time dependent problem

[Hans Pirnay]

Hello everyone,

I am new to hermes, going through the tutorials and the examples. I
would like to solve a nonlinear problem in one space coordinate with
time integration. The tutorial only talks about 2d problems, and the
definition of 1d problems seems to work quite differently. Is there a
small example that shows how to do time integration for 1d-problems?

thanks
Hans

[Pavel Solin]

Hi Hans,

On Thu, May 19, 2011 at 8:25 AM, Hans Pirnay <hansp...@googlemail.com> wrote:

Hello everyone,

I am new to hermes, going through the tutorials and the examples. I
would like to solve a nonlinear problem in one space coordinate with
time integration. The tutorial only talks about 2d problems, and the
definition of 1d problems seems to work quite differently.

This is right. We decided to "freeze" the development of 

H1D and H3D now for some time, in order to be able to 

invest all our manpower into H2D and get to the first 

release and a deb package.

 

Is there a
small example that shows how to do time integration for 1d-problems?

Look for example at system_pendulum. Right now 

it does only one Newton solve (one time step?) 

which I do not quite understand. I will look at it 

in the evening again, try to understand what the 

example does, and change the example to do 

more time steps than one. 

Best,

Pavel

 

thanks
Hans

--
Pavel Solin
University of Nevada, Reno
Home page: http://hpfem.org/~pavel
FEMTEC 2011: http://hpfem.org/events/femtec-2011/
Hermes: http://hpfem.org/
FEMhub: http://femhub.org/

[Pavel Solin]

Hi Hans,

On Thu, May 19, 2011 at 1:46 PM, Pavel Solin <so...@unr.edu> wrote:

Hi Hans,

On Thu, May 19, 2011 at 8:25 AM, Hans Pirnay <hansp...@googlemail.com> wrote:

Hello everyone,

I am new to hermes, going through the tutorials and the examples. I
would like to solve a nonlinear problem in one space coordinate with
time integration. The tutorial only talks about 2d problems, and the
definition of 1d problems seems to work quite differently.

This is right. We decided to "freeze" the development of 

H1D and H3D now for some time, in order to be able to 

invest all our manpower into H2D and get to the first 

release and a deb package.

 

Is there a
small example that shows how to do time integration for 1d-problems?

Look for example at system_pendulum. Right now 

it does only one Newton solve (one time step?) 

which I do not quite understand. I will look at it 

in the evening again, try to understand what the 

example does, and change the example to do 

more time steps than one. 

Hm, I was completely off here. This example is a system

of two ODE. Instead of time-stepping, H1D discretizes

the entire time interval at once using finite elements, hence

only one Newton's loop. 

Nevertheless, I created a new example 

laplace_timedep for the time-dependent heat transfer

equation du/dt - Laplace u - f = 0 discretized using the 

implicit Euler method in time. However, during the 

implementation I realized that the residual needs to 

access the previous time level solution. This is implemented 

in H2D and H3D but not in H1D yet. 

Given the fact that the new object-oriented weak forms in 

H2D are the ultimate standard that will be ported to H1D and 

H3D, I would not do any hacks to H1D at the moment.  

If you like, you can help us port the new object-oriented

weak forms from H2D to H1D. That should not be too difficult

as the forms are almost dimension-independent, and we would

be able to solve time-dependent PDE in H1D.

Best,

Pavel

 

Best,

Pavel

 

thanks
Hans

--
Pavel Solin
University of Nevada, Reno
Home page: http://hpfem.org/~pavel
FEMTEC 2011: http://hpfem.org/events/femtec-2011/
Hermes: http://hpfem.org/
FEMhub: http://femhub.org/