首页> 外文会议>Domain decomposition methods in science and engineering XIX >A Domain Decomposition Method Based on Augmented Lagrangian with a Penalty Term in Three Dimensions
【24h】

A Domain Decomposition Method Based on Augmented Lagrangian with a Penalty Term in Three Dimensions

机译:一种基于带惩罚项的增广拉格朗日域的域分解方法

获取原文
获取原文并翻译 | 示例

摘要

In our earlier work [4], a dual iterative substructuring method for two dimensional problems was proposed, which is a variant of the FETI-DP method. The proposed method imposes continuity on the interface by not only the pointwise matching condition but also uses a penalty term which measures the jump across the interface. For a large penalization parameter, it was proven that the condition number of the resultant dual problem is bounded by a constant independent of both the subdomain size H and the mesh size h. In this paper, we extend the method to three dimensional problems. For this extension, we consider two things; one is the construction of a penalty term in 3D to give the same convergence speed as in 2D and the other is how to treat the ill-conditioning of the subdomain problems due to a large penalization parameter. To resolve these two key issues, we need to be aware of the difference between 2D and 3D in the geometric complexity of the interface. Based on the geometric observation for the difference, we suggest a modified penalty term and a preconditioner aiming at reducing couplings between functions on the interface.
机译:在我们的早期工作中[4],提出了一种二维问题的双重迭代子构造方法,它是FETI-DP方法的一种变体。所提出的方法不仅通过逐点匹配条件在接口上施加了连续性,而且还使用了惩罚项来测量跨接口的跳跃。对于较大的罚分参数,已证明所产生的对偶问题的条件数受一个常数限制,该常数与子域大小H和网格大小h无关。在本文中,我们将方法扩展到三维问题。对于此扩展,我们考虑两件事;一种是在3D中构造惩罚项以使其具有与2D中相同的收敛速度,另一种是如何处理由于惩罚参数较大而引起的子域问题的不适性。要解决这两个关键问题,我们需要意识到2D和3D在界面几何复杂度方面的差异。基于对差异的几何观察,我们提出了一个改进的惩罚项和一个预处理器,旨在减少接口上功能之间的耦合。

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