Abstract:
The thesis deals with the numerical modelling of random wave propagation in the nearshore region. The mathematical modelling is based on the mild - slope equation.
The mild - slope equation has been considered in a linear elliptic form in order to take into account wave refraction, diffraction, and reflections, and therefore to allow for the applicability of the model without constraints even to areas like harbours where reflections play an important role. Owing to the fact that we are dealing with a boundary value problem, the mathematical methods for solving the linear system of equations generated after discretising the governing equation and the boundary conditions imposed are fundamental to developing an efficient numerical model. A range of mathematical methods for solving the linear elliptic form of the mild-slope equation as well as a range of appropriate radiation boundary conditions are analysed, implemented and compared. Comparisons with available results, experimental and numerical, for different cases of study, validate the model for monochromatic waves.
To overcome the difficulty of applying the linear elliptic model over large areas, a multigrid technique was deployed in a numerical model of a non-linear transformed form of the mild-slope equation in order to accelerate convergence. This technique was applied to a range of iterative solvers and satisfactory results were obtained.
A numerical model of a hyperbolic form of the mild-slope equation, an initial value problem, was also developed and extended in order to test a range of radiation boundary conditions. Results show that "sponge layers" provide the best means of dealing with radiation type boundary conditions.
The elliptic model was then further developed by extending it to random waves using the concept of superposition of spectral components. The importance of frequency and directional spreading was investigated. Based on the energy balance equation the phenomenon of breaking of irregular waves in the surf zone was incorporated, using two different approaches, in the model which was applied to cases that had been experimentally studied before. The results from this part of the work are unique and demonstrate the practical usefulness of the model.
