Skip to content
GitLab
Explore
Sign in
Register
Primary navigation
Search or go to…
Project
fem
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Model registry
Operate
Environments
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
GitLab community forum
Contribute to GitLab
Provide feedback
Terms and privacy
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
gmsh
fem
Commits
faf71b72
Commit
faf71b72
authored
5 years ago
by
Anthony Royer
Browse files
Options
Downloads
Patches
Plain Diff
Revert demos/navier
parent
fed6f9c9
No related branches found
No related tags found
No related merge requests found
Pipeline
#6205
failed
5 years ago
Stage: test
Changes
1
Pipelines
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
demos/navier/main.cpp
+156
-1
156 additions, 1 deletion
demos/navier/main.cpp
with
156 additions
and
1 deletion
demos/navier/main.cpp
+
156
−
1
View file @
faf71b72
...
...
@@ -47,7 +47,7 @@ int main(int argc, char **argv)
gmshFem
.
userDefinedParameter
(
pointsPerSWavelength
,
"pointsPerSWavelength"
);
double
lc
=
2
*
pi
/
(
pointsPerSWavelength
*
kS
);
//.............[m] characteristic length
const
double
rExt
=
5
.
;
//.............[m] external radius of the annulus geometry
const
double
rExt
=
2
.
;
//.............[m] external radius of the annulus geometry
const
double
rInt
=
1.
;
//.............[m] internal radius of the annulus geometry
int
ABCOrder
=
2
;
...
...
@@ -203,6 +203,149 @@ int main(int argc, char **argv)
formulation
.
galerkin
(
-
w
*
w
*
rho
*
dof
(
ux
),
tf
(
ux
),
surface
,
"Gauss2"
);
formulation
.
galerkin
(
-
w
*
w
*
rho
*
dof
(
uy
),
tf
(
uy
),
surface
,
"Gauss2"
);
//---------------------------------
//- Absorbing Boundary Conditions -
//---------------------------------
if
(
ABCOrder
==
0
)
{
msg
::
info
<<
"Use zero order Lysmer Kuhlemeyer ABC."
<<
msg
::
endl
;
// On ABC boundary int -i w cp rho In u.u' -i w cs rho It u.u'
formulation
.
galerkin
(
-
im
*
w
*
cP
*
rho
*
nX
*
nX
*
dof
(
ux
),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
w
*
cP
*
rho
*
nX
*
nY
*
dof
(
uy
),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
w
*
cP
*
rho
*
nX
*
nY
*
dof
(
ux
),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
w
*
cP
*
rho
*
nY
*
nY
*
dof
(
uy
),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
w
*
cS
*
rho
*
nY
*
nY
*
dof
(
ux
),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
im
*
w
*
cS
*
rho
*
nX
*
nY
*
dof
(
uy
),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
w
*
cS
*
rho
*
nX
*
nX
*
dof
(
uy
),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
im
*
w
*
cS
*
rho
*
nX
*
nY
*
dof
(
ux
),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
}
else
if
(
ABCOrder
==
2
)
{
msg
::
info
<<
"Use high order ABC."
<<
msg
::
endl
;
//------------------------------------------------------------------------------------------------------------------------------
//- See p 1713 of "A high-order ABC for 2D time-harmonic elastodynamic scattering problems", V. Mattesi, M. Darbas C. Geuzaine -
//------------------------------------------------------------------------------------------------------------------------------
//-------------------------------------------------------------------
//- STEP 1 : Find v in H^1/2(Gamma) such that v = Lambda_{1,eps} u -
//-------------------------------------------------------------------
// On gammaInf int v.v' - i rho w^2 [ 1/kPEps (n vO,v') + 1/kSEps (t v1,v') ] (=0)
formulation
.
galerkin
(
dof
(
vx
),
tf
(
vx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
dof
(
vy
),
tf
(
vy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
rho
*
pow
(
w
,
2
)
*
(
1.0
/
kPEps
)
*
nX
*
dof
(
v0
),
tf
(
vx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
rho
*
pow
(
w
,
2
)
*
(
1.0
/
kPEps
)
*
nY
*
dof
(
v0
),
tf
(
vy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
rho
*
pow
(
w
,
2
)
*
(
1.0
/
kSEps
)
*
(
-
nY
)
*
dof
(
v1
),
tf
(
vx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
rho
*
pow
(
w
,
2
)
*
(
1.0
/
kSEps
)
*
(
nX
)
*
dof
(
v1
),
tf
(
vy
),
gammaInf
,
"Gauss2"
);
// (v0,v0') - sum R_i (h_i,v0') =0
formulation
.
galerkin
(
dof
(
v0
),
tf
(
v0
),
gammaInf
,
"Gauss2"
);
for
(
unsigned
int
i
=
0
;
i
<
padeOrder
;
++
i
)
{
formulation
.
galerkin
(
-
padeR
[
i
]
*
dof
(
h
[
i
]),
tf
(
v0
),
gammaInf
,
"Gauss2"
);
// Sl(hl,hl') - (grad_G/kPEps hl,grad_G/kPEps hl') -(n.u,hl') (=0) forall l=1,...,L
formulation
.
galerkin
(
padeS
[
i
]
*
dof
(
h
[
i
]),
tf
(
h
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
pow
(
1.0
/
kPEps
,
2
)
*
grad
(
dof
(
h
[
i
])),
grad
(
tf
(
h
[
i
])),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
nX
*
dof
(
ux
),
tf
(
h
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
nY
*
dof
(
uy
),
tf
(
h
[
i
]),
gammaInf
,
"Gauss2"
);
}
// (v1,v1') - sum Rl (il,v1') =0
formulation
.
galerkin
(
dof
(
v1
),
tf
(
v1
),
gammaInf
,
"Gauss2"
);
for
(
unsigned
int
i
=
0
;
i
<
padeOrder
;
++
i
)
{
formulation
.
galerkin
(
-
padeR
[
i
]
*
dof
(
ix
[
i
]),
tf
(
v1
),
gammaInf
,
"Gauss2"
);
// Sl( il,il') -(grad_G/kSEps il,grad_G/kSEps il') - (t. u,il') (=0) forall l=1,...L
formulation
.
galerkin
(
padeS
[
i
]
*
dof
(
ix
[
i
]),
tf
(
ix
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
pow
(
1.0
/
kSEps
,
2
)
*
grad
(
dof
(
ix
[
i
])),
grad
(
tf
(
ix
[
i
])),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
(
-
nY
)
*
dof
(
ux
),
tf
(
ix
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
(
nX
)
*
dof
(
uy
),
tf
(
ix
[
i
]),
gammaInf
,
"Gauss2"
);
}
//-------------------------------------------------------------------
//- Step 2 : Find q in H^-1/2(Gamma) st : (I+ Lambda_{2,eps}) q = v -
//-------------------------------------------------------------------
// (q,q') -i[1/ks(grad_G q0,q') - 1/kp (q1 n,q')] - (v,q') (=0)
formulation
.
galerkin
(
dof
(
qx
),
tf
(
qx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
dof
(
qy
),
tf
(
qy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
(
1.0
/
kSEps
)
*
vector
<
std
::
complex
<
double
>
>
(
1.
,
0.
,
0.
)
*
grad
(
dof
(
q0
)),
tf
(
qx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
im
*
(
1.0
/
kSEps
)
*
vector
<
std
::
complex
<
double
>
>
(
0.
,
1.
,
0.
)
*
grad
(
dof
(
q0
)),
tf
(
qy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
im
*
(
1.0
/
kPEps
)
*
nX
*
(
-
nY
)
*
vector
<
std
::
complex
<
double
>
>
(
1.
,
0.
,
0.
)
*
grad
(
dof
(
q1
)),
tf
(
qx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
im
*
(
1.0
/
kPEps
)
*
nX
*
nX
*
vector
<
std
::
complex
<
double
>
>
(
0.
,
1.
,
0.
)
*
grad
(
dof
(
q1
)),
tf
(
qx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
im
*
(
1.0
/
kPEps
)
*
nY
*
(
-
nY
)
*
vector
<
std
::
complex
<
double
>
>
(
1.
,
0.
,
0.
)
*
grad
(
dof
(
q1
)),
tf
(
qy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
im
*
(
1.0
/
kPEps
)
*
nY
*
nX
*
vector
<
std
::
complex
<
double
>
>
(
0.
,
1.
,
0.
)
*
grad
(
dof
(
q1
)),
tf
(
qy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
dof
(
vx
),
tf
(
qx
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
dof
(
vy
),
tf
(
qy
),
gammaInf
,
"Gauss2"
);
//(q0,q0')=sum Rl (jl,q0')
formulation
.
galerkin
(
dof
(
q0
),
tf
(
q0
),
gammaInf
,
"Gauss2"
);
for
(
unsigned
int
i
=
0
;
i
<
padeOrder
;
++
i
)
{
formulation
.
galerkin
(
-
padeR
[
i
]
*
dof
(
j
[
i
]),
tf
(
q0
),
gammaInf
,
"Gauss2"
);
//Sl(jl,jl')-(grad/kSEps jl,grad/kSEps jl') - (q.n,jl') =0
formulation
.
galerkin
(
padeS
[
i
]
*
dof
(
j
[
i
]),
tf
(
j
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
pow
(
1.0
/
kSEps
,
2
)
*
grad
(
dof
(
j
[
i
])),
grad
(
tf
(
j
[
i
])),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
nX
*
dof
(
qx
),
tf
(
j
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
nY
*
dof
(
qy
),
tf
(
j
[
i
]),
gammaInf
,
"Gauss2"
);
}
// (q1,q1')= sum (Rl kl , q1')
formulation
.
galerkin
(
dof
(
q1
),
tf
(
q1
),
gammaInf
,
"Gauss2"
);
for
(
unsigned
int
i
=
0
;
i
<
padeOrder
;
++
i
)
{
formulation
.
galerkin
(
-
padeR
[
i
]
*
dof
(
k
[
i
]),
tf
(
q1
),
gammaInf
,
"Gauss2"
);
// (Sl kl,kl') -(grad_G/kPEps kl,grad_G/kPEps kl') - (t.q, kl') (=0)
formulation
.
galerkin
(
padeS
[
i
]
*
dof
(
k
[
i
]),
tf
(
k
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
pow
(
1.0
/
kPEps
,
2
)
*
grad
(
dof
(
k
[
i
])),
grad
(
tf
(
k
[
i
])),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
(
-
nY
)
*
dof
(
qx
),
tf
(
k
[
i
]),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
(
nX
)
*
dof
(
qy
),
tf
(
k
[
i
]),
gammaInf
,
"Gauss2"
);
}
//----------------------------------
//- Step 3: -tu.u' with t=q+2mu Mu -
//----------------------------------
// - (q,u') - 2 mu (Mu,u')
formulation
.
galerkin
(
-
dof
(
qx
),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
dof
(
qy
),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
2.0
*
mu
*
vector
<
std
::
complex
<
double
>
>
(
1.
,
0.
,
0.
)
*
grad
(
dof
(
un
)),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
2.0
*
mu
*
vector
<
std
::
complex
<
double
>
>
(
0.
,
1.
,
0.
)
*
grad
(
dof
(
un
)),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
2.0
*
mu
*
nX
*
(
-
nY
)
*
vector
<
std
::
complex
<
double
>
>
(
1.
,
0.
,
0.
)
*
grad
(
dof
(
ut
)),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
2.0
*
mu
*
nX
*
nX
*
vector
<
std
::
complex
<
double
>
>
(
0.
,
1.
,
0.
)
*
grad
(
dof
(
ut
)),
tf
(
ux
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
2.0
*
mu
*
nY
*
(
-
nY
)
*
vector
<
std
::
complex
<
double
>
>
(
1.
,
0.
,
0.
)
*
grad
(
dof
(
ut
)),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
2.0
*
mu
*
nY
*
nX
*
vector
<
std
::
complex
<
double
>
>
(
0.
,
1.
,
0.
)
*
grad
(
dof
(
ut
)),
tf
(
uy
),
gammaInf
,
"Gauss2"
);
// (un,un')=(u.n,un')
formulation
.
galerkin
(
dof
(
un
),
tf
(
un
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
nX
*
dof
(
ux
),
tf
(
un
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
nY
*
dof
(
uy
),
tf
(
un
),
gammaInf
,
"Gauss2"
);
// (ut,ut')=(u.t,ut')
formulation
.
galerkin
(
dof
(
ut
),
tf
(
ut
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
(
-
nY
)
*
dof
(
ux
),
tf
(
ut
),
gammaInf
,
"Gauss2"
);
formulation
.
galerkin
(
-
(
nX
)
*
dof
(
uy
),
tf
(
ut
),
gammaInf
,
"Gauss2"
);
}
// Prepro
formulation
.
pre
();
...
...
@@ -214,5 +357,17 @@ int main(int argc, char **argv)
save
(
ux
);
save
(
uy
);
save
(
xComp
(
solution
),
surface
,
"ux_exact"
);
save
(
yComp
(
solution
),
surface
,
"uy_exact"
);
std
::
complex
<
double
>
num
=
integrate
(
pow
(
abs
(
xComp
(
solution
)
-
ux
),
2
)
+
pow
(
abs
(
yComp
(
solution
)
-
uy
),
2
),
surface
,
"Gauss2"
);
std
::
complex
<
double
>
den
=
integrate
(
pow
(
abs
(
xComp
(
solution
)
),
2
)
+
pow
(
abs
(
yComp
(
solution
)
),
2
),
surface
,
"Gauss2"
);
std
::
complex
<
double
>
L2error
=
sqrt
(
num
/
den
);
msg
::
info
<<
"L_2 error = "
<<
L2error
<<
msg
::
endl
;
CSVio
file
(
"convergence.txt"
,
' '
,
OpeningMode
::
Append
);
file
<<
1.
/
pointsPerSWavelength
<<
std
::
real
(
sqrt
(
num
/
den
))
<<
csv
::
endl
;
return
0
;
}
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment