Jump to content



Featured Articles

Check out the latest featured articles.

File Library

Check out the latest downloads available in the File Library.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Low Flow in Pipes- posted in Ankur's blog

0

Hysys Refinery Ccr Simulation Prediction Problem


15 replies to this topic
Share this topic:
| More

#1 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 30 March 2018 - 01:54 PM

Hi,

In Refinery CCR Steady state simulation in hysys, i have feed which contains 39 components pure. Some of them are like single branched paraffins sbp6 etc.. But PR package gives some error of 5 percent in temp , composition. I hv selected only one component for SBP6 out of multiple options available in hysys with same molecular weight.. what could be the solution ?

#2 serra

serra

    Gold Member

  • Members
  • 310 posts

Posted 30 March 2018 - 03:09 PM

if you are modeling a separation a 5% error with temperatures may be acceptable (depending form specs),

there are several parameters as the list of components, BIPs etc. which you can modify (tuning the model),

if an automatic procedure is available,  use that, consider that components with same mw can show different vapor pressure, Tc, Pc...

finally, according my experience (with a different software, Prode, but the models are the same) Peng Robinson is as good as other models for hydrocarbons...



#3 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 01 April 2018 - 12:50 AM

Hi,

 For exmplae, I have one component called as A9 ( aromatics 9 ) for which following components can be selected as there are having same mol wt  but diff Boiling points , Cumene, 1M-3 Ebenzene, 1M-4 E benzene, 135-E benzene. Out os this I have selected only one component as cumene and it 100 percent weightage . Is that correct approach or shall I distribute it equally among all components?

 

How Tc and Pc will affect the results ?

 

How to tune BIPs for PR?

 

I am getting 5 percen deviation in composition, temp and duties of exchangers also?

 

Request your help now.

 

Regards,

Chetan Chavan



#4 serra

serra

    Gold Member

  • Members
  • 310 posts

Posted 03 April 2018 - 02:45 AM

if you adopt a EOS such as Peng Robinson vapor pressure depends from alpha function and Tc, Pc of pure fluids,
given a specified data set (VLE, density etc.) the classic approach (available in almost all simulators) is to define (pseudo) components (characterized in terms of Pc, Tc, Mw), with these (pseudo) compomponents errors, for example with enthalpy (heat exchanger), may be higher,
however alternatives are available,
you can define a list of real components and optimize compositions, BIPs etc. in my copy of Prode Properties the procedure can optimize compositions or BIPs, I prefer to optimize the compositions with BIPs from database,
if your software doesn't include a specific procedure maybe you can adopt external tools (such as Matlab) for optimization...
 



#5 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 03 April 2018 - 01:59 PM

Hi Serra,

I can not play with BIPs as it is project contraint. I m using Aspen Refsys.. i m getting 5 percent deviation in composition , temp, flow in steady state model..There is a limit also on how much i can tune column efficiency to get required composition... What is the alternative ?
SUPPOSE in Compobent A9, I hv 7 similar compounds in hysya with same mol wt but diff BPs, how to distribute total mass among them ? Request your detailed guidance..

#6 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 04 April 2018 - 03:25 AM

SUPPOSE in Component A9, I have 7 similar compounds in hysys with same mol wt but diff BPs, how to distribute total mass among them ?
There is no standard method for that. A9 is a simply mixture of 7 components with unknown mass fractions.

 

You can play with the mass fractions within A9 (and the other isomer mixtures like SBP6, MBP6, et cetera) to tweak the simulation result towards the desired result.

 

To keep it simple assume A9 (and other unknown isomer mixtures) to be a mixture of only 2 components, the isomer with the highest BP and the isomer with the lowest BP so that you only need to tweak one variable in each isomer mixture.



#7 serra

serra

    Gold Member

  • Members
  • 310 posts

Posted 04 April 2018 - 05:55 AM

if you wish to adopt real components (not pseudo) probably the BIPs in database are a reasonable option,

some software allows to optimize compositions automatically (see my previous post) but you can optimize with manual procedure,

select several key components and change compositions as suggested by Ping Pong,

or set some link with an external tool (Matlab etc.)



#8 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 04 April 2018 - 02:23 PM

Hi ,

If i Change the feed comp as u said, whether it will reflect and correct itself in rest of the model? I have reactor section, stabiliser section downstream..
I have total 39 pure components in feed as paraffins, naphthenes, aromatics..My worry is Inlet to reactor contains these PONA componenrs , but at the outlet to reactor, it converts paraffins to isoparaffins, naphthenes to aromtics to some extent.. Now in feed suppose i hv single branched paraffins or mbp6 multi branched ones, and it has 7 sub compounds with same mol wt and can hv isomers also..then how to distribute it ?

#9 serra

serra

    Gold Member

  • Members
  • 310 posts

Posted 05 April 2018 - 01:44 AM

there are not many alternatives,

you can optimize compositions once and accept results

or separate problems and optimize the different sections...



#10 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 05 April 2018 - 04:40 AM

Serra,
Today I will try with giving mol fraction into highest and lowest BP components and see results.. But in steady State, I m getting 5 to 7 percent deviation in H & MB also..Is it acceptable in steady state design case? I ALSO heard that in steady state distillation column, one shall nt converge column with Reflux rate ? Is it true? Also Do u hv any experience in modelling Hysys CCR reactor and tuning it ?

Thanks for your time.

#11 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 05 April 2018 - 08:53 AM

Serra,

Today I have observed one important thing: While doing steady state deethaniser column for CCR plant in hysys wherein in feed , all simple pure components without option of multiple same MW compounds , I am still getting an error or deviation in H & MB. This column feed is simplest : It contains only ethane, butane, propane, ibutane as feed . So ideally I should get correct separation for this column atelast.Can I conclude that fluid package used as PR is having non compatibility for this system of Refinery?



#12 serra

serra

    Gold Member

  • Members
  • 310 posts

Posted 05 April 2018 - 10:20 AM

Peng Robinson is not bad to model a deethanizer column , I have solved columns with GERG 2008 (Prode) model but results are not too different (depending from many variables as overall tray efficiencies etc. etc.),

with PR errors should be within a few %  but (in addition to those mentionded) there are several areas of uncertainty as specified compositions etc. 



#13 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 05 April 2018 - 01:59 PM

Serra,

1)Do you hv any experience in modelling Refinery CCR in steady state in Aspen hysys or any other simulator?

2) Also for simple pure component system without any complex comppunds, i m getting an error may be due to efficiency set equal to 1.

3) i hv now complete integrated model of CCR in hysys but with deviations like 5 to 7 percent in heat duties and mass balance also.
What approach of tuning shall i adopt step by step to come closure to desired results ? I do nt hv freedom to tune BIPs.

Thanks.

#14 serra

serra

    Gold Member

  • Members
  • 310 posts

Posted 05 April 2018 - 04:05 PM

I suppose you need the details about  parameter estimation which is essentially an optimization problem,

for a Continuous Catalytic Regeneration (CCR) process there are several alternatives, I model reaction patterns and reactors with tools as Matlab, perhaps you may find useful this thesis discussing predictive modeling,

https://vtechworks.l...ti_K_D_2011.pdf



#15 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 08 April 2018 - 03:23 AM

Serra,

 

I have a query regarding CCR Reactor modelling. Feed to the reactor is naphtha and recycle hydrogen ( required for Dehydrogenation of Naphthennes). But while configuring reactor in aspen RefSys, shall I take feed as pure naphtha or naphtha + pure hydrogen . How reactor configuration works ?

 

Request your help.

 

Regards,

Chetan Chavan



#16 CHEMSTRONG

CHEMSTRONG

    Gold Member

  • Members
  • 347 posts

Posted 04 May 2018 - 02:34 PM

Serra,
I m tuning CCR reactor in hysys.i m getting more than required yield of aromatics, but h2 yield is nt sufficient. If dehydrogenation to aromatics is taking place then hydrogen should be more. But also i m getting less paraffins as well. That means hydrocracking is consumibg gydrogen and nt producing lower paraffins. Do u hv any idea abt tuning of reactor in calibration environment




Similar Topics