NAJul 2, 2018
A hybrid isogeometric approach on multi-patches with applications to Kirchhoff plates and eigenvalue problemsThomas Horger, Alessandro Reali, Barbara Wohlmuth et al.
We present a systematic study on higher-order penalty techniques for isogeometric mortar methods. In addition to the weak-continuity enforced by a mortar method, normal derivatives across the interface are penalized. The considered applications are fourth order problems as well as eigenvalue problems for second and fourth order equations. The hybrid coupling enables the discretization of fourth order problems in a multi-patch setting as well as a convenient implementation of natural boundary conditions. For second order eigenvalue problems, the pollution of the discrete spectrum - typically referred to as 'outliers' - can be avoided. Numerical results illustrate the good behaviour of the proposed method in simple systematic studies as well as more complex multi-patch mapped geometries for linear elasticity and Kirchhoff plates.
NAMar 2, 2016
Simultaneous Reduced Basis Approximation of Parameterized Elliptic Eigenvalue ProblemsThomas Horger, Barbara Wohlmuth, Thomas Dickopf
The focus is on a model reduction framework for parameterized elliptic eigenvalue problems by a reduced basis method. In contrast to the standard single output case, one is interested in approximating several outputs simultaneously, namely a certain number of the smallest eigenvalues. For a fast and reliable evaluation of these input-output relations, we analyze a posteriori error estimators for eigenvalues. Moreover, we present different greedy strategies and study systematically their performance. Special attention needs to be paid to multiple eigenvalues whose appearance is parameter-dependent. Our methods are of particular interest for applications in vibro-acoustics.
NAJun 3, 2016
Reduced basis isogeometric mortar approximations for eigenvalue problems in vibroacousticsThomas Horger, Barbara Wohlmuth, Linus Wunderlich
We simulate the vibration of a violin bridge in a multi-query context using reduced basis techniques. The mathematical model is based on an eigenvalue problem for the orthotropic linear elasticity equation. In addition to the nine material parameters, a geometrical thickness parameter is considered. This parameter enters as a 10th material parameter into the system by a mapping onto a parameter independent reference domain. The detailed simulation is carried out by isogeometric mortar methods. Weakly coupled patch-wise tensorial structured isogeometric elements are of special interest for complex geometries with piecewise smooth but curvilinear boundaries. To obtain locality in the detailed system, we use the saddle point approach and do not apply static condensation techniques. However within the reduced basis context, it is natural to eliminate the Lagrange multiplier and formulate a reduced eigenvalue problem for a symmetric positive definite matrix. The selection of the snapshots is controlled by a multi-query greedy strategy taking into account an error indicator allowing for multiple eigenvalues.
NAApr 19, 2017
Higher order energy-corrected finite element methodsThomas Horger, Petra Pustejovska, Barbara Wohlmuth
The regularity of the solution of elliptic partial differential equa- tions in a polygonal domain with re-entrant corners is, in general, reduced compared to the one on a smooth convex domain. This results in a best approximation property for standard norms which depend on the re-entrant corner but does not increase with the polynomial degree. Standard Galerkin approximations are moreover affected by a global pollution effect. Even in the far field no optimal error reduction can be observed. Here, we generalize the energy-correction method for higher order finite elements. It is based on a parameter-dependent local modification of the stiffness matrix. We will show firstly that for such modified finite element approximation the pollution effect does not occur and thus optimal order estimates in weighted L2-norms can be obtained. Two different modification techniques are introduced and illustrated numerically. Secondly we propose a simple post-processing step such that even with respect to the standard L2-norm optimal order convergence can be recovered.