5 replies to this topic

[evvrista]

Good morning, I have attached a screenshot of the "monitor" screen of my absorber.

I am doing a sim in HYSYS about the production of lubricant bases and the absorber does not converge. Which I don't understand since, as can be seen in the screenshot, "current value" presents values very similar to those previously specified.

Could someone tell me why it doesn't converge? I need help

Attached Files

•  capt2.JPG   40.35KB   2 downloads
•  capt1.JPG   123.49KB   2 downloads

[Pilesar]

It is a good practice to have some sort of converged model to help troubleshooting. The error messages on unconverged models can be very  difficult to interpret. Some steps i would take (trial and error method): 

1) Set all side draws to 0.1 m3/h rate then try to converge the column with all product to the bottom. If converged, add in side draws one at a time until column breaks again to better troubleshoot.

2) If step 1 gives no convergence at all, then reduce the number of stages significantly until you get convergence. You may have to go down to 4 or 5 stages total. Once you get convergence, then stepwise add features back into your model.

[evvrista]

It is a good practice to have some sort of converged model to help troubleshooting. The error messages on unconverged models can be very  difficult to interpret. Some steps i would take (trial and error method): 

1) Set all side draws to 0.1 m3/h rate then try to converge the column with all product to the bottom. If converged, add in side draws one at a time until column breaks again to better troubleshoot.

2) If step 1 gives no convergence at all, then reduce the number of stages significantly until you get convergence. You may have to go down to 4 or 5 stages total. Once you get convergence, then stepwise add features back into your model.

 

Thank you very much for your help! I'm already testing the simulation only with the absorber part, it's much better. What do you mean by "converge the column with all product to the bottom"? I tried to put all the side streams as you told me and now I get the following messages:

"Liquid on stage 1__Main Tower splits to two phases. Phase fractions: 2.502e-05 (Liquid), 1.000 (Liquid)

    Column Flowsheet C-1 Not Converged"

And after this a screen with the following:

"A consistency error has occurred. This can arise either when two objects calculate differing values for the same variable or one object's calculations are conflicting with existing specifications.

 

In general, this means that the simulation is overspecified.

 

However, if the discrepancy is small and one of the values is calculated by a column flowsheet, tightening the column solver tolerances may eliminate the inconsistency."

 

Could the problem be in the Pump Round?

[Pilesar]

Column convergence can be tricky. Start simple to get a converged run so you can make sense of the error messages. This may mean cutting down to a top and bottom product and simplifying the pumparound as a duty applied to the tray. If you have not gathered much simulation experience yet, this looks like a complex column to start with. You must have vapor and liquid on every stage. Adjust your specs as needed to get convergence. Once you have a converged run, you can better interpret the column profiles and make spec changes to get the model you really want.

 Your error message indicates two liquid phases so you have to account for that in your thermo method, a decanter, or adjusting your stream composition and conditions in some way.

  The software supplier has tech support staff who should be able to help you if you need it.

Edited by Pilesar, 28 February 2023 - 06:37 PM.

[Nikolay_]

Heat/Spec is out of tolerance. Probably pumparound rate spec should be inactive.

[StealthProg]

I suspect its struggling to match the temperature set for the vapour, a comoom problem as there may no composition/bubble combination point that exactly matches that temperature.  I would try setting flow or pumparound return temperature and let it calcuate the vapour temperature.