Suspended Sediment Estimation using Neural Network and

In purpose to performance programs soil protection and reduce sediment, also calculation and design of dam volume in introduction store dams, have necessity that evaluation and calculated the rate of sediment production in a watershed. Generally erosion and sediment transport is of most complex issues the hydrodynamic that not possible simply, determination equations governing because effects of various parameters. About attention in potential artificial intelligence in identify the relationship between variables input and output of a problem without taking the physics of the problem and because swoon of physical models and mathematical in modeling of sedimentary processes too, can be used in modeling sediment transport problem. The purpose of this study was to obtain suitable algorithms with using of artificial neural networks feed-forward back propagation, fitting and Cascade Forward back prop to intent estimate the sediment rate. In this intent for estimate the amount suspended sediment, used of discharge - precipitation and sediment data monthly. Intransitive of recitation that suspended sediment data in the output (Cham anjir station) is more appropriate of the distribution. Among the three networks used in this study was more appropriate to estimate the amount of sediment fitting network. Thirteen the algorithm used in this study was selected TRAINLM as the best algorithm, with a correlation coefficient R = 0.99, RMSE = 0.01.