Synthetic division based integration of rational functions of bivariate polynomial numerators with linear denominators over a unit triangle \(\{0 \leq \xi, \eta\leq 1, \xi + \eta\leq 1\}\) in the local parametric space \((\xi, \eta)\) (Q1970191)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Synthetic division based integration of rational functions of bivariate polynomial numerators with linear denominators over a unit triangle \(\{0 \leq \xi, \eta\leq 1, \xi + \eta\leq 1\}\) in the local parametric space \((\xi, \eta)\) |
scientific article |
Statements
Synthetic division based integration of rational functions of bivariate polynomial numerators with linear denominators over a unit triangle \(\{0 \leq \xi, \eta\leq 1, \xi + \eta\leq 1\}\) in the local parametric space \((\xi, \eta)\) (English)
0 references
25 July 2000
0 references
In finite element analysis, sometimes boundary elements may have a curved boundary. In this paper triangular elements are considered, that have two straight and one curved side (quadratic or cubic). These elements are transformed in a local coordinate system that reduces it to a standard triangle \(\xi\geq 0\), \(\eta\geq 0\), \(\xi+\eta\leq 1\). The Jacobian of this transformation in linear in \(\xi\) and \(\eta\). When these elements appear in the finite element method analysis of second order partial differential equations, one has to integrate products of partial derivatives over these elements. It turns out that these integrals reduce to integrals of rational functions whose numerators are linear in \(\xi\) and \(\eta\), and whose numerators are bivariate polynomials whose degree \(n\) depends on the number of nodes in the curved element. These integrals are decomposed into simpler integrals of rational functions with linear denominator and whose numerator is a univariate polynomial, and the integral of a bivariate polynomial. The form of the rational functions, and hence of the integrals, depends on the position of the node(s) on the curved side of the element. This leads to five different cases for the integrals when the elements considered have 4, 6 or 10 nodes. All the formulas are given explicitly and the method is illustrated numerically with three examples.
0 references
numerical examples
0 references
rational functions
0 references
finite element method
0 references
curved boundary
0 references