Abstract
Wave-induced ship responses have a natural statistical variability, due to the nature of the ocean waves. This needs to be accounted for in ship design. The uncertainty and extreme values of the ship motions and loads depend on the test or simulation duration. This paper presents the experimental relation between uncertainty of root-mean-square (RMS) and most probable maximum (MPM) response values and the test duration, for a sailing ferry in steep bow-quartering waves. The RMSE uncertainty of the RMS of heave, pitch and relative wave elevation was 2–3% after 0.5 h test duration, reducing to 1% after 1 h. The uncertainty of the measured significant wave height was similar. For MPM values with a low uncertainty you need to perform longer tests. The RMSE uncertainty of the MPM of heave and pitch motions was 12% after 2 h, reducing to 5% after 5 h. This uncertainty is lower for relative wave elevation (2–6% after 2 h; 1–3% after 5 h), and higher for green water force on the accommodation (55–65% after 2 h; 20–40% after 5 h). We cannot use linear simulations to predict the RMS or MPM response values themselves, but we can use them to predict the required experimental duration for a given error tolerance of the RMS of (relative) motions.
