Behavioral study of stable channels combined (river) using artificial neural networks

In order to predicting the response of of alluvial rivers applied to to artificial and natural changes Perhaps the first step is to identify and predict the stability of the river s geometry. Most cross-natural streams or rivers is such a complex channel and the main channel rivers under flood and non-flood is different steady. The main channels of rivers due to variations in flow rate and deposition, slope, width and depth to regulate such changes and to reach new mode of dynamic stability. Canal design and optimization of cross sections of river engineering is one of the most sophisticated and foundations due to the complexity and dynamic nature the phenomena of empirical criteria, semi experimental and design and statistics do not adequately determine the dimensions of the regime is stable channel . In this article a new model based on artificial neural network system using the wide field data is instructed and tested. The correlation coefficient of artificial neural networks in stage three parameters for width, depth and slope of the rivers is 0.8845-0.9433-0.9536 respectively. The results of the model with results of experimental and semiexperimental equations and methodology proposed priority on previous empirical methods have been confirmed.

Keywords:

sustainable canals , artificial neural networks , cross rivers