DONATE

IAHR Document Library


« Back to Library Homepage « Proceedings of the 35th IAHR World Congress (Chengdu, 2013)

Rapid Tsunami Simulator on GPU Using Unstructured Mesh System

Author(s): Xinxiang Zeng; Takashi Nakamura; Ryosuke Akoh; Tadaharu Ishikawa

Linked Author(s):

Keywords: Tsunami simulator; GPGPU; Unstructured mesh; Finite volume method; 2011 Tohoku Earthquake

Abstract: The powerful Tsunami triggered by 2011 Tohoku earthquake (also known as the Great East Japan Earthquake) have attacked Japan’s northeast Pacific coast and caused tremendous devastation there. The run-up Tsunami were observed in rivers connecting with Pacific Ocean and even intruded into inland of northeast Japan which was expanded the extent of the damages indeed. In that context, it is necessary to give accurate early warning to safeguard the populations from Tsunami. Meanwhile, detailed numerical analysis of run-up Tsunami in rivers using finer mesh is also a topic should be studied immediately. However, on account of the huge spatial scale, the computational load of Tsunami simulation tends to be too large to afford. Furthermore, it is essential to use finer mesh to reproduce the complex topography of coastline and rivers accurately. In order to overcome these difficulties, a rapid numerical simulator has been constructed by applying the GPGPU technology in solving the 2-dimensional shallow water equations. An explicit finite volume scheme is used to discrete the shallow water equations on unstructured triangular mesh system. Owing to the use of unstructured triangle mesh, the complex boundary such as coastline and river banks is represented well. With the power of GPGPU, real-time simulation of Tsunami movement is achieved, even though using finer meshes, which are small to 15 meters in rivers. Specifically, compared with a CPU, about 50-times speed-up is enabled even using a single GPU. For the sake of validating this simulator, a simulation of run-up Tsunami in Tone River caused by 2011 Tohoku earthquake is performed. Both of the arriving time of first wave and the wave height fit observation data very well.

DOI:

Year: 2013

Copyright © 2024 International Association for Hydro-Environment Engineering and Research. All rights reserved. | Terms and Conditions