Wave run-up and overtopping of coastal structures have been extensively studied over the last 30 years to provide guidance for the construction of sea defences. Numerical models based on fluid flow equations can provide a useful aid in the design of these coastal defences. Computers have now advanced sufficiently to enable programs written to solve the flow equations to run on hardware that is readily available (e.g., desktop or laptop computers), thus giving engineers the ability to conduct multiple runs of an experiment, reconfigure the bathymetry, change the wave conditions and collect data from anywhere in the solution domain. An existing numerical model, AMAZON, based on the non-linear Shallow \Vater Equations (S\VE) was used to give wave height and overtopping discharges for a series of violent overtopping experiments. A second-order accurate highresolution finite-volume method was used to solve the SWE. The source terms that model the bed topography were treated using the Surface Gradient Method (SGM). The numerical model gave overtopping predictions to within 20% of the experimental overtopping discharges for cases where the wave ~onditions at the sea wall were not severely impacting. However, wave height comparisons showed that the SWE could not model wave propagation in intermediate depth water. The Boussinesq class of equations was chosen to extend the numerical modelling of wave propagation, run-up and overtopping into intermediate depth water. A hybrid finite-volumejfinite-difference solver was used to solve two different extended Boussinesq formulations, one of which was chosen to model a range of run-up and overtopping experiments. It was found'that the numerical model was able to model wave propagation where the typical depth to wavelength ratio was less than 0.35 for both regular and irregular waves. However, the numerical model was not able to accurately model breaking waves. Comparisons between overtopping discharges from the physical experiments and the numerical model showed that, in the majority of cases, the numerical model was able to provide predictions to within an absolute relative error of 3. It was found that as the gradient of the seawall increa'3ed, so did the accuracy of the numerical overtopping predictions.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:486867 |
| Date | January 2008 |
| Creators | Shiach, Jonathan Ben |
| Publisher | Manchester Metropolitan University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0075 seconds