Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

question regarding time dependant solver with quasistatic components

Please login with a confirmed email address before reporting spam

Hi,
I have a general question but before asking it, i am poposing the following simple model and associated headache :-)
[ file attached]
there is 2 stationary uncoupled 1d pde with obvious analytical solution
I beleive the attached model solve:
du/dx =-sin(10*pi*x) u(0)=t
dp/dx=0 p(0)=t

note that I am using the weak form.
t of course is representative of the time...
these equations have obvious analytical solutions

OK
Now if I choose the parametric solver with t as a parameter [ as in the attached file]comsol calculate the correct solution
BUT
if I want to use the time dependant solver the solution computed by comsol are GROSSLY wrong.
specifically it seems [ but this is just a feeling that it does not treat correctly the lack of constraint on the right boundary....trying to zero it at any cost....

Now I cannot find anything in the solver setting documentation that tell me how to setup the time dependant solver to calculate the solution.


I think this is an IMPORTANT example because there are a lot of time dependant multiphysic where some parts of the model can be solved in a quasistatic approximation with only the load [ represented by my constraint at 0 in my example depending upon the time and where the solution is then coupled back to another part of the model as a slowly time varying coefficient or loading term.... [ very different timescale for example]

So my questions are
1- can someone explain to me why the time dependant solver with its default setting fail to compute the correct solution providing an obvjously wrong soltion instead while the parametric solver provide the correct solution?
2- what do I need to do on the setting of the time dependant solver to get the correct solution for these equations?
3- In the light of thisexample and more generally what do I need to do in order to couple a static equation with a time varying load in a complex time dependant multiphysic modell?

I have carefully examined the doc and the knowledge base and found nothing relevant regarding this problem.

Thanks a lot in advance
JF


3 Replies Last Post 27.03.2010, 08:41 GMT-4
Daniel Smith COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26.03.2010, 17:30 GMT-4
You have to zero out the terms in the dweak tab.

/ Dan
You have to zero out the terms in the dweak tab. / Dan

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27.03.2010, 05:13 GMT-4
thanks for the quick answer
friday night always difficult :-)

jf
thanks for the quick answer friday night always difficult :-) jf

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27.03.2010, 08:41 GMT-4
Hi

Indeed, that last dweak set the derivative of "u" and "p" to "0" at each node, but only in time dependent mode, if I get this right

Thanks for the info, I learned something again, it's really only with these simple cases that one catch effectively the importance of each entry

Do you have a couple of other of these Jean-François ?
even if you have the reply, we others might learn something mroe ;)

Ivar
Hi Indeed, that last dweak set the derivative of "u" and "p" to "0" at each node, but only in time dependent mode, if I get this right Thanks for the info, I learned something again, it's really only with these simple cases that one catch effectively the importance of each entry Do you have a couple of other of these Jean-François ? even if you have the reply, we others might learn something mroe ;) Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.