A discontinuous-Galerkin-based immersed boundary method with non-homogeneous boundary conditions and its application to elasticity

We propose a discontinuous-Galerkin-based immersed boundary method for elasticity problems. The resulting numerical scheme does not require boundary fitting meshes and avoids boundary locking by switching the elements intersected by the boundary to a discontinuous Galerkin approximation. Special emphasis is placed on the construction of a method that retains an optimal convergence rate in the presence of non-homogeneous essential and natural boundary conditions. The role of each one of the approximations introduced is illustrated by analyzing an analog problem in one spatial dimension. Finally, extensive two- and three-dimensional numerical experiments on linear and nonlinear elasticity problems verify that the proposed method leads to optimal convergence rates under combinations of essential and natural boundary conditions. (C) 2009 Elsevier B.V. All rights reserved.Stanford Graduate FellowshipStanford Graduate FellowshipDepartment of the Army ResearchDepartment of the Army Research[W911NF-07-2-0027]National Institutes of Health (NIH)[U54GM072970]U.S. National Institutes of Health (NIH)NSF[CMMI-0747089]U.S. National Science Foundation (NSF)ONRONR[N000140810852]Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)FAPESP (Brasil)CNPq (Brasil)Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq