controller non convergence

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

Moderator: jasper

controller non convergence

Postby thubelihle » 04 May 2018, 17:37

l set up a controller to set the distillate acetone mole fraction to 0.98 by changing number of stages of the column, and it failed to do so. l did manual guessing till l found out that with 39 stages and feed on stage 31, l obtained the specified mole fraction of acetone. My question is why did the controller not converge? My assumption was that it varied the number of stages, but kept constant the feed stage hence, the non convegence. if that is true, is there anyway it can vary the number of stages and also vary the feed stage while mantaining the fraction of columns in the rectifying section to ones in the stripping section constant? attached is the simulation.
Attachments
CONTROLLER STAGES.fsd
(31.98 KiB) Downloaded 665 times
thubelihle
 
Posts: 73
Joined: 04 March 2017, 16:27

Re: controller non convergence

Postby jasper » 07 May 2018, 13:37

The number of stages is an integer variable. An assumption for the controller is that it is controlling continuous variables. For example, it takes derivatives by perturbation.

The controller will not be able to do this job for you. What is needed here is an optimized, or the mixed integer non-linear programming (MINLP) kind. COFE does not (yet) have one.
User avatar
jasper
 
Posts: 1128
Joined: 24 October 2012, 15:33
Location: Spain

Re: controller non convergence

Postby thubelihle » 09 May 2018, 06:10

thank you
thubelihle
 
Posts: 73
Joined: 04 March 2017, 16:27


Return to COCO (AmsterCHEM)

Who is online

Users browsing this forum: No registered users and 2 guests

cron