Research Group of Prof. Dr. Ira Neitzel
Research Projects
Current
Optimizing Fracture Propagation Using a Phase-Field Approach, Part II
Project SPP 1962 Phase 2, DFG priority program 1962, Phase II.
Sparse controls in optimization of quasilinear partial differential equations
Project C10, DFG SFB 1060.
Completed
Optimizing Fracture Propagation Using a Phase-Field Approach
Project SPP 1962, DFG priority program 1962.
Participants
Institut für Numerische Simulation der Rheinischen Friedrich-Wilhelms-Universität Bonn: Prof. Dr. Ira Neitzel
Fachbereich Mathematik der Technischen Universtität Darmstadt: Prof. Dr. Winnifried Wollner
Description
Within this project, we consider the numerical approximation and solution of control problems governed by a quasi-static brittle fracture propagation model. As a central modeling component, a phase-field formulation for the fracture formation and propagation is considered.
The fracture propagation problem itself can be formulated as a minimization problem with inequality constraints, imposed by multiple relevant side conditions, such as irreversibility of the fracture-growth or non-selfpenetration of the material across the fracture surface. These lead to variational inequalities as first order necessary conditions. Consequently, optimization problems for the control of the fracture process give rise to a mathematical program with complementarity constraints (MPCC) in function spaces.
Within this project, we intend to analyze the resulting MPCC with respect to it’s necessary and sufficient optimality conditions by means of a regularization of the lower-level problem and passage to the limit with respect to the regularization parameter. Moreover, we will consider SQP-type algorithms for the solution of this MPCC in function space and investigate its properties. Additionally, we will consider the discretization by finite elements and show the convergence of the discrete approximations to the continuous limit.
The simultaneous consideration of the inexactness due to discretization and regularization error will allow us to construct and analyze an efficient inexact SQP-type solver for the MPCC under consideration.