Analytic network process for local scour formula ranking with parametric sensitivity analysis and soil class clustering

Abstract

Prediction of scour depth around bridge piers during flood events has been and continues to be regarded as a paramount concern for researchers of local scour, and many empirical formulas have been proposed. Because of the multiplicity and variability of these formulas, it remains extremely delicate to choose the correct formula among the many available. This study aims to develop a new framework to compare the different formulas currently used for the evaluation of local scour. For this purpose, 18 distinctly different formulas are selected and then evaluated using a large set of field-measured scour data. The rating is pronounced along three main streambed granulometric distributions. The validation process is performed using the Analytical Network Process approach (ANP), in which the already available conventional weights are re-evaluated and updated using various analyses, notably in terms of statistics, sensitivity, and correlation. The validation and comparison results of these 18 scour formulas reveal that the efficiency of a given particular formula depends on the type of stream soil studied. Furthermore, the scour formulas that include parameters having a geometrical dimension provide better performances.