Discussion:
[ff3d-users] How to define a moving boundary?
Hao
2007-11-05 02:47:16 UTC
Permalink
Hi Stephane,

I have some problems with ff3d.

1. Is it possible to solve a one dimension problem with ff3d? how to define
a, b, and n?
such as a consolidation problem, du/dt=Cv d(du/dz)/dz

2.Is it possible to solve a moving boundary problem?
such as a(0,0,0), b(10,10,z), where z is changing with time.

3.Is it possible to solve the problem 2 in the following way?

for(int t=0; t<360; t++){
zz1=aplha*t
a(0,0,0)
b(0,0,zz1)
n=(0,0,zz2)
...
}

Thanks,

Hao
Stephane Del Pino
2007-11-06 22:52:14 UTC
Permalink
Dear Hao,
Post by Hao
Hi Stephane,
I have some problems with ff3d.
1. Is it possible to solve a one dimension problem with ff3d? how to define
a, b, and n?
such as a consolidation problem, du/dt=Cv d(du/dz)/dz
Yes, you just need to define only one layer of cells in x and y for instance:
--------------
vector n = (2,2,10);
vector a = (0,0,0);
vector b = (1,1,2);
mesh m = structured(n,a,b);
--------------
Post by Hao
2.Is it possible to solve a moving boundary problem?
such as a(0,0,0), b(10,10,z), where z is changing with time.
Yes.
Post by Hao
3.Is it possible to solve the problem 2 in the following way?
for(int t=0; t<360; t++){
zz1=aplha*t
a(0,0,0)
b(0,0,zz1)
n=(0,0,zz2)
...
}
Again, yes.

Best regards,
Stéphane.
Hao
2007-11-28 14:08:35 UTC
Permalink
Hi Stephane,
Thank you very much for your time. I am still not clear how does it work
with ff3d.

Assumption: there is a block of frozen soil column, a constant
temperature(20C) is applied to the upper surface, the soil will melt
downward. The thaw front is moving at X(t)=f(t); please refer to attached
file.


Could you please give me an example how to define the domain space, and the
lower boundary in ff3d?



Thanks,

Hao
Post by Stephane Del Pino
Dear Hao,
Post by Hao
Hi Stephane,
I have some problems with ff3d.
1. Is it possible to solve a one dimension problem with ff3d? how to
define
Post by Hao
a, b, and n?
such as a consolidation problem, du/dt=Cv d(du/dz)/dz
--------------
vector n = (2,2,10);
vector a = (0,0,0);
vector b = (1,1,2);
mesh m = structured(n,a,b);
--------------
Post by Hao
2.Is it possible to solve a moving boundary problem?
such as a(0,0,0), b(10,10,z), where z is changing with time.
Yes.
Post by Hao
3.Is it possible to solve the problem 2 in the following way?
for(int t=0; t<360; t++){
zz1=aplha*t
a(0,0,0)
b(0,0,zz1)
n=(0,0,zz2)
...
}
Again, yes.
Best regards,
Stéphane.
_______________________________________________
ff3d-users mailing list
http://lists.nongnu.org/mailman/listinfo/ff3d-users
Stephane Del Pino
2007-11-29 00:16:07 UTC
Permalink
Dear Hao.

There is no problem to define a moving mesh: for instance you can do:
-----------------
vector a = (0,0,0);
vector b = (10,1,1);
vector n = (10,2,2);

mesh m = structured(n,a,b);
double dt = 0.2;

double t = 0;

for (double i=0; i<10; ++i) {
save(vtk,"mesh.00".i,m);
t = t+dt;
a = a + (sqrt(t),0,0);
m = structured(n,a,b);
}
-----------------

Does this answers your question?

Best regards,
Stéphane.

Loading...