A Study on the Data Driven Neural Network Model for the Prediction of Time Series Data: Application of Water Surface Elevation Forecasting in Hangang River Bridge

시계열 자료의 예측을 위한 자료 기반 신경망 모델에 관한 연구: 한강대교 수위예측 적용

  • Yoo, Hyungju (Dept. of Civil Engineering, Hongik Univ.) ;
  • Lee, Seung Oh (Dept. of Civil Engineering, Hongik Univ.) ;
  • Choi, Seohye (Korea Institute of Civil Engineering and Building technology) ;
  • Park, Moonhyung (Korea Institute of Civil Engineering and Building technology)
  • 유형주 (홍익대학교 토목공학과) ;
  • 이승오 (홍익대학교 토목공학과) ;
  • 최서혜 (한국건설기술연구원 국토연구보전 본부) ;
  • 박문형 (한국건설기술연구원 국토연구보전 본부)
  • Received : 2019.05.27
  • Accepted : 2019.06.20
  • Published : 2019.06.30


Recently, as the occurrence frequency of sudden floods due to climate change increased, the flood damage on riverside social infrastructures was extended so that there has been a threat of overflow. Therefore, a rapid prediction of potential flooding in riverside social infrastructure is necessary for administrators. However, most current flood forecasting models including hydraulic model have limitations which are the high accuracy of numerical results but longer simulation time. To alleviate such limitation, data driven models using artificial neural network have been widely used. However, there is a limitation that the existing models can not consider the time-series parameters. In this study the water surface elevation of the Hangang River bridge was predicted using the NARX model considering the time-series parameter. And the results of the ANN and RNN models are compared with the NARX model to determine the suitability of NARX model. Using the 10-year hydrological data from 2009 to 2018, 70% of the hydrological data were used for learning and 15% was used for testing and evaluation respectively. As a result of predicting the water surface elevation after 3 hours from the Hangang River bridge in 2018, the ANN, RNN and NARX models for RMSE were 0.20 m, 0.11 m, and 0.09 m, respectively, and 0.12 m, 0.06 m, and 0.05 m for MAE, and 1.56 m, 0.55 m and 0.10 m for peak errors respectively. By analyzing the error of the prediction results considering the time-series parameters, the NARX model is most suitable for predicting water surface elevation. This is because the NARX model can learn the trend of the time series data and also can derive the accurate prediction value even in the high water surface elevation prediction by using the hyperbolic tangent and Rectified Linear Unit function as an activation function. However, the NARX model has a limit to generate a vanishing gradient as the sequence length becomes longer. In the future, the accuracy of the water surface elevation prediction will be examined by using the LSTM model.

HKBJBA_2019_v12n2_73_f0001.png 이미지

Fig. 1. Structure of ANN model

HKBJBA_2019_v12n2_73_f0002.png 이미지

Fig. 2. Structure of NARX model

HKBJBA_2019_v12n2_73_f0003.png 이미지

Fig. 3. The tidal level at Incheon and water surface elevation at Hangang RIver Bridge in July 2011

HKBJBA_2019_v12n2_73_f0004.png 이미지

Fig. 4. Discrete fourier transform of water surface elevation, tidal level and precipitation in July 2011

HKBJBA_2019_v12n2_73_f0005.png 이미지

Fig. 5. Hidden layer sensitivity analysis result

HKBJBA_2019_v12n2_73_f0006.png 이미지

Fig. 6. Time delay sensitivity analysis result

HKBJBA_2019_v12n2_73_f0007.png 이미지

Fig. 7. The water surface elevation distribution (left) and scatter plot (right) according to neural network model

Table 1. Input data characteristics

HKBJBA_2019_v12n2_73_t0001.png 이미지

Table 2. Model setup

HKBJBA_2019_v12n2_73_t0002.png 이미지

Table 3. Comparison of accuracy about ANN, RNN, and NARX model

HKBJBA_2019_v12n2_73_t0003.png 이미지


Supported by : Korea Agency for Infrastructure Technology Advancement(KAIA)


  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. 45(1): 69-79.
  2. Coulibaly, P. and Anctil, F. (1999). Real-time Short-term Natural Water Inflows Forecasting using Recurrent Neural Networks. Neural Networks. 1999. IJCNN'99. International Joint Conference on, IEEE: 3802-3805.
  3. 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. 62: 1375-1380.
  4. 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. 38(2): 121-131.
  5. 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 the Korean Society of Hazard Mitigation. 10(2): 111-122.
  6. Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D. and Veith, T. L. (2007). Model Evaluation Guidelines for Systematic Quantification of Accuracy in Watershed Simulations. ASABE. 50(3): 885-900.
  7. Shen, H. Y. and Chang, L. C. (2013). Online Multistep-ahead Inundation Depth Forecasts by Recurrent NARX Networks. Hydrol Earth Syst. Sci. 17: 935-945.
  8. Song, C. G., Kim, H. J., and Rhee, D. S. (2014). Analysis of Flow Reversal by Tidal Elevation and Discharge Conditions in a Tidal River. Journal of the Korean Society of Safety. 29(6): 104-110.
  9. Thirumalaiah, K. and Deo, M.C. (1998). Real-Time Flood Forecasting Using Neural Networks. Computer-Aided Civil and Infrastructure Engineering. 13(2): 101-111.
  10. 기상청 (2017). 이상기후 보고서.
  11. 김현일, 금호준, 한건연 (2018). 도시침수 해석을 위한 동적 인공신경망의 적용 및 비교. 대한토목학회논문집. 38(5): 671-683.
  12. 한강홍수통제소 (2016). 한강하천예보연감.