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 The information provided may be out of date.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.

spatial derivatives in material properties?

Please login with a confirmed email address before reporting spam

Hi,

I am trying to model a material with non-local response. Particularly, I need a non-local electric permittivity, but I guess it does not really matter for this discussion, as the question is quite general. The easiest way to introduce such non-locality in equations is to replace the usual (local) property epsilon with the operator:

epsilon = epsilon_0 + epsilon_1 (d/dx) + epsilon_2 (d^2/dx^2)

It is easy on a paper, but is it at all possible in COMSOL?

4 Replies Last Post 1 oct. 2013, 07:07 UTC−4
Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30 sept. 2013, 12:52 UTC−4
Why not would just input epsilon as a function of x?
Why not would just input epsilon as a function of x?

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30 sept. 2013, 13:52 UTC−4
No, it's not the same! Non-locality means that the material response (polarization) in any point of space x0 is defined by the external electromagnetic field within a certain area surrounding this point.

Local response: P(x)=epsilon(x) E(x)

Non-local response: P(x)=integral_0^infinity epsilon(x,delta) E(x-delta) d delta

Physically epsilon(x,delta) is localized in its second argument (delta). In this case, the above integral can be replaced with derivatives, as in my original post
No, it's not the same! Non-locality means that the material response (polarization) in any point of space x0 is defined by the external electromagnetic field within a certain area surrounding this point. Local response: P(x)=epsilon(x) E(x) Non-local response: P(x)=integral_0^infinity epsilon(x,delta) E(x-delta) d delta Physically epsilon(x,delta) is localized in its second argument (delta). In this case, the above integral can be replaced with derivatives, as in my original post

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30 sept. 2013, 18:35 UTC−4
Oh, a nonlocal response. That’s not trivial to implement in FEA in my opinion. There is a similar issue in solid mechanics in the implementation of nonlocal elasticity formulations such as “Cosserat” elasticity.
Oh, a nonlocal response. That’s not trivial to implement in FEA in my opinion. There is a similar issue in solid mechanics in the implementation of nonlocal elasticity formulations such as “Cosserat” elasticity.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 1 oct. 2013, 07:07 UTC−4
Hi,

This thread is similar to this one:
www.comsol.com/community/forums/general/thread/25040/
where spatial integrals are also discussed. Look for the "dest" operator at documentation.

Regards.
Jesus.
Hi, This thread is similar to this one: http://www.comsol.com/community/forums/general/thread/25040/ where spatial integrals are also discussed. Look for the "dest" operator at documentation. Regards. Jesus.

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.