Reactor Simulation

[arvind.prasad]

I have trying to simulate a Plug Flow reactor in COCO. The Thermodynamics Pack used is TEA. Which Enthalpy should I use in the drop down list in the reactor GUI? If I use EnthalpyF the reactor temperatures are 2.5 times that that when Enthalpy( no heat of reaction) is selected.

[jasper]

Use EnthalpyF in case no heat of reaction is used:

http://cocosimulator.org/index_help.php?page=COUS/reactionenthalpy.htm

[arvind.prasad]

Thanks Jasper. The understanding of uses of three choices in the drop down list is
Use Enthalpy F if the property package has routines to calculate Enthalpy F. Enthalpy F is the enthalpy of the compound plus its enthalpy of formation. Now the reactor simulator program will calculate the heat generated/required in the reactor based of Enthalpy F.
Use Enthalpy(no heat of reaction) should be used incase the thermodynamic property generator calculates the enthalpy of the compound as the sum of the heat of formation plus the enthalpy required to raise the temperature of the compound. This is not the case with TEA, right.

[jasper]

If EnthalpyF is available, this is always a good choice. Heat of formation however is not always known for all compounds. In this case a property package may not be able to supply EnthalpyF. In this case, heat of reaction may be known and is required to close the heat balance. But now it gets a bit more complex to get things to be consistent.

In case heat of reaction is known (at reference conditions that should match those of the thermo!), you can use an enthalpy that does not contain heat of formation. EnthalpyNF is sure to not contain formation terms. For Enthalpy (without F or NF) it depends on the property package implementation (and perhaps configuration) whether heat of formation is or is not included. For TEA, the property Enthalpy does not include heat of formation.

TEA always has its reference point at 298.15 K, 1 atm, in the ideal vapor phase. So if you configure a heat of reaction, it should be either at the same reference conditions, or you should adjust the material object to match the reference conditions of the reaction package's heat of reaction. COFE can do this; you can change the reference conditions (pressure, temperature and phase) in the flowsheet setup. Effectively this results in COFE calculating enthalpy at the chosen reference conditions for the pure compounds (once) and correcting all calculated enthalpies by the sum of composition times compound reference enthalpties (and the same holds for entropies): http://cocosimulator.org/index_help.php?page=COFE/refstate.htm. Note that with such reference state correction, you can be sure that any formation terms that might or might not be present in Enthalpy are cancelled out. So Enthalpy as seen by the reactor unit operation, in case COFE's reference state correction option is selected, is always at known reference state conditions and without formation terms, no matter which thermo is used.

So in short:

- if formation heats are known, use EnthalpyF, don't use heat of reaction
- if heat of reaction is known, set COFE's reference state correction to the conditions at which heat of reaction is known

It is all a bit complex, and it would help if the reaction terms and the thermodynamics would come from the same source, as then more likely they will be consistent. This will be possible with the next version of the reactions standard, which is in the making (part of the work of the Thermo Special Interest Group (Thermo SIG) of CO-LaN).

[arvind.prasad]

Thanks Jasper.

[jasper]

You are welcome, and have a good start to 2013.

[arvind.prasad]

Thanks for the new year wishes Jasper. I wish you a very exciting 2013

[arvind.prasad]

In many of the papers the equilibrium constants is expressed in the form Kp=10^((A/T)-B).
One such paper is on the link http://www.arpapress.com/volumes/vol5is ... 5_3_02.pdf
I was going through the papers on Methanol synthesis reactor simulation. In the reaction template you have given a textbox to enter the formulae for entering the equilibrium constant. I have been entering it but it does not seem to work. The adiabatic or isothermal conversion is hardly 2%! Lets say the expression for the equlibrim constant for the CO+2H2=CH30H is Kp=10((5139/T)-12.621. Do we enter it in the same way?

[jasper]

Can you send me your fsd file?

[jasper]

It seems to me the equilibrium constant formula is properly entered as 10^((5139/T)-12.621).

In the reaction

CO + 2 H2 => CH3OH

this is interpreted as

Kp=[CH3OH]/[H2]^2/[CO]

where [] denotes mass concentration in kg/s as this is selected for the reaction basis.

The reactor results concentrations

[CO] 5.867786081 kg/m3
[H2] 1.689679154 kg/m3
[CH3OH] 0.003660365 kg/m3

giving Kp=0.003660365/1.689679154^2/5.867786081=0.000218495

which should equal at 573.51995 K:

Kp=10^((5139/573.51995)-12.621)=0.000218502

So it looks like the solution is ok for what is specified. But I reckon that the molar concentrations are to be used (Molality) which gives a different answer altogether.

Now a quick scan of the article reveals that the Kp values are only used in eq (4), so I am not at all sure whether they are intended to be used as equilibrium constants? (it might have been clarified in the text, I did not read the entire article).

The basis issue would have been apparent immediately from the report, which goes lost during persistence, resulting from transferring the unit operation back from the solution thread into the main thread (it was available by selecting Calculate Unit from the unit's context menu). Also from the units of measure of Kp in the reaction package it shows. The full list of reaction basis is documented here:

http://cocosimulator.org/index_help.php?page=CORN/reactions.htm

Now - let us see why the report was not available: ah - yes, this was a persistence problem introduced in COUSCOUS 2.5.0.1 and is now solved (fix available via CUP).

Finally, in your mail you state that selecting fugacity as the basis does not work. This makes the unit of measure of the equilibrium constant 1/Pa^2 (as the fugacities are in Pa), which would make the average order of fugacity about 70 Pa? I can imagine this will indeed not converge at 50e6 Pa. Check the units of measure there please.

Hope that helps - apologies for the report not being there to point out the used units of measure in the first place. Thank you for reporting the problem.