5 replies to this topic

[NavidJon9]

Hello, I am using Aspen Plus V12.1 to create a Methanol production plant with the specifications of:

37500 kg/hr 

99.6 mol% MeOH

 

I am currently attempting to optimise the system using my yearly profit as the objective to maximise. However, whenever I run the system using the specification as constraints I receive the following error.

 

** ERROR

   INITIAL POINT IS INFEASIBLE.  CONSTRAINT 1 IS VIOLATED.

   CONSTRAINT VALUE = -0.35781E-02

 

** ERROR

   Convergence block $OLVER12 did not converge

   normally in the final pass

Any help would be much appreciated. Thank you.

 

I have attached the aspen file below. 



 

Attached Files

•  V9 CHEN21112 - UpScaled - Copy.zip   2.5MB   3 downloads

Edited by NavidJon9, 29 April 2024 - 05:40 AM.

[latexman]

No attachment.

[NavidJon9]

Thank you.

[Pilesar]

Remove or relax your contraints. See what happens. Your constraints may be in units of measure that you do not intend. The starting point for the simulation must be within bounds of the constraints. Did you solve this first without the optimizer?

[NavidJon9]

Thank you Pilesar. I was able to make it work by changing my constraint to be the out flow of the flash separator and creating a convergence block for the optimiser. However, I came across another issue when I included integer variables to the optimiser. I will check my units now to ensure they are consistent.

[Pilesar]

When an optimizer gives trouble, sometimes it is useful to change the parameter manually and solve the model a few times so you will better understand how the model behaves. The optimizer is not magic, it is just able to make the same steps in changing parameters you would make in a systematic way while using mathematics to predict what step to make on the next iteration so that progress is being made toward the solution. Optimizers struggle with non-monotonic systems since they depend on taking baby steps where necessary to stay on the path.