Multi-model method for simulating 2D Surface-Piercing Wave-Structure Interactions

Zhang, X. (2018). Multi-model method for simulating 2D Surface-Piercing Wave-Structure Interactions. (Unpublished Doctoral thesis, City, Universtiy of London)

Text - Accepted Version
Download (7MB) | Preview


Wave-structure interactions play an important role on the design and maintenance of coastal and offshore constructions. Computational Fluid Dynamics (CFD) is a convenient tool for analyzing wave-structure interactions in costal and offshore engineering. The potential model and the viscous model are traditional mathematical models for wave-structure interactions, which have disadvantages in computational robustness, when they are applied individually. Therefore, recently, more and more multi-model methods are used for coupling the viscous model and the potential model together. So far, in the existing multi-model methods, the surface-piercing structure only exists in the viscous domain so that the viscous domain should be large enough. In order to improve the computational efficiency, some multi-model methods are developed, where the structure is considered in both viscous domain and Euler domain.

Firstly, by function-decomposition method, an Euler-viscous hybrid model is proposed. Comparing with the other function-decomposition hybrid models, a surface-piercing structure exists in both a large Euler domain and a small viscous domain. By this, the reflection, diffraction, and radiation waves from the structure can be considered in both two computational domains. Therefore, the computational efficiency can be enhanced remarkably. To couple the Euler model and the viscous model, complementary RANS equations are deduced, with complementary turbulence models. Corresponding boundary conditions are also developed for coupling. A relaxed scheme is proposed for damping the viscous effects and keeping free surface consistent. For wave interactions with moving structures, the transition of total forces acting on the structure from the viscous domain to the Euler domain is used to guarantee the same motion of structures in two domains.

Secondly, the function-decomposition Euler-viscous hybrid model is extended by domain-decomposition method. Then, function-decomposition method and domain-decomposition method are coupled together. The wave generation and propagation is solved in a potential domain. By this, the computational efficiency for wave-structure interactions in a large real wave tank can be improved.

Computational robustness of Euler-viscous hybrid model for surface-piercing wave-structure interactions are studied by some cases. It is found that the size of the viscous domain, the length of transition zone, and mesh resolution can affect computational precision. Computational efficiency is mainly affected by the size of the viscous domain. For extended Euler-viscous hybrid model, the distance before reaching the inlet boundary of the Euler domain plays a crucial role on computational accuracy and efficiency. Validations are done by comparing numerical results based on hybrid models, conventional RANS model and experimental results. It is shown that hybrid models own the same computational accuracy as the conventional RANS model. Furthermore, the computational accuracy can be improved remarkably. In some cases, more than 85% CPU time can be saved.

The hybrid models are applied to simulate wave interactions with a structure subjected to seabed effects. By comparing with numerical simulations based on the conventional RANS model, it is indicated that hybrid models can be also used on complex computational domain. Some properties of wave interactions with a floating structure subjected to a submerged structure are found by numerical simulations.

Item Type: Thesis (Doctoral)
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
T Technology > TC Hydraulic engineering. Ocean engineering
V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering
Divisions: City, University of London theses
School of Engineering & Mathematical Sciences > Engineering
City, University of London theses > School of Mathematics, Computer Science and Engineering theses

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics