Vulnerability criterion of nonlinear coupled resonance for semi-submersible platform using classification algorithm

Abstract

Resonance effect induced by nonlinear coupling is one of the crucial considerations in the design of offshore floating structures. In this study, a vulnerability criterion, which is a decision function used to judge the likelihood of nonlinear resonance of a semi-submersible platform, was established based on a support vector machine (SVM). A mathematical model considering the effects of bracings and nonlinear coupled motion was used as the simulator to generate training and test samples. Semi-analytical solutions of the mathematical model were deduced to speed up the solution process. Key features influencing the vulnerability of the semi-submersible platform were proposed and then selected according to their importance using a recursive feature elimination SVM. Finally, a simple decision function was obtained using the training samples to distinguish unstable platforms, and the performance of the decision function was evaluated using test samples. Furthermore, approaches to avoid nonlinear resonance were proposed based on the decision boundaries of the decision function. The results can contribute to the safety design of semi-submersible platforms and lead to a better understanding of the nonlinear coupled resonance of semi-submersible platforms.