non- convergence with flowrate specification

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

Moderator: jasper

non- convergence with flowrate specification

Postby thubelihle » 13 June 2018, 05:25

Hi. I specified reflux ratio and bottoms flowrate and the column converged. I then specified distillate and bottoms component flowrates, it did not converge. where do l see the output report to see what's the problem causing non- convergence.
Attachments
non- convergence.PNG
non- convergence.PNG (27.63 KiB) Viewed 5152 times
DISTILLATION NONCONVERGE.fsd
(36.22 KiB) Downloaded 682 times
thubelihle
 
Posts: 73
Joined: 04 March 2017, 16:27

Re: non- convergence with flowrate specification

Postby hkooijman » 16 June 2018, 19:41

There can be many reasons for nonconvergence, e.g. infeasible specifications, or an in sufficient number of iterations for the given accuracy. To get an idea, you need to look at the output report of the column. There you see that the iteration count and the log of the error to tolerance ratio:

Iteration log(Err/Tol)
0 4.4452
1 6.5733
2 7.0171
3 7.1796
4 7.2668
5 7.0112
6 6.9685
7 7.0438
...
25 7.2592
26 7.1459
27 7.3711
28 7.2461
29 7.2472
30 7.3241
Run level: Report
Convergence not obtained in 30 iterations

This seems that the solver is not making any progress. This is can be because you specify very specific component flows that may be incompatible with each other. In your case I perceive the problem being that the specification you provide need better initialization. When you are solving inside a flowsheet, it may help to not start from old results. If you are not using component specific specifications, it may also help to use a 2-step method over the Newtons method (see the solve options panel) or to provide more iterations.

In your case however, you are using component specific specs. Typically these require some initialization guesses. As such you need to check the box "Use guess for initialization" and enter a reflux ratio and fraction oof combine feeds as guesses. In your case a value of 5 and 0.75 will enable you to converge the specification set of 191 lbmol/h Propane in the top and 365 lbmol/h Butane in the bottom. Therefore, it is easier to start with reflux ratio and a bottoms flow (or fraction of the feeds). Furthermore, as flows might increase and decrease in a flowsheet, you have to ask yourself if it is not safer to use mole fractions.

Hope this helps - Harry
hkooijman
 
Posts: 51
Joined: 13 April 2013, 23:50


Return to COCO (AmsterCHEM)

Who is online

Users browsing this forum: Google [Bot] and 4 guests