Varying ocean wave statistics emerging from a single energy spectrum in an experimental wave tank

As it strongly impacts the design of offshore structures, the realistic reproduction of sea states in experimental and numerical wave tanks is of great interest to the ocean engineering community. The vast majority of wave qualification procedures rely on the accurate control of i) the wave energy spectrum and ii) the wave crest statistics at a target location in the domain. However the control of the wave field is strongly challenged by nonlinear phenomena such as breaking and high-order nonlinearities, which are at the origin of significant variations of the wave properties along the tank. Considering this issue, the most common industry methodologies focus on reproducing the wave energy spectrum at the target position. The wave crest statistics are compared to reference distributions. The present study aims to explore the limitations of such a practice, investigating in detail the wave field properties for various target locations over a long domain. We address the problem within the framework of deep water long-crested irregular waves. In this respect, using the Ecole Centrale de Nantes (ECN) experimental facility, a specific sea state is consecutively generated at three positions of the wave tank using a dedicated procedure based on wave maker motion iterative corrections. For such nonlinear wave conditions the wave crest statistics are known to be enhanced along the tank by high-order nonlinearities. As a result, configurations characterized by identical wave spectra lead to the generation of different wave crest distributions, revealing an increasing number of extreme events as the spectrum is generated farther from the wave maker. The data yielded by the study provide convincing evidence that controlling the wave field at a target location by correcting the sole energy spectrum is insufficient. Particular attention must be paid to the nonlinear spatial dynamics of the wave field in order to control the wave crest statistics.