I'm trying to optimize a system in Aspen plus. When I tried to get access to the Rplug reactor tube number, it shows the number is 0,which I specify it to 3000 in the reactor set up.
Is anyone know what's going on here cause it is quite weird? Thank you so much.
Best regards,
Francesca
You're going to have trouble trying to optimize the number of tubes with the built-in Aspen optimizer, if you can ever find the proper variable name. The number of tubes has to be integral, and the Aspen optimizer doesn't deal with integers very well at all. You're probably going to end up having to do something like a sensitivity study where you vary the number of tubes (and any other parameters of interest) and then find something near the "optimal" result. You can then fix the number of tubes from this study to its "optimum" value and then go back to the optimizer to find fully optimized values for the continuous variables of interest.
It would also be helpful if toy could attach a .bkp file of your simulation.
I'd just add to Brian''s answer, that you can try to use optimization algorithms with integer handling. I recommend stochastic algorithm NSGA-II (link Matlab Center: https://www.mathworks.com/matlabcentral/fileexchange/71609-real-coded-integer-handling-nsga-ii?s_tid=srchtitle). You can use Matlab NSGA-II algorithm to optimize Aspen model connecting both programs through COM technology. Please, refer to the following example, where I use Aspen Plus and Matlab to carry out model (including integer variables) optimization: https://www.mathworks.com/matlabcentral/fileexchange/74745-aspen-plus-matlab-optimization-of-distillation-column?s_tid=srchtitle
Please, refer to the following link, where you will find an example of Matlab-Aspen Plus connection - basics: https://www.mathworks.com/matlabcentral/fileexchange/69464-aspen-plus-matlab-link?s_tid=srchtitle
Regards, Abril.
Thank you. Brian. That's really helpful to me. Brian Hanley . I will try to optimize in Aspen Plus.
Thank you Andres. I will try this method by connecting Aspen plus with MATLAB first. For the NSGA-II, it will optimize the model directly? Will that converge in Aspen plus easily ?
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders