solver problem when using Chemsep property package

Discusses use of COCO, the process simulation and modelling software suite from AmsterCHEM, downloadable from http://www.cocosimulator.org

Moderator: jasper

solver problem when using Chemsep property package

Postby ghadi123 » 26 July 2023, 08:54

Hello,

am trying to solve a flash separator in a flowsheet. since Tea doesnt include the PSRK EOS, i used chemsep prperty package for the inlet Stream of the flash. But sadly when using the chemsep property package with the PSRK the solver doesnt find a solution.
I actually would like that all the streams in the process to work with the PSRK EOS but this seems impossible.

In many cases i am facing quite difficulties with the solver, am trying to model a recycle flow with a Splitter, if i want just to change the inlet feed streams ( x10 the amount) the solver stops finding solutions. Is there any smart settings to use.
please take in considereation that i am not a COCO expert.

Best regards
Ghadi
ghadi123
 
Posts: 2
Joined: 13 January 2020, 12:09

Re: solver problem when using Chemsep property package

Postby jasper » 26 July 2023, 10:06

It is hard to say anything based on a generic description like that - can you attach a simple example of something that you expect should converge but does not?
User avatar
jasper
 
Posts: 1128
Joined: 24 October 2012, 15:33
Location: Spain

Re: solver problem when using Chemsep property package

Postby ghadi123 » 04 August 2023, 09:19

Hello Jasper,

Thanks for responding.

Mainly what i am trying to solve is a some sort of a brayton cycle for supercrtical CO2 Process with a Recycle of CO2 mixture.

The Flash Seperator is used to seperate the CO2 from Water. For that i want to use the PSRK.

After i adjusted the Process, the solver worked. But as i Change some factors it stops working, and never workes again till i keep trying to adjust and make the flowsheet much easier.

Attached is the COFE Flowsheet.

my main perpuse of the flowsheet is to evaluate the effect of the Recycle CO2 (After the Flash Seprator and massflow Splitter splitting faktor) on the efficiency and the Temepratures.

Best Regards
Attachments
Alam Cycle_ohneZwischenVerdichtung_Ausgeweitet_mitPSRK - Kopie.fsd
(106.38 KiB) Downloaded 6363 times
ghadi123
 
Posts: 2
Joined: 13 January 2020, 12:09

Re: solver problem when using Chemsep property package

Postby jasper » 04 August 2023, 11:19

At the basis of this seems to lie a flash problem. This problem is easy to isolate, but I have to check with the ChemSep authors what is going on. It is being investigated.
User avatar
jasper
 
Posts: 1128
Joined: 24 October 2012, 15:33
Location: Spain

Re: solver problem when using Chemsep property package

Postby jasper » 05 August 2023, 11:30

It has been identified by the ChemSep authors that the issue is specific to the PSRK model, but it will take a bit to repair it. In the mean time, selecting any other model should work around the issue without trouble. PR and SRK would be suitable for the problem at hand.
User avatar
jasper
 
Posts: 1128
Joined: 24 October 2012, 15:33
Location: Spain

Re: solver problem when using Chemsep property package

Postby hkooijman » 02 January 2024, 17:28

Ghadi,
I suspect the PSRK fitted group interaction parameters (GIPs) does not extrapolate to temperatures above 300 degrees C. Your incoming stream is 1400 C! I checked and the PSRK model flash worked up to 500 C but then its limits are reached, most likely because one of the GIPs approaches to values that are not realistic. In particular, I discovered the CO2-H2O parameters are the problem here. When these are set to zero the flash solves. In particular, it is parameter C that has the value 0.00131 which is too large. If it is assigned a value of 0.000131 it already works. Of course, this means that the whole temperature dependence of this binary needs to be refitted to work at such high temperatures. Alternatively, you may set the enthalpy model to ideal, but then the excess enthalpies are not computed and this reduces the fidelity of your model. The thing is, the quadratic temperature dependencies should be limited to some value but it is not apriori possible to say at what temperature the fit will nor work properly anymore. For practical reasons it may be much better to switch to the SRK-UMR altogether as that method performs very similar as the PSRK and since the GIPs are not fitted with a quadratic model, it does not suffer from the extrapolation error at temperatures >500 C for the CO2-H2O pair.
-Harry
hkooijman
 
Posts: 51
Joined: 13 April 2013, 23:50


Return to COCO (AmsterCHEM)

Who is online

Users browsing this forum: No registered users and 5 guests