理化学研究所 計算科学研究センター

メニュー
メニュー

第190回 第3部

第190回 第3部
日時: 2020年2月3日(月)、14:35 - 15:30
場所: R-CCS 6階講堂

・講演題目:Scalable fixed mesh method for structure and fluid simulations
・講演者:西口 浩司(複雑現象統一的解法研究チーム)
※発表・スライド共に英語

講演要旨: 詳細を見る

This presentation focuses on a scalable Eulerian finite volume method for structure and fluid-structure interaction (FSI) problems that are difficult to simulate in industry. In conventional simulation methods used in industry, structure problems are computed with Lagrangian methods, while fluid problems are computed with Eulerian methods. However, this approach poses three computational problems.
Firstly, we have to spend lots of labor and time to generate a computational mesh and modify the low-quality mesh for complex geometries. Secondly, data mapping/transfer between different mesh/solvers are required in fluid-structure interaction problems. Thirdly, Lagrangian methods using unstructured mesh are challenging to obtain good scalability in the massively parallel environment.
Considering the background as mentioned above, in my work, an Eulerian finite volume method using a scalable hierarchical Cartesian mesh method has been developed. The present method has the following three advantages. The first one is good scalability in a massively parallel computing environment. Secondly, we can quickly generate the computational mesh of complex geometries, such as a car body. In my talk, I will demonstrate the structural simulation of a car body, which is spatially discretized by approximately 200 million cells and was computed using 104,520 cores on the K computer. Thirdly, the proposed Eulerian method can also compute fluid-structure interaction problems in a unified manner.
In future work, we plan to conduct car crash simulations using many patterns of multi-material vehicle structures to study ultralight vehicle structures, and fluid-structure interaction simulation of buildings under high-wind conditions for disaster prevention and mitigation.