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.

Smoothing Curve in Comsol 3.5

Please login with a confirmed email address before reporting spam

Hi Comsol Experts,

I have a parameter pp that is discontinuous between two subdomain at a boundary along the r-coordinate. I would like to smooth this parameter at that boundary (as shown in the attached figure) as I will use it in my simulation. Does anyone know how to do that in comsol as it continues to simulates?

I have tried to make the elements smaller but still I obtain the same pp figure.

Your reply is greatly appreciated.


3 Replies Last Post 2012/06/24 11:10 GMT-4
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 2012/06/24 2:44 GMT-4
Hi
if your pp is discontinuous, does that mean you have an "up" value and a "down" value that you might read in and average ?

Or do you want to generate an average curve on your noisy data ?, in which case you could enter it into an interpolation function, or use an optimisation run to get a smooth parabolic or polynomial fitting

--
Good luck
Ivar
Hi if your pp is discontinuous, does that mean you have an "up" value and a "down" value that you might read in and average ? Or do you want to generate an average curve on your noisy data ?, in which case you could enter it into an interpolation function, or use an optimisation run to get a smooth parabolic or polynomial fitting -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012/06/24 10:45 GMT-4
Hi Ivar,

Thanks for the quick reply.


You are right in both cases: pp is discontinuous meaning below the boundary, in which I plotted pp, it has high and above that boundary it is low value. So, I am guessing that there is an up value and a low value but not located in the boundary that I have plottted. However, I do not want to read the average pp. I want to remove the noise and obtain a smooth average curve as COMSOL is solving.

So, I want COMSOL to solve my problem, extract pp, smooth it by "optimisation run to get a smooth parabolic or polynomial fitting", and use the result to solve the next step. Can you please tell me how to do that, or send me the link/title to the document that explains it in Comsol 3.5.

Truely appreciate your help

Suzan
Hi Ivar, Thanks for the quick reply. You are right in both cases: pp is discontinuous meaning below the boundary, in which I plotted pp, it has high and above that boundary it is low value. So, I am guessing that there is an up value and a low value but not located in the boundary that I have plottted. However, I do not want to read the average pp. I want to remove the noise and obtain a smooth average curve as COMSOL is solving. So, I want COMSOL to solve my problem, extract pp, smooth it by "optimisation run to get a smooth parabolic or polynomial fitting", and use the result to solve the next step. Can you please tell me how to do that, or send me the link/title to the document that explains it in Comsol 3.5. Truely appreciate your help Suzan

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 2012/06/24 11:10 GMT-4
Hello

Sorry, but I do not know where to look for something like that, what I do not understand is how you can get such a noisy variable from something solved, as I believe it should be locally smooth, possibly with stair steps, but not noise like that

Then if pp is a value derived from the domain then the boundary will get, by heritage, an average value from the "up" and "down" domains. If you are in union mode then you have normally continuity mode, hence pp_up=pp_down=pp_average but if you have a true discontinuity (t"hin" boundary property or assembly mode) then the up and down may be different and hence pp_average in between. So be sure you know if its pp_u, _dovn, or _average you get as result.

Next for me my first impression is that you are seeing numerical noise, perhaps even from a mesh density issue
.
Now what could perhaps be done is to add "pp" as a new local dependent variable and request that its second or third derivative is =0. But how to do that depends on your model set-up and the pp equation, anyhow not something I do everyday

--
Good luck
Ivar
Hello Sorry, but I do not know where to look for something like that, what I do not understand is how you can get such a noisy variable from something solved, as I believe it should be locally smooth, possibly with stair steps, but not noise like that Then if pp is a value derived from the domain then the boundary will get, by heritage, an average value from the "up" and "down" domains. If you are in union mode then you have normally continuity mode, hence pp_up=pp_down=pp_average but if you have a true discontinuity (t"hin" boundary property or assembly mode) then the up and down may be different and hence pp_average in between. So be sure you know if its pp_u, _dovn, or _average you get as result. Next for me my first impression is that you are seeing numerical noise, perhaps even from a mesh density issue . Now what could perhaps be done is to add "pp" as a new local dependent variable and request that its second or third derivative is =0. But how to do that depends on your model set-up and the pp equation, anyhow not something I do everyday -- Good luck 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.