A WAVE MODEL FOR SIMULATING VESSEL EFFECTING SHALLOW WATER WAVES IN A MARITIME SIMULATOR

Abstract

One important area of Maritime Simulations is the Wave e ects of the ocean. Whilst there is signi cant work done in modeling deep ocean waves, the area of shallow water wave modeling has taken precedence in recent times. The objective of this thesis is, to study the possibilities of overcoming the barrier of high computation shallow water wave models that cannot be used for real time applications. Quoting from literature there exists three main approaches to model ocean waves: 1) Geometrical Description Models(constructed using periodic functions), 2) Spectral Models (using empirical data from Oceanic researches) & 3) Physical based models (from Computational Fluid Dy- namics(CFD) based on numerical models). One key interest of this research is towards models that are able to achieve high accuracy in terms of the wave properties; thus guiding the user in making critical decisions and to predict close to real wave energy e ects. It is noteworthy that only the approach of numerically solved physical based models (from the 3 approaches mentioned above) can provide wave parameters with high accuracy. Hence, this experiment is constrained to such physically based numerical models. All such models available for shallow-water simulations are inherently limited to not considering the depth of the water volume. Thus, the main focus of this thesis is towards a wave model that considers "depth" as an integral parameter in its calculations. The challenge imposed by such accurate models is the computation complexity that results in long processing time. Hence, they are not the most suitable choice for real-time simulators. This thesis experiments the possibility of a solution to this problem by restricting the simulation area to only that which has an impact on the vessel and by introducing cell reductions. The results obtained within the duration of this study, reveal that maintaining an optimal accuracy with such mesh-restrictions is not feasible and further e orts needs to be put in terms of parallel/gpu processing, pro ling & etc.