Dispersion Coefficient Prediction Using Empirical Models and ANNs

Abstract

The concentration of a conservative pollutant is changed along a river, as a result of transport processes. The dispersion coefficient is the most important parameter of mass transport in rivers. In this paper, the dispersion coefficient was estimated in a section of Axios River, with the analytical procedure of Fischer method, under different hydrological and hydrodynamic conditions. An empirical equation and a model of artificial neural networks (ANNs) for dispersion coefficient were proposed, based on the data estimated with analytical Fischer method. The dispersion coefficients predicted by the proposed models and other empirical equations reported in earlier studies were compared to the coefficients obtained in the present study. The most accurate equations for dispersion coefficient were used to predict the concentration of conservative toxic pollutants released instantaneously in Axios River upstream of the border of Greece-Former Yugoslav Republic of Macedonia (FYROM).