한국건설관리학회논문집 (Korean Journal of Construction Engineering and Management)

  • 제20권6호
  • /
  • Pages.34-43
  • /
  • 2019
  • /
  • 2005-6095(pISSN)
  • /
  • 2465-9703(eISSN)
한국건설관리학회 (Korea Institute of Construction Engineering and Management)
권호 표지
DOI QR코드

DOI QR Code

도로포장의 유지관리 계획 수립을 위한 딥러닝 기반 열화 예측 모델 개발

Development of Deep Learning Based Deterioration Prediction Model for the Maintenance Planning of Highway Pavement

  • 이용준 (한국건설기술연구원 인프라안전연구본부) ;
  • 선종완 (한국건설기술연구원 인프라안전연구본부) ;
  • 이민재
  • Lee, Yongjun (Department of Infrastructure Safety Research, Korea Institute of Civil Engineering and Building Technology) ;
  • Sun, Jongwan (Department of Infrastructure Safety Research, Korea Institute of Civil Engineering and Building Technology) ;
  • Lee, Minjae (Department of Infrastructure Safety Research, Korea Institute of Civil Engineering and Building Technology)
  • 투고 : 2019.07.19
  • 심사 : 2019.09.23
  • 발행 : 2019.11.30
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

도로연장의 지속적인 증가와 공용기간이 상당히 경과한 노후 노선이 늘어남에 따라 도로포장에 대한 유지관리비용은 점차 증가하고 있어, 예방적 유지관리를 통해 비용을 최소화 하는 방안에 대한 필요성이 제기되고 있다. 예방적 유지관리를 위해서는 도로포장의 정확한 열화 예측을 통한 전략적 유지관리 계획 수립이 필요하다. 이에 본 연구에서는 고속도로포장 열화예측 모델 개발을 위해 딥러닝 모델 중 가장 보편적으로 많이 사용하는 심층신경망(DNN)과 시계열 데이터 분석에 강점을 가진 순환신경망(RNN)을 사용하였으며, 두 개의 모델의 성능을 비교 분석하여 우수한 모델을 제안하였다. RNN의 Vanishing Gradient Problem을 해결하기 위해 좀 더 복잡한 형태의 RNN구조인 LSTM(Long short-term memory circuits)을 사용하였다. 학습 결과, RNN-LSTM 모델의 RMSE 값이 0.102로 DNN모델보다 낮아 성능이 더 우수하였다. 또한, 대상구간의 시간경과별 평균 도로포장 상태 예측치와 실제 도로포장 상태 실측치의 비교를 통해 RNN-LSTM 모델의 높은 정확도를 검증하였다. 따라서 향후 고속도로 콘크리트 포장의 유지관리 계획 수립시 유지보수 수요 추정을 위한 열화 예측 모델로는 DNN 모델보다 시계열 분석에 강한 RNN-LSTM의 모델을 제안한다.

The maintenance cost for road pavement is gradually increasing due to the continuous increase in road extension as well as increase in the number of old routes that have passed the public period. As a result, there is a need for a method of minimizing costs through preventative grievance preventive maintenance requires the establishment of a strategic plan through accurate prediction of road pavement. Hence, In this study, the deep neural network(DNN) and the recurrent neural network(RNN) were used in order to develop the expressway pavement damage prediction model. A superior model among these two network models was then suggested by comparing and analyzing their performance. In order to solve the RNN's vanishing gradient problem, the LSTM (Long short-term memory) circuits which are a more complicated form of the RNN structure were used. The learning result showed that the RMSE value of the RNN-LSTM model was 0.102 which was lower than the RMSE value of the DNN model, indicating that the performance of the RNN-LSTM model was superior. In addition, high accuracy of the RNN-LSTM model was verified through the comparison between the estimated average road pavement condition and the actually measured road pavement condition of the target section over time.

키워드

참고문헌

  1. Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., and Ghemawat, S. (2016). "Tensorflow: Large-scale machine learning on heterogeneous distributed systems."arXiv preprint arXiv:1603.04467.
  2. Choi, S. (2018). "Development of Road Asset Management System based on Artificial Intelligence using Visual Information." Doctoral Thesis.
  3. Do, M. (2011). "Comparative analysis on mean life reliability with functionally classified pavement sections." International Journal of Highway Engineering, 14(5), pp. 11-19. https://doi.org/10.7855/IJHE.2012.14.5.011
  4. Do, M., Lee, Y., Lim, K., and Kwon, S. (2011). "Estimation of Performance and Pavement Life using National Highway Pavement Condition Index." KSCE Journal of Civil Engineering, 15(2), pp. 261-270. https://doi.org/10.1007/s12205-011-1065-4
  5. DOMITROVIC, J., DRAGOVAN, H., RUKAVINA, T., and DIMTER, S. (2018). "Application of an Artificial Neural Network in Pavement Management System." Tehnicki vjesnik, 25(2), pp. 466-473.
  6. Gharaibeh, N., and Darter, M. (2003). "Probabilistic analysis of highway pavement life for Illinois." Transportation Research Record 1823, No.03-4294, pp. 111-120. https://doi.org/10.3141/1823-13
  7. Kobayashi, K., and Do, M. (2010). "Estimation of Markovian transition probabilities for pavement deterioration forecasting." KSCE Journal of Civil Engineering, 14(3), pp. 343-351. https://doi.org/10.1007/s12205-010-0343-x
  8. Korea Expressway Corporation (2010). "Guideline for Exposure to Environment."
  9. Korea Expressway Corporation (2018). "2017 Investigation and analysis of highway pavement condition."
  10. Kwon, S., Jeong, K., and Sun, Y. (2012). "A Study on Decision Criteria of traffic volumes for Choosing of Modified Asphalt Pavement in Korea National Highway." International journal of highway engineering, 4(3), pp. 25-33.
  11. Han, D., Yoo, I., and Lee, S. (2017a). "Improvement of Multivariable, Nonlinear, and Overdispersion Modeling with Deep Learning: A Case Study on Prediction of Vehicle Fuel Consumption Rate." Intl. Journal of the Highway Engineers, 19(4), pp. 1-7.
  12. Han, D., Do, M., and Kim, B. (2017b). "Internal Property and Stochastic Deterioration Modeling of Total Pavement Condition Index for Transportation Asset Management." International journal of highway engineering, 19(5), pp. 1-11. https://doi.org/10.7855/IJHE.2017.19.5.001
  13. Hochreiter, S., and Schmidhuber, J. (1997). "Long Short-Term Memory." Neural Computation archive, 9(8), pp. 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735
  14. Lee, I., Lee, Y., Park, S., Cho, H., and Lee, M. (2018). "A Study on Utilization of Private Capital for Efficient Highway Pavement Management." Korean Journal of Construction Engineering and Management, KICEM 19(1), pp. 3-11. https://doi.org/10.6106/KJCEM.2018.19.1.003
  15. Lee, Y. (2013). "A Study on the Method of Establishing Road Maintenance Strategy Considering the Forecasting Traffic Demand." Master Thesis.
  16. Lee, Y., and Lee, M. (2016). "A Study on Estimating of Probability Distribution and Mean Life of Bridge Member for Effective Maintenance of the Bridge." Korean Journal of Construction Engineering and Management, 17(4), pp. 57-65. https://doi.org/10.6106/KJCEM.2016.17.4.057
  17. Lee, Y. (2019). "A Study on Construction of Highway Pavement Asset Management System based on Big Data." Doctoral Thesis.
  18. Loizos, A., and Karlaftis, M.G. (2005). "Prediction of pavement crack initiation from in-service pavements: A duration model approach." Journal of the Transportation Research Board, 1940, TRB, pp. 38-42. https://doi.org/10.1177/0361198105194000105
  19. MOLIT (2011). "Road Pavement Structure Design Manual."
  20. Park, J. (2013). "A Study on the Improvement of Business Process Efficient for Expressway Pavement Management." Master Thesis.
  21. Park, J. (2018). "Estimation of Electrical Loads Patterns by Usage in the Urban Railway Station by RNN." The Transactions of the Korean Institute of Electrical Engineers, 67(11), pp. 1536-1541. https://doi.org/10.5370/KIEE.2018.67.11.1536
  22. Suman, S., and Sinha, S. (2012). "Pavement Condition Forecasting Through Artificial Neural Network Modelling." International Journal of Emerging Technology and Advanced Engineering, 2(11), pp. 474-478.
  23. Yang, J., Gunaratne, M., Lu, J.J., and Dietrich, B. (2003). "Application of Neural Network Models For Forecating of Pavement Crack Index and Pavement Condition Rating." Florida Department of Transportation.
  24. Yang, J., Gunaratne, M., Lu, J.J., and Dietrich, B. (2005). "Use of recurrent Markov chains for modeling the crack performance of flexible pavements." Journal of Transportation Engineering, 131(11), pp. 861-872. https://doi.org/10.1061/(ASCE)0733-947X(2005)131:11(861)
  25. You, P., and Lee, D. (2002). "Methodology of a Probabilistic Pavement Performance Prediction Model Based on the Markov Process." International Journal of Highway Engineering, 4(4), pp. 1-12.