Is a PSV required if there are no credible scenarios? 3

[ipocoyo]

Hi All,
My site consists of 3 air compressors connected in parallel which are connected to an air receiver. The air compressors each have a maximum operating pressure and cut off pressure of 10 bar(g). All the equipment in the compressed air system are rated with a design pressure of 16bar(g). According to API 521, the only credible scenario of over-pressurisation happening is due to an external fire since the air compressors are located in a room beside the boiler. From my understanding, fitting a PSV for the event of an external fire is not regarded as sufficient protection, but does buy extra time in the event of an over-pressurisation. The compressor system will still be lost due to the external fire.

My question is, is it still a requirement to fit a PSV if the only credible scenario is due to an external fire (which does seem unlikely with the different controls in place)? My local standards states that “each pressure vessel shall be protected by a pressure-relieving device that shall prevent pressure from rising to more than 110% of the design pressure of the vessel.” This does seem a bit ridiculous since even the small filters (16bar(g)) are considered as a pressure vessel and thus requires a PSV. Is there anything that I can reference which states that if an event of over-pressurisation is unlikely then a PSV is not required?
Thank you!


Still an intern

 

[The Obturator]

You mention ".....external fire (which does not seem likely......", you have essentially revealed a potential risk. The whole point of a PRV is to protect loss of property and life.

Plant location and/or local jurisdiction dictates the requirement of PRV.
In USA for pressure vessels, the ASME code requires a PRV for pressures above 15 PSIG
In EU the PED requires a Safety Valve* for pressures above 0.5 barg

ISO-4126 A Product STANDARD, wherever applied, applies to pressures 0.1 barg and above.

(* - In EU the term Safety Valve covers all PRD's (relief valve, pilot operated etc).

So Yes, I guess you need to spec a PRV.

Safety goes a long way these days. I was once involved in a PRV application, which required the PRV to have steam jacket fabricated aroud it's body to keep it warm due to high viscosity of protected fluid. Since the steam jacket had a working pressure of 2.1 barg, it had to have it's own PRV protection uder PED. rules.

Per ISO, only the term Safety Valve is used regardless of application or design.

 

[RVAmeche]

Many inspectors will require a vessel to have a small PSV just to be safe. "Safe by design" is a hard concept to approve and requires a lot of documentation. Do the calculations for a fire case and install the required relief valve, but you're correct that it won't actually protect the vessel in this case. The owner (you?) should be aware that other actions are required if a fire occurs.

 

[LittleInch]

What do you mean by "From my understanding, fitting a PSV for the event of an external fire is not regarded as sufficient protection"

A fire case PSV should allow the vessel to maintain its integrity in a fire case scenario.

Having vessels exploding in a fire case is potentially lethal to fire fighters and can escalate an event, especially in this case where the additional oxygen could create a fireball.

In general, unless you can show that an overpressure event CANNOT occur under any feasible circumstances, then the consequence of failure is normally many time greater than the small cost of a PRV, especially when it is air in the system which you can free vent in most locations, just away from any personnel.

Something which is just "unlikely" won't or shouldn't get past the person certifying and approving the particular installation.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

Ipocoyo,

Read this recent thread. thread1203-454905

Good luck,
Latexman

Engineers helping Engineers

 

[RVAmeche]

For air receivers/non-wetted vessels, API 521 indicates that a relief valve won't be sufficient to protect the vessel because the heat & lack of heat transfer could result in the vessel wall failing before the relief valve goes off.

 

[ipocoyo]

Hi everyone,

Thanks for replying to my thread. I guess a PSV is always required to be attached on anything that's considered a pressure vessel.
I will always size the PSV to a fire case scenario even though there is no definite identifiable source of fire.

Thank you!

Still an intern

 

[Latexman]

I guess a PSV is always required to be attached on anything that's considered a pressure vessel.

Most of the time, in my experience, that's the economic answer to the Code requirements. In your case, air, I'm sure it is. But, if you had a pressure vessel with gaseous "Methyl Ethyl Death", you might not want any significant releases whatsoever. UG 140 in ASME Sec VIII gives the Engineer the flexibility to use "system design" and not install a relief device. But, you need to follow UG 140 and it's quite thorough and onerous, as it should be!

Good luck,
Latexman

Engineers helping Engineers

 

[ipocoyo]

Hi everyone,

I am having trouble with sizing the PSVs. API 521 clause 4.4.13.2.3 provides a formula to find the estimate of the required effective discharge area in the event of external fire. However I get a non-real result because T1 is greater than Tw. Is Tw the maximum operating temperature of the vessel (120 degrees)? The definition of Tw is very vague and is defined as the maximum wall temperature of vessel material.

Help is appreciated thank you!

Still an intern

 

[The Obturator]

Think you may be using the wrong formula. From what you advised, the vessel is "unwetted" and filled with air only. Since we are looking at external fire, the top formula in your sheet should be used. This calculates area of PRV against surface area, environmental factor etc. Based on a bare vessel you'll be calculating an area of approx 23.23 mm2/0.036 in2 in which case your best economical solution would be a small 3/4" x 1" threaded valve (since your stated operating pressure is sufficiently lower than set pressure to suit longer blowdown of the PRV). You'll need a lever if the PRV is to be UV Code Stamped.

Note that I have taken 120'C in calculation for the flowing temperature (due to fire)

Note also your own calculation sheet has an incorrect metric & USCS value for atmospheric pressure.

Don't take my sizing as gospel. Satisfy yourself with the notes and data and review he formula use.

Hope that helps.

Per ISO, only the term Safety Valve is used regardless of application or design.

 

[The Obturator]

Apologies. Was meant to say I used the top formula using standard default of F1 (Environmental Factor) of 1.0

Per ISO, only the term Safety Valve is used regardless of application or design.

 

[ipocoyo]

Hi avalveman,

Thank you for pointing out that mistake with the atmospheric pressure.

How did you obtain the environmental factor F1 of 1.0? Did you mean 0.01 by any chance?

Thanks!

Still an intern

 

[The Obturator]

Hi. I should have made that point clearer. For F1 a conservative value of 0.045 was used (It's just the way the programme I used had it 1.0 as the default). I checked manually and the sizing comes up as I stated earlier. Using 0.01 of course gives a smaller result (In any case a 1/2 x 1 or 3/4 x 1 is as small for PRV as you can get).

Per ISO, only the term Safety Valve is used regardless of application or design.

 

[ipocoyo]

Hi avalveman,

Thanks for the reply.

If i use a conservative value of 0.045, would I still be required to calculate the required relief rate?

Since 0.045>0.01 equation 12 should be used. However, i obtain the same scenario of a non-real result.
Is it recommended to use equation 13 instead?

Thanks!

Still an intern

 

[The Obturator]

From what I understand at this point, I don't think it's not going to make any difference in this case since the fluid is air. We are already using worst case F1 = 0.045. You can by all means use equation 13 to determine the required (minimum) flow rate for the PRV, then simply use the formulas in API-520 to calculate the minimum nozzle area of PRV required.
From the data I used earlier, my calculations result in a required flow rate of 308 Kg/hr air.
A typical 1/2 or 3/4" inlet PRV will flow some 715 Kg/hr at those conditions (rated flow) dependant on the valve manufacturer/model selected.

Per ISO, only the term Safety Valve is used regardless of application or design.

 

[MortenA]

You could go for a Bursting disc. It Will fully relieve the Vessel (no implication When the vessel gers to hot) and Will require less maintenance and its Also cheaper?

 

[Ehzin]

If you've determined that a fire, while unlikely, is possible then I don't think UG-140 would be appropriate. Regardless, you mention that local standards require a relief device so that's fairly straightforward, unsure if this is legal jurisdiction or company standards.

In regards to meeting any code requirements refer to the code it was built to, it appears it references API-520/521 based on your responses. In regards to actually protecting your vessel find options that could purge the vessel before you have significant and localized heat input that could lead to weakening of the vessel wall. It appears there's papers out there on the subject such as one titled "On the Adequacy of API 521 Relief-Valve Sizing Method for Gas-Filled Pressure Vessels Exposed to Fire" by Andreasen/Borroni/Etc... I don't have the whole paper before me but it indicates the logical conclusion that localized heat input into the vessel wall can lead to premature failure before pressure ever builds to lift a relief device. The easiest way to visualize this is for a large system protected by a single relief device where the fire is only occurring on a localized area. While the heat input tries to achieve equilibrium with the vapor on the other side it leads to very small increases in pressure while having a notable increase on vessel wall temperature. If the vapor can't dissipate this heat before it increases pressure enough to open the relief device then it could lead to a rupture.

The above doesn't really answer your question. It actually makes it more convoluted but you need to determine minimum code requirements that need to be met and then find adequate means to protect your facilities and personnel in the event of a fire that go beyond simple code requirements as needed.

Thanks,
Ehzin

 

[ipocoyo]

Hi All,

I just realised that in the specifications of this filter, it states that is complies with 2014/68/EU pressure vessels. How is this possible since in the EU PED is states that "for fluids of group 2(non hazardous), with a volume greater than 1 L and a product of PS and V is greater than 50barL, or with a pressure PS greater than 1000bar a PSV will be required.

How did this filter conform to the EU PED if a PSV was not installed?? It does not even have any connections for a PSV to be installed.

Specifications for the filter can be found here.

https://www.google.com/url?sa=t&sou...FjABegQIBBAB&usg=AOvVaw0qQNWEysZhpAvclTBWecl0

Thanks!

 

[Caloooomi]

Hi All,

I just realised that in the specifications of this filter, it states that is complies with 2014/68/EU pressure vessels. How is this possible since in the EU PED is states that "for fluids of group 2(non hazardous), with a volume greater than 1 L and a product of PS and V is greater than 50barL, or with a pressure PS greater than 1000bar a PSV will be required.

How did this filter conform to the EU PED if a PSV was not installed?? It does not even have any connections for a PSV to be installed.

Specifications for the filter can be found here.

https://www.google.com/url?sa=t&source=web&...

Thanks!

Are you suuuuure?

Recheck 2.10 of Annex I of PED and the manufacturers IOM. I suspect it will say somewhere that safety accessories are to be installed by others.

 

[don1980]

There are two steps in defining the "rules" for a proper relief design. One is to define what is needed for compliance (governmental and code compliance). The other part is to define what is needed for the safety of the specific application. You may think those are the same, but that's no always the case, so you need to consciously think about both when designing relief protection. For governmental and code compliance, that's obviously dependent on the geographic location, and the code(s) acceptable in that location. If you're in an ASME jurisdiction, then those requirements are prescriptive - if it's a pressure vessel, then it needs a relief device (or documentation per paragraph UG140 if you're applying "protection by system design"). In PED jurisdictions (most European countries), then the requirements are less prescriptive. As cited by Calooomi in the post above, the high-level requirement is stated in Annex I, paragraph 2.10 (copied below). Unlike ASME, PED doesn't prescriptively mandate a relief device. Instead, it requires one in cases where "...under reasonable foreseeable conditions, the allowable limits could be exceeded...". And when relief device are used, you have a broader range of optional devices, as defined by the PED term "pressure accessory". That includes PRDs that we all recognize as relief devices, but it can also includes "limiting devices" (i.e.instrumentation) used to correct the pressure excursion or shutdown the equipment.

So, it's possible the manufacturer of this vessel didn't foresee a need for a PRD, due to absence of an credible overpressure scenario (and assumed it would be installed in a PED jurisdiction). If so, that explains why they didn't provide a relief nozzle on the vessel. Or, possibly they expected that the owner would protect from overpressure by use of instrumentation (and assumed it would be installed in a PED jurisdiction).

The bottom line is that you need a relief device if you're in an ASME jurisdiction, and you may or may not need one if you're in a PED jurisdiction.