85 replies to this topic

[PingPong]

To see whether your numbers make sense I would have to setup a complete simulation myself, for which I have no time now.

Moreover you do not provide PA flowrates and duties.

 

One thing though: for each PA you need two theoretical stages (3 or 4 actual trays).

PA is drawn from the lowest stage and returned to the highest stage, and the duty is withdrawn from the highest stage.

[NAP]

When you say PA is drawn from lowest stage and returned to highest, you mean between two draws, right?

 

For ex. if Tray 7 is my Kero draw tray, I would draw PA from Tray 7 or Tray 6 and return it to Tray 2. Is that correct?

[PingPong]

Lowest and highest stage refer to the bottom and top stages of each PA section, not between PA sections.

 

Each PA circulation is over 3 or 4 actual trays, which are to be simulated as 2 theoretical stages.

 

For example: if Kero PA is drawn from theoretical stage 7 then it is returned to theoretical stage 6.

The corresponding PA duty is then removed from stage 6. At least, in PRO/II you would then specify a DUTY on stage 6, I don't know how a PA is best modelled in HYSYS.

 

Note that by adding an extra stage for each PA you probably need to increase theoretical stages in each section to account for that, as PA stages do not really contribute to fractionation. So kero draw and kero PA draw will not be from stage 7 anymore, but probably from stage 9, due to the 2 extra stages added for top PA and kero PA.

Edited by PingPong, 17 October 2013 - 06:25 AM.

[anuragsharma762002]

I have fixed my flash zone temperature based on two things.

 

1. The condenser pressure fixed by naphtha dew point (on air coolers) and then adding the pressure drop across all the stages gives me pressure in flash zone. At the flash zone pressure (slightly above atm.), I looked at the gas oil bubble point as GO is the heaviest side draw from the column.

 

2. I have the product specs such as Kero D86 90% temp and GO D86 90% temp to which I need to design CDU. So I looked at the TBP/ASTM D86 curve of the crude I charecterized and considered the D86 90% gas oil temp to be near the flash zone temp

 

Anything else I need to think in deciding flash zone temp?

this may not be accurate method.

u need to fix the COT in such a manner that atthe flash zone pressure fixed by you gives 2-5% wash/overflash ,this is irrespective of the throughput of your unit.

[Neelakantan]

long time back, it was 1999, i did a mini refinery simulation with a CDU 10000 bbl/d to 40000 bbl/d using HYSIM, apart from the column, the simulation had a furnace (pre-heater/reboiler), oil/oil exchangers, side strippers; the overhead is air-cooled; because it was in russia, the air-cooler was winterised to provide internal recirculation to avoid very low temperature; the furnace heater was supplied by famous canadian supplier; the CDU produced straight run naptha, kero and diesel cut and had summer and winter operations; the feed crude was extensively characterised by russian design institute providing detailed cuts to enable proper cuts; 

 

it is a tough job to simulate; but remember simulation is just a tool; finally engineering "idea" is required to specify and obtain all the equipment.

 

regards

neelakantan

 

(PS: i know the reply is generic!)

[Neelakantan]

The CDU processing capacity is 200000 bpd. My target kero rate is 20000 bpd and GO rate is 60000 bpd.

 

I believe this BPD targets are at standard conditions, right?

 

My first run of the simulation is based on specification of cut point distribution in the column.

Naphtha - 165 C

Kero - 220 C

Gas Oil - 360 C

 

The cut point specs are based on target product specs which are

Kero ASTM D86 90% T of 230 C

GO ASTM D86 90% T of 370 C

 

My simulation is converging. However, I am not able to get the total quantity of the Kero and GO cuts from the column when I am trying to meet quality.

 

What do I change in the simulation to slip more naphtha in kero and more AR in Gas oil?

 

you cannot specify the rate of KERO and GO randomly; based on the feed oil characterisation, split the stream manually first to see the range of the components and see percentage of lighters, middles and heaviers; then obtain the maximum possible in each product category.

 

regards

neelakantan

 

(to explain further, if the oil is heavy you will get more of furnace oil, right? so the ratio or the amount of the products is to be calculated based on the feed characterestics)

[ZHI]

Hi

 

I had a query regarding the PAs in Crude Column. I thought it may be better to discuss it over here then starting a new thread.

 

Can someone please let me know how do we fix the pump around duties. I know we take care of the L/V so that it doesnt fall below a certain value (0.1 as much as I know). I wanted to understand the emphasis of this no. Does this no ensure decent separation or is it to avoid tray drying or something else.

[PingPong]

The purpose of a PA is to recover heat at a higher temperature level than the column condensor. That heat is used to preheat the crude, and/or to reboil other distillation columns, and/or to generate steam.

 

But everything has a price: the disadvantage of any PA is that it reduces the vapor/liquid traffic above its location.

So you need more fractionating stages (trays or packing) in the column part above it, so as to obtain the desired separation between, for example, kero and LGO when the LGO drawoff is combined with an LGO PA.

 

It all depends on the required properties of the kero and LGO, and the required ASTM gap/overlap between them, how much LGO PA duty can be removed at a certain number of stages.

 

The L/V ratio of minimum 0.1 that you mentioned, can be a reasonable starting point for the optimisation process of PA duty versus number of stages, while meeting all product requirements.

[Prabhat123]

Getting 3-5 % of overflash is very tough in CDU simulation , which may call for very high number of theoretical stages than the typical exists. You please check whether over flash is in range or not? Overflash is how much liquid is leaving the tray from the last product to flash zone. This percentage must be calculated on the quantity of vapor in column inlet and not on Crude.I have faced so much difficulties in simulation in getting 3-5 % of overflash.

 

Prabhat

[Hamo0od92]

Hello everyone!
Is there some one in this forum that can help me in designing a deethanizer for my university project?
please check my blog for the project statement.
I will really be grateful for you help.
thanks a lot!!

[mohamedellejmi]

The discussion is too long, I couldn't read all of it.

I have designed a CDU 10 years ago... and here some recommendations:

- with HYSYS make sure you get the right SGs in your oil characterization.

- in HYSYS, do not rely on Peng Robinson for predicting the VLE, it will give you headache to get a converged column with PAs and side strippers, use Chao Seader EOS.

- the PA and SS may force you to change the draw-off position in order to get the maximum of the required cut

- get some help from previous design works in your company, don't try to reinvent the wheel

Good luck!