Numerical Simulation Of Wind-wave Based On CFD

The simulation of waves which transform severely during its propagating process overan irregular topography in the coastal regions, is of great importance for the study ofhydrodynamics, in which wind-wave is a basic style for wave generation. In consideration ofthe fact that there no conclusion of the mechanism for wind-wave, as well as the complexityfor water-wave transformation and propagation that concern lots of dynamic mechanisms, it isessential to a combination of some models developed now in order to simulate the wind-wavesuccessfully. The SWAN model has improved a lot compared to the original style and itpossess enormous advantage to describe the wind effect exerted to waves. An extendedhyperbolic equations are developed in this paper, which can not only predict the combinedeffect of water wave transformation such as refraction, diffraction, shoaling, and reflection,but also take into account a rapidly varying topography, the nonlinear dispersion relation, thewind-energy input, the bottom friction dissipation and wave-break dissipation. The newlydeveloped model is obtained by the addition of some modifying terms corresponding to thedesired physical mechanism. Hence, we got the idea to combine SWAN model with thisdeveloped hyperbolic equations to solve the problem encountered.In this paper we firstly review the introduction and improvement of various numericalmodels used for wave simulation. A comparison of advantages and deficiency for them areprovided in details. We have to point out that the advent of CFD has accelerated the progressof wave simulation. Our second chapter provides a clear description for the mild slopeequation and the SWAN model. The original model developed by Berkhoff is quite rightunder the mild-slope assumption, which can consider the combined effect of water wavetransformation talked above effectively. However this elliptic model has to be extended todepict other physical mechanisms neglected by the original and the numerical scheme need tobe improved too. Based on the dynamic spectrum balance equation, SWAN model is able toinclude wave dynamic mechanisms by means of linear addition of resource terms and itprocesses the wind-wave effect appropriately. In the third chapter we developed an extendedhyperbolic mild-slope equations with the inclusion of modifying terms. To get numericalsolution for this extended equations successfully, we propose an way of combining the ADIscheme and the C-N scheme with the help of relaxation factor to speed up convergence. Wavedirections are unknown in the radiation boundary condition, a good candidate for suitableboundaries. Thus we give a P-C scheme to solve the irrotational equation for wave numbervector, which in turn provides all the angles. We have verified this extended equations in our forth chapter. The results computed from the derived model and the original hyperbolicequations (by Compland) are compared with the data got from physical experimentsrespectively. All of them are illustrated here and the figures demonstrate that our extendedmodel conforms more to physical experiments. At last, for the simulation of wind wave, wedevelop an effective scheme, that is, the combination of a self-embedded SWAN model withour newly derived equations. This method can take advantage of both strength in the SWANmodel which computes the wind-energy input exactly and the extended equations that depictwave transformation and propagation practically.