The variability of wave-in-deck loading in random seas: a data-driven approach

This study investigates wave-in-deck (wid) loading in random sea-states. These are among the most severe forces experienced by offshore structures and a critical factor in assessing their reliability. Earlier research demonstrated that wid loading is fundamentally determined by a momentum-exchange process governed by the properties of both the incident wave and the topside structure. However, observations made in quasi-deterministic (qd) focused waves did not fully explain the larger variability (uncertainties) of the loading observed in realistic random sea-states, which is an essential consideration for a full reliability analysis.

To address this gap, wid loading measurements were conducted across a broad range of random sea-states using a realistic model topside structure. The resulting large dataset was analysed using a data-driven methodology to identify the key factors influencing maximum wid loads. The findings confirm the significance of incident wave momentum. Importantly, the results highlight the increased influence of the spatial and temporal evolution of the incident waves relative to the topside. It is evident that the more diverse forms of wave evolution in random seas account for the increased variability in wid loads. This was further demonstrated through a novel data-driven model, which accurately reproduced short-term wid loading distributions using only a small set of inputs derived from free-field wave data. The general applicability of the model across different sea-states without re-training allows more efficient reliability assessments, significantly reducing the need for extensive model testing.