Free roll decay modeling of a trimaran based on least square support vector machine

Abstract

A trimaran is a multihulled boat consisting of a main hull and two smaller outrigger hulls, attached to the main hull with lateral struts. There have been many studies highlighting that a trimaran takes many advantages over other types of hulls especially the roll performance, because of the greater resistance to rolling that the outrigger hulls offer. In order to study the rolling motion of a trimaran, an accurate roll motion model is needed. The rolling motion of a trimaran is inherently nonlinear due to both nonlinearities in the restoring moment and damping moment. This paper presents an estimation method to predict the linear and nonlinear roll damping parameters based on free roll decay experiments of a trimaran, whose outrigger hulls are movable both along the longitudinal and transverse direction. The estimation method is based on least square support vector regression (LS-SVR) theory. In order to verify the performance of this new estimation method, the least square (LS) estimation method is also applied to identify the parameters. Both methods show satisfying performances, however models generated by LS-SVR shows less accumulated errors than models generated by LS.Index Terms-trimaran; free roll decay; support vector regression; least square