2D urban flood modelling using a numerical model of Delft3D flexible mesh

Abstract

In this additional thesis a 2D urban flood model in the downtown area of Shanghai was investigated. Therefore a numerical model of Delft3D flexible mesh (DFM) was used. Since the model was investigated from scratch, a lot of data analysis and pre-processing was required to have reasonable input data. As input data a surface elevation map, a map with roughness coefficients, the drainage capacity and the infiltration capacity was required. Furthermore data of two typhoon events was available, one event on September 13, 2013 and the other on August 8, 2012 which also was used as input data. For the model in DFM a rectangular grid with a resolution of 50 m was used. The same resolution was used for the input data. A simulation for both rainfall events was done and it was found that the locations with the most intense rainfall and the locations with the highest inundation depths agreed with each other. Furthermore it was found that the highest inundation depths occurred at the same time when the highest rainfall intensities occurred. For most locations the resulted inundation depths were smaller than 0.1 m but there were also several locations with higher inundation depths. Furthermore a sensitivity analysis for all input data was done. It was found that the roughness values had only very few influence on the inundation depths and it depends on the location if the depths increase or decrease with changed roughness values. A higher infiltration leaded to slightly smaller inundation depths, but most of the differences were smaller than 5 cm The highest influence on the inundation depths was found when the drainage capacity was changed. When no drainage system was available, the inundation depths almost doubled. Hence, inundation depths up to 1 m occurred. This can be explained by the fact, that the drainage of rainwater is one of the main functions of a sewer system which was completely neglected in this case. For Jingan district a more detailed unstructured grid was established. Therefore curvilinear grids for the roads and triangular grids between the roads were combined. It was found that this grid leaded to better results because of the more accurate flow directions.