1 reply to this topic

[Catherine Kwan]

Hi all, 

 

We are currently using HYSYS to simulate a styrene plant. We have a three-phase separator in which vapor, water and ethylbenzene-styrene mixture are separated. We do realise that Peng-Robinson thermo package that we currently using for our simulation are most possibly not accurate in simulating this three-phase separator. However, as there seems to be none other thermo package more suitable in our HYSYS to do the job, we intend to find liquid-liquid equilibrium data on EB-Styrene, EB-water and Styrene-water, perform some regression and compare to our simulated data. However, those we found on the internet like Dechema required payment. Is there any resources where we can actually obtain these LLE data? 

And once we obtained these data, is there any good way to perform the regression so that we can perform a comparison of data? Is using Matlab appropriate? 

Thanks in advance for the help. 

 

Best Regards

[MrShorty]

However, those we found on the internet like Dechema required payment. Is there any resources where we can actually obtain these LLE data?

Since you seem to dislike the idea of paying for Dechema's service, I'm guessing you are looking for a "free" databank. I am not aware of any good database like this that is free. I seem to recall that the Korean government may have put something out there without charging for it, but, if they have, I have not used it. I know that it often seems in the "information age" that information like this should be freely available over the internet, but it seems that those who control the compilation and publishing of this information do not share that sentiment.

 

Posting in the Student forum, can I safely assume that you students at a university? I would expect that your university library or department library would have access to a databank like this. I would suggest you ask your professor, department, or library what databanks they have access to for student use.

 

And once we obtained these data, is there any good way to perform the regression so that we can perform a comparison of data? Is using Matlab appropriate?

Yes, there are good ways to perform these regressions, and Matlab would likely be one good "application"/"programming language" to do it. However, you must recognize that this is not a trivial programming exercise. You will need algorithms for solving the EOS, algorithms for finding phase boundaries (setting fugacity for each component in each phase equal), and an algorithm for finding best fit interaction parameters. I'm not familiar with Matlab, so I'm not sure how difficult these will be to program into Matlab. If you have the skill and experience, it should be doable.