Journal of Korea Water Resources Association (한국수자원학회논문집)

Korea Water Resources Association (한국수자원학회)
권호 표지

Hydrological Forecasting Based on Hybrid Neural Networks in a Small Watershed

중소하천유역에서 Hybrid Neural Networks에 의한 수문학적 예측

  • 김성원 (동양대학교 토목환경공학과) ;
  • 이순탁 (영남대학교 토목도시환경공학부) ;
  • 조정식 (대구대학교 건설환경공학부)
  • Published : 2001.08.01
Download PDF
⟨ Previous Next ⟩

Abstract

In this study, Radial Basis Function(RBF) Neural Networks Model, a kind of Hybrid Neural Networks was applied to hydrological forecasting in a small watershed. RBF Neural Networks Model has four kinds of parameters in it and consists of unsupervised and supervised training patterns. And Gaussian Kernel Function(GKF) was used among many kinds of Radial Basis Functions(RBFs). K-Means clustering algorithm was applied to optimize centers and widths which ate the parameters of GKF. The parameters of RBF Neural Networks Model such as centers, widths weights and biases were determined by the training procedures of RBF Neural Networks Model. And, with these parameters the validation procedures of RBF Neural Networks Model were carried out. RBF Neural Networks Model was applied to Wi-Stream basin which is one of the IHP Representative basins in South Korea. 10 rainfall events were selected for training and validation of RBF Neural Networks Model. The results of RBF Neural Networks Model were compared with those of Elman Neural Networks(ENN) Model. ENN Model is composed of One Step Secant BackPropagation(OSSBP) and Resilient BackPropagation(RBP) algorithms. RBF Neural Networks shows better results than ENN Model. RBF Neural Networks Model spent less time for the training of model and can be easily used by the hydrologists with little background knowledge of RBF Neural Networks Model.

본 연구에서는 중소하천수계에서 수문학적 예측을 위하여 Hybrid Neural Networks의 일종인 반경기초함수(RBF) 신경망모형이 적용되었다. RBF 신경망모형은 4종류의 매개변수로 구성되어 있으며, 지율 및 지도훈련과정으로 이루어져있다. 반경기초함수로서 가우스핵함수(GKF)가 이용되었으며, GKF의 매개변수인 중심과 폭은 K-Means 군집알고리즘에 의해 최적화 된다. 그리고 RBF 신경망모형의 매개변수인 중심, 폭, 연결강도와 편차벡터는 훈련을 통하여 최적 매개변수의 값이 결정되며, 이 매개변수들을 이용하여 모형의 검증과정이 이루어진다. RBF 신경망모형은 한국의 IHP 대표유역중 하나인 위천유역에 적용하였으며, 모형의 훈련과 검증을 위하여 10개의 강우사상을 선택하였다. 또한 RBF 신경망모형과 비교검토하기 위하여 엘만 신경망(ENN)모형을 이용하였으며, ENN 모형은 일단게 할선역전파(OSSBP) 및 탄성역전파(RBP)알고리즘으로 이루어져 있다. 모형의 훈련과 검증과정을 통하여 RBF 신경망모형이 ENN 모형보다 양호한 결과를 나타내는 것으로 분석되었다. RBF 신경망모형은 훈련시키는데 시간이 적게 들고, 이론적 배경이 부족한 수문학자들도 쉽게 사용할 수 있는 신경망모형이다.

Keywords

References

  1. 건설교통부(1982-1992). 국제수문개발계획(IHP) 대표유역 연구조사 보고서
  2. Battiti, R.(1992). 'First and second order methods for learning : between steepest and descent Newton's method' Neural Computation, Vol. 4, No.2, pp. 141-166
  3. Broomhead, D.S. and Lowe, D.(1988). 'Multivariable functional interpolation and adaptive networks.' Complex Systems, Vol. 2, pp. 321-355
  4. Daniell, T.M.(1991). 'Neural networks- application in hydrology and water resources engineering. Int. Hydro. and Water Resour. Symp., Inst. of Engineers, Perth, Australia, pp. 797-802
  5. Demuth, H., and Beale, M.(199S). Neural network toolbox : for use with MATLAB user's guide, The Math Works Inc
  6. Duda, R.O., and Hart, P.E.(1973). Pattern Classification and Scene Analysis, Wiley, New York, NY
  7. Fernando, D.A.K, and Jayawardena. A.W.(l99S). 'Runoff forecasting using REF networks with OLS algorithm.' J. of Hydro. Eng., Vol. 3, No.3, pp. 203-209 https://doi.org/10.1061/(ASCE)1084-0699(1998)3:3(203)
  8. Hush, DR, and Home, B.G.(1993). 'Progress in supervised neural networks.' IEEE Signal Processing Magazine, Vol. 10, No. 1, pp. 6-39 https://doi.org/10.1109/79.180705
  9. Isobe, I., Ohkado, T., Hanyuda, H., and Gotoh, y(l994l. 'The development of a forecasting system of the water levels of rivers by neural networks.' J. Japan Soc. of Hydro. and Water Resour., Vol. 7, No.2, pp. 90-97
  10. Jain, A.K., and Dubes, R.C.(l988). Algorithms for clustering data, Prentice Hall, New Jergey
  11. Kim, S., and Lee, S.(2000). 'Forecasting of flood stage using neural networks in the Nakdong river, South Korea.' Proc., Watershed Management & Operations Management 2000, ASCE/EWRI, Fort Collins, CO
  12. Liong, S.Y., and Chan, WT.(l993l. 'Runoff volume estimates with neural network.' Proc., 3rd Int. Conf. in Application of AI to Civ. and Struct. Engrg. : Neural Networks and Combinational Optimization in Civ. and Struct. Engrg., Topping, B.H.V., and Khan, AI., eds., Civil Comp Press, U.K, pp. 67-70
  13. Lin, C.T., and Lee, C.S.G.(l996). Neural Fuzzy Systems : A Neuro-Fuzzy Synergism to Intelligent Systems, Prentice Hall, New Jersey
  14. Mason, J.C., Price, R.K., and Tem'me, A (l996). 'A neural network model of rainfall-runoff using radial basis functions' Hydro. Sci. J., Vol. 41. No.3, pp. 399-417
  15. Minns, AW., and Hall, M.].(l996l. 'Artificial neural network as rainfall runoff models.' Hydro. Sci. J., Vol. 43, No.3, pp. 281-294
  16. Moody, J.E., and Darken, C.J. (l989). 'Fast learning in networks of locally-turned processing units. ' Neural Computation, Vol. 1, pp, 281-294
  17. Nguyen, D.H. and Widrow, B,(1990). 'Neural network for self-learning control systems.' IEEE Contrl Systems Magazine, pp. 18-23 https://doi.org/10.1109/37.55119
  18. Poggio, T., and Girosi, F.(l990). ''Networks for approximation and learning.' Proceedings of the IEEE, Vol. 78, pp. 1481-1497 https://doi.org/10.1109/5.58326
  19. Powell, M.J,D.(1987). 'Radial basis functions for multivariable interpolation: A review.' in Algorithms for the Approximation of Functions and Data, Mason, J,C., and Cox, M.G., eds., Oxford, England : Clarenden Press, pp. 143-167
  20. Riedmiller, M., and Braun, M.(l993). 'A direct adaptive method for faster backpropagation learning The RPROP algorithm' Proceedings of the IEEE International Conference on Neural Networks, San Francisco https://doi.org/10.1109/ICNN.1993.298623
  21. Raman, H., and Sunilkumar, N.(l995). 'Modeling water resources time series using artificial neural networks.' Hydro. Sci. J., Vol. 40, No.2, pp, 145-162
  22. Smith, J., and Eli, R.N.(l995). 'Neural network models of rainfall-runoff process.' ]. Water Resour. Plng. and Mgmt., ASCE, Vol. 121, No.6, pp. 499-508 https://doi.org/10.1061/(ASCE)0733-9496(1995)121:6(499)
  23. Tou, J.T., and Gonzalez, RC.(l974). Pattern Recgnition Principles, Addison-Wesley, Reading, MA
  24. Tsoukalas, L.H., and Uhrig, R.H.(l997). Fuzzy and neural approaches in engineering, John Wiley & Sons Inc
  25. Wasserman, P.D.(l993) Advanced Methods in Neural Computing, Van Nostrand Reitrand, New York