This preview shows page 1. Sign up to view the full content.
Unformatted text preview: at smooth solutions of the weak problem (7.4.19) satisfy
the Neumann problem
;r prE + qE = r pEn = R (x y) 2 e (x y) 2 @ e : (7.4.20a)
(7.4.20b) Solutions of (7.4.20) only exist when the data R and r satisfy the equilibrium condition
ZZ r(x y)dxdy + e Z @e R(s)ds = 0: (7.4.20c) This condition will most likely not be satis ed by the choice of = 1=2. Ainsworth and
Oden 5] describe a relatively simple procedure that requires the solution of the Poisson
; !e = r @!e = R
@n (x y ) 2 (x y) 2 @ e ; @ !e = 0 (7.4.21a) e
E (x y) 2 @ E : (7.4.21b)
(7.4.21c) The error estimate is
kE kA =
e=1 Ae(!e !e): (7.4.21d) The function is approximated by a piecewise-linear polynomial in a coordinate s on
@ e and may be determined explicitly prior to solving (7.4.21). Let us illustrate the
e ect of this equilibrated error estimate.
Example 7.4.4. Oden 16] considers a \cracked panel" as shown in Figure 7.4.4 and
determines u as the solution of A(v u) = ZZ (vxux + vy uy )dxdy = 0: 7.4. A Posteriori Error Estimation 29
u = r1/2 cos θ/2 r
u=0 ΩR x θ
u y= 0 Figure 7.4.4: Cracked panel used for Example 7.4.4...
View Full Document
This document was uploaded on 03/16/2014 for the course CSCI 6860 at Rensselaer Polytechnic Institute.
- Spring '14