Softwares used:
- Ansys Design Modeler (for geometry)
- Ansys Mesh (for meshing)
- OpenFoam v1912 (for simulation setup and calculus)
- Paraview 5.12.1 (for post-processing)
About the geometry:
- Follows the standart Rushton turbine stirred tank
- Domain was cut in parts to facilitate meshing with hexahedral elements
About the mesh:
- Most domain has hexahedral elements
- MRF zone, such as impeller, were done with tetrahedral elements
- 662.051 elements
About the simulation setup:
- All the domain walls have no slip boundary condition
- The tank top has symmetry boundary condition to simulate an open tank
- Zero gradient pressure B.C. was applied for all parts, except for tank top
- Fluid is water with 1.0e-06 m²/s of kinematic viscosity
- Turbulence model is Realizable k-epsilon
- Turbulence parameters were calculated based on turbulence intensity, characteristic velocity and turbulence length scale (estimated)
- MRF approach was used for impeller motion
- Impeller angular speed is 200 rpm
- Solver pisoFoam with 2nd order schemes for all variables, Δt = 0.0001
- Simulation stopped when all variables residuals were at least in 10⁻⁵ order
- Courant number mean less than 0.0155 for entire simulation
Additional commentary: It's hard to see transient simulations with MRF approach, which is mostly used for steady-state situations. That's why I tried to simulate MRF mixing tank with a transient solver. The final results are very similar to a steady-state solver, like simpleFoam