Query on Property Calculation in the Two-Phase Region

[Sush1090]

Hello,

I am currently working on computing bulk properties, such as density, volume, temperature, entropy, and specific heats (Cp, Cv), using enthalpy and pressure as inputs.

In most cases, my local solver finds the temperature via the Tproperty function and then computes the required property. This approach works well outside the two-phase region. However, I am looking for a robust method to handle calculations in the two-phase region as well.

Specifically, I am searching for a tool similar to PropsSI that computes properties accurately using pressure and enthalpy. Is such functionality available in Clapeyron, or is there a recommended method for handling this?

Your guidance would be greatly appreciated.

Thank you,
Sushrut

[longemen3000]

Hello,
technically, what you need is a T-H (T-S and T-V) flash. At the moment, we only have P-T flash implementations. I think we need to track this, is ok if we use this issue for that purpose?

[Sush1090]

Yes indeed :)

[longemen3000]

i think i got an idea on how to do this.

• on Clapeyron 0.6.5, i introduced some modules (PT,PH,PS,VT), the idea is to do something like:

VT.enthalpy(model,v,T,z) #equivalent to Clapeyron.VT_enthalpy(model,v,T,z)
PT.enthalpy(model,p,T,z) #equivalent to enthalpy(model,p,T,z)
PS.enthalpy(model,p,s,z) #equivalent to calling TProperty and then evaluating enthalpy at p,T conditions

the idea is to add flash functions in the following way

result = PT.flash(model,p,T,z) #equivalent to a modified Clapeyron.tp_flash2(model,p,T,z), changing the output of tp_flash is breaking
comps,fractions,volumes,data = result #if you need specific data
hmix = enthalpy(model,result) #if you need to calculate a weighed sum of property for all phases, the "glide".

A good thing with Tproperty is that identifies equilibria conditions (when prop_bub < prop < prop_dew), so we can modify it to also return the condition (as you already saw in #309) Now, the idea is to hook an adequate flash for each pair of specifications, using the output of Tproperty as an starting point.

[Sush1090]

Hello,
I saw the new modules in the latest update, I was locally trying to make this work as well.
Correct me if I am wrong, if we rely on Tproperty internally for our calculations we can achieve this functionality only when we are gas or liquid. We probably will not be able to locate property when the state point is in "two-phase" region.

Because (p,T) cannot represent point in two-phase region (at least for single component systems).
Would we be able to have something like the following in CoolProp:

using CoolProp
fluid = "Water"
p = 101325
H_g = PropsSI("H","Q",1,"P",p,fluid);
H_l = PropsSI("H","Q",0,"P",p,fluid);
H_ = (H_g + H_l)/2 # in TwoPhase region

@show PhaseSI("H",H_,"P",p,fluid)
@show PropsSI("S","H",H_,"P",p,fluid)

Here we can compute entropy in two-phase. But if we do the same with Tproperty we will internally return saturation temperature and then we will end up returning entropy at liquid phase.

Unless there is something different in tp_flash2 I am not sure if Tproperty can handle this.

Do correct me if I am wrong here.

[longemen3000]

You are correct, p, T is not enough for exactly determining the state point, but if we know that we are in a equilibria state, then we can call a flash. In the latest master I added PS and PH flashes for single components, expecting to add the multicomponent versions soon

[Sush1090]

Ah I see. I'll check this too. Thanks

[longemen3000]

You are correct, p, T is not enough for exactly determining the state point, but if we know that we are in a equilibria state, then we can call a flash. In the latest master I added PS and PH flashes for single components, expecting to add the multicomponent versions soon.

[longemen3000]

on the current master:

julia> model = cPR(["ethane","propane"],idealmodel=ReidIdeal)
PR{ReidIdeal, TwuAlpha, NoTranslation, vdW1fRule} with 2 components:
 "ethane"
 "propane"
Contains parameters: a, b, Tc, Pc, Mw

julia> p = 101325.0
101325.0

julia> h = -13831.0 #enthalpy at T = 0.5*T_bubble + 0.5*T_dew
-13831.0

julia> z = [0.5,0.5]
2-element Vector{Float64}:
 0.5
 0.5

julia> Clapeyron.ph_flash(model,p,h,z)
([[0.17317188972407677, 0.8268281102759232], [0.6229562865518485, 0.37704371344815146]], [0.4753594797398013, 0.5246405202601987], [6.610779468339207e-5, 0.017026792382129566], Clapeyron.PTFlashData{Float64}(101325.0, 213.38726377112155, -162.11074278396245))

julia> enthalpy(model,result)
-13830.999999999996

[Sush1090]

I get this error when running the script

ERROR: UndefVarError: `spec` not defined in `Clapeyron`
Suggestion: check for spelling errors or missing imports.
Stacktrace:
 [1] Clapeyron.FlashSpecifications(spec1::typeof(pressure), val1::Float64, spec2::typeof(enthalpy), val2::Float64)
   @ Clapeyron C:\Users\sush9\.julia\packages\Clapeyron\splf7\src\methods\property_solvers\multicomponent\flash.jl:133

I added the master branch using

add https://github.com/ClapeyronThermo/Clapeyron.jl.git

status

[7c7805af] Clapeyron v0.6.5 `https://github.com/ClapeyronThermo/Clapeyron.jl.git#master`

[longemen3000]

huh, weird, it works in my REPL right now, let me check.

btw, the formulation of the ph flash was made totally generic, so the next steps are adding a ps flash (and other specifications)

[Sush1090]

This is actually amazing. This is very close to CoolProp or RefProp

[Sush1090]

There is a difference in your version or the version on the master branch?
Because I cannot precompile it:

(ClapeyronFlash) pkg> precompile
Precompiling project...
  ✗ Clapeyron
  0 dependencies successfully precompiled in 5 seconds. 121 already precompiled.

ERROR: The following 1 direct dependency failed to precompile:

Clapeyron

Failed to precompile Clapeyron [7c7805af-46cc-48c9-995b-ed0ed2dc909a] to "C:\\Users\\sush9\\.julia\\compiled\\v1.11\\Clapeyron\\jl_CD2A.tmp".
ERROR: LoadError: UndefVarError: `helmholtz_energy` not defined in `Clapeyron`
Stacktrace:
  [1] top-level scope
    @ C:\Users\sush9\.julia\packages\Clapeyron\GVLnz\src\methods\property_solvers\multicomponent\flash.jl:212

[longemen3000]

yes, fixing that right now.

by the way, the current result type from ph_flash could change between master and the release, but the way to use it (result = ph_flash(args...), h = enthalpy(model,result) will be the same

[Sush1090]

It works! thank you

[Sush1090]

Can we close this issue?

[longemen3000]

This needs a release and a lot of polish, so not now.