From ec117e3735c53854cef5b3873230c1a00cd98117 Mon Sep 17 00:00:00 2001 From: Mohd Afeef Badri <52162083+mohd-afeef-badri@users.noreply.github.com> Date: Fri, 23 Aug 2024 11:39:08 +0200 Subject: [PATCH] move to md for doc --- doc/ImposingDirichletConditions.html | 746 --------------------------- doc/ImposingDirichletConditions.md | 83 +++ 2 files changed, 83 insertions(+), 746 deletions(-) delete mode 100644 doc/ImposingDirichletConditions.html create mode 100644 doc/ImposingDirichletConditions.md diff --git a/doc/ImposingDirichletConditions.html b/doc/ImposingDirichletConditions.html deleted file mode 100644 index ff54378..0000000 --- a/doc/ImposingDirichletConditions.html +++ /dev/null @@ -1,746 +0,0 @@ - - -
- - -To enforce Dirichlet boundary conditions in FEM many methods are available, some of the common ones include:
Method | Reference | In ArcaneFEM |
---|---|---|
Weak penalty method | Babuška, 1973a | YES |
Penalty method | Babuška, 1973a | YES |
Row elimination | SOON | |
Row/Column elimination | SOON | |
Lagrange multiplier method | Babuška, 1973b | NO |
Nitsche’s method | Nitsche, 1971 | NO |
The first four are made available, so that one can choose the most appropriate method for Dirichlet boundary condition implementation. The word 'appropriate' in the preceding sentence is tricky, it depends on the physics, size of the problem, conditioning of the problem, used linear-solver, etc. So choose wisely.
What remains common in these methods is these are applied after one assembles the FEM linear system. As such one could see these methods as a post-processing step to be applied to the assembled linear-systems
These methods are not exact rather weak sense of applying Dirichlet boundary condition, however these are the most straightforward way to apply the Dirichlet boundary condition. These methods remain popular among practitioners and can be found in FEM packages such as MOOSE, FreeFEM, etc. Main benefits of these methods include, simplistic implementation, conserved non-zero structure (sparsity) of
In a nutshell, to apply weak penalty method for a DOF
here,
Similarly to apply penalty method for a DOF
here,
The logic is simple as
Generally,
To apply a Dirichlet condition
First is on the matrix
Second operation is on the vector
This method is more exact way of imposing Dirichlet boundary conditions , however the matrix
To apply a Dirichlet condition
First is on the matrix
Second operation is on the vector
This method is more exact way of imposing Dirichlet boundary conditions, however it is algorithmically more challenging to implement as