Speaker: Alexis Mérigaud, Maynooth University Department of Electronic Engineering.
Title: "Real-time forecasting of ocean waves: a statistical perspective".
Abstract: Real-time prediction of free-surface elevation or wave
excitation force is necessary for a variety of prospective applications,
including noncausal optimal control of wave energy converters and the
detection of quiescent periods for offshore operations. Most approaches
rely on a deterministic, physical description of ocean waves, assuming
that the wave field comprises a finite number of wave components. In
this work, it is shown more rigorous and informative to adopt a
statistical perspective on ocean waves, assuming that the wave field is
composed of infinitely many, independent random contributions. With such
an approach, which has been well-known for decades by oceanographers,
the sea state is described as a Gaussian random field, of which the
spectrum can be relatively easily measured, estimated, or predicted. It
is shown that the statistical information contained in the spectrum can
be employed, to derive a statistically optimal predictor of the sea
surface, as well as a complete statistical characterisation of the
prediction error. Preliminary experiments using real sea wave data are
carried out.