Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
Small quantity and numeric zero
Posted Jan 12, 2011, 1:51 p.m. EST Studies & Solvers Version 3.5a 2 Replies
Please login with a confirmed email address before reporting spam
Hi all,
I have just a question about Comsol.
The dimensions of my problem are very small
(<10E-9 m = 1nm). Because of these dimensions,
the different quantities calculated by comsol are close
to 10E-40 SI...My problem from particle physic and
all these very small numeric values are natural in
this domain...
All my quantities are < numeric zero (=10E-14) of
the software.
Is there a solution to work with these small quantities
in comsol ? Is it dangereous to work with these values ?
Thank you for your answer !
S
I have just a question about Comsol.
The dimensions of my problem are very small
(<10E-9 m = 1nm). Because of these dimensions,
the different quantities calculated by comsol are close
to 10E-40 SI...My problem from particle physic and
all these very small numeric values are natural in
this domain...
All my quantities are < numeric zero (=10E-14) of
the software.
Is there a solution to work with these small quantities
in comsol ? Is it dangereous to work with these values ?
Thank you for your answer !
S
2 Replies Last Post Jun 17, 2011, 3:19 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
then it is a question of scaling, to get them back around "1" for the simulations. Normally the smallest number that is useful is "eps" (try it out as a parameter) and as most expressions in physics are 2nd order the ruling RATIO is around sqrt(eps). for Double floating point number in todays binary format, ad this is not very large.
But if all your values are small then by scaling them you can get out of this zone. However this might mean that you need to scale manually, for that take a course on the solvers or ask your COSMOL support.
Another way is to do a transform (if all values are positive) i,e via a y=log(x) but then you must also rewrite somewhat the physics equations too
--
Good luck
Ivar
then it is a question of scaling, to get them back around "1" for the simulations. Normally the smallest number that is useful is "eps" (try it out as a parameter) and as most expressions in physics are 2nd order the ruling RATIO is around sqrt(eps). for Double floating point number in todays binary format, ad this is not very large.
But if all your values are small then by scaling them you can get out of this zone. However this might mean that you need to scale manually, for that take a course on the solvers or ask your COSMOL support.
Another way is to do a transform (if all values are positive) i,e via a y=log(x) but then you must also rewrite somewhat the physics equations too
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hello mr. Ivar,
I think that I am experiencing similar problems. I am basically following the boiling water tutorial, however on a much smaller scale - the nozzle or nucleation site in my simulation is 2e-5 m. When I run the model, the model seems to run smoothly as the initial gas pocket nucleates and forms a bubble on the surface.
However, when the bubble reaches a certain size it seems to become static and does not depart from the heater surface.
I suspect that this could be a numerical issue as the geometry of the nozzle is very small with respect to the geometry of the liquid pool...
Does this seem reasonable?
Kind Regards
Søren
NB. I have set the scaling to automatic
I think that I am experiencing similar problems. I am basically following the boiling water tutorial, however on a much smaller scale - the nozzle or nucleation site in my simulation is 2e-5 m. When I run the model, the model seems to run smoothly as the initial gas pocket nucleates and forms a bubble on the surface.
However, when the bubble reaches a certain size it seems to become static and does not depart from the heater surface.
I suspect that this could be a numerical issue as the geometry of the nozzle is very small with respect to the geometry of the liquid pool...
Does this seem reasonable?
Kind Regards
Søren
NB. I have set the scaling to automatic