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.

Stabilization in PDE mode: Convection problem

Please login with a confirmed email address before reporting spam

Hello everybody,

I'm trying to solve a 2D compressible convection problem using the PDE coefficient form [div(-alpha*u) = f]. I would like to include SUPG and Crosswind diffusion by means of a weak contribution. I tried to reproduce the stabilization terms that appear in the "Convection-Diffusion Mode", but I couldn't manage to make it work. I obtain error 6250 because the Jacobian of alpha cannot be evaluated.

I'm not sure if the variable names are the problem or if there is something else. I tried: (al1, al2) and (alx, aly) for the alpha variable; and the derivatives as: al11, alxx, diff(alx,x), etc. According to the theoretical background this variable should be named "alx", but in the PDE section is mentioned as "aluly" and includes the "u" contribution (alpha*u).

Does anybody have a clue to cope with this problem? Which is the correct variable naming?

Thanks in advance.

2 Replies Last Post 25.04.2010, 12:09 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22.04.2010, 19:59 GMT-4
I implemented 2 week ago a 2D compressible fluid flow in the "weak" form by myself, and I'm also facing the same problem as yours: I'm not able (don't know how) to implement the stabilization terms in my home made fluid flow appliquation. The stabilization terms are nativelely implemented in comsol fluid appliquations.

My goal was to implement fluid flow by myself, which is in my opinion, the power of comsol.

If ever you find the theory behind thoses stabilization terms, contact me!

Regards,
I implemented 2 week ago a 2D compressible fluid flow in the "weak" form by myself, and I'm also facing the same problem as yours: I'm not able (don't know how) to implement the stabilization terms in my home made fluid flow appliquation. The stabilization terms are nativelely implemented in comsol fluid appliquations. My goal was to implement fluid flow by myself, which is in my opinion, the power of comsol. If ever you find the theory behind thoses stabilization terms, contact me! Regards,

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25.04.2010, 12:09 GMT-4
In the Model Library there is an example showing how to implement SUPG using the discontinuous Galerkin shape function shdisc(...). It's in Equation-Based Models -> Transport Problem. Look also in the documentation for this Model for a description of how the weak term is implemented in terms of a domain ultraweak (bd.weak) expression. It's fairly complicated, but once you figure out the definitions and usage of up(...), down(...), unx, uny, dnx, dny, etc., it all makes sense. In the same section of the documentation there is also a description of regular streamline diffusion, whose weak expression is somewhat simpler than SUPG.

I haven't been able to find a description of the weak expression for cross wind diffusion, but it ought to be similar to streamline diffusion, except operate orthogonal to the convection vector instead of parallel to it. I've requested further info from Comsol support but so far have not heard back from them.

In the Model Library there is an example showing how to implement SUPG using the discontinuous Galerkin shape function shdisc(...). It's in Equation-Based Models -> Transport Problem. Look also in the documentation for this Model for a description of how the weak term is implemented in terms of a domain ultraweak (bd.weak) expression. It's fairly complicated, but once you figure out the definitions and usage of up(...), down(...), unx, uny, dnx, dny, etc., it all makes sense. In the same section of the documentation there is also a description of regular streamline diffusion, whose weak expression is somewhat simpler than SUPG. I haven't been able to find a description of the weak expression for cross wind diffusion, but it ought to be similar to streamline diffusion, except operate orthogonal to the convection vector instead of parallel to it. I've requested further info from Comsol support but so far have not heard back from them.

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.