Application des réseaux neurones artificiels dans la relation "débit- hauteur " cas des oueds du Nord Algérien

Abstract

The estimate of discharge in natural rivers is an essential component of the majority of engineering analysis. For that purpose, a relationship between stage and discharge is established using field measurement. The found relations "Stage- discharge" are time-dependent and very often they exhibit random fluctuations, their relationship is not always unique. For that, artificial neural network method and multiple regression linear are examined on three hydrometric stations at North Algeria. The results show the interest of the two approaches, in particular in extrapolation, when the relation "Stage- discharges" is not univocal