Meandering open channel flow roughness coefficient estimation via support vector machine

Natural channels such as rivers serve as a major source of water for drinking, irrigation and industrial uses. Almost all the natural channels meander. Reliable estimation ofdischarge capacity of a natural channel depends on selection of proper value of roughness in terms of Manning’s n. Evaluation of Manning’s n for a meandering channel is a complex procedure because of its dependence on many geometrical, hydraulic and surface parameters of the channel. In this paper the capability of Support Vector Machine (SVM) as a data driven approach was investigated in roughness coefficient estimation in meandering open channels. The results were compared with well-established methods available in the literature and statistical error criteria were used for evaluating the accuracy of the models. The obtained results revealed that the proposed technique performed quite well compared to commonly used formulas. It was deduced that, in practice SVM model can be used as a suitable and effective method to predict the non-linear relationship between roughness coefficient and the nondimensional factors affecting it.

Keywords:

Empirical formulas , Meandering open channel , Roughness coefficient , SVM.