Discussion Forum

Sir, Madam,

1st problem :

I use moving mesh ALE formulation (A solid cylinder rotating in liquid), for time dependant diffusion calculations.

I need to calculate a time derivative for a dependant variable that we use in a coupling for another physics

Let's say C is the dependant variable concentration inside the rotating cylinder.

I want to determine the TIME MATERIAL DERIVATIVE : d(C,TIME) (and not d(C,t) )

I define a Variable on the rotating domain with d(C,TIME) as expression, and comsol doesn't recognise the expression, but the expression d(C,t) (which is the eulerian derivative) is recognised.

I think the problem comes from the variable TIME which is supposed to be added with the ALE moving mesh.

When we make the postprocessing of d(C,TIME), we have a result

In comsol 3.5 and 3.5a this calculus used to work (We defined the d(C,TIME) as a a domain expression).

I also noticed, the variable TIME

The question is : How to define a MATERIAL DERIVATIVE with Variable TIME as an expression on subdomain ?
We have a comsol 4.1 license in our lab.


2nd problem :
I have a lot of problems in defining Identity pairs, in my old models, Comsol used to recognise pairs of the assembly directly, now I have to define them by hand ( on 200 geometric elements, it's quite long), and finally, comsol won't calculate on the rotating domain, (it's like ther were no identity pairs).

Once again I had no problem on version 3.5, except a convergence issue.

Is there a trick to define more easely the identity pairs and make Comsol use the continuity condition.


Thank you for your help.

Nabil.

PS, I cannot include any model as this work concerns a patent application.
I am really thinking going back to Comsol 3.5.