Analysis of penalty parameters for interior penalty Galerkin methods. Issue 5 (2nd September 2019)

• Record Type: Journal Article
• Authors: Straßer, Sebastian; Herzog, Hans-Georg
• Is Part Of: Compel. Volume 38:Issue 5(2019)
• Page Start: 1401
• Page End: 1412
• Publication Date: 2019-09-02
• Subjects: Boundary value problems -- Finite element method -- Numerical analysis; Electrical engineering -- Data Processing -- Periodicals; Electrical engineering -- Mathematics -- Periodicals; Electrical engineering -- Periodicals; Electronics -- Data Processing -- Periodicals; Electronics -- Mathematics -- Periodicals; 621.3
• DOI: 10.1108/COMPEL-12-2018-0514
• Languages: English
• ISSNs: 0332-1649

Abstract

Abstract : Purpose: The purpose of this paper is to analyse the influence of penalty parameters for an interior penalty Galerkin method, namely, the symmetric interior penalty Galerkin method. Design/methodology/approach: First of all, the solution of a simple model problem is computed and compared to the exact solution, which is a periodic function. Afterwards, a two-dimensional magnetostatic field problem described by the magnetic vector potential A is considered. In particular, penalty parameters depending on the polynomial degree, the properties of the elements and the material are considered. The analysis is performed by varying the polynomial degree and the mesh sizes on a structured and an unstructured mesh. Additionally, the penalty parameter is varied in a specific range. Findings: Choosing the penalty parameter correctly plays an important role as the stability and the convergence of the numerical scheme can be affected. For a structured mesh, a limiting value for the penalty parameter can be calculated beforehand, whereas for an unstructured mesh, the choice of the penalty parameter can be cumbersome. Originality/value: This paper shows that there exist different penalty parameters which can be taken into account to solve the considered problems. One can choose a global penalty parameter to obtain a stable solution, which is a sharp estimation. There has always to be the consideration to guarantee the coercivity of the bilinear form while minimising the number of iterations.