Numerical Analysis Seminar
Seminar information archive ~01/21|Next seminar|Future seminars 01/22~
| Date, time & place | Tuesday 16:30 - 18:00 002Room #002 (Graduate School of Math. Sci. Bldg.) |
|---|---|
| Organizer(s) | Norikazu Saito, Takahito Kashiwabara |
2026/02/10
16:30-18:00 Room #122 (Graduate School of Math. Sci. Bldg.)
Franz Chouly (Universidad de la República)
Finite Element Methods for the Elasto-Plastic Torsion Problem (English)
https://sites.google.com/g.ecc.u-tokyo.ac.jp/utnas-bulletin-board/
Franz Chouly (Universidad de la República)
Finite Element Methods for the Elasto-Plastic Torsion Problem (English)
[ Abstract ]
We will describe finite element methods for the elasto-plastic torsion problem, as a prototype of variational inequality with constraints on the gradient of the solution. This problem has been an object of study from the theoretical viewpoint since the 1970s. From the numerical analysis viewpoint, it remains difficult to obtain optimal a priori error estimates. We will present the numerical analysis and numerical results for three methods based on standard Lagrange finite elements of orden 1 or 2: a direct discretization and a Nitsche discretization of a modified variational inequality, when the source term is constant, and a penalty technique for a non constant source term. Particularly the Nitsche technique yields fully optimal a priori error bounds. This is a joint work with Patrick Hild (Université Paul Sabatier, Toulouse, France) and Tom Gustafsson (Aalto University, Finland).
[ Reference URL ]We will describe finite element methods for the elasto-plastic torsion problem, as a prototype of variational inequality with constraints on the gradient of the solution. This problem has been an object of study from the theoretical viewpoint since the 1970s. From the numerical analysis viewpoint, it remains difficult to obtain optimal a priori error estimates. We will present the numerical analysis and numerical results for three methods based on standard Lagrange finite elements of orden 1 or 2: a direct discretization and a Nitsche discretization of a modified variational inequality, when the source term is constant, and a penalty technique for a non constant source term. Particularly the Nitsche technique yields fully optimal a priori error bounds. This is a joint work with Patrick Hild (Université Paul Sabatier, Toulouse, France) and Tom Gustafsson (Aalto University, Finland).
https://sites.google.com/g.ecc.u-tokyo.ac.jp/utnas-bulletin-board/


Text only print
Full screen print

