[Fatnes]

Hi.

I am doing a master thesis in multiphase flow. My task is to create an Eulerian - Eulerian CFD model (using CFX) to study hydrate transport in a turbulent flow (picture of geometry is attached). The last few months I have been running many simulations without much success. The results do not comply with the experimental results at all and I have had problems with solver errors. I find this a bit strange because the same setup is used in Star-CD with great success.

In short text I shall simulate hydrate deposition in the horizontal section of my pipe. Experiments show that from 0.4 m/s -0.1 m/s hydrate will deposit. Up to now I have tried performing simulations with use of the Kinetic theory together with k-epsilon and SST.

I have attached an out-file giving my setup. I will be wery thankful if anyone can have a look and give me feedback if they have any suggestions about what to do to get reasonable results.

[ghorrocks]

One coefficient loop per time step can cause problems. I know second order timestepping does not work properly with only one coef loop per iteration and it would not surprise me if there are others. Read the documentation on obtaining convergence - it recommends 3-5 coefficient loop iterations per timestep, or 5-10 for tricky multiphase stuff. I suspect you may well be in tricky end of multiphase.

So do adaptive timesteps homing in on 5-10 coefficient loops per timestep.

[Fatnes]

Thank you very much for your suggestion, I will sure try that tomorrow morning!