Nodal Discontinuous Galerkin Methods-Misprints

Jan S Hesthaven and Tim Warburton
Nodal Discontinuous Galerkin Methods:

Algorithms, Analysis, and Applications
– LIST OF CORRECTIONS AND CLARIFICATIONS –
August 12, 2009
Springer
List of corrections and clarifications
Notation: First number is page number, second number line number with a ”+” counting from the
top and a ”-” counting from the bottom. Line counts include equations.
Thanks to C. Bahls, A. Panizza, X. Zhu, C. Rohde, S. Field, J. Thorenson, J. Li, A.Engsig-
Karup, A. Jameson for pointing out these misprints.
50, -5 : Last expression should be
Vij = P̃j−1 (ri ).
58, -1 : Matrix FToV should be

 
1 0 0 0 0
0 1 0 0 0
 
0 1 0 0 0
 
0 0 1 0 0
FToV = 
0
.
 0 1 0 0
0 0 0 0 1
 
0 0 0 0 1
0 0 0 1 0
59, +3 : This also impacts (FToV)(FToV)T which becomes

 
10000 0 0 0
0 1 1 0 0 0 0 0
 
0 1 1 0 0 0 0 0
 
0 0 0 1 1 0 0 0
(FToV)(FToV)T = 
0 0 0 1 1
.
 0 0 0
0 0 0 0 0 1 1 0
 
0 0 0 0 0 1 1 0
00000 0 0 1
This also implies that [59, +5] should read rows 1 and 8 indicate.
4
67, -2 : x ∈ [−2, 2] should be x ∈ [−1, 1].

68, +2 : E(−2, 0) = E(2, 0) = 0 should be E(−1, t) = E(1, t) = 0.
68, +15 : Should be
1
= ...
J k µk
70,-7 : Software line

[Nv, VX, K, EToV] = MeshGen1D(-2.0, 2.0, 80);
should be
[Nv, VX, K, EToV] = MeshGen1D(-1.0, 1.0, 80);
71, -20 : Wrong exact solution. Should be
h i
H (m) (x, t) = n(m) A(m) exp(iωn(m) x) + B (m) exp(−iωn(m) x) exp(iωt)
B (2) = exp(i2n(2) ω)A(2)
A(2) = exp(−iω(n(1) + n(2) ))
79, +13 : The expression for |ṽn | is correct but it should be clarified by write it as.
”We recover
p Z
1
|ṽn | ≤ (Lv)p P̃n (r) dr ,
n(n + 1) I
by integration by parts 2p times. Recalling (4.2) and combining this with . . .”

85, +17 : a(ux , v)Ω should be a(ux , φ)Ω
87, +3 : kεN k2Ω,h should be kεh k2Ω,h
89, +16 : Should be

iωh
− M + aS . . .
2
5
118, Sec 5.2 : Clarification: In this section, fhk has the meaning of the interpolation of the pro-
jection of f , i.e.,
Np
X
fhk (x, t) = (PN f )(xi , t)`ki (x).
i=1
This is equivalent to assuming that all integrals are exact.

119, -8 : ”from” should be ”form”.
122, +5 : Eq.(5.8) should be
1 d
kuh k2Ω,h ≤ 0.
2 dt
123, -8 : Last expression should be

Np
X
fhk (x, t) = fhk (xki , t)`ki (x),
i=1
124, +3 : ukh (xi , t) should be ukh (xki , t).

130, +10 : Eq.(5.16), lower limit should be ηc < η ≤ 1.
134, -3 : The statement is only true for n̂ = 1. The general statement should be
2a+ a−

1
n̂ · (au)∗ = + {{u}} + [[u]]
a + a− 2
151, -9 : Should be
ul = uh; ul(3:Np,:) =0; ul = V*ul;
162,-1 to 163, +2 : Statements
drho (mapI) = ...
drhou(mapI) = ...
dEner(mapI) = ...
should be
drhof (mapI) = ...
drhouf(mapI) = ...
dEnerf(mapI) = ...
163, +6-8 : Statements
drho (mapO) = ...
drhou(mapO) = ...
dEner(mapO) = ...
6
should be
drhof (mapO) = ...
drhouf(mapO) = ...
dEnerf(mapO) = ...
165, +7-8 : The purpose of this piece of code is to describe initial conditions at the cell-centers
as is required for discontinuous initial conditions. The correct statement should be
cx = ones(Np,1)*(sum(MassMatrix*x,1))/2;
If the initial conditions are smooth, cx, can be replaced by x in the codes in lines 9-11.
175, +20ff : Should be ”Thus, w(r) is an N-th-order polynomial approximation to a function
which measures the difference between the equidistant points and the Legendre-Gauss-Lobatto
points; ...”
194, +9 : ”each of length 3KNfp .
195, +1 : Remove one ”find”.
245, +9 : ”discretize” instead of ”discretizing”
245, +19 : ”types” instead of ”type”
246, +9 : Should be ”As in Section 5.3, we have ...”
271, +3 : In caption – n should bd N .
294, -8ff : Until p295, line 4, uh should be exchanged with PN u to avoid the confusion that uh is
used for a numerical solution elsewhere. In these few lines, uh is the projection of the solution
and these results are basic approximation results.
294, -7 : Last term should be
kh−1/2 (u − uh )k2Γb
295, +14 : Remove one |u|Ω,σ,h .

312, -12 : ”compare” instead of ”compared”
353, -19 : Remove ”which” to have ”spectrum and appears to be isolated.”
373, +6 : ”complex situations.”
376, -14 : Software statement, line 16. Should be
theta1 = atan2(VY(v1)-yo, VX(v1)-xo);
theta2 = atan2(VY(v2)-yo, VX(v2)-xo);
399, -1 : ”information”
447, -9 : Software statement
x(1) = (alpha-beta)/(alpha+beta+2);
7
should be
x(1) = -(alpha-beta)/(alpha+beta+2);
(2i+2j+2,0) (2i+2j+2,0)
449, -1 : Pk (b) should be Pk (c)
459 : CorrectBCTable.m fails if an element has two faces on the boundary. An improved version
but with a different calling sequence is
function BCType = CorrectBCTable_v2(EToV,VX,VY,BCType,fd,BCcode)
% function BCType = CorrectBCTable(EToV,BCType,fd,BCcode);

% Purpose: Setup BCType for boundary conditions in 2D
%
% EToV : Element-To-Vertice table
% VX, VY : (x,y)-coordinates of mesh vertices
% BCType : Table with types of faces for BC’s
% fd : handle to distance function
% BCcode : Integer for specific boundary type
%
% By Allan P. Engsig-Karup
Globals2D;
VNUM = [1 2;2 3;3 1]; % face orientations
pxc = 0.5*(VX(EToV)+VX(EToV(:,[2 3 1])));

pyc = 0.5*(VY(EToV)+VY(EToV(:,[2 3 1])));
dc = abs(fd([pxc(:) pyc(:)])); % distances to boundaries from face centers
tol = 1e-4; % tolerance
idx = find(dc<tol);
BCType(idx) = BCcode;
return
464, +11 : Should be ”interior values, u− = u(vmapM).”
464, +13 : Should be ”exterior values, u+ = u(vmapP).”
464, +14 : vmapB is a one-vector but the size is problem dependent.
467, +12 : Software line
[TRI,xout,yout,uout,interp] PlotField2D(2*N, x, y, vort);

should be
[TRI,xout,yout,uout,interp] = PlotField2D(2*N, x, y, vort);
467, -3 : Software line
PlotContour2D(TRI, xout, yout, vortout, linspace(-6, 6, 12));
8
should be
PlotContour2D(TRI, xout, yout, vort, linspace(-6, 6, 12));
481, ref 134 : ”applications”

Nodal Discontinuous Galerkin Methods-Misprints

Uploaded by

Copyright:

Available Formats

Nodal Discontinuous Galerkin Methods-Misprints

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nodal Discontinuous Galerkin Methods-Misprints

Uploaded by

Copyright:

Available Formats

Jan S Hesthaven and Tim Warburton

Nodal Discontinuous Galerkin Methods:

– LIST OF CORRECTIONS AND CLARIFICATIONS –

August 12, 2009

Vij = P̃j−1 (ri ).

58, -1 : Matrix FToV should be

59, +3 : This also impacts (FToV)(FToV)T which becomes

67, -2 : x ∈ [−2, 2] should be x ∈ [−1, 1].

70,-7 : Software line

71, -18 : Wrong exact solution. Should be

B (2) = exp(i2n(2) ω)A(2)

71, -16 : Wrong exact solution. Should be

A(2) = exp(−iω(n(1) + n(2) ))

by integration by parts 2p times. Recalling (4.2) and combining this with . . .”

This is equivalent to assuming that all integrals are exact.

123, -8 : Last expression should be

124, +3 : ukh (xi , t) should be ukh (xki , t).

295, +14 : Remove one |u|Ω,σ,h .

% function BCType = CorrectBCTable(EToV,BCType,fd,BCcode);

VNUM = [1 2;2 3;3 1]; % face orientations

pxc = 0.5*(VX(EToV)+VX(EToV(:,[2 3 1])));

[TRI,xout,yout,uout,interp] PlotField2D(2*N, x, y, vort);

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.