Efficient and Globally Convergent Minimization Algorithms for Small- and Finite-Strain Plasticity Problems

Publikation: Hochschulschrift/AbschlussarbeitDissertation


We present efficient and globally convergent solvers for several classes of plasticity models. The models in this work are formulated in the primal form as energetic rate-independent systems with an elastic energy potential and a plastic dissipation component. Different hardening rules are considered, as well as different flow rules. The time discretization leads to a sequence of nonsmooth minimization problems. For small strains, the unknowns live in vector spaces while for finite strains we have to deal with manifold-valued quantities. For the latter, a reformulation in tangent space is performed to end up with the same dissipation functional as in the small-strain case. We present the Newton-type TNNMG solver for convex and nonsmooth minimization problems and a newly developed Proximal Newton (PN) method that can also handle nonconvex problems. The PN method generates a sequence of penalized convex, coercive but nonsmooth subproblems. These subproblems are in the form of block-separable small-strain plasticity problems, to which TNNMG can be applied. Global convergence theorems are available for both methods. In several numerical experiments, both the efficiency and the flexibility of the methods for small-strain and finite-strain models are tested.


QualifizierungsstufeDr. rer. nat.
Gradverleihende Hochschule
Betreuer:in / Berater:in
  • Sander, Oliver, Hauptbetreuer:in
  • Meyer, Christian, Gutachter:in, Externe Person
Datum der Verteidigung (Datum der Urkunde)21 Aug. 2023
PublikationsstatusVeröffentlicht - 21 Aug. 2023
No renderer: customAssociatesEventsRenderPortal,dk.atira.pure.api.shared.model.researchoutput.Thesis