Ocean Wave Prediction Using Large-Scale Phase-Resolved Computations

Abstract

A direct phase-resolved simulation tool for large-scale nonlinear ocean wavefield evolution, which is named SNOW, has been developed. Unlike the phase-averaged model, it solves the primitive Euler equation and preserves the phases of the wavefield during its nonlinear evolution. Therefore, the detailed descriptions of the free surface and the kinematics of the wavefield are obtained. To provide realistic and representative wavefields for ship motion analyses, we have computed an ensemble of three-dimensional (3D) wavefields (of typical domain size of O(103~4) km2) based on initial JONSWAP spectra. The statistical properties of the synthetic wavefields are computed and compared with theory and experimental measurements to study long-time sea spectrum evolution. SNOW simulations have been used to identify and characterize the occurrence statistics and dynamical properties of extreme wave events. We confirm that linear theory significantly under predicts the probability of large rogue wave events, especially for sea states with narrow spectra bandwidth and narrow directional spreading angle. A new phase-resolved wave prediction capability, with the incorporation of multiple hybrid (satellite/radar/lidar/wave-probe) sensed wave data as initial input, for deterministic short time O(Tp) prediction of ocean waves in deep water close to real time in a region with relatively small scale (~O(1) km×O(1) km) for a single ship handling is also developed. The validity and efficacy of SNOW in reliably predicting nonlinear ocean wavefield evolution is demonstrated and verified.