Mixture Not Working in Matlab

[depcik]

Description

I have been using REFPROP 10 well with single fluids, however it is not working with a mixture

Steps to Reproduce

Code

clear all;
RP = py.ctREFPROP.ctREFPROP.REFPROPFunctionLibrary('C:\Program Files (x86)\REFPROP10');
MOLSI = RP.GETENUMdll(int8(0), 'MOLAR BASE SI').iEnum;
MASSI = RP.GETENUMdll(int8(0), 'MASS BASE SI').iEnum;
iMass = int8(0);                            % 0: molar fractions; 1: mass fractions
iFlag = int8(0);                            % 0: don't call SATSPLN; 1: call SATSPLN
ierr = int8(0);
z = {1.0};                                  % Composition
% Set the flags for the desired behavior of the program
r = RP.FLAGSdll('Bounds', int8(1));
WF = 'R502.MIX';
% Properties
p1 = 1.5168e+05;    % Pressure at State 1: Pa
x1 = 1;             % Saturated vapor at State 1
% Get the enthalpy given the quality
r = RP.REFPROPdll(WF, 'PQ', 'H', MOLSI, iMass, iFlag, p1, x1, z);       % Specify the pressure and quality
o = double(r.Output);
h1 = o(1);      % [J/kg]
% In REFPROP for R502 - the enthalpy would be 326.16 kJ/kg at this pressure
r = RP.REFPROPdll(WF, 'PQ', 'T', MOLSI, iMass, iFlag, p1, x1, z);
o = double(r.Output);
T1 = o(1);      % [K]
% The temperature should be 236.89 K (get -9999990)
r = RP.REFPROPdll(WF, 'PQ', 'S', MOLSI, iMass, iFlag, p1, x1, z);
o = double(r.Output);
s1 = o(1);      % [J/kgK]
% And the entropy should be 1.5458 kJ/kgK

Behavior: I expect to get a value of 326.16 kJ/kg at that pressure, but I end up getting -9999990. Same thing with temperature (should be 236.89 K) and entropy (should be 1.5458 kJ/kgK)

Versions

REFPROP Version: 10 Operating System and Version:
Microsoft Windows 10 Access Method:
Python in Matlab

Additional Information

If possible, please post examples and/or screenshots of the issue.

[ianhbell]

Some comments:

• You no longer need to call GETENUMdll to retrieve the unit classes, they are now exposed as attributes of the REFPROPFunctionLibrary, so you could do: RP.MOLAR_BASE_SI
• There are some bugs dealing with predefined mixtures in the high-level interface. Thus I would recommend to call SETMIXdll to get the names and molar compositions, and then pass the compositions (but not the names, leave hFld as the empty string ''. That should always work (in theory).

[ianhbell]

Also:

• I added the matlab tag to your code block to get the right formatting
• I thought initially this was a problem with reference states (as is the case 99% of the time when someone reports "wrong" entropies), but it seems something else is happening. In general you should check the returned value of ierr at each call. If it is something other than 0, you should print herr, which will give an indication of why the call failed (and returned the default bad value)

[depcik]

OK, I have now been able to get values for h1, T1, and s1, but they are not correct. If you look at REFPROP10 for R-502 at a saturation pressure of 1.5168e+05 Pa (0.15168 MPa) you get:

My code now is:

WF = 'R502.MIX';
sm = RP.SETMIXdll(WF, 'HMX.BNC', 'DEF');        % See SETUPdll for specifics
% Properties
p1 = 1.5168e+05;    % Pressure at State 1: Pa
x1 = 1;             % Saturated vapor at State 1
% Get the enthalpy given the quality
r = RP.REFPROPdll('', 'PQ', 'H', MASSI, iMass, iFlag, p1, x1, sm.z);       % Specify the pressure and quality
o = double(r.Output);
h1 = o(1);      % [J/kg]

That gives me an enthalpy value of 3.3779e+05 J/kg or 337.79 kJ/kg which is different than the REFPROP value of 326.16 kJ/kg. The temperature and entropy provide different values too.

[ianhbell]

So then it comes to reference states. Try changing the reference state in the GUI; with the appropriate selection the DLL and the GUI should agree (the default reference states in the GUI and the DLL are different, a historical decision that we have been forced to live with into the present).

[depcik]

If you look at the image I uploaded, it indicates that: "Ref. State: Default" which is what I am running with SETMIXdll. I just tried the other reference states in the GUI and I get different answers than Matlab.

[depcik]

Wait. If I apply the reference state to "Each pure component" in the GUI, then it looks like I get the same answers between the GUI and Matlab.

[ianhbell]

Right, I was about to type that response to you.

[depcik]

OK, it seems like I am good to go at this point. Thanks!

[ianhbell]

Yes, I am hoping we can get some docs together on the reference states, and precisely how they operate, as it is a mystery to me still.

[ianhbell]

But glad to hear you are resolved for now.