10 replies to this topic

[flangecoupling]

Dear All,

 

These days I am working in Refinery dynamic simulations and currently I have to deliver a project involving simulation of an OLEFIN CONVERSION UNIT. Metathesis reaction occurs here and propylene is formed. The reactions involved are equilibrium in nature, but the issue is that we do not have equilibrium reactors in our simulator.

 

We only have kinetic reactors, to come across this issue I have devised the following approach.

 

1. I am specifying all the reactions as double reactions, like one forward and other reverse.

2. I am taking all the reactions to be elementary in nature

3. I am gathering the data of variation in equilibrium constant with feed temperature from hysys as a mathematical relation.

4. I am using this above mentioned relation to calculate how much the forward reacion will occur and how much reverse reaction will occur.

 

I want to know your expert advise on the approach I am following.

 

Plus I want to incorporate the "approach temperature" function as in hysys into my simulation, please any suggestions of how can I come across this issue.

 

Any literature on modeling equilibrium reactors from scratch will be really really halpful.

 

Thanks & Regards

 

[thorium90]

Might I suggest calculation of equilibrium constants by Gibbs energy minimization?

 

Although doing the calculations for complex catalytic metathesis reactions would be a pretty advanced level in chemical engineering.

 

I can show you the simple (complex) chemical equilibrium minimization methods.

Otherwise, I would suggest getting real plant data, applying standard filtering techniques and curve fitting to get your equilibrium constants for modelling.

 

Im not familiar with an approach temperature function for the equilibrium reactor module in hysys. Maybe you could elaborate?

Edited by thorium90, 14 May 2013 - 10:45 AM.

[flangecoupling]

Dear Thorium,

 

Thank you very much for your considerate reply.

 

If you can show me a simple (complex) chemical equilibrium minimization method, it would be realy helpful and it will give me a hint to go onwards.

 

I have the real plant Heat & Material Balance data, Process flow diagrams and P&IDs, can you suggest that how will I be able to calculate equilibrium constants from this data ?.

 

By Approach I meant Fractional Approach, mistakingly I wrote temperature approach, my apologies. This function comes when we view reactions specified in hysys and go to "Approach" tab there we can specify Co, C1 & C2, the equation of fractional approach is as given below

 

Approach %: Co + C1T + C2T^2.

 

Also I am attaching a small excel file which will clarify my present modeling approach. Any comments on it will be realy helpful.

 

Thanks & Regards

Attached Files

•  Eq Const Calc.xlsx   25.93KB   73 downloads

[thorium90]

I would suggest reading Introduction to Chemical Engineering Thermodynamics by Smith, Van Ness, Abbott, the chapter on chemical reaction equilibrium.

 

If you have Matlab, you can look through my example here.

http://www.mathworks...ked-bed-reactor

 

 

Conversion % = Co + C1T + C2T^2

The above is not called fractional approach either. It allows you to put in a value for conversion that can depend on the temperature of the reaction.

 

You can read about the temperature dependence of the equilibrium constant below. Its a nice website.

http://www.chem.ufl....ures/lec_v.html

 

How are you getting these equilibrium constants from hysys? Doesnt that imply you have already completed your project? You have already completely simulated your reactors?

Attached Files

•  Conversion Reaction.jpg   56.92KB   9 downloads

Edited by thorium90, 15 May 2013 - 04:52 AM.

[flangecoupling]

 

Dear Thorium,

 

Thank you again for your reply.

 

I do not have matlab, but I will see if I can work on it with some friend. I will go through the refrences which you suggested. Thanks

 

I actually have to do the reactor modeling in a software that does not have equilibrium reactor, I have simulated the reactor to some extent in hysys and I am taking eq constants from hysys. I will put the relation of the eq constant with feed temp deducted from this data of hysys in the other simulator, as described in my first post, but I am not sure if this is the correct approach.

 

I called it fractional approach because this is what hysys is saying, please see the thumbnail. If this is only conversion then it means that if we give a value of less then 100 to it, then it will alter the equilibrium composition? but will it change the calculation of eq const too ?.

Edited by flangecoupling, 15 May 2013 - 06:17 AM.

[flangecoupling]

The thumbnail is as attached

Attached Files

•  Fractional Approach.jpg   56.09KB   12 downloads

[thorium90]

Ah, my mistake, you were referring to the equilibrium reactor module. This is what I got from the manual.

 

The following will be very useful in understanding the method. One may use excel for the calculation. Also, one does not necessarily need to use goal seek or solver. If you are good with numerical methods, you can use the generalized newton method or a simpler forward difference method.

 

http://www.google.co....46471029,d.bmk

 

 

http://www.google.co....46471029,d.bmk

 

 

If you have some real data, you can use a Kalman filter on your plant data. It would be preferable if you have data for different flows and temperatures for your reactor. It would be useful if you had Simulink with signal processing toolboxes to perform this. It would be even better if you could supplement it with the curve fitting toolbox in matlab.

 

If you are really familiar with matlab already, you can build your entire olefin unit in simulink and run dynamic simulations in it.

Attached Files

•  equilibrium fractional conversion.jpg   155.2KB   8 downloads

Edited by thorium90, 15 May 2013 - 07:02 AM.

[flangecoupling]

Dear Thorium,

 

Thank you very very much for all the insights. Your replies have been really helpful. People like you are a blessing for us all.

 

Regards

[thorium90]

I would like to point out that the first link I gave in post 7 is from this book.

 

http://www.amazon.co...H/dp/0131482041

 

All examples mentioned in the book are available for download from the author's website.

 

http://www.problemso....com/bookowner/

Edited by thorium90, 16 May 2013 - 11:55 AM.

[thorium90]

Hi,

 

Further to what I have posted, I would like to share a very recent webinar from PSE titled "Design and Scale-up of Multitubular Reactors". It is about the use of process modelling software to design fixed bed catalytic reactors. Some stuff of note were the use of first principle models, in built parameter estimation, solvers, optimization, and integration with FLUENT for cfd modelling. I felt the practical tips given were very helpful and I would encourage anyone with an interest an process modelling to spend roughly 40mins to register and view the free webinar.

 

http://www.psenterpr...rs/archive.html

Attached Files

•  Design and scale-up of a high-performance multitubular reactor.pdf   10.14KB   45 downloads

Edited by thorium90, 18 May 2013 - 01:54 AM.

[flangecoupling]

Thanks alot, looke very interesting, will definitely go through it.

 

Regards