Safety Relief Valve Chattering due to oversized valve 3

[processrj]

thread1203-186657
I have seen throughout this forum many threads dealing with PSV chattering. One of the causes for this phenomena being valve over sizing. Accordingly, to avoid it Required capacity should be greater than 25 % the Rated capacity. Does anyone know and can provide a reference to this rule? Thanks!

 

[The Obturator]

There is no rule only users experience. There are too many factors besides oversizing to cause chattering; valve trim, ring setting, pressure drop, correct valve design for application, allowable blowdown, whether fluid is 100 % liquid or multi phase etc., etc. From my years, a figure of 1/3 rated capacity was talked about, but it is no guarantee. A good valve can be a modulating pilot operated SRV, dependant on application.

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

 

[gocougs1]

There is no standard saying "thou shall not oversize PSVs", but API 521 has an excerpt saying:

sect. 5.8.2.2
"These studies verified that vapors released from PRVs through their individual stacks are safely dispersed even when
the valves were operating at only 25 % of their full capacity, which corresponds to the reseat level of the valves."

One could argue that from this, one is not following recommended practice if a PSV's sizing case is less than or equal to 25% of full capacity. Also note that a PSV's rated capacity has a 10% conservatism factor included, so this is then equal to 27.78% of available area, and I've heard PSVs requiring less than 30% of available area are generally considered oversized.

For the minimum size 3/4"x1" PSVs, they have much less force/weight flopping around, so oversizing one of these valves is not a concern. On the other hand, having a 4L6 PSV (2.853 sq. in.) where only a 2H3 (0.785 sq. in.) is required is pretty questionable... personally I would recommend reducing PSV size for this instance, but API does not address this issue and the costs may not be justified. In general it is very difficult to reduce PSV size unless clearly justified, as having an undersized PSV is much worse than an oversized PSV, and all PSVs are oversized to some degree.

Installing multiple PSVs to avoid chattering is a good practice, but not that common from my experience, probably because it is slightly cheaper. I'm actually quite curious why 2 PSVs aren't used more often instead of only 1 to protect a vessel. Seems like two smaller weight PSVs would make them easier to install, cheaper, and easier to maintain. Also for spared service, one could have 3 half size PSVs vs 2 full size PSVs.

In terms of weight, a 300# RF 4L6 weighs ~220 lb while a 2H3 weighs ~65 lb. The thought of that amount of weight chattering around makes me uncomfortable.

 

[met11]

There is rarely a single overpressure scenario that you could pick an orifice for. There will usually be multiple cases. Take the case a relief valve sized for a control valve or regulator failure. You usually size the relief valve for the full flow at the rated Cv of the failed device. What if the valve only fails partially? Or there's an obstruction in the line, limiting the flow to the vessel? Now your relief valve is inherently oversized.

Take another relief valve with multiple cases: fire, failed control valve, overfilling vessel via PD pump. It's almost guaranteed to be oversized for the non-fire cases, which are probably the more likely cases anyway.

Short story long, relief valves will always be oversized.

 

[moltenmetal]

This is frankly a very silly topic. Met11 beat me to it: a relief valve will ALWAYS be oversized for everything except the worst-case fire or other worst-case sizing case. All real cases including (with hope to your favourite deity) any REAL fire, will require the valve to relieve at rates vastly lower than that!

If relief events are frequent enough that post-relief inspection and maintenance of the relief valve is inadequate to keep the valve available with high reliability, then the system is improperly designed and requires some other means of overpressure protection. Yes, you absolutely can get into trouble when you start using safety relief valves as process pressure controls- that's not what they're intended for!

Large valves may require some study of resonance behavior, with the extra prudence being justified by the magnitude of the risk the valve is protecting against. The 3% inlet rule must also be given due consideration to avoid making things worse- but otherwise, people should just plain accept that relief valves will be oversized. The only real downside to over-sizing a relief valve is the impact on the size of any downstream treatment system (flare, scrubber etc.) the cost of the valve and header system, necessary to handle relief flows that may be intense, intermittent and likely brief.

 

[don1980]

Hear, hear! I'm lifting a pint to ya, liquid metal man!

 

[georgeverghese]

PSV chattering itself is not a safety concern, it is what it does to the vessel nozzles and relief piping. Large relief valves can cause very high reaction forces on nozzles and on piping supports during lift.

After some years of operation, relief piping can break free of their supports due to general corrosion / galvanic corrosion during a relief event with a single large size RV.

Installing smaller RVs' with lower set pressure in parallel with a larger one with a higher set pressure may help to avoid a loss of containment due to coincident pipe support failure, if a single modulating type pilot operated RV is not acceptable. Talk to your piping engineers to see where the breakpoint (in terms of the upper limit on RV size and set pressure/ flowing momentum at rated flow) may be in terms of tolerable reaction forces on relief piping in case at least one major piping support fails.

In most plants, piping supports are often in inaccessible locations, and it is unreasonable to expect 100% reliability on these, especially on high velocity blowdown / relief piping runs, over a 20year life cycle. Else you need sufficient redundancy in the spacing and design of these supports to account for partial failure somewhere along the run.