| Title |
A Development of Flood Mapping Accelerator Based on HEC-softwares |
| Authors |
김종천(Kim, JongChun) ; 황석환(Hwang, Seokhwan) ; 정종호(Jeong, Jongho) |
| DOI |
https://doi.org/10.12652/Ksce.2024.44.2.0173 |
| Keywords |
홍수; 홍수지도; HEC-RAS; RAS Mapper; 인공지능 Flood; Flood map; HEC-RAS; RAS mapper; AI |
| Abstract |
In recent, there has been a trend toward primarily utilizing data-driven models employing artificial intelligence technologies, such as machine learning, for flood prediction. These data-driven models offer the advantage of utilizing pre-training results, significantly reducing the required simulation time. However, it remains that a considerable amount of flood data is necessary for the pre-training in data-driven models, while the available observed data for application is often insufficient. As an alternative, validated simulation results from physically-based models are being employed as pre-training data alongside observed data. In this context, we developed a flood mapping accelerator to generate flood maps for pre-training. The proposed accelerator automates the entire process of flood mapping, i.e., estimating flood discharge using HEC-1, calculating water surface levels using HEC-RAS, simulating channel overflow and generating flood maps using RAS Mapper. With the accelerator, users can easily prepare a database for pre-training of data-driven models from hundreds to tens of thousands of rainfall scenarios. It includes various convenient menus containing a Graphic User Interface(GUI), and its practical applicability has been validated across 26 test-beds. |