On the robustness of the \(h\)- and \(p\)-versions of the finite-element method (Q1176802): Difference between revisions

An efficiency study of the finite element method is made between the classical \(h\)-version and the \(p\)-version, using results already known relative to the notions of locking and robustness for finite elements. The study is realized for two particular problems of heat transfer through highly orthotropic and nearly incompressible materials. These two examples show the superiority of the \(p\)-version. The \(h\)-version requires polynomials of fixed degree \(p\) which appear in the test functions and a refining of the mesh. The \(p\)-version requires the mesh to be fixed and an increasing of the number of nodes on the element, that is the increasing of the degree of polynomials.