Deriving and Implementing a Contact Formulation in an Open-Source Finite Elements Framework
Abstract:
Contact mechanics is a wide class of physical phenomena present in a variety of life situations,
as well as engineering applications. This, in turn, triggers the motivation for continuously devel-
oping a reliable mathematical model, and implementing a robust & efficient numerical simulation
scheme for such phenomena. In the scope of this work, a formulation derivation and an imple-
mentation of a segment-to-segment contact scheme based on collocation method for numerical
integration and standard Lagrange multipliers was presented. The contribution of this work is
proposing a reliable contact approach –the segment-to-segment approach –, which employs an
exact contact constraint enforcement scheme –Lagrange multipliers –, and aiming at reducing
the overhead of interface pre-processing for numerical integration by using the collocation inte-
gration method. For the sake of completeness, a review of solid mechanics concepts has been
first introduced to the reader, as well as discussions about the constraint enforcement meth-
ods and different approaches for numerical integration adopted in contact mechanics numerical
implementations. Moreover, the necessary linearizations related to the formulation, which are
necessary for the Newton-Raphson solution method were appended in full detail. The formulation
was implemented in a C++ open-source framework, Kratos Multiphysics, and different aspects
are particularly tested through the common benchmarks in the literature. Additional engineering
test cases serve as supplementary verification of implementation capabilities. The focus of the
work was dedicated to investigating the implementation of 2D in full detail, nevertheless, an ex-
tension to 3D using the proposed method has been also included. For the 2D case, the benchmarks
has been fulfilled successfully, and the method has demonstrated notable robustness bearing in
mind its innate simplification of the collocation integration method. In 3D, the influence of said
simplifications required large number of integration points to fulfill the benchmark test studied,
which reflected a notably expensive computational cost. Finally, the foreseen future outlook for
this work has been proposed, based on the results and discussion.