Author(s): Wentao Gong; Yasuyuki Shimizu; Toshiki Iwasaki
Linked Author(s): Yasuyuki Shimizu
Keywords: Adaptive mesh refinement (AMR); CIP; Flood modeling; Shallow water flow model; Computational time
Abstract: In this study, we propose an urban flood modeling framework using Adaptive Mesh Refinement (AMR) technique, which has been developed as an effective way of reducing the computational cost but guaranteeing the computational accuracy. AMR is able to refine or de-refine the computational grid resolution in space and time when necessary. We implement the AMR method to a two-dimensional shallow water flow model, which is solved by using the separation method and constrained interpolation profile (CIP) technique, for flood simulation. The proposed model is then applied to a case of urban flood in Sapporo, Japan. We invest igate various refinement criteria to invest igate the performance of flood modeling with AMR (i.e., accuracy and efficiency). The fixed fine grid model and fixed coarse grid model are also tested as a comparison with the AMR. The results show that the AMR saves the 47%-68.2% computational time with a reasonable accuracy of flood characteristics.