Professor, Civil Engineering Department, National Institute of Technology, Kurukshetra
Online published on 22 July, 2014.
Estimation of maximum local scour on a bridge site is necessary for the safety and economical design of the bridge. Numerous formulae are available and almost the entire formulae are developed based on the laboratory data. A number of equations have been developed over the years to predict local scour depth using experimental data and a few have formed pier equations using field data taken from Landers and Muller (1999). This paper presents an alternative in the terms of linear regression and SVM (Poly & Rbf) techniques to the conventional regression based equation (HEC-18), regression equation proposed by (Azmathualla et al. 2010). Total 352 data sets of the field measurements were used to train and test the program. The outcome of this study suggest that the SVM based modeling can be employed as an alternate to regression equations HEC-18 and the equation proposed by Azmathualla et al. (2010) in predicting scour depth of bridge piers. Further, a sensitivity analysis is also carried out to study the influence of one parameter over the other.
Pier scour, regression, modeling, prediction, Support Vector Machines
