Calibration of low-fidelity ship motion programs through regressions of high-fidelity forces

Abstract

Lower-fidelity or reduced-order models based on constant coefficient ordinary differential equations provide computationally efficient means to generate ship motions. In these models, forces acting on a ship are typically approximated through closed-form expressions involving fixed coefficients that need to be set for the wave and operational conditions of interest. Here, this problem is addressed through the regressions of forces from higher-fidelity, for example, potential flow-based, solvers, that the lower-fidelity models aim to emulate. Numerical simulations are provided to illustrate the considered approach.