Conference Agenda
Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).
|
Daily Overview |
| Session | |
|
SES 2-4-2: Model development and model performance Location: HSB1 Session Chair: Nils Kändler Session Chair: Margherita Evangelisti | |
| Presentation 2 | |
4:00pm - 4:15pm
Effective GPU use on a 1D-2D model for real-time flood prediction Taisei Corporation, Japan For rapid prediction of urban flood that occurs within a short period of time from the onset of heavy rainfall, the use of GPUs is considered effective. For 1D-2D flood simulation based on the shallow water equations, CPU/GPU hybrid models, in which the sewer (1D) flow is solved by the CPU and the ground surface (2D) flow is done by the GPU, have been used in previous studies. However, a full GPU model that solves both 1D and 2D parts on the GPU has not been examined. In this study, we created the full GPU model and compared the computational speed with the hybrid model under various calculation conditions. In computational domains with varying the numbers of the 1D and the 2D cells, the full GPU model became up to about 7 times faster by increasing the number of the 1D cells and decreasing that of the 2D cells. In calculations for an actual urban drainage area, the full GPU model was able to complete calculations within the 5-minute delivery interval of rainfall forecast data, while the hybrid model took about 15 minutes. In addition, the calculated flood areas generally covered observed ones. | |
