جستجو در تالارهای گفتگو

Investigation and Evaluation of Artificial Neural Networks in Babolroud River Suspended Load Estimation Original Article, C29 Kia E, Emadi A.R, Fazlola R. Journal. Civil Eng. Urban. 3(4):183-190. 2013 ABSTRACT:Estimation of sediment load is a priority of the river management, dam's reservoirs and generally water projects. Because of nonlinear structure of sediment phenomena, the classical and common methods like sediment rating curve is not able to estimate sediment rate correctly. In recent years, artificial intelligence methods such as ANNs are recommended for solving the nonlinear problems and to achieve closer result to the actual data. In the present study, various combinations of flow discharge and sediment rate in present and past days were used as input parameters, while the suspended sediment load was used as output of the model. Then, by use of MATLAB software and different Neural Network such as MLP, RBF and GRNN, optimum architecture of networks is obtained based on four statistical indices viz. mean square error, mean bias error, modeling efficiency and determination coefficient for Ghoran Talar station in Babolroud River to compare with sediment rating curve. The results showed that MLP with combination of current discharge for estimate of current sediment has a better precision than other two neural networks. Also, the use of artificial neural network has a better performance than sediment rating curve method and recommended for river suspended load estimation. Keywords: Suspended Load, Perceptron, Neural Network, Rating Curve منبع: [Hidden Content] دانلود: [Hidden Content].,29-183-190.pdf J. Civil Eng. Urban.,29-183-190.pdf