[ishan_ae]

Hello.

I am simulating subsonic compressible flow through a duct.

I plan to initialize the rhoSimpleFoam solution with that of simpleFoam.

I have a couple of questions for this approach which have confused me and slowed down my progress for quite some time:

1. Would this be the correct way to do things?
   • Solving for turbulence, the usual u,v,w,P and then for T seems correct to me.
2. Since the pressure field solved using simpleFoam is not pressure but kinematic pressure, how can I use this as an input to initialize the rhoSimpleFoam solution?
   • I ask this because for simpleFoam we will solve for kinematic pressure and not pressure. The units are different from that of pressure.
   • Do I need to multiply my simpleFoam kinematic pressure solution with the reference density before I input it to rhoSimpleFoam solver? This sounds complicated to me and I don't know how to do it.
3. Mass flow rate or mass flow rate divided by density for flowRateOutletVelocity ?
   • My setup BC contain totalPressure at inlet and flowRateOutletVelocity at outlet. For simpleFoam solving, the flowRateOutletVelocity requires an input for volumetric flow rate and not mass flow rate. I have mass flow rate available with me.
   • At the outlet, I have estimated the density using isentropic flow relations. Would it be wise to divide the mass flow rate with this density, and then specify this as volumetric flow rate to the BC for simpleFoam solution or, should I supply mass flow rate as it is? The latter seems incorrect to me but I am not able to convince myself why so.

Some information on this would be helpful.

[dlahaye]

One possible way out is (?) to use rhoSimpleFoam at low Mach number instead using simpleFoam. That is, to use the ramping of parameters of the problem (e.g. mass inflow rate) to generate the initial guess. The foamDictionary scripting tools facilitates the process. Queries on boundary conditions can possibly be resolved by looking into tutorials.

[ishan_ae]

Hello.

I was not able to clearly understand what you meant by ramping of parameters. Do you mean steadily changing the parameters of the flow or the solution ? With regards to the Mach number, it is supposed to reach a maximum of 0.7.

Another option which is coming to my mind right now is that I can switch of turbulence solving in rhoSimpleFoam run and then once u,v,w,T have acceptable convergence value, I can solve for turbulence. But again, I am not sure if this is the correct approach. Someone with more experience and knowledge can throw some light on it.

[dlahaye]

gradually increase mass inflow rate.

[ishan_ae]

Ok. Thanks for the inputs. I will keep that approach as one of the options.

[Carlo_P]

Hey ishan_ae, what I normally do is to run a simpleFoam simulation, then use the U and turbolence to inizialize the rhoSimpleFoam simulation.


The pressure and temperature will start from zero.


This is the simplest way.


Another way is to convert pressure from simple to rho.


The formula can be find here:


https://github.com/Carlopasquinucci/fromPtoPComp


This script is not wokring perfectly, but it is a good starting point.

[ishan_ae]

Thanks Carlos. I am also thinking of using U field for initialization.

I will take a look at your script. Thanks for sharing.

[Carlo_P]

You can also have a look into his discussion:
Calculate a new filed from variables in the controldict