Return to search

An efficient high-performance computing based three-dimensional numerical wave basin model for the design of fluid-structure interaction experiments

Fluid-structure interaction (FSI) is an interesting and challenging interdisciplinary area comprised of fields such as engineering- fluids/structures/solids, computational science, and mathematics. FSI has several practical engineering applications such as the design of coastal infrastructure (such as bridges, levees) subjected to harsh environments from natural forces such as tsunamis, storm surges, etc. Development of accurate input conditions to more detailed and complex models involving flexible structures in a fluid domain is an important requirement for the solution of such problems. FSI researchers often employ methods that use results from physical wave basin experiments to assess the wave forces on structures. These experiments, while closer to the physical phenomena, often tend to be time-consuming and expensive. Experiments are also not easily accessible for conducting parametric studies. Alternatively, numerical models
when developed with similar capabilities will complement the experiments very well because of the lower costs and the ability to study phenomena that are not feasible in the laboratory.

This dissertation is aimed at contributing to the solution of a significant component of the FSI problem with respect to engineering applications, covering accurate input to detailed models and a numerical wave basin to complement large-scale laboratory experiments. To this end, this work contains a description of a three-dimensional numerical wave tank (3D-NWT), its enhancements including the piston wavemaker for generation of waves such as solitary, periodic, and focused waves, and validation using large-scale experiments in the 3D wave basin at Oregon State University.

Performing simulations involving fluid dynamics is computational-intensive and the complexity is magnified by the presence of the flexible structure(s) in the fluid domain. The models are also required to take care of large-scale domains such as a wave basin in order to be applicable to practical problems. Therefore, undertaking these efforts requires access to high-performance computing (HPC) platforms and development of parallel codes. With these objectives in mind, parallelization of the 3D-NWT is carried out and discussed in this dissertation. / Graduation date: 2011

Identiferoai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/18829
Date11 October 2010
CreatorsNimmala, Seshu B.
ContributorsYim, Solomon C.
Source SetsOregon State University
Languageen_US
Detected LanguageEnglish
TypeThesis/Dissertation

Page generated in 0.0019 seconds