Adapting cartesian cut cell methods for flood risk evaluation

Morris, Andrew Gordon (2013) Adapting cartesian cut cell methods for flood risk evaluation. Doctoral thesis (PhD), Manchester Metropolitan University.

Preview

Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (5MB) | Preview

Abstract

Assessing the risks of flooding, and the effectiveness of mitigation strategies, is an important part of any river management strategy. This is improved greatly by the accurate simulation of surface hydraulics, and moving to two-dimensional simulations that can capture the dynamics of surface processes has clear advantages. Only with the increased availability of accurate topographic data, has this become practical for many cases. In chapter two the methodologies of river flood modeling are described. A number of concerns peculiar to that field are discussed. These include roughness parameterisation and heterogeneity of features on the domain. Finite Volume (FV) methods can simulate shallow water flows effciently. Their shock-capturing ability makes them especially useful for flash-flood events. A particular FV package AMAZON-CC is adapted, which uses an approximate twoshock Riemann solver over a regular orthogonal grid. The Cartesian cut-cell method allows solid regions to be included as local modifcations to individual cells. Wetting and drying causes particular difficulties with FV methods. The Volume to Free-Surface Relationship (VFSR) method provides a framework in which a variety of mitigation strategies can be adopted. A modifed form of this was used, adapted to the rectangular grid with a piecewise level interpretation of topography. Several strategies are tested, and the most successful adopted thereafter. The Cartesian cut-cell method was extended to include large scale but complex features as vector, or polygon data-sets. The approach here concentrates on the fluxes across boundaries, to represent linear features such as hedgerows and fences. Preliminary results are presented and analysed and compared to classical results for headloss from interaction with structures. Test cases based on physical expreriments and real-life events are successfully reproduced. These demonstrate the suitability of AMAZON as a tool to model inland flooding.