National Technical University of Athens (NTUA)
A posteriori error analysis and optimal control for phase field equations
Abstract
dc:descriptionFirst, we prove a posteriori error estimates for the Allen-Cahn equation for low-order (implicit Euler) schemes in time combined with conforming linear finite elements (conforming FEM) in space. In particular, a posteriori estimates on the natural energy norm induced by the Ginzburg-Landau functional are proved using a carefully constructed test function. We apply elliptic reconstruction techniques in combination with continuity and energy arguments and spectral estimate of the linearized Allen-Cahn operator around the discrete solution. The main result also leads to an estimate in the natural energy norm, while the error estimate requires that a dependence condition of the discretization parameters with respect to the parameter 'ε' , the interface width, is satisfied. Then, a posteriori error estimates are proved for the numerical scheme consisting of the discontinuous Galerkin method in time and the interior penalty discontinuous Galerkin method in space. The condition, which must be satisfied, exhibits equally "good" 'e'-dependence compared to the corresponding condition of the simpler fully discrete scheme. A key role in the proof of the above estimates is the use of space-time reconstruction techniques for higher order problems. The main motivation for studying this method is the potential for substantial reduction of computational complexity by making use of few but "large" general shape elements in bulk regions (regions of different phase), while strongly refined elements in the vicinity of the interface layers.Finally, a related optimal control problem was studied. In particular, a tracking functional is minimized subject to the Allen-Cahn equation using distributed controls that satisfy point-wise control constraints . Part of the thesis is to study the existence of a minimum as well as the necessary and sufficient first- and second-order conditions. The main contribution is the proof of error estimates of the difference between local optimal controls and their discrete approximations and also estimates for the differences between the corresponding state and adjoint state and their discrete approximations under the limited regularity assumptions imposed by our optimal control setting.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Plaka, Dimitra
- Πλάκα, Δήμητρα
Subjects
dc:subject × 21- Εκ των υστέρων ανάλυση σφάλματος
- Allen-Cahn
- Εξισώσεις πεδίου φάσεων
- Βέλτιστος έλεγχος
- Υπό συνθήκες εκτιμήσεις
- Ασυνεχής μέθοδος Galerkin
- Μέθοδος πεπερασμένων στοιχείων
- Εκ των προτέρων εκτίμηση σφάλματος
- A posteriori error estimates
- Phase-field equations
- Optimal control
- Conditional assumptions
- Discontinuous time-stepping scheme
- Discontinuous Galerkin methods
- A priori error estimates
- Μαθηματικά
- Φυσικές Επιστήμες
- Εφαρμοσμένα μαθηματικά
- Mathematics
- Natural Sciences
- Applied Mathematics
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/51657