Discussion Forum

Hai,

I have tried to model my problem using PDE module in both COMSOL 3.5a and 4.2a versions. It have 4 dependent variables. I got reasonable results using 3.5a and verified them analytically. When I tried to do the same model using 4.2a , I am getting different results. I have difficulty in finding the reasons for this behavior. Although results are following the same behavior qualitatively, their quantitative values were different. Analytical verification tells me that results were not correct.

I used Identity pairs(IP) and periodic boundary conditions(PBC) in my model.

I suspect that problems with 4.2a as follows.

a). PBC doesn't override the default zero flux condition, it only contributes with it.

For this, I applied a 'continuity' using IP at both boundaries, so that it overrides the default zero flux and contributes with PBC. Results has not changed.

b) Bondary conditions using IP:
In 3.5a, boundary conditions using pairs shows like n.(gamma1- gamma2)= *. If this * equals to 0, it means continuity at given IP. If * equals some value, it means some jump in flux is there at that boundary. (For eg. withdrawing current at electrode).

But in 4.2a its showing n.(gamma)= * instead n.(gamma1- gamma2)= *, I have no other option then to give the same value for * as in 3.5a. It could be the one reason for other results.

Moreover, I have tried linear extrusion coupling parameter alternative to PBC, still I am getting the same results.

I believe that all my inputs are correct and verified many times.

The reason I came to conclude that the problem is with new version is I tried the 3.5a model file with 4.2 and i have some other results than 3.5a, but I remodeled everything new in 4.2a, I am getting the same results as upgraded results, not the results obtained in 3.5a.

It would be great help for me if someone throws light on my issues.

Thank you
Ganesh