Sensitivity analysis of post-disturbance roll stability for a high-speed planing craft

High-speed planing craft are susceptible to short-duration disturbance moments from waves, wind, or asymmetric loading. Compared to the instantaneous peak during disturbance, the roll evolution in the free-response phase after disturbance removal more accurately reflects navigation safety: the hull may exhibit persistent mean bias or insufficient oscillation decay. This study establishes a computational fluid dynamics method based on URANS/VOF/DFBI using the Naples Systematic Series planing craft C1 model. A short-duration disturbance moment is applied and removed, yielding the post-disturbance roll response under six-degree-of-freedom motion. Latin hypercube sampling generates samples in a five-dimensional parameter space (mass, beam, deadrise angle, longitudinal and vertical centre of gravity). Two indicators are extracted: time-averaged roll angle (mean bias) and roll oscillation amplitude (decay features). A Kriging response surface reconstructs patterns in the parameter space. Main effects and interaction analyses reveal significant differences: mean bias is more sensitive to beam and deadrise angle; mass modulates the effects of deadrise angle and vertical centre of gravity on bias. Roll oscillation amplitude is governed by longitudinal centre-of-gravity-centred interactions with mass and beam, highlighting the role of longitudinal mass distribution in oscillation decay. This study provides quantifiable patterns of post-disturbance roll stability assessment and adverse parameter identification.