Numerical modeling of dam break induced flow through multiple buildings in an idealized city

Abstract

This study presents two-dimensional dam break numerical modeling through an urban area. There are three approaches to defining a building in numerical modeling, as a wall boundary, contour, and Manning, which have been widely used. However, there is no research that compares and evaluates the accuracy of each approach. Thus, this study aims to determine the best approach, in term of accuracy, and evaluate its reliability by comparing the numerical modeling approach's results to experiment results. The model was developed based on the Saint-Venant equations, then solved using the MacCormack method with a numerical filter. The authors first utilized a single-building experiment to determine the best approach. The results indicated that the wall boundary approach had the highest accuracy based on MDAPE, with the lowest value of 13.47%, followed by Manning with 16.48%, and Contour with 21.46%. A multiple-building case is then used to evaluate the best approach's reliability. The results show high similarity to the experiment results. However, the wall boundary approach cannot be used when the building is submerged or has a height variation, since these approaches cannot define the building's height. Moreover, Manning approach cannot be considered as a building approach, since the approach does not reflect the behaviour of the building. Thus, the authors recommend the wall boundary approach when the building does not have a height variation and is not submerged, and the contour approach otherwise.