You can change tray efficiency or you can create one or more new columns with a different number of trays,

https://www.youtube.com/watch?v=iyg_5rBBEFs&list=PLKHdoX3DsDGNbCFWBgoUePnrPnBToWtii

Dear Haris,

Please follow the link,

Article Optimization and Dynamics of Distillation Column Using Aspen Plus®

Thanks Syed A. Taqvi , Asad Ullah and Rick Manner however, the value of stage comes out to be zero. I have attached the simulation file. Number of stages are defined in constraint, but the answer of the constraint is zero!! Please help.

What mixture you want to separate by means of rectification. Maybe you do not need Aspen. Regards,

Muhammed, I cannot open the apwz file. A few screen shots may be useful to help us understand your problem

Rick Manner , Sir, please attached .bkp file. Objective function is to maximize the concentration/purity of top product. I hope this file will open.

The only simulator I currently have access to is Petrofine/Hysys. Screen shots or a description of the set up might help.

If you really think you have a very low number of stages try simulating a 1 stage separation with a flash and 2 or 3 stages with 2 or 3 flashes and a recycle. That may be all you need to model your separation.

If the mixture is highly not ideal (otherwise use shortcut methods) , I recommend you, change the approach to the problem. Choose the cost of the column (fixed cost + operating costs) as an objective function, the purity of the distillate (product) as a restriction (>99.5%) and Reflux ratio, # of stages, reboiler duty and feed stage as optimization variables.

However, the optimizer of Aspen Plus doesn't handle integer variables (# stages and feed stage), therefore, I suggest use metaheuristic techniques (not available in Aspen plus) as genetic algorithms, among others. For that, it may be useful to use the connection between Aspen Plus and Matlab. An example of the connection can be found in the following link:

la.mathworks.com/matlabcentral/fileexchange/69464-aspen-plus-matlab-link

Another way is by performing sensitivity analysis by setting the number of stages and optimizing only by varying the reflux. Then, change the number of stages and again optimize by varying the reflux. Continue until obtaining a 3D graph (purity, reflow and # stages) and graphically establish the values closest to the optimum. However, maximizing purity can be achieved with a very large number of stages (at minimum reflux) or a very large reflux (at a minimum number of stages). Therefore, I recommend changing the approach to the optimization problem.

You can use sensitivity analysis by fixing the number of stages and varying the reflux ratio.

Dear Andres Felipe Abril Duran , thank you for a comprehensive guidance. I was facing the issue, as you mentioned. However, can you please suggest some book or article to learn the GA in MATLAB for linkage with Aspen Plus?

Books:

1) Optimization of Process Flowsheets through Metaheuristic Techniques

Link : https://www.springer.com/us/book/9783319917214

2) Multi-Objective Optimization in Chemical Engineering: Developments and Applications

Link: https://www.wiley.com/en-us/Multi+Objective+Optimization+in+Chemical+Engineering%3A+Developments+and+Applications-p-9781118341667

3) Chemical Process Retrofitting and Revamping: Techniques and Applications

Link: https://www.wiley.com/en-us/Chemical+Process+Retrofitting+and+Revamping%3A+Techniques+and+Applications-p-9781119016335

4) Stochastic Process Optimization using Aspen Plus®

Link: https://www.crcpress.com/Stochastic-Process-Optimization-using-Aspen-Plus/Segovia-Hernandez-Gomez-Castro/p/book/9781498785105

Matlab:

1) GA function in Matlab

Link: https://www.mathworks.com/help/gads/ga.html

Best Regards,

The number of stages is a discrete variable so you cannot use case studies or inbuilt sensitivity analysis feature of Aspen. However, you can automate the process by interfacing ASPEN and MATLAB and then vary the number of stages along with other parameters. You may look at the following paper for more information regarding interfacing simulator with MATLAB and perform optimization using metaheuristic techniques:

Article Heating Value Reduction of LNG (Liquefied Natural Gas) by Re...

One way to change the number of stages is to use the stage efficiencym,which can be changed in the sensitivity block. if you have 40 stages and vary the eff from 1 to 0.5, if will provide a similar answer as running from 20-40 stages. You can also have the stripping and rectification sections vary independently.

Thanks Dear Sir @Bruce Murphree.

Doing an optimization using the Aspen Sensitivity block can be tricky. You will generally get some runtime errors, and/or faulty results if you don't vary the total number of stages and the feed stage accordingly (per Bruce's comment above). There should be some examples of this posted in the Aspen Knowledge base area.