Application of artificial intelligence models in water quality forecasting

  • Yeon I
  • Kim J
  • Jun K
  • 17

    Readers

    Mendeley users who have this article in their library.
  • 13

    Citations

    Citations of this article.

Abstract

The real-time data of the continuous water quality monitoring station at the Pyeongchang river was analyzed separately during the rainy period and non-rainy period. Total organic carbon data observed during the rainy period showed a greater mean value, maximum value and standard deviation than the data observed during the non-rainy period. Dissolved oxygen values during the rainy period were lower than those observed during the non-rainy period. It was analyzed that the discharge due to rain fall from the basin affects the change of the water quality. A model for the forecasting of water quality was constructed and applied using the neural network model and the adaptive neuro-fuzzy inference system. Regarding the models of levenberg-marquardt neural network, modular neural network and adaptive neuro-fuzzy inference system, all three models showed good results for the simulation of total organic carbon. The levenberg-marquardt neural network and modular neural network models showed better results than the adaptive neuro-fuzzy inference system model in the forecasting of dissolved oxygen. The modular neural network model, which was applied with the qualitative data of time in addition to quantitative data, showed the least error.

Author-supplied keywords

  • ANFIS (adaptive neuro-fuzzy inference system)
  • Forecasting
  • Monitoring
  • Neural networks
  • Water quality

Get free article suggestions today

Mendeley saves you time finding and organizing research

Sign up here
Already have an account ?Sign in

Find this document

Authors

  • I. S. Yeon

  • J. H. Kim

  • K. W. Jun

Cite this document

Choose a citation style from the tabs below

Save time finding and organizing research with Mendeley

Sign up for free