Separation and quantitative evaluation for the nonlinear effects in the motion response of floating structures

Abstract

Considering the influence of structural types and environmental conditions on the operation of floating structures, studying the nonlinear effects of their motion responses is a necessary prerequisite for conducting characteristic analyses, comprehensive risk assessment, motion control, et al. In this paper, an approach to separating nonlinear motion components is proposed to examine the nonlinear effects on the motion responses. To quantitatively analyze these nonlinear effects, a relationship between the operational environment and the structural responses is established to separate nonlinear components independently of structural model information. The main contributions of this study are listed as follows: (1) the memory depth of the nonlinear model is comprehensively determined by the Akaike-Bayesian joint information criterion; (2) an improved Kalman filtering method is developed to improve the identification effectiveness and accuracy of nonlinear kernel functions. Two numerical examples, including a nonlinear polynomial and a semi-submersible platform, are used to verify the correctness and applicability of the proposed method for the separation of nonlinear components. Results demonstrate that good agreement between the separated components obtained by the proposed method and the theoretical solutions is achieved at a degree of more than 98%, with the maximum normalization error marked in millimeters. Finally, physical experiment of a semi-submersible platform subjected to regular and irregular waves is carried out to further validate the proposed approach. Experimental results show that the proposed approach can effectively separate and evaluate the nonlinear components from the motion responses of the floating structures under operational states, paving the way for the development of efficient quantitative techniques to assess nonlinear effects widely present in various ocean engineering applications.