Convergence problems in multiphase flow!
 

Hello,

I´m doing simulations for a helically coiled tube containing a liquid phase along with granules in the Euler- Euler multiphase solver. The problem is that I can't get it to converge. I started out by solving the flow without the volume fraction and got the residuals down to 10e-4 and then they flatened out. After that I added the volume fraction and got the residuals to decrease some without occilations. But after a while residuals starts to occillate very much. I have tried to lower the under relaxation but it didn't help. Somebody who knows what could be the problem?

Another thing is: is it necessary to use mass flow inlet, can't one use velocity inlet instead (the fluent manual says one should use mass-flow-inlet)?

//Mikael

Re: Convergence problems in multiphase flow!
 

Are you running transient? If not you probably should.

Re: Convergence problems in multiphase flow!
 

I have tried solving transient and steady without any results. Same behaviour in both cases. I can add that I'm solving laminar because I got a Re-number of around 70.

Re: Convergence problems in multiphase flow!
 

it is difficult to converge this kind of problems in both steady and unsteady state i feel if your residuals do not diverge, then you can accept them as acceptable results.

before that, check for your under-relaxation parameters (do not decrease them too much even) and method of discretization

keep us posted..this is a genuine problem in multiphase flow

Re: Convergence problems in multiphase flow!
 

Hi,

you can check the main variables at some strategic points in the geometry. If their values are not evolving with iterations or has a periodic evolution (transient), then it can be considered as converge. Cheers

Re: Convergence problems in multiphase flow!
 

Multiphase flow problems in FLUENT is not well solved,so how about modify your mesh file or boundary conditions and try it again? Some problems are grid-sensitive.

Re: Convergence problems in multiphase flow!
 

Thank you for the help, I have now tried to change boundary conditions between velocity inlet and mass flow inlet without any result. Changed discretization also without results. However I havn't tried to change the grid so maybe I can try that out (my grid has very low skewness thou). So far I have gotten the massbalance to match between inlet and outlet to 89% accuracy, not that god...