• Keine Ergebnisse gefunden

Problem 10.1: A posteriori error analysis for general smooth, symmetric coefficients cont.

N/A
N/A
Protected

Academic year: 2022

Aktie "Problem 10.1: A posteriori error analysis for general smooth, symmetric coefficients cont."

Copied!
1
0
0

Wird geladen.... (Jetzt Volltext ansehen)

Volltext

(1)

IWR – Universit¨at Heidelberg Prof. Dr. Guido Kanschat Alec Gilbert

Due date:21.12.2018

Homework No. 10

Finite Elements, Winter 2018/19

Problem 10.1: A posteriori error analysis for general smooth, symmetric coefficients cont.

LetΩ ⊂ Rd be a bounded, convex, polyhedral domain and letV = H01(Ω). Letai,j ∈ C(Ω) fori, j = 1,2, . . . , dbe symmetric coefficients (ai,j=aj,i) such that for allx∈Ωthe coefficient matrixA(x) = [ai,j(x)]satisfies

ξTA(x)ξ ≥ C0|ξ|2 for allξ∈Rd. Consider the bilinear forma:V ×V →Rgiven by

a(u, v) =

d

X

i,j=1

Z

ai,j(x)∂iu(x)∂jv(x) dx,

and define the correspondingenergy normonV by

kukV = kuka := p a(u, u).

Forf ∈V, the variational equation corresponding to the above bilinear form is to findu∈V such that

a(u, v) =f(v) for allv∈V. (10.1)

Further, let{Vh}h>0be a family of conforming finite element spaces each defined on a shape regular meshTh, and letPk(T)⊂ Vh|T for all cellsT ∈Th.

(a) Let alsoThbe locally quasi-uniform. Formulate and then prove an estimate of the form (2.89) from Lemma 2.3.12 in the notes.

Hint:Recall from the previous class that forv, w∈V the strong form of the residual is

a(Rv, w) = X

T∈Th

Z

T

rT(v)wdx− X

F∈Fih

Z

F

2{{n·A∇v}}wds , (10.2)

whererT(v) := f +∇ ·(A∇v), and we proved that forv∈V the weak residual satisfies ku−vka = kRvk−a = sup

kwka=1

|a(Rv, w)|. (10.3)

(b) As in Definition 2.3.15, define the residual based error estimatorηa,hfor (10.1).

(c) Formulate and then prove an estimate of the form (2.95) from the notes for (10.1) and the energy norm.

(d) Formulate and then prove an estimate of the form (2.96) from the notes for (10.1) and the energy norm.

Referenzen

ÄHNLICHE DOKUMENTE

The authors gratefully acknowledge partial financial support by the project Reduced Basis Methods for Model Reduction and Sensitivity Analysis of Complex Partial Differential

This paper establishes a unified a posteriori error estimator for a large class of conforming finite element methods for the Reissner-Mindlin plate problem.. The analysis is based

(a) Write the residual for a(w, v) = f (v) in weak (Definition 2.3.8) and strong (Definition 2.3.11) form.. Compare to

IWR – Universit¨at

IWR – Universit¨at

In each case a discrete Lagrange multiplier is designed, which is efficient in the sense of Theorem B (with the exception of CFEM for the Bingham problem) and allows for a

Daher legt jeder Prozessor zun¨achst einen std::vector<bool> vertexmask an, in dem letztendlich stehen soll, ob er f¨ur den Knoten mit dem Index i zust¨andig ist oder nicht

F¨ ur q = 1/2 n¨ ahert sich die Strecke nur beliebig nah dem Brunnenrand an, ohne ihn aber in endlicher Zeit zu erreichen (muß frustrierend sein).. Aufgabe 23 (a) Die Idee (hier