Check Valve In Relief Valve Discharge Line 2

[barnesed]

A survey of an existing steam relief valve installation shows the 4" valve outlet connected to a 4" discharge line that connected to an 8" vent header to the roof. (There are other lines tied into the 8" vent header.) There is a 4" check valve in the 4" relief discharge line where it connects to the 8" vent header. Aside from the backpressure questions, is a check valve in a relief discharge line allowed? I would have thought that there could be a situation where the check valve fails to open.

 

[BenThayer]

check valves are not allowed in installations i approve/sign off on.

 

[psafety]

Agree with BenThayer.

 

[zdas04]

Me too. Just too many ways that they could go terribly wrong.

David

 

[KVVS]

yes dear........i am fully agreed with zdas04.

 

[chemebabak]

API 520, Part II 5th Ed. Section 6.3 Isolation Valve Requirements does not allow check valves. It only allows valves that can be locked open.

Aside from the code requirements prohibiting check valves please know that check valves have been known to fail in the closed position. Imagine that happening in a relieving situation.

 

[maddocks]

Despite the many concerns listed above, how would you prevent the reverse migration of sour vapours from a flare system back into a sweet package without a check valve?

 

[zdas04]

You don't comingle sweet and sour streams in your flare header.

David

 

[maddocks]

ok, this is just not a practical answer. Not every facility has isolated flare systems - and if they are, they're often isolated from a LP or HP flare. If you're lucky, you might get an acid gas flare, but that's often used in lieu of a LP flare. Bottom line, segregation just isn't always possible in the real world.

 

[mbt22]

I wonder if continuous inert purging of a header and careful piping layout might let you mix sour and sweet in header?

I'm with David on this - I can think of no situation where a non-return valve would be acceptable in a relief header.

Matt

 

[maddocks]

ok, don't get me wrong. I'm not advocating the use of check valves in PSV discharges. I was trying to make a point that we don't always have a choice over how flare systems are segregated when working in existing plants and was looking for some input on how this is dealt with.

The only time I've used a check valve in a flare system is on gas engine starter exhaust that ties into the flare header. You're correct - no other use of check valves would be appropriate in a code based relief system.

 

[zdas04]

Let's look at your gas engine starter. Say the PSV is set at 125 psig. Now let's say that you are observing a problem developing in the plant that might be solved by putting another compressor in service. While staging to start the compressor a big PSV lifts and raises the flare header pressure to 200 psig at the point where the engine starter enters. Now lets say that the starter system exhausts into the flare header between the PSV and the check valve (I know it sounds dumb by itself, but I've seen the near-religous fervor to reduce "greenhouse" gases lead people to interesting decisions). When you turn the starter on, it will come up against the check valve held closed by the big PSV. A few seconds later, the starter PSV will lift with no place to go. A few seconds later you have a 60% overpressure on the starter system and the check valve starts to open.

That scenario is credible and doesn't require unrelated events to happen concurrently.

Flare-header analysis is very complex and needs to consider a really wide range of possible scenarios. When a new stream is tied into it, the entire analysis is required again. In a plant where management has decided to bring in a fundamentally different stream, one of the costs of bringing that stream into the plant is an analysis of the flare header and the associated costs of modification. If you have a sweet-gas plant and the decision is made to add an amine system to handle sour gas, you need to build a new flare header as part of the cost of conversion.

David

 

[koba78]

Interconnecting piping between protected equipment and its pressure relief device shall be free of block valves (unless by locks or car sealing), control valves, and equipment that may fail or stop in a closed in a restrictive position, suchas check valves, expanders, etc. Use of maintenance blinds for equipment isolation upstream of a relief device is acceptable.

 

[Duwe6]

For Sweet vs. Sour isolation, if necessary, put a Sour-rated rupture disk right at the inlet to the Flare header. Now your Sweet PRV and discharge piping will not see Sour service.

End of the day, its cheaper to rate everything for Sour service, or run two Flare headers.