Segregated Group Solver different results

Hi folks,

I have been using COMSOL for a while now and, sometimes, when facing a difficult convergence problem I do mess around with the solvers a bit.

I have seen some posts here on the forum regarding the segregated vs fully coupled and my question/issue
is somewhat of the same nature, I presume.

When doing CFD for solid-liquid mixtures I tend to have some issues with the dispersed phase, with 3 different segregated groups. I can get around this by putting the dispersed phase variables in the same group as the velocity and pressure, and then the convergence is considerably smoother.

The problem here is that the results have changed by comparison when I don't mess with the segregated groups. The pressure and velocities resulting from putting the dispersed phase variables in the same group as velocity and pressure. Can anyone explain why this happens?? I would think that since the fully coupled and segregated solvers are to provide with the same solution, then altering the segregated groups to emulate a quasi-fully coupled solver wouldn't mess with the results...

I'll say that my understanding of the solvers is somewhat superficial resulting from COMSOL documentation, and if someone can point me to some papers that better clarify the difference I would be very grateful.

Hope someone has some insight on this subject.

Best Regards,

Rui Silva