Learning Forum

[sjohn]

I am using k-omega turbulence viscous physics (along with mulitphase VOF) to model steady state flow through a presure swirl atomizer. I specify an inlet flowrate and outlet pressure (near ambient) and get good correlation between the steady state inlet pressure obtained from simulation and the pressure obtained using the design value for water. I don't get a good correlation when the fluid is changed from water to a different fluid which has roughly 40% higher density and 30 times the visocsity of water. The model over estimates steady state inlet pressure by roughly 1.8 times.

The roughness height is set to zero but do I need to set the roughness constant to zero. Is the roughness constant only applicable for k-omega viscous physics? In the Fluent theory guie,  roughness constant is said to be applicable for k-EPSILON model.

Because of the change in viscosity and density, should the model be using a different viscous physics? Thanks.

[Rob]

I'd not expect the roughness to do too much, so won't explain a factor of 1.8 on pressure. The physics should be OK too, but, is the thicker material a Newtonian fluid? Is there a thermal effect that could result in a lower viscosity in the nozzle in reality, but isn't included in the CFD model? How does the surface tension change between the two materials?

[sjohn]

The second fluid has surface tension similar to water but the visocsity is more sensitive to temperature. For example, vicosity reduces to 1/3rd of the initial value when the temperature rises from 20 to 50 C. My simulation is being run at room temperature so unsure if viscosity dependence of temperature is a big in sirl atomizer modeling. 
We suspect the fluid is slightly shear thining based on test data. Is there a way to account for this in material properties or the viscous model? Thanks.