Paper Details

[ Downloads: 0 | Abstract Viewed: 62 | Pages: 10 ]

Abstract of the Article

Note: English CIVILICA is in its Trial Period so Full Texts can not be provided! Persian users can download it here

Artificial neural networks in river flow modelling (goals and limitations analysis)

Author:
[ Dastorani ] - Assistant professor, University of Yazd, Iran

Abstract:

In this paper it has been tried to analyze and evaluate the abilities and applicability of Artificial Neural Networks (ANN) in river flow modelling. This evaluation is carried out on the strength of four research projects completed by the writer in this field during last few years. Therefore the purpose here is to address the characteristics, strengths and limitations of this new computer technique for this specific application. The completed research projects includes: The application of ANN in ungauged catchments flow prediction; Real-time river flood prediction using different types of ANN; Application of ANN for hydrodynamic modelling results optimisation; and Evaluation of the applicability of ANN in hydrological data gap filling. It is clear that applicability of ANN varies depending on its structure as well as the type of problem in hand. For river flow modelling, dynamic type of ANN showed superior abilities in real-time flow prediction. It was also become clear that in most cases ANN can play an important role when it is coupled with some other modelling software, rather than using it alone. These points some other important points and characteristics of this technique is addressed in this paper.

Keywords: