Dynamically adaptive tree grid technique for flood simulation

Abstract

The purpose of this research is to develop a computational algorithm with using adaptive tree grid technique for flood simulation. For the algorithm, we applied the finite volume method to solve the shallow water equations, and we designed the grid adaptivity method based on general rectangular tree grids. In order to test a potential of the algorithm in terms of computational costs, the model is tested and applied to simulate flood in Thailand. The results are compared with experimental data and the simulation on uniform grids. The comparison shows that the model can reduce number of grid cells and computational times without losing much accuracy in the results. Moreover, the algorithm is applied to simulate a simple model based on a diffusion equation for flood risk area assessment occurred by rainfall. The results show the accuracy of the simple model that converges to the shallow water model, and show the efficiency of the adaptive grid technique for the flood risk area assessment.