Sedimentary flows like water flow follows various statistical distributions. Many factors impact the sedimentary transfer and quantitative equations.
The estimation of rivers Sedimentary is one of the significant problems in designing hydraulic structures. Moreover this estimation is used in decreasing the harmful effects of Sediment in water reservoirs, purifying drinkable water, and so on.
The main purpose of this study is the prediction of flood Sedimentary in Merreq River in Kermanshah State using artificial neural networks. The rainfall of four stations, relative humidity percent, and maximum and minimum temperatures are the input of this suggested ANN. The output will be the daily amount of suspense sedimentary.
Scrutinizing various neural networks Topology, the best one was selected.
It becomes clear that a neural network with 13 nodes at a hidden layer produces the least errors. Goodness of fit, Root means square error, Mean absolute error and Maximum error are used to control the accuracy of calculation.
The calculation shows that the accuracy of neural network meets an acceptable level of estimation of daily flood sediment. Moreover comparing with other methods this approach needs less information and is done in a short time. Consequently a neural network is a good device for estimation of river sediments.
A sensitivity analysis is done for determining the impact of each input node, which influence the final results. Sensitivity analysis leads us to omit less important input. This action, though has little effect on the network accuracy, decreases the amount of input information as well as the cost of data gathering.