Multiphase flow in stirred vessel
 

Hi, everyone

Trying to simulate multiphase flow in stirred vessel: continuous phase (Water) + dispersed phase (Particles).

Physics:
- Eulerian Segregated Multiphase, two phases: Water and Particles, both are turbulent (k-ε or k-ω), Gravity and Solid Pressure are set;
- Multiphase Interaction: Continous-Dispersed, Particle Diameter is set (Interphase Length Scale) to 0.2 mm, Turbulent Dispersion Force, Virtual Mass Force, Drag Force (Gidaspow);
- Two regions: stationary and rotating (1000 rpm);
- Initial conditions: zero velocity, dispersed phase is settled (with Volume Fraction of 0.5-0.6) on the bottom occupying ~10% of tank height;
- Volume Fraction Convection = 2nd-order, Temporal Discretization = 2nd-order;

Stopping criteria for inner iterations:
- Maximum fluid velocity asymptotic;
- Volume Integral of Volume Fraction of Particles asymptotic.

Mesh is fine, prism layer is fine, but i encounter problems:
- In steady case (regardless under-relaxation factors even with ramp): Volume Integral of Volume Fraction of Particles decreases significantly during computation, thus dispersed phase is lost. Can't fix it at all.
- In transient case situation with dispersed phase is much better, but requires physical time is high and rpm is high (leads to small time step).
- Also i get huge Turbulent Viscosity of dispersed phase (and Y+ on the mixer body is up to 10000-50000) around the mixer body although all the dispersed phase is still on the bottom of the tank.

To shorten the post I might have omitted some points that are important - ask if there are any.

Can anyone please suggest any idea/advice on the matter?