15 replies to this topic

[cabeghe]

I would like some thoughts on when a PSV must be sized for a flare knock out drum (FKOD). I have a FKOD with a PSV to a 2-phase pop tank. However, the pop tank is open to atmosphere and the process fluid is sour. 3% H2S. I am considering getting rid of the PSV and directing all vapor to the flare stack. The FKOD is maintained at 20-35 psi while the flare line is 2 inch. The fluid is 99% vapor but there could be some liquid phase in the gas that could carry over to flare without the pop tank. However, since the pop tank is open to atmosphere, it is basically releasing all this sour gas to the surrounding. Against regulations.

The problem is, I think getting rid of the PSV leaves only one protection for the FKOD. The FKOD has a high level alarm and is set to shutdown on LSHH. I have a PCV set to control D/S pressure to 3 psi.

There is the fear of potential fire rain if there is liquid phase in the flare gas. The piping to the FKOD goes underground for some distance before coming above ground to the FKOD. This may hold liquid pockets that would eventually get in to the FKOD and flare.

Can I get rid of the PSV? What do I need to do to ensure adequate protection once the PSV is gone?


Will appreciate other view points on this.

Many thanks!

[paulhorth]

Cabeghe,
You must be in a specialised area of process engineering, because although I have spent a career in oil and gas, I do not understand your post at all.

What is a pop tank?

I have never seen a flare knockout drum with a PSV on it. They are always open directly to the flare without any closure.
If you could send in a diagram of the system, it would help me and others to understand your problem.

Paul

[cabeghe]

Hey Paul,
Thanks for taking the time to post a comment. Much appreciated. Perhaps I did not do a good job on the problem statement. This task is aimed at replacing an existing FKOD that has a PSV on it. When the PSV pops, it relieves to a POP tank. This is how it was set up.

A pop tank is a vessel typically used to contain fluid relieved from a system when the PSV pops. These pop tanks are commonly used in facilities in Canada.

I consulted the provincial boiler and pressure vessel regulator and their view is that there is no need to have a PSV on a vessel where pressure will not exceed 15 psig. This can free flow to the flare. Apparently the regulators require that a PSV be installed on vessels with operating pressures greater than 15 psig.

[paulhorth]

Cabeghe,
Thanks for your reply, but I am still puzzled by your system description. In my experience (which covers North Sea offshore oil and gas, as well as onshore gas plants around the world), all PSVs in a plant are routed to one or more headers which discharge to a Flare knockout drum, which then is vented directly to the flare stack. There are no pop tanks. The purpose of the flare KO drum is to separate the liquid coming from the PSVs from the gas which passes to the stack. This function of the flare KO drum is what you say your "pop tank" is for. And furthermore, as I said, the flare KO drum itself never has a PSV on it.
Are you saying that without a pop tank, the PSV would discharge directly to atmosphere without going into a flare header and from there to a KO drum?

Your comment about no need for a PSV at 15psig is baffling. Don't you have to adhere to vessel design codes in Canada, is it really up to the views of an individual in a regulator's office?
The point is not what is the "normal operating" pressure of a vessel, but what pressure it could be exposed to in an upset condition. If your vessel was at 10 psig, but had a feed at 500 psig, and could be blocked in, you MUST fit a PSV!

Paul

[GS81Process]

I work in Canada as well.

Just to clarify, a "pop tank" is an atmospheric tank used for secondary containment, particularly at remote oil and gas sites. That is why is does not require a PSV- it has an open vent. Tthe only reason a "pop tank" exists is because there is no easy way to store relief fluid at these remote sites so it must be kept in a "pop tank" until it can be trucked away.

cabeghe, I disagree with your statement that any vessel operating at less than 15 psig does not require a PSV. A vessel operating at less than 15 psig does not require registration with your local Boiler Safety Association, however that does not mean that the vessel does not require a PSV. That is up to the professional engineer to determine as your are responsible for the safe design of the vessel. I have sized many PSVs for non registered vessels. Also, proper design standards must still be adhered to (eg. atmospheric tanks to API 2000, etc.)

Edited by GS81Process, 22 November 2011 - 07:02 PM.

[paulhorth]

GS81,
Thank you for that explanation. It is a little clearer. As I understand, you are saying that the pop tank is used for individual relief valves at a remote location in place of a full flare system, and the PSV vapour discharges to atmosphere. Would the location be unmanned? Or would people be exposed to the risk?
I have a further question: what happens when the relief flow is sustained, not just a "pop", due to some upset condition which is not stopped? Is there on the pop tank a high level shutdown which shuts down the wellheads or other sources? This would be the case on a conventional flre KO drum, to prevent the liquid continuing to flow, eventually discharging to atmosphere.

My comment to cabeghe about the requirement for PSVs stil stands.

Thanks
Paul

Edited by paulhorth, 24 November 2011 - 09:00 AM.

[GS81Process]

I must have had a few too many beers before my first post because now I see that it was not too clear.

The "pop tank" will receive multiple relief streams at a remote location. The site will also have a flare knockout drum (FKOD) and flare stack, but the following relief streams will be typically directed only to the pop tank: relief from the free water knockout (FWKO), flow from the pressure reducing valve on the inlet if the wellhead pressure switch fails, and inlet PSV.

The Pop Tank should be equipped with a high level switch to trip an ESD valve, and a flow switch in each relief line (alarmed).

It is used as a secondary containment device to avoid spills to the environment. There are government regulations (by Province) on the minimum required size of a pop tank. As pointed out by Cabeghe there are also strict government regulations on emissions (especially H2S) that are allowed from a pop tank, both at the point of discharge and at the perimeter of the facility.


A few comments on this specific application:

- Have you confirmed that the H2S and other emissions at the pop tank vent and facility boundary do not meet provincial regulations? I bet that you're correct and it does not...

- Why are you concerned about liquid carryover to the flare stack if you eliminate the pop tank? If you eliminated the pop tank would you not tie the relief streams into the others going to the FKOD, where the liquids would knockout?

- I agree with Paul Horth's comment that a PSV should not be required on the FKOD, but you have to ensure in the design that there is no way to block in between the FKOD and the flare stack. I cannot guarantee this without seeing detailed design documents.

- My previous comment stating that PSVs are often required for vessels operating at less than 15 psig still applies- even in Canada. Vessels operating at less than 15 psig are not required to be registered with the Provincial boiler safety association, but they still need overpressure protection if required. If an inspector told you that they do not- yikes. That is wrong and dangerous.

- Is the underground piping you described located between the FKOD and the flare stack? Is it possible to get a liquid pocket in this section? If so, you would already have a risk of rain out from the flare stack.

- There are obviously a few key design checks for your proposal. Is the FKOD large enough for additional flare rate? New heating value ok? Is the 2" line large enough based on mach no., pressure drop etc. You will need to do a detailed flare hydraulic analysis.

Edited by GS81Process, 24 November 2011 - 02:02 PM.

[cabeghe]

Many thanks Paul & GS81 for the lively debate. I want to provide some additional points. I know it would have been best to have provided drawings. Apologies...

I am getting rid of the PSV on the FKOD which releases to the pop tank. This will effectively render the pop tank to demolition. This is to prevent release of H2S to the atnosphere. All the FKOD fluids will therefore go to flare now. FKOD and Flare system including piping are more than adequately sized for the additional fluid vols.

The flare system has been designed to API RP 521 as far as I know. I am actually not a process guy but mechanical. Hence my subscription to this forum so I can to pick from your collective knowledge base. As per regulations we have a high level alarm to call-out, level indication as well plant shutdown on high level. Shutdown closes the two inlet EDSs from the FWKO and Treater that supply fluids to the FKOD.

To prevent freeze up or block-in of the flare line, we plan to insulate and electrically heat trace this line. Also this line is above ground and is sloped (1:200) back to the FKOD. Furthermore, I am insulating and heat tracing the FKOD.

The underground lines are the supply lines from the FWKO and Treater. These are fibre glass materials not steel.

There is a point that both Paul and GS81 have raised here on how to ensure there is no overpressuting of the FKOD that is very interesting. However, no matter how much I think about it I cannot seem to come up with a scenario given the steps taken above where we might block-in the flare line and over pressure the FKOD to an unsafe level. Your comments/thought on this point is very much appreciated.

If we add a high pressure shut down in addition to the above controls, what do I set it to? 14 PSI or much higher? I can see to that either? Wont this lead to an over design? Wont want the plan to be shutting down frequently.Seems to make sense to free flow to flare under this situation. Anyway, hope these details help.

Will await your thoughts! Many thanks guys!!

[GS81Process]

Sounds like you do not need a PSV on the FKOD based on your design. This is normal.

I also do not believe that you required a high pressure shutdown.

[cabeghe]

Thanks GS81. What would you consider in an underground piping to an above ground FKOD? What safety considerations would you put in place with that design to avoid liquid pockets?

A team member has provided an opinion on putting the FKOD underground rather than above ground. This sounds pretty unusual/complicated but could allow for sloping back to the FKOD from the underground FWKO and Treater lines?. What design codes should one reference on this? API RP 521 and the local boiler regulations do not specify whether an above ground or below ground FKOD location. The piping to the flare has remained above ground. Thanks!

[paulhorth]

cabeghe,
I understand that this installation is in Canada, and things seem to be done differently there. However I would say that for any other location which I have designed for, underground routing of piping to the flare KO drum would be simply unacceptable. Faced with it being already installed, I think that most operating companies would pull it up and replace it with correctly sloped free draining lines on a piperack.
Pockets in a flare line are dangerous for a number of reasons, even if not underground.I will list three, but there will be others.

• Liquid will get trapped in the pocket.If there is any water, this can freeze (especially in Canada) blocking the line.
• Liquid will get trapped in the pocket. When a relief valve opens, this liquid will be propelled at the speed of the gas, and will hit piping elbows with a force of many tons.I have seen photos of 36 inch flare piping hanging off the rack after such an incident.They were lucky.
• Liquid will get trapped in the pocket. Corrosion will take place due to water and CO2, leading to perforations. When the line is used, there is the risk of a dangerous undetected gas release.

That's why underground flare drums are not discussed in API RP 521, good design practice would never give rise to them.You mention local boiler regulations. Why should you expect steam boiler regulations to hydrocarbon flare KO drums? These industries are completely different.

Paul

[GS81Process]

I agree with PaulHorth that underground piping to the FKOD is unacceptable because you can never guarantee there will not be liquid pockets. This is a bad design. Throw it in the garbage.

Paulhorth, in many Canadian jurisdictions the regulatory body responsible for ALL pressure vessels is the Boiler Safety Association. Confusing I know.

Edited by GS81Process, 29 November 2011 - 03:02 PM.

[cabeghe]

Paul & GS81,
This layout is existing and we are looking to make it safer by removing pockets. That is why we are examining safer options and your comments have been very helpful in steering me in the right direction. The preferred solution is to have an above ground piping sloping to the FKOD to avoid pockets while the 2nd option is to have an underground FKOD with the existing underground piping sloping back to the underground FKOD.

I do not personally like the 2nd option as it makes the design more complicated with lots of issues to consider. Do you guys have an experience with this 2nd option for the design? Any learnings you can share? I am not sure there is any standard for this option.

Cheers!

[GS81Process]

Go with the first option.

[cabeghe]

Thanks GS81. This is the option am looking to at this time. Where are you in Canada? Am in Alberta. Have a nice evening.

[GS81Process]

Also in Alberta.