In the PIC simulations, there is a mention of Solids slip velocity scale factor.
What is that and how is it defined?
Also, what does scale factor of 1 mean and scale factor of 0 mean?
Also related to this:
What is the PARCEL FRACTION THRESHOLD? It defines it as below: but I do not understand how it gets involved in the pic calculations. ANy more documentation here would be great:
The “Solids slip velocity scale factor” is controlling the keyword “MPPIC_VELFAC_COEFF”. If the documentation supplied is not sufficient, I suggest you examine the source code and see how this variable is used. “Use the source, Luke”
– Charles
Also note that controls in the “Advanced” subsection generally do not need to be modified.
grep -ir mppic_velfac leads pretty quickly to the file
des/pic/integrate_time_pic.f90
where we see:
...
! VEL_FAC = empirical velocity scaling factor; programmer's control
! for testing only
DOUBLE PRECISION :: VEL_FAC, MFP
...
! VELFAC is an empirical parameter that damps the slip velocity
VEL_FAC = MPPIC_VELFAC_COEFF
...
! Note: Musser modified SLIPVEL with an empirical velocity factor on the
! bulk, i.e. "VEL_FAC *" but should not be necessary
! This appears to act as a damping factor for decision tree, sort of like
! in Barracuda manual where the "fraction of average velocity" is accounted
...
! Calculate the velocity differential between bulk solid vel and parcel
! u bar sub p - u tilda sub p
slipvel(1) = VEL_FAC*U_s(IJK,1) - VEL(1)
slipvel(2) = VEL_FAC*V_s(IJK,1) - VEL(2)
slipvel(3) = VEL_FAC*W_s(IJK,1) - VEL(3)
I hope this answers your question.
Thanks Charles!
Yes, I was looking for a physical interpretation of this scale factor rather than the code:
Basically, what does a scale factor of 1 imply in practice?
Scale factor of 1 means no scaling.
Got it thanks! This was very helpful!
That said, when slip factor is 1, and a certain drag model is chosen (say Gidaspow), the U_s will be calculated based on the parcel acceleration, drag force, all other body forces etc. right?
But if we have a scale factor of 0.5, that means the solid_velocity calculated from all the necessary equations will be modified by 0.5X, and then the slip velocity will be calculated, right?
Yes, that’s my interpretation of the code snippet I posted. And that is the only place in MFiX that variable is ues.
1 Like
Sorry to bother again @Charles!
One last question, we had was what is signficance of pic_CFL_parcel_fraction?
Based on the cod, it seems to be a condition used to update the pic_time_step.
Could you kindly elaborate a bit more?
Please examine the code in model/des/pic/integrate_time_pic.f90 to see how this variable is used.
1 Like
Here slipvel is a slip velocity between the bulk of the solids phase and the current parcel. It is not the slip velocity between parcel and gas phase. The scale factor is used to damp this term. It is used when updating the parcel velocity, not when computing drag. It is recommended to keep the default value of 1 (meaning the slip velocity is exactly the velocity difference between the bulk of the solids phase and the parcel).
pic_CFL_parcel_fraction is used as a threshold when adjusting the PIC time step with a CFL number. Say the CFL number is set to 0.1 and pic_CFL_parcel_fraction is set to 0.01. this means the PIC time step will only be adjusted if more than 1% of the parcels have a CFL larger than 0.1.
1 Like
Thanks Jeff! So basically always keeping it as 1 will not include any damping.
Is it advisable to keep it less than 1 for packed beds or dense solid regions, where we might have damping on the parcel due to neighboring particles?