[njdyck]

I'd like to understand what the difference is between the reference pressure specified in the Domain tab and the reference pressure specified in the Material tab.

I have searched the documentation but I can't see a clear difference. In particular I've looked at section 1.2.8 in the CFX Solver Modeling guide, which discusses the influence of reference pressure on computations, but doesn't clarify my question.

Furthermore, I've read through many posts on this forum discussing reference pressure in CFX, including this forum post where something similar is asked, but the answer isn't very clear.

I'm running compressible flow simulations of the Ranque-Hilsch vortex tube (using Air Ideal Gas as my working fluid). Calculations are performed in a stationary reference frame, and buoyancy has been neglected, a mass flow inlet has been used with pressure outlet BCs. I'm using a range of turbulence models.

Currently I've set the reference pressure to 14.7 psi in both the domain and the air ideal gas material. Absolute Pressure varies throughout the domain between 16.7-100 psi.

1. Where in the computation are each of these reference pressures used?
2. Does changing either of the reference values make any difference in the computation (aside from reducing roundoff errors as discussed in section 1.2.8 of the Solver Modeling guide)?
3. Should these values be the same or are there cases where it's advantageous to specify different reference pressures?

Thanks!

[Opaque]

Careful here.

Reference Pressure in the materials definition is for property calculations. For example, say the calculation for entropy at a given (p,T) conditions is given for an ideal gas (constant Cp) as

S = S_ref + Cp * ln (Temperature / T_ref) - R * ln( Absolute Pressure / P_ref)

As long as the value for S_ref is consistent with T_ref, P_ref, the calculation should work,

Reference Pressure in the domain only shifts the Absolute Pressure to a relative pressure. Say you decide to use P_ref_domain then when evaluating the properties

Absolute Pressure = Pressure + P_ref_domain

Material Reference Pressure is completely independent of the domain reference pressure.

[njdyck]

Thanks for the response Opaque!

Originally Posted by Opaque

S = S_ref + Cp * ln (Temperature / T_ref) - R * ln( Absolute Pressure / P_ref)

As long as the value for S_ref is consistent with T_ref, P_ref, the calculation should work,

What do you mean by "is consistent with" here? Do you mean that

S_ref = - Cp * ln(T_ref) + R * ln(P_ref)

This also has sparked a follow-up question. The static enthalpy of an ideal gas in CFX is defined as (according to section 1.1.3.9.2 of the CFX Solver Modelling guide)

h_static - h_ref = integrate Cp(T) * dT from T_ref to T

The guide says:

When the solver calculates static enthalpy, either directly or from total enthalpy, you can back static temperature out of this relationship.

Now for a perfect gas (constant heat capacities) this formula reduces to:

h_static - h_ref = Cp * (T - T_ref)

So if I choose anything other than h_ref = Cp * T_ref that would mean the static temperatures (and contour plots) are off by a constant value?

[njdyck]

Further to my above post, The ideal gas equation depends on the thermodynamic (static) temperature (using the notation from section 1.2.2.2):

rho = w * P_abs / (R_0 * T)

Does anyone know if the Temperature T here is calculated using the equation from my above post? If so, then the density is incorrect whenever h_ref is not equal to Cp*Tref (for perfect gases anyway). As the density is used in the momentum equation, the velocities and pressures will be incorrect, and all the computed quantities are wrong.

I'm going to do some testing to see if there's a difference.

[Opaque]

All the above seems correct. Temperature can only be computed from the equation you described since the known thermodynamic state is defined by P, and h. For ideal gases, h is not a function of P; therefore, only h is needed.

As you described, the reference state consistency matters.