31 replies to this topic

[paulinaote]

x

Edited by paulinaote, 19 August 2015 - 01:18 AM.

[Zauberberg]

What do you want to do? Correlate end point (D86) of diesel with pour point?

[Zauberberg]

You can recover more diesel and still maintain the required pour point specification, by increasing separation efficiency between diesel and reduced crude (atmospheric residue). This has been discussed in another thread recently, and the ways to achieve this are:

 

- Increase the heater outlet temperature during winter season, while maintaining constant draw-off rates of products.

- Increase stripping steam rate

- Combination of the two from above

 

During winter, usually there is more condensing capacity available in the overhead system (fin-fan coolers assumed) to compensate for increased heater load. If the diesel fractionation trays will not exceed the flooding point, this is one effective way to keep both the yield and the specification of diesel product unchanged.

[PingPong]

That other thread http://www.cheresour...illation-hysys/ was also by paulinaote about the same problem. It would have been better to simply continue there.

 

Apart from increasing the heater outlet temperature, and increasing the bottoms stripping steam, you could consider to draw off the overflash (dirty washoil) by installing a new drawoff tray just below the washbed packing, and recycle it to the feed heater.

 

The most effective way however to recover gasoil from the reduced crude is to install a (mild) vacuum column.

Edited by PingPong, 08 April 2015 - 06:53 AM.

[PingPong]

-As I can inject the  stream red directly to the flash zone of the topping?
-As I calculate the plate which should enter?
-What is the difference vapor between mixing upstream and inyect directly to the flash zone?

You can simulate both options and see what is best.

Without doing a simulation I think it is best to leave as it is now. It seems to me that it is better to inject the red stream (preflash vapor) in the transfer line from the heater to the column, than to inject it directly into the flash zone. This preflash vapor is superheated at the T and P at the column inlet so it helps vaporizing the crude entering the column. That is assuming that there is good mixing in the transfer line. So the closer the injection point is to the heater, the better. The temperature drop between before and after injection is mainly due to the heat of vaporization of extra vaporized crude.

 

Note that the diesel pumparound should not be taken from the sidestripper bottoms but from the sidestripper feed.

 

The dirty wash oil its founded in the 21 tray?

No, the dirty washoil is the liquid dripping from the packed washbed directly above the flash zone, say tray 20.

Edited by PingPong, 09 April 2015 - 04:16 AM.

[PingPong]

It seems to me that it is better to inject the red stream (preflash vapor) in the transfer line from the heater to the column, than to inject it directly into the flash zone.This preflash vapor is superheated at the T and P at the column inlet so it helps vaporizing the crude entering the column. That is assuming that there is good mixing in the transfer line. So the closer the injection point is to the heater, the better. The temperature drop between before and after injection is mainly due to the heat of vaporization of extra vaporized crude.

I did a quick and dirty simulation of this, by modifying a CDU simulation model that I recently used, although that simulation uses a heavier middle east crude, not Tierra del Fuego 40API.

 

It turns out that injecting the preflash vapor in the transfer line, instead of directly  into the flash zone, does increase the amount of vaporization at the column inlet by about 1 t/h, as I expected yesterday, but ....... the stripout from the reduced crude stripping (trays 21 - 24) decreases by almost the same amount (at the same amount of stripping steam). That is because the reduced crude temperature is lower than if the preflash vapor is sent directly to the flash zone.

 

As the net effect on the gasoil yield will be negligable, it may be best to send the preflash vapor directly to the flashzone (instead of injecting it into the transfer line), because that will increase the reduced crude temperature (at the same amount of stripping steam) by roughly 10 ^oC. The diesel pumparound duty will however decrease somewhat, but nevertheless there should be some positive effect on the heat recovery in the crude preheat train.

 

If you decide to include a mild vacuum column to recover some gasoil from the reduced crude, then a higher reduced crude temperature is certainly a benefit.

 

But, as I said, this is based on simulating a heavier crude than Tierra del Sol, so you should compare both options in your own model and see whether sending the preflash vapor directly to the flash zone is indeed best.

Edited by PingPong, 10 April 2015 - 06:32 AM.

[PingPong]

I don't speak Spanish, so I don't know what you exactly mean by "Torre de Repaso", but I guess it is a Vacuum Column.

 

As it is the intention to only recover the atmospheric gasoil that is still present in the reduced crude from the Topping Column, you could use a simple Vacuum Flasher. The hot reduced crude from the Topping Column is simply sent to a Vacuum Flasher. The vapors are washed in a packed bed that is fed by reflux from the top pumparound. The top PA also condenses the gasoil that you want to recover.

 

 Simple Vacuum Flasher.jpg   23.49KB   5 downloads

 

You can add some superheated steam to the topped crude feed line, and you can also use some superheated stripping steam to the bottom of the vacuum flasher. Whether you use steam, and how much is a matter of economic optimization: steam costs money and produces sour water in the overhead condensors, but increases gasoil yield.

 

Selection of the absolute pressure in the vacuum flasher is also a matter of optimization.

 

Vacuum is maintained by an ejector package to remove any leak air as well as light hydrocarbons that might still be present in the topped crude.

Edited by PingPong, 11 April 2015 - 10:23 AM.

[Zauberberg]

ZIP is the magic word

[PingPong]

I don't use Hysys so I will not comment on your input file.

 

I dont understand your simulation flow scheme. You mix the product streams in MIX-100 to create the crude, but I don't see C1 - C4 product(s).

And then you mix more in MIX-101? What is that about?

And then you again add Kero and Diesel to the column?

And Steam Fondo to the feed line? Is that supposed to be the stripping steam? But that does not go into the feed line.

 

And you seem to have a reboiler on the column? That is not right. You should use stripping steam to the lowest stage.

 

It is correct this small amount of flow on the plate 8?

What is plate 8 in your simulation? Is that the dirty washoil dripping from the washbed?

 

but it it worries me the low production of kerosene

If you don't make enough distillates, or not enough washoil, you need to incresae the heater outlet temperature (column feed temperature). For the moment ignore the actual measured column feed temperature, worry about that later.

Edited by PingPong, 15 April 2015 - 10:55 AM.

[PingPong]

According to their work experiences, what are the requirements of REDUCED CRUDE to be treated in a SIMPLE VACCUM FLASHER?? (heavy metals,sulfur,etc etc etc...any requirement that you know.

There are no requirements for the reduced crude, except of course that it is properly steam stripped in the topping column.

 

What would be the price (a reference of course) of an SIMPLE VACUUM FLASHER??

That obviously depends on the size of the thing, and the amount and type of packing in it. Note moreover that the installed cost of the column, or any other equipment, is much higher than alone the cost of the equipment, due to its foundation, instrumentation, insulation, attached piping, labor cost, et cetera.

 

Engineering companies do not publish how they estimate the cost of equipment, or complete plants.

You can readrelevant chapters in Perry or Coulson&Richardson Volume 6 to get some idea, but those data are rather old. You can adjust them to todays prices using Cepci. Via google you can find additional cost estimating sources.

 

If you have access to Aspen Icarus software you could use that.

 

Read also: http://www.cheresour...n-column-price/

and there may be more topics about cost or price on this forum worth reading.

Edited by PingPong, 30 April 2015 - 07:48 AM.

[PingPong]

I don't have a PFD of a vacuum flasher specific for your refinery, but it is very similar to a normal VDU in many refineries.

 

It is like any other VDU, except that there is no feed heater, and there is only one PA with one product draw of gasoil, and below that one packed bed for fractionation. See the simplified flow scheme I posted in message #12.

 

The vacuum system is the same as any other VDU, with ejectors, condensors, a separator and pumps for condensed water and slop oil. Any gas (mainly leak air) from the separator is sent, via a flame arrestor, to the firebox of the crude feed heater.

 

The process data follow from your Hysys simulation. They depend on the flash zone pressure that you select. That is a matter of optimization: the lower the flash zone pressure, the more gasoil you can recover, but the heavier that gasoil becomes, and the bigger the vacuum flasher diameter needs to be. Try flash zone pressures of 0.10 bara, 0.09 bara, ..... down to 0.03 bara and see what gasoil you get. Use top pressure 0.01 bar lower than flash zone pressure.

 

You need to determine how far you can go without that gasoil becoming too heavy to be of much use or value. Refinery has to advise you about that.

 

You probably have to add some leak air (a few kg/h) to the flashzone in the simulation, as otherwise the column may not converge.

[PingPong]

- Why I must add air?? Only to converge my simple vacuum flasher? I treat it as a spec?

Every vacuum column leaks air into it through flanges of piping, manholes and instrument connections.

When you simulate the overhead system with the ejectors/condensors/separator you need to add an amount of leak air to get a realistic amount of offgas to be incinerated, normally in the firebox of a heater.

If the vacuum column does not have stripping steam you should add that leakair already to the flashzone otherwise there may be no vapor leaving the column top and the column will probably not converge. Some clients, or design companies (such as Shell Global Solutions) prefer dry vacuum column with very deep vacuum. I wrote that air comment with this in the back of my mind.

But in your case you will probably decide to use stripping steam so then it does no matter whether you add the leak air to the column itself.

 

Stripping steam quantity is really for you to optimize, you can use anything from 0 to 5 wt% on vacuum flasher bottoms.

Start with 2 wt% and adjust if deemed necessary.

Start with 2 theoretical stripping stages and adjust if deemed necessary.

 

Use 2 theoretical stages for the top PA section.

 

For the fractionation section (wash section) start with 4 theoretical stages, and increase/decrease if deemed necessary. Watch gasoil properties.

 

Flash zone temperature is resulting from the simulation: it is resulting from the reduced crude being flashed at the flash zone pressure (your optimized choice) in the vacuum flasher.

 

Tower top temperature in the simulation is resulting and affected by the return temperature of the top PA, which is dependent on the last exchanger in the PA loop. That new exchanger can be on crude, or cooling water, or aircooling, I don't know.

Your job to figure out at what temperature the PA will enter the tower top, but do not make the PA so cold that it might condense some stripping steam.

 

Do you have the ASTM D1160 laboratory data of the reduced crude, both in summer and in winter?

Edited by PingPong, 01 May 2015 - 12:29 PM.

[PingPong]

there is a procedure for estimating this temperature calculation?

For the moment assume that the top PA return temperature of the vacuum flasher is 45 ^oC.

After you determined how the top PA is going to be cooled you can adjust that number in the final simulation run.

 

Lamentably, I don´t have the ASTM 1160 of reduced crude. For some reason I do not understand, this refinery lab does not make.

So you have no data about the reduced crude, which is 45 v% of the crude, and which has an impact on the simulation of the lower part of the topping column, and on the vacuum flasher.

And you have no data on the C4 and lighter in the crude, which have an effect on the vaporization in the preflash drum, on the vaporization at the topping column inlet, on the temperature profile over the topping column, on the top temperature of the topping column.

 

Any study starts with the composition of the feed. However after nearly two months (since the beginning of your previous topic) you still don't really know what that crude feed is.

 

In your previous topic you wrote that the crude is Tierra del Fuego with API gravity 39.7

Just now I noticed that you have also posted on another forum, saying that the crude feed is a mixture of 3 crudes.

 

So what are the names and API gravities of these three crudes? And are there any crude assays of those crudes?

Edited by PingPong, 02 May 2015 - 05:40 AM.

[PingPong]

- Wich is  the PA flow? Any relationship between x and y who can you suggest?

PA flow is set to get the column overhead temperature that you want. I suggest you start with overhead temperature 20 oC higher than the PA return temperature, and adjust if deemed necessary.

 

- The z stream returning to the column  comes from a stripper??

No, comes directly from the PA pump (not shown).

[PingPong]

- The z stream returning to the column  comes from a stripper??

No, comes directly from the PA pump (not shown).

 

Instead of taking the underreflux from the PA pump, it is also possible to do it like in the topping column where there is a drawoff tray above the wash bed from which the diesel PA and diesel product are taken, while the underreflux, flows on gravity into the liquid distributor above the washbed. What is actually done depends on the type (and pressure drop) of the selected liquid distributor.

In any case, in the simulation that underreflux is always an internal stream that is not taken out of the column.

 

drop pressure in the stripping zone?.

That depends on what kind of internals you are going to put into that stripping zone.

To start just take 0.05 bar pressure drop, and adjust later.

 

Stripping Steam = [0,5]%wt * reduced crude feed ?

You can take as much or as little stripping steam as you like.

I suggested something in the range of 0 to 5 wt% of vacuum column bottoms, say 2 wt% to start with.

Edited by PingPong, 05 May 2015 - 02:40 PM.

[PingPong]

In de pdf page 10 its say that stripping zone is "VG AF trays equiped with SVG-H valves". I consider pressure drop 0,05 bar?

Total pressure drop depends on the number of trays you intend to install, and the design of the trays (outlet weir height, number of sieve holes or fixed valves). I suggested to start the first simulation with 0.05 bar, but you can take a higher number if you like.

In the final simulation you then put what it actually is based on the final number and type/design of the trays.

 

For the packing in the rest of the column I suggested earlier (in message #22) to take 0.01 bar pressure drop to start with. If, after you optimised the number of theoretical stages, select the type(s) of packing(s), et cetera, you calculate that the pressure drop will actually be 0.02 bar, instead of 0.01 bar, then you enter that in the final simulation.

 

You should not worry too much when you do the first simulation. Just put in numbers that are realistic and adjust them as you proceed in the optimisation of the design.

Edited by PingPong, 05 May 2015 - 04:02 PM.

[PingPong]

Delta T of pumparound is not a problem.

 

When I suggested a pressure drop of about 0.01 bar over the top part of the column, I was expecting that you would run the column at 0.1 bar or less. If you run it at 0.46 then the pressure drop is likely to be in the order of 0.03 to 0.05 bar.

Also if you would run at 0.46 bar it is better to increase the PA return temperature from 45 to say 65 ^oC to avoid possible water condensation.

 

You seem to heat the reduced crude feed to 350 ^oC, which would require a fired heater. However the whole idea of a simple vacuum flasher, to recover some gasoil, is to not needing a fired heater.

 

You seem to process only 20 % of the reduced crude in the vacuum flasher, and you try to recover 40 m³/d SDGO from that 258 m³/d RC. That would correspond with a TBP cutpoint between SDGO and RC of roughly 355 ^oC. The TBP endpoint is always higher than that due to imperfect fractionation. The SDGO that you make will have quite a high pour point.

 

You should process all the RC in the simple vacuum flasher, without a feed heater, and with a much lower pressure than 0.46 bar. In that case the 40 m³/d SDGO will have a lower TBP endpoint and D86 endpoint, which is what you seem to want according to the spreadsheet. Moreover this SDGO will have a lower pour point (I guess about 5 degrees lower) and also a lower viscosity, compared to what you are trying now.

 

You seem to use 4 theoretical stripping stages. That would require many trays and is unrealistic. Take 2 theoretical stages at the most.

 

You seem to draw the SDGO product from stage 1, but in reality it will be taken from the PA.

 

What parameter should I change so that the curve Swg.Dsl.Go  looks like Light stream? (excel)

Which stream in the flow scheme is the Light Stream in the spreadsheet ?

 

For my blending only need 40 - 45 m3/d of Swg.Dsl.Go (swinger diesel-gas oil)

What stream are you going to blend with the SDGO ?

Edited by PingPong, 08 May 2015 - 07:08 AM.

[PingPong]

As I said: you should process all the RC in the simple vacuum flasher, without a feed heater, and with a lower pressure than 0.46 bar.

The RC will partly vaporize simply by flashing over the level control valve of the Topping bottoms. You must set the pressure of stream 110 equal to that of the flash zone of the SVF. The vapor part 110V of stream 110 shall go to stage 7, the liquid part 110L to stage 8. The stripping steam (Steam2) to stage 9. Remove stages 10 and 11.

 

It is a matter of playing with the SVF pressure, stripping steam quantity (Steam2). If necessary add some transfer line steam (Steam1) but I don't think that is necessary.

 

It should not be a problem to obtain 40 m³/d of SDGO meeting your requirement that TBP endpoint <= 360 ^oC, provided that you send all RC to the SVF.

If you try to get 40 m³/d SDGO from only 258 m³/d RC then obviously it is impossible that the SDGO TBP endpoint will be below 360 ^oC. 

[LCPS]

thank you very much sir, I'll do it says.

Best Regards

 

Hola Paulina, Sigues con los problemas / Hey, are You still solving the problems?

 

Estoy tratando de revisar toda la información que das para ver como te ayudo / I'm trying to check all Your information provided to try to help.

 

Que tal si le agregas una Especificación a la Columna en la que le pidas la cantidad exacta que quieres del producto, y si te corre la Torre, puedes comparar las caracteristicas nuevas. Teniendo estos datos, puedes ir cambiando especificaciones, Platos de alimentación y Equipos por agregarle a la Columna buscando obtener el mismo resultado. / What if You Add a "Column Specification" to the Column forcing it to extract the desire quantity of X Product, if the simulation runs, You can compare the new calculated specifications. Having those specifications, You can try changing configurations, Feeding Trays and even adding Equipments to the column to simulate the results obtained before.

[PingPong]

You should read again what I wrote about tghe PA return temperature and the column top temperature.

Your PA return temperature of 100 ^oC is much too high.

Your Top temperature of 160 ^oC is very much too high. With this huge top temperature you loose too much light gasoil components that will end up in the wet slops oil from the ejector set, instead of in the SDGO.

 

Your stripping steam quantity (Steam2 = 97 kg/h) is very low with only 0.2 wt%. I am very surprised you managed to draw 50 m³/d SDGO at 0.33 bar with such tiny amount of stripping steam.

 

Your Underreflux of only 10 m³/d is much too low. By use more stripping steam that will increase.

 

Your SVF flash zone pressure of 0.33 bar is too high if you now want to recover 50 m3/d SDGO instead of 40 m³/d. You need to drop that pressure so as too increase the underreflux, and as a result of that you get better TBP Endpoint (<360 ^oC) and ASTM D86 Endpoint (< 340^oC).

 

I think you can produce 50 m³/d SDGO (TBP EP <360 ^oC & D86 EP<340 ^oC) with:

flash zone pressure of 0.25 bar (top 0.23 bar)

2 wt% stripping steam,

PA return temperature 55 ^oC

Top temperature 75 ^oC (20 degrees higher than PA return)

 

Or with:

flash zone pressure of 0.10 bar (top 0.09 bar)

 1 wt% stripping steam

PA return temperature 50 ^oC

Top temperature 70 ^oC (20 degrees higher than PA return)

 

There are many combinations of flash zone pressure and stripping steam quantity possible.

 

Don't worry about the outlet pressure of the feed control valve. At the SVF inlet the pressure of stream 110 is equal to that of the flash zone, and that is what matters in this simulation.

 

An ejector set is always necessary because of leak air into the SVF system. As long as you keep the flash zone pressure at 0.10 bar or higher you can use a precondensor upstream the first ejector, which makes the ejectors smaller.

[PingPong]

A PA return temperature of 65 ^oC and a top temperature of 85 ^oC would be OK, but It seems that in your simulation the SVF top temperature is not 85 but 236 ^oC. That is even worse than you had before.

 

Look also at the material balance:

RC feed = 1306 m³/d

SDGO = 50 m³/d

Vac Resid = 1069 m³/d

 

So there is 187 m³/d missing, which is the lighter part of the SDGO going overhead due to that very high top temperature.

 

So you are not really producing 50 m3/d SDGO but 50 + 187 = 237 m³/d. Obviously that means that the FBP of the SDGO becomes very high.

 

You must spec the top temperature in Hysys and let it adjust the PA flow to achieve that spec.

[PingPong]

At least the top temperature is now down to 92 ^oC. Still a little high but for the moment good enough.

 

But now you draw 90 m³/d SDGO instead of 50 or 40 m³/d.

I can not believe that that is possible at 0.25 bar and only 0.5 wt% stripping steam.

 

I think that the composition of your Reduced Crude is too light.

Two months I warned about the unrealistic crude oil TBP, with an end point of only 397 ^oC, that you listed in one of your spreadsheets. http://www.cheresour...sys/#entry93499

 

If you are still using a similar unrealistic crude oil TBP then your Vacuum Flasher simulation results will also be unrealistic.

Edited by PingPong, 13 May 2015 - 12:49 PM.

[PingPong]

I improved characterization of the Feed. I explained a few paragraphs up. It's the best I could get with the resources that I have.

This is the TBP of the feed from HYSYS.

That TBP curve looks fairly realistic, except the part between 65 v% (392 ^oC) and 85 v% (439 ^oC), which is much too flat compared to other crudes of similar gravity, as you probably already noticed yourself.

The part between 20 and 65 v% has a TBP Slope of 5.1 ^oC/v% but between 65 and 85 v% the TBP Slope is only 2.4 ^oC/v%

Normally the TBP slope does not decrease after the atmospheric gasoil part, but gradually increases towards the FBP.

See also Perry Fig. 13-105 (8th Edition) or Fig. 13-85 (7th Edition).

 

Production m3/d

                                                SIMULATION          REAL

NAFTA                                     574,5                      570                                                                                 

KERO                                       268,4                      450

DIESEL                                    502,2                       520

REDUCED CRUDE                  1306                       1110

Due to that big difference between Simulated Kero and Real Kero there is a big difference between the Simulated RC and the Real RC.

That has a big impact on the simulation of the SVF because that 1306 m³/d Simulated RC contains about 200 m³/d too much gasoil components. Note that you want to recover 40 - 50 m³/d SDGO but in your Simulated RC there is 200 m³/d GO that should not be there in the first place!

Obviously that makes it very easy to recover a lot of SDGO with little effort, but that is not real, because the SDGO that you are recovering in your SVF simulation is not really present in the real RC. Do you understand?

 

You need to redo the Topping simulation to produce only 1110 m³/d RC and feed that to the SVF.

Edited by PingPong, 13 May 2015 - 03:24 PM.

[PingPong]

I don't know what you mean by "defense" but I have not had time to look in detail at all your numbers yet.

To justify the TBP you use you could also refer to other crudes of similar gravity, and the Perry figure that I mentioned.

 

Because I found your SVF simulations very optimistic I was afraid that your RC was too light due to unrealistic crude TBP, but now it is clear that the RC is too light simply because it contains far too much gasoil (MGO).

 

You must therefor draw more Kero and/or MGO(diesel) in the Topping simulation untill you get the Real 1110 m³/d RC quantity to feed the SVF.

For the SVF simulation it does not matter where you cut between Kero and MGO, as long as the cut between MGO and RC is correct.

 

Consider matching the Topping simulation with the Topping reality as a separate problem to solve on another day.

Edited by PingPong, 13 May 2015 - 04:13 PM.

[PingPong]

2) increase mass flow of main steam

What do you mean by main stream ?

 

- overflash (1 tray above the feed tray) of 1134 m3/d.

Do you mean that the liquid from stage 13 is 1134 m³/d ? If so, that would be absurdly high, and you should decrease the heater outlet temperature.

 

The number of stages that you use in the Topping column is unrealistically high, but let's worry about that later.