As of 2018, 81.2% of the 17,240 reservoirs in Korea are aging facilities which are 50 years after construction. Reservoir failure often occurs since climate change and meteorological disasters recently have increased. For that reason, the flood hazard...
As of 2018, 81.2% of the 17,240 reservoirs in Korea are aging facilities which are 50 years after construction. Reservoir failure often occurs since climate change and meteorological disasters recently have increased. For that reason, the flood hazard maps were developed for some enormous dams to rank maintenance priority, but they reflect few geological characteristics and fluid dynamic attributes. When the flood is numerically analyzed, dynamic changes of fluids by geological characteristics have enormous impacts on estimating wave travels. Therefore, new reservoir failure model was conducted in this study to precisely predict downstream flood damage due to the failure of agricultural reservoirs. The prediction accuracy of the model was assessed by comparing the actual wave travel domain with the result. Limits of the previous model and novel flood wave analysis prediction method were presented by comparing the new model with the previous model. First, Structure from Motion (SfM) method, getting the 3D structure from 2D continuous images, and Unmanned Aerial Vehicle (UAV) method effective to gain close-up images were combined to develop topological data. The topological reproducibility rate of point cloud model according to photographing condition was examined. The optimum photography was gained when the resolution was 2.5 cm/pixel, the frontal overlap was 80%, side overlap was 72%, and the photography angle was 55°. Second, the inflow hydrograph of each stream section was estimated and set as boundary conditions of each point, to compare reservoir water level about to failure with the baseflow of the stream. Third, reservoir failure cause, failure duration, and breach shape were examined to analyze breach discharge hydrograph. Flood wave travels of each model according to reservoir water level about to failure were compared. The 1D model estimated the overrated flood wave travel domain calculated because the flood wave was interpreted only 1D propagation. Therefore, the 1D model could not calculate the flood wave travel domain in receptors. The 2D model used the digital surface model as topological data, combined 2D pixels and the highest elevation of each points, so streamline calibration was required. Several calculations were required to precisely make model, due to weir coefficient was empirical value. The 2D model result was 57% similar to the actual flood wave travel domain, which represents the dynamic analysis limit of the 2D model. 3D model required a few calibration tasks for some area and no repetitive calculation to set coefficients. The 3D model result was 95% similar to the actual flood wave travel domain, which represents the great accuracy of the model. Piping breach scenario result of the dam was less similar to the actual flood wave travel domain, so it could be assumed that the dam was breached immediately. In addition, it is considered that the analysis of the dam break flood wave for a large reservoir will show a greater difference from the existing model. Limits of the previous numerical models were determined to base on this study, and a new methodology was presented to make precise flood risk maps to rank reservoir maintenance priority.