Design pressure for Vent scrubber

[mucour]

Guys,

I was reviewing a set of P & Id (PEFS) and I discovered that a design pressure of 100 psig was recommended for the vent scrubber which vents to the atmosphere.

But I found out that a relief valve which has a set pressure of 1000 psig reliefs to this vent scrubber. It is a relief valve which was originally relieving to the sea but has now been re-designed to relief to the vent scrubber.

Other relief valves and equipment connected to this vent scrubber are 1) relief valves set at 230 psig (2) a liquid sump pump with max. discharge pressure of 1800psig

Please note that this vessel is almost at atmospheric pressure because it is opened to the atmosphere through a vent tip. So there will not be a need for selecting balanced-type relief valves.

Could you explain to me why the vent scrubber was designed to be of design pressure of 100 psig? Why not 40 or 60 psig as an example.


which was previously relieving to the sea now is
Could somebody explain to me the reason why

 

[TD2K]

I can't answer that with the information you've given. Design pressures can be set a couple of ways. The minimum DP is from the process requirements, what was the maximum pressure the original designer envisioned this vessel having to operate at (doesn't sound like much above atmospheric) or what was the maximum pressure necessary to handle an upset case, ie. high vent rates to atmosphere.

Design pressures can also be based on the vessel's capability. Depending on the vessel's diameter, the typical plate thickness used for the shell and heads may be more than adequate to support a 100 psig design pressure and they may simply have used that. Some companies also dictate the minimum design pressure though 100 psig seems a little high for that.

 

[mucour]

TD2K,

As a process designer, I would not want to go into the plate thickness used for the vessel body.

I want to limit myself to the process conditions and as you pointed out the worst upset conditions.

I have looked at all the process lines feeding the vent scrubber and beside the relief valves that are designed for atmospheric reliefs (although routed through the vent scrubber to capture entrained liquid), the other high pressure line is the line from the sump pumps with actual operating discharge pressure of 500 psig.

So I am worried of the 100psig design pressure recommended for the vesel when in actual sense it is really not a pressurised vessel. The vessel is opened to the atmosphere.

Could you think of where a high pressure could result from as an upset-condition? Because I cannot think of any.

The PEFS has been HAZOPed and there is no problem of a blocked vent-tip envisaged that could create a pressurised scenario.

So why can't the design pressure of the vessel be 40 of 60 psig. Let assume that the client does not require a minimum of 100 psig for min. des. pressure.

Could you make other suggestions or am I missing something?

Thanks

 

[TD2K]

What's the benefit of dropping the design pressure at this point? Are you going to save money on the vessel fabrication? If you are early in the project, this is the time to do it. If you are farther along, you might need to consider the cost to make this change on the impacted documentation (even if you are working on just P&ID you may have to update data sheets, drawings, line lists, etc).

You mentioned that there is a relief valve discharging to this vessel set at 1000 psig. If it's in vapor service, then the maximum allowable backpressure would be 10% or 100 psig. May have nothing to do with how this scrubber's design pressure was set but it's a coincidence.

Alternatively, if you've looked at all the possible relieving cases and don't see a need for the higher design pressure then it can be lowered. Have you talked to the person who set this 100 psig design pressure as to what reasons they had?

 

[mucour]

Thanks TD2K.

I have confirmed from the design contractor, and I was informed that they were guided by 2 reasons.

1) The vent tip can experience air ingress and this could lead to explosion in the vessel because it will contain hydrocarbon. As a result, from their calculation this will be within the 100psig presure developed by the explosion.

2) There was an original design similar to the one in question and they used 100psig design pressure for the vent scrubber.

TD2k or any guys out there, how can we determine the pressure resulting from an explosion in a vessel?

Thanks

 

[TD2K]

Sorry, I have no experience in that. Might try a company like Fike who handle explosion venting, they would have more information.

 

[galil]

I,m studing about scrubber used in dap unit in petrochemical(IRAN)
and i have a problem with it,s packing because they cracked soon.
we use a saddle packing.
i think it,s reason is the pressure drop in the scrubber.
please help me
gol180@yahoo.com

 

[mucour]

Galil, what is a dap unit?

 

[krossview]

Mucour,

I would like to elaborate on a good point made by TD2K , when he mentions that the 100 Psig design pressure could be based on the 10% maximum allowable back pressure in vapour service. For a conventional relief valve this is typical 10% of the relief valve's set pressure, but can be 30% if a balansed seal (bellows) type relief valve is installed. (See API 520).

You may will find your answer when you check the pressure drop in the flare system and calculate what the build-up back pressure in the vent scrubber will be. Be sure to use the actual piping configurations and include all fittings, valves, flow restriction, sudden expansion/enlargements as this will make quite an impact on the overall pressure drop. BTW, this would be a good check to ensure that the existing design can still handle the current/new process conditions.

There is also a good chance you may find more problems as you go through the relief system and start checking relief valve capacities, relief scenarios and pressure drops in the flare system. For instance, you mentioned that relief valves, with a set pressure close to atmospheric, discharge into the vent scrubber. Can these valves still relieve into the system, if the relief valve set at 1000 Psig pops and causes the back pressure in the vent scrubber to exceed the maximum allowable back pressure on the relief valves? These relief valves would then be blocked and prevented from relieving the pressure. This would only apply if the same upset causes both valves to pop. Also, are there any blowdown valves, who discharge into this vessel? What is their relief capacity? I have seen a case of grossly oversized blowdown valves.

Coincidently, I am working on a project that involves explosion prevention in a flare system using a detonation arrestor and explosion isolation valves. We have asked the same question if the explosion pressure could be predicted and sofar did not receive a conclusive answer. Only testing based on the existing system would provide the actual pressure. The reason is that the explosion pressure depends on several variables, like gas composition, intial temperature and pressure, distance from ignition source, piping configuration, etc, etc. Vendors like Fike/Fenwal use a rule of thumb of 10x the intial pressure to estimate deflagration explosion pressures and locate their explosion isolation devices close to the ignition source. In case of detonations explosions much higher values are experienced. For a very good description on this subject read this document

http://www.enardo.com/fatech.pdf

on the Enardo website.

Mucour, based on the rule-of-thumb method the design pressure of the vent scrubber should be 147 Psig (10x Atmospheric pressure) and not 100 Psig , if the explosion pressure was a major criteria in the design of the vessel.

Sounds like flame flash back was considered in the initial design and the system may have a flame arrestor to quench and stop the flame front from traveling back to the process. Check, if this is ULA/US Coast Guard certified deflagration or detonation arrestor and not an end-of-line arrestor. The Enardo document notes the distinct differences between arrestors.

Finally, I have to agree with TD2K that I do not see any reason to lower to vessel's design pressure unless you want to save some money. Fortunately, your are in a much better position with an overdesigned vessel :-> then an underdesigned vessel!

Hopes this helps.

Krossview

 

[mucour]

Dear Krossview,

Firstly, your use of those symbols are quite amazing as they also serve as self explanatory indicators.

Let me discuss on some particular points you made.

For various pressure safety valves (PSV) discharging to a common vent scrubber (similar to Flare Liquid knock-out veesel) it is an acceptable design practice as far as the outlet of the vessel is operating at atmospheric pressure (or close-by), However, what is important is that the headers where the PSVs are connected to must be segregated (API-520) based on there different set pressures to avoid back pressure greater than 10% for conventional PSV (pilot) and 30% for balanced type PSVs.

If this PSVs segregation is ensured it is possibe to still tie them to a big veesel like the vent scrubber which operates at atmospheric presure. What this means is that the different headers are connected to different inlet nozzles of the vent scrubber to avoid back-pressure if they are connected on the same header before tie-in into the vent scrubber.

The design I am reviewing ensured the separate connections of the PSV outlets according to there set pressures.

The only thing of concern, which you also recognised, is the design pressure selected for the vent scrubber. The question is what is the basis?

It is quite illuminating that in a similar exercise you also experienced the difficulty of calculating explosing pressure in a vessel resulting from flammable flash-back into the vessel.

Could you give me a reliable reference to your recommended "rule of thumb" of 10x the atmospheric pressure. I will appreciate if you can source it out.

TD2K point of 10% of the PSV set presure (1000psig) was just coincidental. TD2K also recognised this he was just wondering whether that was the basis for the design pressure selection by the contractor.

The flame arrestor has not been selected yet. The client is evaluating the use of a vent-tip as aagainst a flame-arrestor. The vent tip is meant to ensure good air/gas mixing and dispersion to level that is below the Lower Explosive Limit (LEL).

The vessel has not been fabricated yet, that was the reason this question was raised in the first instance. I quite agree that there would not have been any need for re-fabricating if it is an existing vessel.

 

[krossview]

Mucour,

Good to see someone using due diligence in his work.
A couple of comments I wanted to make.
Segarating PSV discharge lines is very effective in reducing back pressure on the PSV, but is not a garantee in particular if you can have simultaneous releases from two or more PSV's. I assume the vent scrubber is the colletion point of all those PSV lines with one common line going to the flare or vent stack. API Section 5.4.1.3.1 states that build-up back pressure of a relieving PSV should not reduce the relief capacity of another PSV. The following example may illustrate this.
1. HP PSV set at 1000 Psig pop and reliefs vapours.
2. Relief flow build-up back pressure in vent scrubber.
3. Calculated back pressure is 3 Psig.
(This is not uncommon)
4. Maximum allowable back pressure of LP PSV set at 10 PSIG is 1.0 PSIG (10 Psig * 10% of conventional PSV)
5. Back pressure in vent scrubber exceeds 1.0 PSIG MABP and reduces relief capacity of LP PSV.

Regarding the rule-of-thumb of estimating the explosion pressure I need to make the following comment. API 521 section 3.12 states the simple "rule-of-thumb" should not be used as is can lead to improper designs". Predicting explosion pressures depends on many variables and the engineer designing such systems better understand the impact of the explosion risks (See link to Enardo document) and quickly explosion pressure can build-up in a contained system. The reason why I made the distinction between deflagration and detonation explosion pressures. For example, an end-of-line flame arrestors installed about 100' away from the ignition source will likely fail and not protect the system as the arrestor is not designed and tested to withstand the developed explosion pressure. A detonation arrestor, designed, tested and certified, would be required here. The above comments are for fellow engineers reading this thread.

Now, regarding the mentioned "rule-of-thumb" of 10x the initial pressure only applies to explosion pressures in the deflagration region . The 10x "rule-of-thumb" factor came from discussions with Fike/Fenwal after we had questions on the anticipated explosion pressure in our current design of the flare system. However, this did not tells us anything about what the actual deflagration/detonation explosion pressure could be.

Mucour, I advise you contact Fike/Fenwal directly to confirm this and have them review your system.

To show due diligence and protect my engineering title I advise that everything in this e-mail should be checked [flush] carefully and assume no responsibility.

 

[krossview]

Dear Mucour,

Here is another reference for estimating the maximum deflagration pressure is giving in section 4.2.2.2 of NFPA 68, Guide for Venting of Deflagrtions, 2002 Edition.

Section 4.2.2.2 states:
"The maximum deflagration pressure, Pmax, and rate of pressure rise, dP/dt are determined by test over a range of fuel concentrations (See annex B.) The value of Pmax for most ordinary fuels is 6 to 10 times the absolute pressure at the time of ignition.

Hopes this helps

Krossview.

 

[mucour]

Dear Krossview,

I was reading the ENARDO article through out yesterday.

i must admit, honestly, that the article is very rich and educating. Now I fully understand the intricacy of the two consequences of ignition in a flammable gas system viz: deflagration and detonation.

I believe the case we are looking at bothers more on deflagration since it is a vent from a scrubber/vessel. So the biggest potential of ignition is from lightning. The chances of detonation (i.e, at or above sonic speed) is very remote since the flame is not confined and the flame can burn or expand limitlessly at the vent tip in the atmosphere.

But in the event of flame propagation (resulting from flash-back) through the pipe into the vessel/scrubber, then there can be an explosion inside the vessel.

SO I believe the max. pressure resulting from deflagration is the value given ranging from 6 - 10 times the absolute pressure in the vesel at the time of burning.

Thanks for the article, I am indebted. I recommend it to other engineers who made contributions to this thread.

Mucour