Ship Speed Loss Estimation Using Wave Spectrum of Encounter

An accurate estimation of ship speed loss is required to verify ship propulsion performance in actual sea conditions. In this study, ship speed loss was estimated using a neural network. Firstly, the directional spectra were extracted from the ocean wave hindcast database of Japan Weather Association. Then, the spectra were converted into wave spectra of encounter. Secondly, ship speed through water was modeled using ship monitoring data and spectra with a neural network. Thirdly, speed-power curves were estimated using the model. Effects of the wind and waves on ship speed were checked by considering the sensitivity of the model's outputs. Finally, ship speed loss was calculated using this model. The estimated ship speed loss was consistent with the ship speed loss obtained using ship monitoring data.