Discussion Forum

Hi

it could come from the interpolation COMSOL is doing when you select an arbitrarian line, to map mesh values to your line. Is the difference great ?

--
Good luck
Ivar

Hi

Thanks 4 your reply.

Yes, the difference is great. The boundary condition is 1(dirichlet). But when I plot the values at boundary using cross-section plot parameters dialogue box or export the data at the boundary, the boundary values range from 0.8 to 64. Actually that trend is similar to the results of the PDE inside the domain. The plotted boundary values are changing with time also.

Does COMSOL take the boundary condition which I specify (i.e. 1) for all times or is it changing the boundary values after each time step? So I am totally confused about what is happening.

For simpler problems, COMSOL is showing the correct boundary values when I plot.

Can you please tell what is the solution for this? It will be of great help.

Regards,
Seetha

Hi

normally a BC defind with a constant, or a parameter that is not updated by any sweep solver, remains constant. But, you might have a very steep gradient of your dependent varable just after the BC, then you mesh needs too to be very fine to correctly follow the values.

But I'm talking here in all generality, there are so many things that can happen, it's difficult to discuss without knowing realyy all details of your case

--
Good luck
Ivar

Hi

Thank you very much for your reply.

Yes, there is a steep gradient in the concentration between the boundary and the first node.

I refined the mesh to very fine one near the boundary using boundary layer mesh parameters. But then also the boundary values are changing with time. It is oscillating.

Is it possible to check the problem, if I send my COMSOL model?

Regards,
Seetha

Hi

take a look at the COMSOL Reference Guide Numerical Stabilization methods (p496 in v4.2a) this can also potentially be part of you issue

--
Good luck
Ivar