【24h】

GMRES acceleration of subiteration for fluid-structure interactions

机译:GMRES加速子迭代的流固耦合

获取原文
获取原文并翻译 | 示例

摘要

The numerical solution of fluid-structure interaction problems is of great relevance in many disciplines of engineering and science. Monolithic solution methods for fluid-structure interaction typically employ subiteration, i.e., multiple fluid-structure iterations. Although for certain problems the subiteration method is an excellent solver, for other problems it converges only slowly or even diverges. In this paper, we analyze the convergence behaviour of the subiteration method and investigate the application of under-relaxation and GMRES acceleration in order to improve its convergence behaviour. The Krylov acceleration constructs the search directions from available intermediate solution vectors already calculated in the subiterations. Therefore our proposed strategy is cheap and easily implemented in existing codes which use subiteration as a solver. The subiteration method can then be considered as a preconditioner to the Krylov subspace method. Numerical results for a model problem show that convergence difficulties are nicely mitigated by under-relaxation and GMRES acceleration, which demonstrates that our proposed strategy is capable to render the subiteration method more robust and efficient.
机译:流固耦合问题的数值解在工程和科学的许多学科中都具有重要的意义。用于流体-结构相互作用的整体解决方案方法通常采用子迭代,即多次流体-结构迭代。尽管对于某些问题,子迭代法是一种出色的求解器,但对于其他问题,它仅收敛缓慢甚至发散。在本文中,我们分析了子迭代方法的收敛行为,并研究了欠松弛和GMRES加速的应用,以改善其收敛性能。 Krylov加速度根据子迭代中已经计算出的可用中间解矢量来构造搜索方向。因此,我们提出的策略价格便宜,并且易于在使用子迭代作为求解器的现有代码中实现。然后可以将子迭代方法视为Krylov子空间方法的前提。模型问题的数值结果表明,欠松弛和GMRES加速可以很好地缓解收敛困难,这表明我们提出的策略能够使子迭代方法更加健壮和有效。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号