Session: 06-17-02 AI Technology for Ocean Engineering

Submission Number: 156465

AI-Powered Down-Scale Wave Hydrodynamic Analysis for Coastal Marine Structures 

Coastal marine structures are exposed to complex wave fields after nonlinear wave transformations due to the varying bathymetry and irregular coastlines. In the effort to define the coastal design sea state and analyze hydrodynamic loads on coastal structures, the NORA-SARAH procedure has been developed. The procedure devises a down-scale process, bringing the met-ocean information in the offshore hind-cast database NORA3 to the hydrodynamic load calculations of coastal marine structures in the complex coastal wave fields. The phase-averaging wave model SWAN, phase-resolving wave model REEF3D, the arbitrary Eulerian-Lagrangian (ALE) fast force calculation method in a fully nonlinear potential flow framework and the hydrodynamic coupling (HDC) are deployed to complete this cascade down-scale process. This work presents a data-driven study on complex coastal wave and wave load prediction based on the NORA-SARAH procedure. The work automates the procedure to produce a large quantity of down-scale simulations with various offshore met-ocean inputs in southern Norway close to Kristiansand. These synthetic simulation data of the coastal waves, including phase-resolving surface elevations and the ALE-based force estimations are used to train a neural network-based machine learning (ML) algorithm. Predictions of the coastal wave properties and wave loads derived from the ML are validated with the validation datasets. Tentative forecasting of coastal waves is made based on meteorological forecasting. This artificial intelligence (AI)-powered automated procedure, NORA-SARAH.AI, examines the capability, efficiency and possibility for operational phase-resolved ocean wave and wave load forecasting for coastal marine structures. 

Presenting Author: Widar Weizhi Wang Norwegian University of Science and Technology

Presenting Author Biography: Widar Weizhi Wang is a researcher in the development team of the open-source hydrodynamics suite REEF3D at the Norwegian University of Science and Technology where he is dedicated to the research on numerical wave modelling at different scales for a wide variety of topics on ocean engineering, coastal engineering, coastal protection and adaptions and marine renewable energy. He has also been working with data-driven wave prediction approaches with numerical wave simulations for accelerated efficiency.