Author(s): Haruki Matsui; Kenji Kawaike; Kazuki Yamanoi; Takahiro Koshiba
Linked Author(s): Kenji Kawaike
Keywords: Numerical model; Inundation flow; Hydraulic experiment; Infiltration into buildings
Abstract: In recent years, precise flood simulations in urban areas have become more crucial to deal with increasing severe water-related disasters. The current standard methods of flood simulations proposed by the Japan’s national government represent buildings with void ratios and drag force coefficients of each grid, assuming instantaneous inundation inside the buildings. However, since inside the buildings would be gradually inundated in actual situations, this method might fail to accurately represent the synthetic flood situations. Therefore, in order to develop a precise simulation model of fluvial inundation in highly urbanized areas, taking into account the inundation inside the buildings, one new simulation method has been proposed and compared with experimental measurement results and conventional fine unstructured grid model in this study. The conventional method, the Building Gap (BG) model, employs unstructured grids aligning building boundaries with grid edges. The Trichel’s theorem is applied to represent the inundation inside the building through small openings in building walls. One new method, the Building in Mesh (BM) model, is proposed here, assuming virtual building area depending on building occupancy rates within a grid. It divides a grid into the virtual building portion and the other area, and the exchanging discharge between them is also calculated using the Trichel’s theorem to express gradual inundation inside the buildings. To validate these simulation models, we constructed a physical urban model for hydraulic experiments following a real building arrangement using individual building models with small holes to represent the inundation inside the buildings. The experiments were conducted at a 1/100 scale of a densely populated urban area, measuring water depth at representative points and capturing the inundation area from the above the experimental set-up. Comparing the results of the simulation models with those of the hydraulic model experiments, the BG model, despite having larger computational load, reproduced the experimental results better than BM model, while the BM model, with lower computational load, demonstrated reasonable results. In addition to that, adopting unstructured grids and separating road grids from residential blocks ones, it was confirmed that the simulation results of the BM model can be much improved.
Year: 2024