Journal of Korea Water Resources Association (한국수자원학회논문집)

Korea Water Resources Association (한국수자원학회)
권호 표지
DOI QR코드

DOI QR Code

Prediction of water level in a tidal river using a deep-learning based LSTM model

딥러닝 기반 LSTM 모형을 이용한 감조하천 수위 예측

  • Jung, Sungho (Department of Disaster Prevention and Environmental Engineering, Kyungpook National University) ;
  • Cho, Hyoseob (Water Resources information center of Han River Flood Control Office, Ministry of Environment) ;
  • Kim, Jeongyup (Water Resources information center of Han River Flood Control Office, Ministry of Environment) ;
  • Lee, Giha (Department of Disaster Prevention and Environmental Engineering, Kyungpook National University)
  • 정성호 (경북대학교 과학기술대학 건설방재공학과) ;
  • 조효섭 (환경부 한강홍수통제소 수자원정보센터) ;
  • 김정엽 (환경부 한강홍수통제소 수자원정보센터) ;
  • 이기하 (경북대학교 과학기술대학 건설방재공학과)
  • Received : 2018.09.11
  • Accepted : 2018.10.23
  • Published : 2018.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Discharge or water level predictions at tidally affected river reaches are currently still a great challenge in hydrological practices. This research aims to predict water level of the tide dominated site, Jamsu bridge in the Han River downstream. Physics-based hydrodynamic approaches are sometimes not applicable for water level prediction in such a tidal river due to uncertainty sources like rainfall forecasting data. In this study, TensorFlow deep learning framework was used to build a deep neural network based LSTM model and its applications. The LSTM model was trained based on 3 data sets having 10-min temporal resolution: Paldang dam release, Jamsu bridge water level, predicted tidal level for 6 years (2011~2016) and then predict the water level time series given the six lead times: 1, 3, 6, 9, 12, 24 hours. The optimal hyper-parameters of LSTM model were set up as follows: 6 hidden layers number, 0.01 learning rate, 3000 iterations. In addition, we changed the key parameter of LSTM model, sequence length, ranging from 1 to 6 hours to test its affect to prediction results. The LSTM model with the 1 hr sequence length led to the best performing prediction results for the all cases. In particular, it resulted in very accurate prediction: RMSE (0.065 cm) and NSE (0.99) for the 1 hr lead time prediction case. However, as the lead time became longer, the RMSE increased from 0.08 m (1 hr lead time) to 0.28 m (24 hrs lead time) and the NSE decreased from 0.99 (1 hr lead time) to 0.74 (24 hrs lead time), respectively.

본 연구는 물리적 수리 수문모형의 적용이 제한적인 감조하천에서의 수위예측을 목적으로 하고 있으며, 이를 위해 한강 잠수교를 대상으로 딥러닝 오픈소스 소프트웨어 라이브러리인 TensorFlow를 활용하여 LSTM 모형을 구성하고 2011년부터 2017년까지의 10분 단위의 잠수교 수위, 팔당댐 방류량과 한강하구 강화대교지점의 예측조위 자료를 이용하여 모형학습(2011~2016) 및 수위예측(2017)을 수행하였다. 모형 매개변수는 민감도 분석을 통해 은닉층의 개수는 6개, 학습속도는 0.01, 학습횟수는 3000번로 결정하였으며, 모형 학습 시 학습정보의 시간적 양을 결정하는 중요한 매개변수인 시퀀스길이는 1시간, 3시간, 6시간으로 변화시키며 모의하였다. 최종적으로 선행시간에 따른 모의 예측능력을 평가하기 위해 LSTM 모형의 예측 선행시간을 6개(1 ~ 24시간)로 구분하여 실측수위와 예측수위와의 비교 분석을 수행한 결과, LSTM 모형의 최적의 성능을 내는 결과는 시퀀스길이를 1시간으로 하였을 때로 분석되었으며, 특히 선행시간 1시간에 대한 예측정확도는 RMSE는 0.065 m, NSE는 0.99로 실측수위에 매우 근접한 예측 결과를 나타내었다. 또한 시퀀스길이에 상관없이 선행시간이 길어질수록 모형의 예측 정확도는 2017년 전기간에 걸쳐 평균적으로 RMSE 0.08 m에서 0.28 m로 오차가 증가하였으며, NSE는 0.99에서 0.74로 감소하였다.

Keywords

SJOHCI_2018_v51n12_1207_f0001.png 이미지

Fig. 1. Recurrent neural network structure (Geron, 2017)

SJOHCI_2018_v51n12_1207_f0002.png 이미지

Fig. 2. LSTM Structure (Greff, 2017)

SJOHCI_2018_v51n12_1207_f0003.png 이미지

Fig. 3. Jamsu bridge location

SJOHCI_2018_v51n12_1207_f0004.png 이미지

Fig. 4. Time series of water level at the jamsu bridge in 2017

SJOHCI_2018_v51n12_1207_f0005.png 이미지

Fig. 5. Time series data sets for LSTM applications

SJOHCI_2018_v51n12_1207_f0006.png 이미지

Fig. 6. Water level time series and scatter plots for lead time of 1 hr ~ 6 hrs

SJOHCI_2018_v51n12_1207_f0007.png 이미지

Fig. 7. Water level time series and scatter plots for lead time of 9 hrs ~ 24 hrs

SJOHCI_2018_v51n12_1207_f0008.png 이미지

Fig. 8. Comparison of water levels for the specific rainfall events

Table 1. RMSE results of different sequence length cases

SJOHCI_2018_v51n12_1207_t0001.png 이미지

Table 2. NSE results of different sequence length cases

SJOHCI_2018_v51n12_1207_t0002.png 이미지

References

  1. Chen, W. B., Liu, W. C., and Hsu, M. H. (2012). "Comparison of ANN approach with 2D and 3D hydrodynamic models for simulating estuary water stage." Advances in Engineering Software, Vol. 45, No. 1, pp. 69-79. https://doi.org/10.1016/j.advengsoft.2011.09.018
  2. Chiang, Y. M., Chang, L. C., Tsai, M. J., Wang, Y.F., and Chang, F.J. (2010). "Dynamic neural networks for real-time water level predictions of sewerage systems-covering gauged and ungauged sites." Hydrology and Earth System Sciences, Vol. 14, No. 7, pp. 1309-1319. https://doi.org/10.5194/hess-14-1309-2010
  3. Geron, A. (2017). "Hands-on machine learning with Scikit-Learn and TensorFlow." O'Reilly Media.
  4. Greff, K., Srivastava, R. K., Koutnik, J., Steunebrink, B.R., and Schmidhuber, J. (2017). "LSTM: A search space odyssey." IEEE transactions on neural networks and learning systems, Vol. 28, No. 10, pp. 2222-2232. https://doi.org/10.1109/TNNLS.2016.2582924
  5. Hidayat, H., Hoitink, A. J. F., Sassi, M.G., and Torfs, P. J. J. F. (2014). "Prediction of discharge in a tidal river using artificial neural networks." Journal of Hydrologic Engineering, ASCE, Vol. 19, No. 8, 04014006. https://doi.org/10.1061/(ASCE)HE.1943-5584.0000970
  6. Hochreiter, S., and Schmidhuber, J. (1997). "Long short-term memory." Neural computation, Vol. 9, No. 8, pp. 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735
  7. Jung, S. H., Lee, D. E., and Lee, K. S. (2018). "Prediction of River Water Level Using Deep-Learning Open Library." Journal of Korean Society of Hazard Mitigation, Vol. 18, No. 1, pp. 1-11.
  8. Kim, S., and Tachikawa, Y. (2017). "Real-time river-stage prediction with artificial neural network based on only upstream observation data." Annual Journal of Hydraulic Engineering, JSCE, Vol. 62, pp. 1375-1380.
  9. Lee, E. R., Kim, W., and Kim, S. H. (2005). "Effect of Flood Stage by Hydraulic Factors in Han River." Journal of Korea Water Resources Association, Vol. 38, No. 2, pp. 121-131. https://doi.org/10.3741/JKWRA.2005.38.2.121
  10. Lee, G. H., Jung, S. H., and Lee, D. E. (2018). "Comparison of physicsbased and data-driven models for streamflow simulation of the Mekong river." Journal of Korea Water Resources Association, Vol. 51, No. 6, pp. 503-514. https://doi.org/10.3741/JKWRA.2018.51.6.503
  11. Lee, J. K., and Lee, J. H. (2010). "A Study on Water Level Rising Travel Time due to Discharge of Paldang Dam and Tide of Yellow Sea in Downstream Part of Paldang Dam." Journal of Korean Society of Hazard Mitigation, Vol. 10, No. 2, pp. 111-122.
  12. Lee, S. J., Ryoo, K. S., Lee, B. S., and Yoon, J. S. (2007). "Development of Regression Equation for Water Quantity Estimation in a Tidal River." Journal of Korean Society on Water Quality, Vol. 23, No. 3, pp. 385-390.
  13. Maier, H. R., and Dandy, G. C. (2000). "Neural networks for the prediction and forecasting of water resources variables: a review of modelling issues and applications." Environmental modelling & software, Vol. 15, No. 1, pp. 101-124. https://doi.org/10.1016/S1364-8152(99)00007-9
  14. Ministry of Land Infrastructure and Transport, Han River Flood Control Office (2017). Han River Forecast Yearbook.
  15. Oh, C. Y., Yang, D. H., Lee, J. H., and Won, Y. S. (2014). "Application of Artificial Neural Network Model for Flood Forecasting." Water for Future, Vol. 47, No. 5, pp. 50-62.
  16. Olah, C. (2015). "Understanding lstm networks." GITHUB blog, http://colah.github.io/posts/2015-08-Understanding-LSTMs/.
  17. Park, C. G., and Baek, K. O. (2017). "Reconsideration of evaluating design flood level at Imjin River estuary." Journal of Korea Water Resources Association, Vol. 50, No. 9, pp. 617-625. https://doi.org/10.3741/JKWRA.2017.50.9.617
  18. Song, C. G., Kim, H. J., and Lee, D. S. (2014). "Analysis of Flow Reversal by Tidal Elevation and Discharge Conditions in a Tidal River." Journal of the Korean society of Safety, Vol. 29, No. 6, pp. 104-110. https://doi.org/10.14346/JKOSOS.2014.29.6.104
  19. Seo, I. W., Song, C. G., and Lee, M. E. (2008). "Flow and Mixing behavior at the Tidal Reach of Han River." Journal of the Korean Society of Civil Engineers, KSCE, Vol. 28, No. 6B, pp. 731-741.
  20. Supharatid, S. (2003). "Application of a neural network model in establishing a stage-discharge relationship for a tidal river." Hydrological processes, Vol. 17, No. 15, pp. 3085-3099. https://doi.org/10.1002/hyp.1278
  21. Shin, Y. K., and Yoon, K. S. (2005). "The Spatial Distribution of Water Quality and Sediments Characteristics in the Han River Estuary." Journal of the Korean Geomorphological Association, Vol. 12, No. 4, pp. 13-23.