# Where the superscript 2 is used to identify

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: the form x( y( 3 3 ) = X X xij N 2 ( ) yij i j i=1 j =1 ) (5.4.4) where Ni2j ( ), i j = 1 2 3, is given by (5.4.3). This transformation produces an element in the (x y)-plane having curved (quadratic) edges as shown in Figure 5.4.3. An isoparametric approximation would be biquadratic 5.4. Coordinate Transformations y 19 η 3 11 00 3 11 00 11 00 6 11 00 11 00 11 00 11 00 1 11 00 5 11 00 11 00 11 002 11 00 11 00 11 00 11 00 6 11 00 5 11 00 11 00 11 00 11 00 11 00 11 00 4 x 1 11 00 ξ 11 00 4 2 Figure 5.4.4: Quadratic mapping of a triangle having one curved side. and have the form of (5.4.3). The interior node (3,3) is awkward and can be eliminated by using a second-order serendipity (cf. Problems 4.3.1) or hierarchical transformation (cf. Section 4.4). Example 5.4.3. The biquadratic transformation described in Example 5.4.2 is useful for discretizing domains having curved boundaries. With a similar goal, we describe a transformation for creating triangular elements having one curved and two straight sides (Figure 5.4.4). Let us approximate the curved boundary by a quadratic polynomial and map the element onto a canonical right triangle by the quadratic transformation x( y( 6 ) = X xi N 2 ( ) yi i i=1 ) (5.4.5a) where the quadratic Lagrange shape functions are (cf. Problem 4.2.1) Nj2 = 2 j ( N42 = 4 j ; 1=2) N52 = 4 12 j=1 2 3 23 N62 = 4 (5.4.5b) 31 (5.4.5c) and 1 =1; ; 2 = 3 =: (5.4.6) Equations (5.4.5) and (5.4.6) describe a general quadratic transformation. We have a more restricted situation with x4 = (x1 + x2 )=2 y4 = (y1 + y2)=2 x6 = (x1 + x3 )=2 y6 = (y1 + y3)=2: 20 Mesh Generation and Assembly This simpli es the transformation (5.4.5a) to x( y( ) = x1 N 2 + x2 N 2 + x3 N 2 + x5 N 2 ^ ^ ^ ^ ) y1 1 y2 2 y3 3 y5 5 (5.4.7a) where, upon use of (5.4.5) and (5.4.6), ^ N12 = N12 + (N42 + N62 )=2 = 1 = 1 ; ; (5.4.7b) ^ N22 = N22 + N42 =2 = (1 ; 2 ) (5.4.7c) ^ N32 = N32 + N62 =2 = (1 ; 2 ) (5.4.7d) ^ N52 = N52 = 4 : (5.4.7e) From these results, we see that the mappings on edges 1-2 ( = 0) and 1-3 ( = 0) are linear and are, respectively, given by x =...
View Full Document

## This document was uploaded on 03/16/2014 for the course CSCI 6860 at Rensselaer Polytechnic Institute.

Ask a homework question - tutors are online