Artificial neural network for bedload estimation in alluvial rivers

Abstract

A talented soft computational technique is applied to predict bedload sediment discharge in rivers. The feedforward–backpropagated (Levenberg– Marquardt algorithm) Artificial Neural Network (ANN) architecture is employed without any restriction to an extensive database compiled from measurements in 16 different rivers. Following the assessment of several possible models, two dimensionless parameters were selected from an initial set of five for the prediction of dimensionless bedload discharge. The ANN method demonstrated an encouraging performance compared to other standard methods. The mean value and standard deviation of the bedload predictions of theANN model differ only slightly from the measured values. The coefficient of determination and the efficiency coefficient of the ANN method are higher than those of the traditional methods. The performance of the currently used ANN method demonstrates its predictive capability and the possibility of generalization of the modeling to nonlinear problems for river engineering applications.

Keywords:

• Alluvial channels
• artificial neural network
• bedload
• Levenberg–Marquardt algorithm
• river engineering
• sediment transport