SVR-Based Parametric Identification for Parametric Roll Resonance of Ships in Longitudinal Regular Waves

Abstract

Parametric roll resonance, as a nonlinear phenomenon related to ship stability, is particularly apt to happen when a ship is sailing in longitudinal waves. It can generate sudden oscillation with large amplitude up to 30–40 degrees of roll and put the ship and crew in danger. To predict the parametric roll resonance of ships, a suitable model for describing this phenomenon is needed. In this paper, a nonlinear mathematical model considering the strong nonlinear coupling among the heave, roll, and pitch motions of ships is established, and support vector regression (SVR) is applied to identify the unknown damping and restoring coefficients in the mathematical model. To verify the accuracy and validity of SVR in parametric identification, a container ship is considered, and the coupled heave, roll, and pitch motions of the ship in longitudinal regular waves are simulated. Based on the simulated responses, the unknown coefficients in the mathematical model are identified by SVR. Then the coupled heave-roll-pitch motion of the container ship in regular waves is predicted by using the identified coefficients in comparison with the simulated data, and satisfactory agreement is achieved. From this study, it is concluded that SVR can be applied to identify the unknown coefficients in the nonlinear mathematical model for predicting the parametric roll resonance of ships in longitudinal regular waves.