I am interested in a posteriori error estimation for the finite element method.
There are so many different types of a posteriori error estimation:
averaging-based (ZZ) estimates
equilibrated fluxes estimates
equilibrated residual estimates
I first worked with the residual-based a posteriori error estimation for the potential formulations in electromagnetism for my PhD-thesis, using the industrial code Code_Carmel.
Then I work in Inria in a postdoc position under the direction of Martin Vohralík, with the a posteriori estimate based on the equilibrated flux recontruction, using the scientific software Freefem++ .
That is why I have this occasion to work with Frédéric Hecht and why Martin would like to create this website.
Here you can find some implementation work about the a posteriori error estimation based on the equilibrated flux reconstruction in collaboration with Frédéric Hecht, as well as Martin Čermák.
The details about the theoretical part of the implementations work can be found here for the general a posteriori error estimation for the Laplace problem and the Stokes problem.
For the adaptive stopping criterion part, which is MORE INTERESTING !!!!
YOU CAN find it here for the $p$-laplace problem and here for the Stokes problem using an Uzawa algorithm.
A posteriori estimator for the discretization error
A posteriori estimator distinguishing different error components and adaptive stopping criteria
- The $p$-Laplace problem: distinguishing the algebraic, discretization, and linearization error components
- The Stokes problem with Uzawa algorithm: distinguishing the algebraic, discretization, and the Uzawa error components
return to main page