22 a superscript 1 has been added to emphasize that

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

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ear polynomials.) The equation of the line connecting Nodes 1 and 3 of the triangular element shown on the left of Figure 4.2.4 is N21 = 0. Likewise, the equation of a line passing through Node 2 and parallel to the line passing through Nodes 1 and 3 is N21 = 1. Thus, to map the line N21 = 0 onto the line = 0 in the canonical plane, we should set = N21 (x y). Similarly, the line joining Nodes 1 and 2 satis es the equation N31 = 0. We would like this line to become the line = 0 in the transformed plane, so our mapping must be = N31 (x y). Therefore, using (4.2.2) 2 3 2 3 1xy 1xy det 4 1 x1 y1 5 det 4 1 x1 y1 5 1 x3 y3 1 x2 y2 3 3 : (4.2.6) = N21 (x y) = 2 = N31 (x y) = 2 1 x2 y2 1 x3 y3 det 4 1 x1 y1 5 det 4 1 x1 y1 5 1 x3 y3 1 x2 y2 As a check, evaluate the determinants and verify that (x1 y1) ! (0 0), (x2 y2) ! (1 0), and (x3 y3) ! (0 1). Polynomials may now be developed on the canonical triangle to simplify the algebraic 4.2. Lagrange Shape Functions on Triangles 7 1 0 1 03 1 0 11 00 11 00 11 00 8 90 1 1 N 1= 0 2 1 0 1 0 1 0 1 0 1 0 11 0N 1= 1/3 1 0 10 0 N 1= 0 1 1 1 0 1 06 1 0 1 0 1 0 1 50 1 02 1 0 1 0 N 1= 2/3 1 4 1 0 1 0 1 0 1 0 1 07 1 0 Figure 4.2.5: Geometry of a triangular nite element for a cubic polynomial Lagrange approximation. complexity and subsequently transformed back to the physical element. 2. Transformation using triangular coordinates. A simple procedure for constructing Lagrangian approximations involves the use of a redundant coordinate system. The construction may be described in general terms, but an example su ces to illustrate the procedure. Thus, consider the construction of a cubic approximation on the triangular element shown in Figure 4.2.5. The vertex nodes are numbered 1, 2, and 3 edge nodes are numbered 4 to 9 and the centroid is numbered as Node 10. Observe that the line N11 = 0 passes through Nodes 2, 6, 7, and 3 the line N11 = 1=3 passes through Nodes 5, 10, and 8 and the line N11 = 2=3 passes through Nodes 4 and 9. Since N13 must vanish at Nodes 2 - 10 and be a cubic polynomial, it must have the form N13 (x y) = N11 (N11 ; 1=3)(N11 ; 2=3) where the constant is determined by normalizing N13 (x1 y1) = 1. Since N11(x1 y1) = 1, we nd = 9=2 and N13(x y) = 9 N11 (N11 ; 1=3)(N11 ; 2=3): 2 The shape function for an edge node is constructed in a similar manner. For example, in order to obtain N43 we observe that the line N21 = 0 passes through Nodes 1, 9, 8, and 3 8 Finite Element Approximation the line N11 = 0 passes through Nodes 2, 6, 7, and 3 and the line N11 = 1=3 passes through Nodes 5, 10, and 8. Thus, N43 must have the form N43(x y) = N11N21 (N11 ; 1=3): Normalizing N43 (x4 y4) = 1 gives Hence, = 27=2 and 21 N43 (x4 y4) = 3 3 ( 2 ; 1 ): 33 N43 (x y) = 27 N11 N21 (N11 ; 1=3): 2 3 must vanish on the boundary of the triangle and is, Finally, the shape function N10 thus, determined as 3 N10(x y) = 27N11N21 N31 : 3 The cubic shape functions N13 , N43 , and N10 are shown in Figure 4.2.6. The three linear shape functions Nj1 , j = 1 2 3, can be regarded as a redundant coordinate system...
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