LS-SVM and FCM assessment approach in estimation of pipeline scouring on a river bed

Abstract

The mechanism of flow around a circular structure such as pipeline is very complicated so that it is difficult to establish a general empirical model to provide accurate estimation for scour. River bed deformations due to underwater pipeline scour have been determined from an analysis of the least square support vector machine (LS SVM) and Fuzzy C-mean system (FCM). Finding the best data input combinations using Gamma Test (GT) was done. It was very clear from model identification in GT that four input variables have better prediction for modeling in this case. Also, sensitivity analysis shows that the flow depth and pipe diameter have a greater influence on equilibrium scour depth than the other parameters. Finally, it was indicated that sugeno modeling of FCM method has better prediction in comparison with LS SVM.