Imposition of time-dependent boundary conditions in FEM formulations for elastodynamics: critical assessment of penalty-type methods

Abstract

This paper is concerned with the variationally consistent incorporation of time dependent boundary conditions. The proposed methodology avoids ad hoc procedures and is applicable to both linear as well as nonlinear problems. An integral formulation of the dynamic problem serves as a basis for the imposition of the corresponding constraints, which are enforced via the consistent form of the penalty method, e.g. a form that complies with the norm and inner product of the functional space where the weak formulation is posed. Also, it is shown that well known and broadly implemented modelling techniques such as “large mass” and “large spring” methods, arise as limiting cases of the penalty formulation. Further extension of the proposed methodology is provided in the case where a generalized weak formulation of the dynamic problem with an independent velocity field is assumed. Finally, a few examples serve to illustrate the above concepts.