15 replies to this topic

[peeyar]

Hi
I am new to this group but have noted very experinced people in this group
I am sizing a relieve valve on desalter which is a liquid filled vessel The discharge goes to the flash section of crude coloumn.My question are
What contingencies need to be considered--- blocked out let or fire or both

The discharge line has to be free draining to the crude coloumn which is some 40 feet from grade level this means that the RV has to be at a height near to crude coloumn this would require a long inlet line with static head The 3 % criteria may be difficult to achive

Iwould like to have your knowledgeable comment
thanks
PEEYAR

[Zauberberg]

Hello Peeyar,

I do not have experience in desalter PSV sizing, but here are my experiences in desalter/PSV operation while I was working as process engineer in crude distillation unit.

1. Desalter PSV discharge is routed to main column flash zone and is accompanied with bypass (globe valve), for steaming and circulation purposes. Usually, the operating pressure difference between desalter and column flash zone is 5-10 bar. PSV is placed at the top of desalter vessel and the elevation difference between RV and column flash zone is 4-5m (depending on unit/column capacity). With this arrangement, RV was functioning perfectly. Do not forget that desalter contains 100% liquid at operating conditions.

2. Blocked discharge is the only operating case I encountered in practice.

Hope this helps,

[pleckner]

First, without seeing your P&IDs no one can answer you with any certainty and if they try, they are opening themselves and you to one big liability issue.

However we can make some general comments. I think a fire would be a credible scenario. Blocked-in condition could or could not be and I can't tell without seeing the P&IDs. I think you should re-think the relief valve location. Perhaps you can put in a catch pot with a small pump. I would not want the relief valve 40' in the air, above the source of over pressure.

[peeyar]

Hello Peeyar,

I do not have experience in desalter PSV sizing, but here are my experiences in desalter/PSV operation while I was working as process engineer in crude distillation unit.

1. Desalter PSV discharge is routed to main column flash zone and is accompanied with bypass (globe valve), for steaming and circulation purposes. Usually, the operating pressure difference between desalter and column flash zone is 5-10 bar. PSV is placed at the top of desalter vessel and the elevation difference between RV and column flash zone is 4-5m (depending on unit/column capacity). With this arrangement, RV was functioning perfectly. Do not forget that desalter contains 100% liquid at operating conditions.

2. Blocked discharge is the only operating case I encountered in practice.

Hope this helps,

Hi 'Zauberberg

Thanks for your response My intial thought was to place it on the vessel nozzle but in that case it cant be free draining to the main coloumn as the line would rise again
I asked a friend who works for a large refinery designed by world renowned lincesor and detailed engineered by another large contractor They have have the safety valve near the coloumn a large line is connected to the desalter nozzle and this goes to inlet of the RV located at a height this way the discharge is free draining
But as you have worked on a desalter with RV on the drum it appears that both arrangement are being used

Any further comment
best regds
PEEYAR

[peeyar]

Thanks
The scheme is very simple and very standard in a refinery
most deslater RV that ihave seen only show blocked out let with discharge as free draining The flash vessel scheme is something i have not seen besides the size of the pump will be about the size of crude feed pump with tight NPSHA
Peeyar

[JoeWong]

Thanks
The scheme is very simple and very standard in a refinery
most deslater RV that ihave seen only show blocked out let with discharge as free draining The flash vessel scheme is something i have not seen besides the size of the pump will be about the size of crude feed pump with tight NPSHA
Peeyar

You may not get certain & reliable answer without submitting a sketch ( if not P&ID) with necessary info like block valve, PSV location, etc.

[Zauberberg]

Hello Peeyar,

CDU plant where I was working in the past has been comissioned in year 1978. There was no single one ESD valve in the plant safeguarding system (only process and manual shutdown), and fire case definitely was not considered as a potential hazard for the personnel, environment and process equipment. That is why I stated that "blocked discharge" was the only case I encountered in practice.

Raising desalter pressure was the mechanical issue (blocked discharge or downstream booster pump failure). In such case, I think that 5m hydrostatic head is of no importance - since you are discharging the fluid to the main column flash zone operating at 1-2barG (5 to 10 times lower pressure) and the resulting overpressure is well below the vessel design pressure.

As I said, I do not have experience in desalter RV sizing, but in the case of external fire I think you will need thermal relief valve. Since there is no vapor space in desalter, increasing the temperature will not generate any vapors. Maybe I am wrong; this statement is derived only by using my common sense logic. I hope you will have more useful answers from experienced professionals such are Phil Leckner, Art Montemayor, Latexman and Joe Wong.

Best of luck,

[pleckner]

I too do not have experinece in refinery design so someone will have to tell me why a vessel filled with oil (OK, and some water with impuruties) would not be considered for a fire scenario?

As the vessel is heated liquid will be burped-out first due to swell. That will create space for vapor formation.

[Zauberberg]

Good evening Phil,

I am very interested to hear more about this burping and swelling scenario of 100% liquid filled vessel. I think that people who designed crude unit long time ago (the end of '60s) were not aware of this possibility. And it still looks quite a bit mysterious to me.

If I got the point, the sequence of events is as follows:

1. Increase in vessel temperature will cause PRV to pop-open, due to liquid thermal expansion effect.
2. This relief will generate sufficient space for vapor generation (I am suspicious about this issue; I think we will have 100% liquid filled vessel again, but with lower liquid density due to increased temperature).
3. This vapor space will be "responsible" for further pressure increase, if the vessel is exposed to external heat source?

Please shed some additional light on this subject, I think it will be extremely useful for both Operations and Design engineers.

Regards,

[peeyar]

I too do not have experinece in refinery design so someone will have to tell me why a vessel filled with oil (OK, and some water with impuruties) would not be considered for a fire scenario?

As the vessel is heated liquid will be burped-out first due to swell. That will create space for vapor formation.

HI
This problem was also highlighted by Mr Wing Y.Wong in Chemical Engineering of may 2000 but does not offer any soloution as the long inlet line to RV may also require heat tracing otherwise for high pour point crude during winter may pose hydraulic problem a by pass globe valve culd be provided but again poses its own problem--any comment
PEEYAR

[Zauberberg]

Hello Peeyar,

As I said, placing RV directly on the desalter nozzle does not pose any additional problems - no requirement for heat tracing, amongst others (if line is filled with liquid from the column flash zone, pour point of the liquid in the line will be much higher than for the crude oil itself). This has been proved more than 10 times while I was working as operations engineer in primary distillation facilities, and I don't need any engineering procedure in order to feel sure about this particular matter of concern. As long as opening of your PRV is governed by the pressure being raised in some other way than the one in fire scenario, there is no single one thing that you should worry about.

Regarding the fire case, I would really like to hear opinions from other forum experts. Using conventional PRV in 100% liquid filled service was always a mystery to me. And it is still not solved.

[peeyar]

Hi thanks again
The problem arose as the discharge line to flash zone is shown in the Pand ID as free draining
question is how to make discharge line free draining without raising the RV location

pl also see Chemical Engineering of May 2000 article titled
"Fires,Vessels And Pressure Relief Valve"
regds
PEEYAR

[jprocess]

We had a desalting plant project earlier for which the desalter contains high boiling point crude such that if fire occures for it, the crude will not vaporize even after a long time!
So we routed the psv discharge line to the HC closed drain system to protect the desalter against fire.

[pleckner]

@Zauberberg

The case of a 100% liquid filled vessel (no vapor space whatsoever) during a fire has been studied in some depth since the commissioning of DIERS. It's a complex dynamic because at the onset of a fire the liquid will expand and start the pressure increase just as you would have in a blocked-in heat exchanger tube. The PSV will open to relieve this pressure and as soon as the pressure is relieved and you reach blowdown the valve will close, creating the vapor space. Bubbles form in the liquid as it is heated by the fire but the consnsus is that mixing occurs in the vessel between the cool liquid (being heated) and the bubbles. At first this mixing is poor but the pressure will build to open the PSV once again. At this point you will probably get a two-phase relief creating a larger vapor space. You have not achieved full relief yet because there is a finite time from the onset of the pressure relief from the liquid heating and boiling until you reach 121% of MAWP. During this time you loose more liquid until a sufficient vapor space develops that allows full disengagement of the bubbles from the liquid and relief becomes 100% vapor.

If the valve does not close after the initial pressure relief due to the thermal expansion of the liquid, you will begin to have this two-phase relief because bubbles are formed within the liquid and begin to rise but because the level is high, there is no room for proper disengagement. But again, it takes time to reach full relieving pressure and by the time this happens, you usually achieve the necessary vapor space to convert the relief into 100% vapor.

It has been observed that because of this time delay from the onset of boiling until the vessel reaches full relief pressure, you can safely size the relief for 100% vapor rather than two-phase unless the system is either foamy and/or very viscous. If you run into this type of system, two-phase relief is almost a certainty.

[Zauberberg]

Thanks Phil,

The explanation is very good and it has an engineering logic. Somehow, it still sounds too strange for me (which does not mean it cannot be true). I remember I had a very intensive discussion about placing PRV on 100% liquid filled hydrocyclone skid operating @ 5barG pressure (and apart from specifying PRV for that particular service, relieving temperature is specified to be 100C - which convinced me even more that there is a serious mistake in PRV design approach).

I am trying to imagine that scenario - having pure water and 100% liquid vessel. What will be the relieving temperature in this case?

Regards,

[pleckner]

Why would you have a problem with a PSV on at 100% liquid filled vessel? Todays PSVs are designed to handle both liquid and vapor/gas flows.

The relieving temperature will depend on the scenario. Obviouslly, non-fire or non-heat related cases will not increase the temperature of the water in the vessel so that will govern the temperature. Heat related scenarios and fire will can the water to boil so that will be the relief temperature. Relief temperature is always coincident with the relieving pressure. You might also get flashing (two-phase flow) through the PSV depending on the relief conditions.