A data-driven vessel motion model for offshore access forecasting

Gilbert, Ciaran and Browell, Jethro and McMillan, David; (2019) A data-driven vessel motion model for offshore access forecasting. In: OCEANS 2019 - Marseille. IEEE, FRA. ISBN 9781728114507 (https://doi.org/10.1109/OCEANSE.2019.8867176)

[thumbnail of Gilbert-etal-OCEANS-2019-A-data-driven-vessel-motion-model-for-offshore-access-forecasting]
Text. Filename: Gilbert_etal_OCEANS_2019_A_data_driven_vessel_motion_model_for_offshore_access_forecasting.pdf
Accepted Author Manuscript

Download (448kB)| Preview


Access forecasting for offshore wind farm operations is concerned with the prediction of conditions during transfer of personnel between offshore structures and vessels. Currently dispatch/scheduling decisions are typically made on the basis of single-valued forecasts of significant wave height from a numerical weather prediction model. The aim of this study is to move beyond the significant wave height metric using a data-driven methodology to estimate vessel motion during transfer. This is because turbine access is constrained by the behaviour of crew transfer vessels and the transition piece in the local wave climate. Using generalised additive models for location, scale, and shape, we map the relationship between measured vessel heave motion and measured wave conditions in terms of significant wave height, peak wave period, and peak wave direction. This is explored via a case study where measurements are collected via vessel telemetry and an on-site wave buoy during the construction phase of an east coast offshore wind farm in the UK. Different model formulations are explored and the best performing trained model, in terms of the Akaike Information Criterion, is defined. Operationally, this model is driven by temporal scenario forecasts of the input wave buoy measurements to estimate the vessel motion during transfer up to 5 days ahead.