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

Cdu Simulation


85 replies to this topic
Share this topic:
| More

#26 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 19 September 2013 - 07:00 PM

It seems to me that you are using the first definition of Overflash (see my message #12). In that definition 5 % Overflash is extremely high.

 

As I have advised already twice before: Dirty Washoil flowrate should be about 10 wt% on hydrocarbon vapors from flash zone.

 

Moreover I is not clear to me how you get to the 52 % (is that liquidvolume% or weight% ? ) crude vapourization required for distillate separation as per TBP curve.

 

Flashed crude feed is entering above 39th tray. The preflash vapors are also entering at this location in a separate nozzle

 

Crude, kg/hr - [477000 (total vap) - 1200 (water vapor)] + 638600 (liq)

Preflash hydrocarbon vapor - 52280 kg/hr - 5253 (water) = 47027 kg/hr

Hydrocarbon vapor up from tray 40 to 39 = total vapor - stripping steam rate = 128926-43000 = 85926 kg/hr

 

Total hydrocarbon vapor in flash zone = 477000-1200+52280-5253+128926-43000=608753 kg/hr

 

Liquid from tray 38 down to tray 39 - 36734.4 kg/hr

 

Dirty wash oil/hydrocarbon vapor ratio = 36734/608753 = 6%........Is this calculation right?

 

I am already maxed out on crude inlet tempearture of 370 C!

 


Edited by NAP, 19 September 2013 - 08:41 PM.


#27 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 20 September 2013 - 05:59 AM

Dirty wash oil/hydrocarbon vapor ratio = 36734/608753 = 6%........Is this calculation right?
It is correct, provided that all other numbers are correct.

 

It seems your crude is heavier than the Arab Light you mentioned in message #13.

200000 BPSD Arab Light should be about 2 % less than your 1161427 kg/h crude (dry basis).

Arab light should yield about 48 wt% 360+ residue and 52 wt% distillates, but your simulation produces 50.7 wt% residue and only 49.3 wt% distillates.



#28 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 23 September 2013 - 12:54 AM

If my product draws from the column are meeting the specs of flash point, D86 90% temp. etc etc., do I still need side strippers?

 

Is it common to have a CDU column without a side stripper?



#29 Padmakar Katre

Padmakar Katre

    Gold Member

  • ChE Plus Subscriber
  • 992 posts

Posted 23 September 2013 - 02:24 AM

If my product draws from the column are meeting the specs of flash point, D86 90% temp. etc etc., do I still need side strippers?

 

Is it common to have a CDU column without a side stripper?

 

Hi,

Would be pleased to say "Yes", but unfortunately answer is usually "No". In absence of side strippers, required number of trays in main fractionator will go up. You may get flash point spec i.e. D86 10% recovery temperature, but what is extent of overlap you have? is it acceptable to client or can it be justified? Addition to AL crude, what are other crudes likely to be processed, have you checked the column performance for these check cases? What are the product yields? what is 360 minus material in RCO?

Addition to above points, downstream configuration should be checked before finalizing i.e. routine of these products e.g Kero and Diesel if routed together to a common hydtrotreater, requirement of diesel stripper can be checked in this case. You need to analyze each issue with lots of thoughts and only you can do it, as you know in & out of this project.

Good luck.



#30 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 23 September 2013 - 02:45 AM

Is there a thumb rule or guideline to decide which tray will be the draw-off tray for a particular side product?

 

Obviously, it will depend on the temperature profile of the column and the required side-product spec. but when doing the simulation the first time, when the results are not known, any way to estimate the number of the draw tray?



#31 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 23 September 2013 - 03:44 AM

If my product draws from the column are meeting the specs of flash point, D86 90% temp. etc etc., do I still need side strippers?

 

Is it common to have a CDU column without a side stripper?

I have never seen a CDU column without sidestrippers. They are required for control of product flash points.

 

I belief your present simulation does not yet have pumparounds, so the column now has a very big reflux ratio from the top down. Moreover you have rather high number of trays for only two sidedraws.

 

After you add pumparounds part of those trays will be for heat transfer only and will no longer contribute to separation. Moreover due to those added PA's the top relux will then decrease and consequently L/V ratio in all section above lowest PA will decrease which has negative effect on separation.

 

Note also that flash point prediction using ASTM data is very inaccurate. There are several methods floating around. I do not know which method your Hysys simulation uses but none of them are very good.

 

A CDU column with only two sidedraws is already strange enough. Don't make it even stranger by not having sidestrippers.


Edited by PingPong, 23 September 2013 - 03:46 AM.


#32 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 23 September 2013 - 04:00 AM

When we say distillate yield, do we count the light end part of the crude in the distillate yield as well?

 

I mean HYSYS yield fractions based on specified cut points for my crude oil is,

 

Naphtha  0.1937

Kerosene 0.1017

Atm Gas Oil 0.2201

Residue 0.4842

 

So is the distillate yield a sum of first three cuts, or only sum of second and third cut?

 

around 5% of preheated crude has already flashed in preflash drum and enters the column as vapors in seperate nozzle. So is this fraction a part of CDU column yield? (I am sorry if this is a dumb question!)

 

By the way, I confirmed with my client, the CDU has to be designed based on cut points and not draw rates in terms of BPSD. The cut points will decide whatever yield that comes out of the column. I have to meet the cut points and then the final product specs!



#33 Padmakar Katre

Padmakar Katre

    Gold Member

  • ChE Plus Subscriber
  • 992 posts

Posted 23 September 2013 - 04:30 AM

When we say distillate yield, do we count the light end part of the crude in the distillate yield as well?

 

I mean HYSYS yield fractions based on specified cut points for my crude oil is,

 

Naphtha  0.1937

Kerosene 0.1017

Atm Gas Oil 0.2201

Residue 0.4842

 

So is the distillate yield a sum of first three cuts, or only sum of second and third cut?

 

around 5% of preheated crude has already flashed in preflash drum and enters the column as vapors in seperate nozzle. So is this fraction a part of CDU column yield? (I am sorry if this is a dumb question!)

 

By the way, I confirmed with my client, the CDU has to be designed based on cut points and not draw rates in terms of BPSD. The cut points will decide whatever yield that comes out of the column. I have to meet the cut points and then the final product specs!

Hello,

Distillate yield In crude unit = (Feed rate - RCO rate)/Feed Rate * 100.



#34 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 23 September 2013 - 04:38 AM

Padmakar:

 

 But before you converge your simulation, you wouldnt know the actual RCO rate.

 

I am presuming that you are estimating RCO rate from the RCO fraction predicted by HYSYS oil charecterization x crude feed rate.

 

Is that right?



#35 Padmakar Katre

Padmakar Katre

    Gold Member

  • ChE Plus Subscriber
  • 992 posts

Posted 23 September 2013 - 04:48 AM

NAP, Hello.

 

Yes you are right, once you have converged the simulation meeting all the specifications then you can calculate the actual distillate yield.  Anyways, what is the spec for RCO?  Is it maximum allowable gas oil slip %?



#36 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 23 September 2013 - 05:03 AM

By the way, I confirmed with my client, the CDU has to be designed based on cut points and not draw rates in terms of BPSD. The cut points will decide whatever yield that comes out of the column. I have to meet the cut points and then the final product specs!
Did you also confirm with the client what crude assay to use for the design: Arab Light or Arab Medium or What?

 


I mean HYSYS yield fractions based on specified cut points for my crude oil is,

 

Naphtha  0.1937

Kerosene 0.1017

Atm Gas Oil 0.2201

Residue 0.4842

Are those volume or weight fractions?

Based on 165 - 220 - 360 cutpoints?

Based on which crude?



#37 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 23 September 2013 - 07:09 AM

 

PingPong:

 

The crude assay was actually provided by the client way before. (There was a BIG communication gap between me and the guy who left)

 

I charecterised the crude to a full TBP vs Vol% plot and showed the volume % yield as the estimated yield of the CDU column based on the cut points I gave earlier.

 

The crude is the API 29.9 crude I gave information earlier which happens to be a Kuwait crude sample.

 

So the design basis is the cut points of 165, 220 and 360 for the API 29.9 crude and meeting the product specs as I gave before,

 

The fractions I have given above are vol. %

 

My next task is to add pumparounds.

 

My existing guideline talks about a 50 C delta between pump around supply and return temperature. Does that sound about right? What should I start with for pumparound rates in the beginning?



#38 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 23 September 2013 - 07:32 AM

The crude is the API 29.9 crude I gave information earlier which happens to be a Kuwait crude sample.
Kuwait, okay, I'm glad we finally got the design crude right.

Is similar to Arab Medium, contains also about 2 wt% asphaltenes, so I suggest you use TBP FBP in the order of 850 oC.

 


My existing guideline talks about a 50 C delta between pump around supply and return temperature. Does that sound about right? What should I start with for pumparound rates in the beginning?

The purpose of every pumparound is to provide duty to other users (crude preheat, reboiler of stabilizer, steam generation, or whatever) at the temperature level they require. That also determines the drawoff tray(s) of the PA(s) in the CDU column.

To start with, a delta of 50 oC is as good as any other number, but to finalise the CDU column simulation you will have to look at the users of the PA duties.



#39 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 23 September 2013 - 07:12 PM

I understand that pumparound is a way to effectively use heat from CDU column and use it to preheat crude (typically). This also provides reflux to the column above the side product draw-off trays.

 

But I believe, how to distribute the pump around duty in the crude HEN is a regorous pinch analysis exercise where we plot the requried heat gain (T vs H) by the consumer and rejected heat by the supplier on the same chart and try to optimize the heat distribution across the network.

 

At this stage of my simulation, wouldnt it be wise to estimate the heat required for crude from 31 C to ~360 C, provide  PA coolers as a representation of Crude HEN and congifure the PA flow and duties to match the estimated crude preheating duties?

 

I understand that later, I will have to fine tune the simulation to match the PA heat distribution across the HEN to the actual required pre-heat duties.



#40 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 23 September 2013 - 08:36 PM

 

 

 

When we say distillate yield, do we count the light end part of the crude in the distillate yield as well?

 

I mean HYSYS yield fractions based on specified cut points for my crude oil is,

 

Naphtha  0.1937

Kerosene 0.1017

Atm Gas Oil 0.2201

Residue 0.4842

 

So is the distillate yield a sum of first three cuts, or only sum of second and third cut?

 

around 5% of preheated crude has already flashed in preflash drum and enters the column as vapors in seperate nozzle. So is this fraction a part of CDU column yield? (I am sorry if this is a dumb question!)

 

By the way, I confirmed with my client, the CDU has to be designed based on cut points and not draw rates in terms of BPSD. The cut points will decide whatever yield that comes out of the column. I have to meet the cut points and then the final product specs!

Hello,

Distillate yield In crude unit = (Feed rate - RCO rate)/Feed Rate * 100.

 

 But before you converge your simulation, you wouldnt know the actual RCO rate.

 

I am presuming that you are estimating RCO rate from the RCO fraction predicted by HYSYS oil charecterization x crude feed rate.

 

Is that right?

 

Hello,

Yes you are right, once you have converged the simulation meeting all the specifications then you can calculate the actual distillate yield. Anyways what is spec for RCO is it maximum allowable gas oil slip %.

 

 

There is no spec on RCO. Spec only on Naphtha, Kero & GO

 

This is going to be a RFCC based refinery with an integrated petrochem complex.

 

Objectives are to maximise naphtha for petrochemical production and fuels like kero and diesel.

 

The AR will be sent to a RHDS and then to RFCC to make polypropylene.



#41 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 27 September 2013 - 04:14 AM

@ PingPong, Padmakar

 

I have a converged base case simulation with following salient features. Please hightlight if something looks unual in your experience!

 

Crude heat gain from amb. to 300 C (Heater inlet) = 212 mw

4 PA's + 3 Product streams' heat available to preheat crude = 232 mw (An opportunity to increase heater inlet temp and reduce heater duty)

 

Crude vaporization at preflash inlet = 5% by wt.

Heat COT = 364 C

Overflash = 10.8% wt (DWO from tray above FZ/hydrocarbon vapor from FZ) OR 5.6% (liquid vol flow from tray above FZ/Crude vol flow)

 

Tower configuration (45 trays - for now - opportunity for optimization)

Tray 1 - 4: PA 1 draw and return

Tray 5 - 13: Separation

Tray 14-17: PA 2 draw and return, Kero draw and Kero SS vapor return

Tray 18 - 22: Separation

Tray 23 - 26: PA 3 draw and return, GO draw and GO SS vapor return

Tray 27 - 30: PA 3 draw and return

Tray 31 - 38: Separation

Tray 39: Feed inlet, Flash Zone

Tray 40 - 45: AR stripping

 

Flash Zone

FZ Pressure = 1.154 barg

FZ Temperature = 352 C

 

Stripping Steam

AR - 25000 kg/hr @ 45 tray

Kero SS - 2500 kg/hr

GO SS - 500 kg/hr

 

Fractionation Spec

Cut point 1 = 165 C

Cut point 2 = 222 C

Cut point 3 = 360 C

Kero 5% - Naphtha 95% gap = 13 C (Spec limit = 5 C min.)

 

Product Specs

Naphtha spec : 90% D86 T = 143 C (Spec limit = 155 C)

Kero spec: 90% D86 T = 217 C (Spec limit = 230 C)

Kero flash point: 51 C (Spec = 38 C min)

Gas oil spec: 90% D86 T = 361 C (Spec limit = 360 C) (Need for fine tuning)

Gas oil flash point: 97 C (Spec = 55 C min)

 

Product Quantities

LPG - 136.4 kg/h (Doubtful?)

Naphtha - 185300 kg/hr (40000 bpsd @ std cond)

Kero - 94140 kg/h (17840 bpsd @ std cond)

Gas Oil - 290200 kg/h (50790 bpsd @ std cond)

AR - 590900 kg/hr (88010 bpsd @ std cond)



#42 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 27 September 2013 - 10:15 AM

 

Product Quantities

LPG - 136.4 kg/h (Doubtful?)

Very doubtful. Obviously completely wrong. Input error in crude assay input?

 

Middle east crudes typically have roughly 1.5 wt% C4 minus, so it seems your quantity is at least a factor 100 too small.

 

This might also contribute to the ease with which you seem to be able to exceed meet the flashpoint specs, with only little steam on the kero stripper and almost no steam on the GO stripper.

ASTM D86 prediction from TBP of simulated products depends on method used, and flash point prediction also, as that uses some of the calculated ASTM D86 points. I do not know what methods Hysys uses.

 

 

 

Gas Oil - 290200 kg/h (50790 bpsd @ std cond)

AR - 590900 kg/hr (88010 bpsd @ std cond)

Unless you moved to a different crude assay, I do not see how you can produce so much GO with 222-360 cutpoints, and so little AR from that Kuwait assay. I would then expect only 21.5 vol% GO or 43000 BPSD.

 

 

What is not clear to me anymore is what the column overhead system looks like. In message #13 you stated:

7) What pressure and temperature do you use in the reflux drum? Based on naphtha dew point temp., condenser pressure 0.7 barg and temp 58 C (air coolers).

which implies use of a double condensation system: first condensor condenses only reflux, second (partly) condenses Naphthaminus product.

 

But now you mention a PA on the column top. Using a top PA plus a double condensation system seems overdone. Do you still have the double condensation system?

 

A PA at the GO drawoff, as well as a PA just beneath it, looks also strange. It is not impossible, but a column with only 2 sidedraws and nevertheless 4 PA's looks strange.



#43 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 29 September 2013 - 07:41 PM

@PingPong,

 

HYSYS uses API 2B7.1 (Pennsky Martin Closed Cup) method as default for flash point calc.

 

HYSYS oil charecterization indeed is showing GO cut fraction as 21.4% @ 222 C cut point (matches to your estiamate of 21.5%)

 

Does this mean my simulation is slipping too much kero in GO or too much AR in GO?

 

This also means my GO flash point spec is not being met even if the HYSYS results are saying so (too much kero in GO cut would bring the flash point down and too much AR in GO would raise the flash point up, right?)



#44 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 30 September 2013 - 03:48 AM

According to that Kuwait assay, at a cutpoint of 360 oC your simulation should produce about 48.2 vol% AR, or 96400 BPSD.

 

However your simulation produces only 88010 BPSD AR, or only 44.0 vol% on crude.

 

The TBP cutpoint in that simulation is therefor not 360 oC but about 390 oC. That is a very heavy GO.



#45 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 30 September 2013 - 04:22 AM

oh shit...i just realized what i did...

 

when I input the specs in HYSYS for cut points, they were cut points for ASTM D86 temp and not TBP temp....damn, I am dumb!

 

Thank you for the valuable post above which made me see my mistake!



#46 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 30 September 2013 - 04:29 AM

The TBP cut point spec is associated with a vol% distillied.

 

Should that be 90% or 95%?

 

Since the 95 - 5 gap is also a requirement of fractionation spec, I would expect the cut point to correspond to 95% spec but just confirming...



#47 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 30 September 2013 - 05:12 AM

TBP cut point has nothing to do with 90% or 95% or 95 - 5 gap.

 

Product yield matching certain TBP cutpoints simply follows from the crude assay data.

 

One does not need Hysys or any other simulator to determine how much gasoil will be produced when the TBP cutpoints are 222 and 360 oC. Just put the assay data in a graph and read the vol% at 222 and at 360 oC, and the difference is then the GO yield: about 21.4 vol%. And the AR yield is then about 48.2 vol%. It is that simple, no computer required.



#48 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 30 September 2013 - 05:35 AM

Yeah, i understand the logic of estimating yield of a particular cut from TBP curve at a corresponding cut point specified.

 

For some reason, HYSYS requires the user to enter vol% value when I am specifying the cut point temperature.

 

 

 

 

 



#49 PingPong

PingPong

    Gold Member

  • Members
  • 1,446 posts

Posted 30 September 2013 - 05:52 AM

You lost already far too much time trying to let Hysys produce the right amount of GO. Simply specify the intended product quantities based on the desired TBP cvutpoints. Forget about trying to specify cutpoints in Hysys.

 

For example: specify intended quantity BPSD of AR, GO and Kero (naphtha minus resulting), those are 3 specs, and let Hysys vary draw rates of Kero and GO, and Condensor Temperature, those are 3 variables.



#50 NAP

NAP

    Gold Member

  • Members
  • 62 posts

Posted 01 October 2013 - 03:08 AM

Fotunately, I have some spare time at my disposal...the schedule pressure has not yet kicked-in. This study hasnt started as a full fledged project. The client is sponsoring this on a limited scope as this is a tier-2 priority work for him.

 

I am curious to find out the reason for mismatch between yield fraction predicted by oil charecterization and the actual yield fraction resulted from simulation.

 

Is this because, assay charecterization is on a dry basis meaning we dont include vol% of water in crude as one of the light end components?

 

While defining the assay, if I enter water vol% as one of the light ends, the yield fractions change significantly for the same cut points.

 

P.S: I now also have Pro-II at my disposal...






Similar Topics