3 replies to this topic

[suhas]

Dear all
I am trying to simulate FCC main fractionator using aspen Plus. I am trying to model with actual trays and efficiency. I have configured all the PA with the specs of PA duty and flows. I am trying to run the columns with specs of all the product withdrawal flow rates are fixed. Cuurenty I am generating FCC rx effluent by using product backblending. (feed is generated using mixer of FG, LPG,gasoline, LCO and Slurry oil/decant oil)
Currently I am not able to match the Column internal traffic with actual observed one. Simulation is showing much higher Reflux, LCO IR, HCO IR flows than actual one. (almost 1.8-2 times than actual one) I have already rechecked all the input values.

Has anybody observed the same? Can you please suggest something (some thermodynamic option/ may be pseudo component property estimation option) to minimize this mismatch.. Currently I a m using Grayson Street as thermodynamic option, and aspen defaults methods for pseudo component property estimation.

Thanks in advance
Suhas

[Zauberberg]

A few things I can think of:

1. Improper characterization of feed components (Naphtha, LCO, HCO, Slurry). Why did you use pseudocomponents? The best approach is to employ TBP distillation curves, density curves and viscosity curves.
2. It's better to use ideal trays rather than actual. There's a lot convergence/equilibrium issues when actual trays are used. Switch to ideal trays and adjust location of side-draws, if required.
3. I believe Peng-Robinson should be just fine for this type of simulation. We used to do the same for our Main Frac/Gas Plant while I was working in refinery.

Good luck,

Edited by Zauberberg, 15 December 2009 - 12:36 PM.

[suhas]

Dear Zauberberg
Thanks for your reply, I think there is some misinterpretation


Thanks for your reply, I think there is some misinterpretation of my query regarding feed definition. I am using ASTM-D86 (gasoline and LCO) and D2887 (decant oil) to define this assays in simulation and not the pseudo components. I was saying that when aspen generates the pseudo components from this curves, there is option for applying different methods to find properties of the pseudo components.
I have also tried with therotical stage approach but it didn’t results in change in internal traffic flows.

Thanks
Suhas

A few things I can think of:

1. Improper characterization of feed components (Naphtha, LCO, HCO, Slurry). Why did you use pseudocomponents? The best approach is to employ TBP distillation curves, density curves and viscosity curves.
2. It's better to use ideal trays rather than actual. There's a lot convergence/equilibrium issues when actual trays are used. Switch to ideal trays and adjust location of side-draws, if required.
3. I believe Peng-Robinson should be just fine for this type of simulation. We used to do the same for our Main Frac/Gas Plant while I was working in refinery.

Good luck,

[Padmakar Katre]

Dear all
I am trying to simulate FCC main fractionator using aspen Plus.

Dear,
Is it a revamp or just preparing the as built model of FCC main fractionator? Try with the thoeratical trays rather than actual trays with the efficiencies as suggested by Zaubeberg.

I am trying to model with actual trays and efficiency. I have configured all the PA with the specs of PA duty and flows.

PA has the specifications like,
1. PA Duty
2. PA Flow (Mass,Molar,Vol.)
3. Return temperature
4. Delta Temperature
Check with any both of the above 4 for one PA rather than the fixed PA Duty and Flow only.

I am trying to run the columns with specs of all the product withdrawal flow rates are fixed.

You can't run any column with all prodcut flows as active specs.

Cuurenty I am generating FCC rx effluent by using product backblending. (feed is generated using mixer of FG, LPG,gasoline, LCO and Slurry oil/decant oil)
Currently I am not able to match the Column internal traffic with actual observed one. Simulation is showing much higher Reflux, LCO IR, HCO IR flows than actual one. (almost 1.8-2 times than actual one) I have already rechecked all the input values.

What is the basis for comparison? Do you any other similar model output for column traffic or the column PDS? I hope you have the same column ovhd operating pressures, column pressure drop is accoring to the operating plant data and revised feed has no increased yield of lighters/distillate compare to the design case.

Has anybody observed the same? Can you please suggest something (some thermodynamic option/ may be pseudo component property estimation option) to minimize this mismatch.. Currently I a m using Grayson Street as thermodynamic option, and aspen defaults methods for pseudo component property estimation.

Thanks in advance
Suhas

GS or PR would be ok in this case.You can compare the results inorder to validate the model by with references like plant operations data (operating parameters as well the products/feeds lab analysis reports. As you already told that you have rechecked the input data(feed charecterization part) just one thing more you need to check is column feed/prodcuts/PA locations. Waiting for your comments.