. Abdeljaber, O., Avci, O., Kiranyaz, S., Gabbouj, M. and Inman, D. J. (2017). Real-time Vibration-based Structural Damage Detection Using One-dimensional Convolutional Neural Networks, Journal of Sound and Vibration, 388, 154-170. https://doi.org/10.1016/j.jsv.2016.10.043.
  2. Arnaiz, D., Alarcon, E., Moll, F. and Vilajosana, X. (2023). On Anomaly-Aware Structural Health Monitoring at the Extreme Edge, IEEE Access, 11, 90227-90253. https://doi.org/10.1109/ACCESS.2023.3306958.
  3. Bank, D., Koenigstein, N. and Giryes, R. (2021). "AutoEncoders", arXiv:2003.05991, https://doi. org/10.48550/arXiv.2003.05991
  4. Chen, S., Yu, J. and Wang, S. (2020). One-dimensional Convolutional Auto-encoder-based Feature Learning for Fault Diagnosis of Multivariate Processes, Journal of Process Control, 87, 54-67. https://doi.org/10.1016/j.jprocont.2020.01.004.
  5. Kim, S., Choi, Y., Park, S., Kwon, O. and Shim, H. (2022). Nuclear Power Plant Severe Accident Diagnosis Using Deep Learning Approach, J ournal of Korea Society of Industrial Information Systems, 27(6), 95-103. http:// dx.doi.org/10.9723/jksiis.2022.27.6.095.
  6. Korea Institute of Civil Engineering and Building Technology(KICT)(2023). Yearbook of Road Bridge and Tunnel Statistics, Seoul, Ministry of Land, Infrastructure and Transport.
  7. Korea Infrastructure Safety Corporation (2019). Development of AI-based Dynamic Response Evaluation Model Using Big Data to Ensure Reliability of Accelerometer Frequency Analysis, Seoul, Korea Infrastructure Safety Corporation.
  8. Lee, D., Lee, J. and Yang, S. (2023). Comparative Study of AI Models for Reliability FunctionEstimation in NPP Digital I&C System Failure Prediction, Journal of Korea Society of Industrial Information Systems, 28(6), 1-10. http://dx.doi.org/10.9723/jksiis.2023.28.6.001.
  9. Lee, K., Jeong, S., Sim, S. and Shin, D. (2019). A Novelty Detection Approach for Tendons of Prestressed Concrete Bridges Based on a Convolutional AutoEncoder and Acceleration Data, Sensors, 19(7), 1633. https://doi.org/10.3390/s19071633.
  10. Lee, K., Shim, S. and Shin, D. (2021). Analysis of The Influence of The Number And Position of Accelerometers on Deep Learning-Based Bridge Damage Estimation, The Journal of the Korean Society of Hazard Mitigation, 21(5), 183-190. https://doi.org/10.9798/KOSHAM.2021.21.5.183.
  11. Lee, M., Nam, K. and Lee, C. (2019). Crack Detection on the Road in Aerial Image using Mask R-CNN, Journal of Korea Society of Industrial Information Systems, 24(3), 23-29. http://dx.doi.org./10.9723/jksiis.2019.24.3.023.
  12. Li, L., Liu, G., Zhang, L. and Li, Q. (2021). FS-LSTM-Based Sensor Fault and Structural Damage Isolation in SHM, IEEE Sensors Journal, 21(3), 3250-3259. https://doi.org/10.1109/JSEN.2020.3022099.
  13. Lin, S., Wang, S., Liu, T., Liu, X. and Liu, C. (2023). Accurate Measurement of Bridge Vibration Displacement via Deep Convolutional Neural Network, IEEE Transactions on Instrumentation and Measurement, 72, 1-16. https://doi.org/10.1109/TIM.2023.3291786.
  14. Li, Z., Lin, W. and Zhang, Y. (2023). Real-time Drive-by Bridge Damage Detection Using Deep Auto-encoder, Structures, 47. https://doi.org/10.1016/j.istruc.2022.11.094.
  15. Moravvej, M. and El-Badry, M. (2024). Reference-Free Vibration-Based Damage Identification Techniques for Bridge Structural Health Monitoring-A Critical Review and Perspective, Sensors, 24(876) https://doi.org/10.3390/s24030876.
  16. Oh, B. K. and Kim, J. (2021). Multi-Objective Optimization Method to Search for the Optimal Convolutional Neural Network Architecture for Long-Term Structural Health Monitoring, IEEE Access, 9, 44738-44750. https://doi.org/10.1109/ACCESS.2021.3057687.
  17. Park, J., Lee, T., Hong, Y. and Youn, J. (2023). Intelligent Bridge Safety Prediction Edge System, KIP S Transactions on Computer and Communication Systems, 12(3), 357-362. https://doi.org/10.3745/KTCCS.2023.12.12.357.
  18. Sun, H., Song, L. and Yu, Z. (2023). A Deep Learning-based Bridge Damage Detection And Localization Method, Mechanical Systems and Signal P rocessing, 193. https://doi.org/10.1016/j.ymssp.2023.110277.
  19. Talaei, S., Zhu, X., Li, J., Yu, Y. and Chan, T. H. T. (2023). Transfer Learning Based Bridge Damage Detection: Leveraging Time-frequency Features, Structures, 57. https://doi.org/10.1016/j.istruc.2023.105052.
  20. Wah, W. S. L., Chen, Y., Roberts, G. W. and Elamin, A. (2017). Damage Detection of Structures Subject to Nonlinear Effects of Changing Environmental Conditions, Procedia Engineering, 188, 248-255. https://doi.org/10.1016/j.proeng.2017.04.481.
  21. Yang, J. and Peng, Z. (2020). Beetle-Swarm Evolution Competitive Algorithm for Bridge Sensor Optimal Placement in SHM, IEEE Sensors Journal, 20(15), 8244-8255. https://doi.org/10.1109/JSEN.2019.2934996.
  22. Zhang, Y. and Yuen, K-V. (2022). Review of Artificial Intelligence-based Bridge Damage Detection, Advances in Mechanical Engineering, 14(9). https://doi.org/10.1177/16878132221122770.
  23. Zou, J. Z., Yang, J. X., Wang, G. P., Tang, Y. L. and Yu, C. S. (2021). Bridge Structural Damage Identification Based on Parallel CNN-GRU, IOP Conference Series: Earth and Environmental Science, 626, 012017. https://doi.org/10.1088/1755-1315/626/1/012017.