首页> 外文会议>International workshop on meshfree methods for partial differential equations >Exact Bounds for Linear Outputs of the Convection-Diffusion-Reaction Equation Using Flux-Free Error Estimates
【24h】

Exact Bounds for Linear Outputs of the Convection-Diffusion-Reaction Equation Using Flux-Free Error Estimates

机译:对流扩散反应方程的线性输出的精确界,采用无通量误差估计

获取原文
获取外文期刊封面目录资料

摘要

The Flux-free approach is a promising alternative to standard implicit residual error estimators that require the equilibration of hybrid fluxes. The idea is to solve local error problems in patches of elements surrounding one node (stars) instead of in single elements [1]. The resulting local problems are flux-free, that is the boundary conditions are natural and hence their implementation is straightforward. This allows precluding the computation and the equilibration of fluxes along the element edges. The domain decomposition is performed using a partition of unity strategy. The resulting estimates are much simpler from the implementation viewpoint, especially in the 3D cases, and provide upper bounds of the energy norm of the error (as well as the standard implicit residual estimators with equilibration of hybrid fluxes).In the past, the local flux-free problems have been solved using a finite element mesh inside each local subdomain. Consequently, the resulting estimates were asymptotic upper bounds (w.r.t. a reference solution) rather than exact upper bounds (w.r.t. the exact solution). Some effort has been devoted to recover exact upper bounds using the equilibrated hybrid fluxes approach. The idea is to solve the local problem using a dual formulation and to minimize the complementary energy [2], In this work, the same idea is employed to obtain exact upper bounds using the flux-free approach. The resulting estimates have similar features as their asymptotic version, while providing a guaranteed upper bound. This strategy is applied both to the primal and adjoint problem to recover guaranteed bounds for the quantity of interest.
机译:无通量方法是需要平衡混合通量的标准隐式残留误差估计量的有希望的替代方法。这个想法是要解决围绕一个节点(星形)的元素补丁而不是单个元素的局部错误问题[1]。由此产生的局部问题是无磁通的,也就是说边界条件是自然的,因此实现起来很简单。这允许排除沿单元边缘的通量的计算和平衡。使用统一策略的分区执行域分解。从实现的角度来看,得出的估计值要简单得多,尤其是在3D情况下,并提供误差能量范数的上限(以及混合磁通量均衡的标准隐式残差估计值)。 过去,使用每个局部子域内部的有限元网格解决了局部无通量问题。因此,所得的估计值是渐近上限(w.r.t.参考解),而不是精确的上限(w.r.t.精确解)。已经采取了一些努力来使用平衡的混合通量方法来恢复精确的上限。这个想法是使用对偶公式来解决局部问题,并使互补能量最小化[2]。在这项工作中,同样的想法被采用无通量方法来获得精确的上限。所得的估计值与其渐近版本具有相似的功能,同时提供了有保证的上限。该策略既适用于原始问题,也适用于伴随问题,以恢复所关注数量的保证范围。

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