Author(s): Mohammad Zakermoshfegh; S. Ali Akbar Salehi Neyshabouri

Linked Author(s): Seyed Ali Akbar Salehi Neyshabouri

Keywords: Bridge Scour; Cohesive Sediment; Clayey Sand; Artificial Neural Network

Abstract: Scour at bridges has been studied extensively in the past for non-cohesive sediments. No model for scour depth estimation is addressed to account for the presence of cohesive materials in cases where bridges are founded in clayey sands. The presence of cohesive materials has made the scour depth prediction too complex and demanding. That is why the traditional data analysis methods have undesirable performance compared to those applied to non-cohesive materials in order to extract the complex and non-linear relationship governing the process. Artificial Neural Network is a distributed processing method, capable of recognizing the nonlinear relations and generalizing the learned knowledge in other situations. In this paper, the feed forward Multi Layer Perceptron Neural Network is applied to estimate the bridge pier scour depth in the clayey sands as a function of the clay content (CC) and approach Froude number (Fr). Results indicate the superior performance of this intelligent approach compared to the conventional methods. A design chart is also developed in which one of the three effective parameters can be estimated using two other predefined parameters.

DOI:

Year: 2010