Set pressure for relief valve 2

[MechMad1]

I want to fix the set pressure for a relief valve (in a pressure vessel).
Let's say that the design pressure of the vessel is 100 barg. As far as I know, it is common practise to fix the set pressure of the relief valve to the design pressure of the equipment (100 barg).
But relief valves have a 10% of overpressure, so the pressure will reach 110 barg.
So the vessel will attain a pressure higher than its design pressure.
Is this acceptable?
Thanks.

 

[zdas04]

This really is both more complex and simpler than you think.

First the complex bit. You need to look at the system and come up with "credible scenarios". "Scenarios" are events like "shut block valve", "runaway reaction", "failure of controls", etc. and API 521 has a list of over 20 categories of scenarios. Let's say that it is possible for something like a "shut block valve" to happen. Then you have to look at what pressure could be seen in the vessel if a "possible scenario" happened. If the maximum pressure that could be developed is greater than the vessel MAWP, then the scenario moves from "possible" to "credible". For every credible scenario you need to determine the flow rate out the PSV that would prevent an overpressure event. Generally an overpressure event is defined as exceeding 110% of MAWP (but there are special cases like the Fire Case which allows 122%).

Now the simple bit. Size and set your PSV such that in the worst cast credible scenario you will not exceed the limiting overpressure. You can set a PSV at 120% of MAWP if your only credible scenario is Fire Case and AND your are absolutely certain that it will open quick enough AND you are absolutely certain that it will not lift a tiny bit higher than the set point AND you are foolish.

My default setting is MAWP. I have had specific cases where I set the PSV well below MAWP. I've had other cases where I set the PSV at 105% of MAWP (for example when I have two credible scenarios, one with very high flow requirements and one with very low flow requirements, I will use two PSV's, a very small one set at or below MAWP and a large on set 5-10% higher for the high flow requirement scenario--NOTE: never install two PSV in parallel with the "same" setpoint)

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.

 

[moltenmetal]

Mechmad1: remember that relief events are supposed to be RARE by design. You should not be challenging this relief valve during "normal"operation- if you are, you need to revisit your overall design and render it safer.

Remember also that your vessel has been hydrotested well beyond its MAWP, or pneumatically tested to at least 110% of its MAWP.

For these reasons, 10% accumulation is tolerated by the code for "normal" relief events, and 21% for fire. After a fire intense enough to lead to a relief event, it's unlikely that the vessel will simply be put back into service without a 2nd look.

David's advice in relation to selecting set pressures is, as usual, excellent.

 

[BenThayer]

the only comment i will add is that for some "runaways" such the decomposition of hydrogen peroxide, we have found we need to set the device below the MAWP in order to keep the pressure spike from exceeding the 110% during an event.

but this is effectively covered by David in the "scenario" development portion of his note above.

 

[don1980]

The safety factor for pressure vessels is high. Unless the vessel has been damaged or improperly maintained, there's little risk of it failing until the pressure exceeds 2xMAWP.

Pressure relief design rules and practices are very conservative - as they should be.
That includes the ASME accumulations limits (110% for process upsets and 121% for fire exposure). Those limits are prefectly safe, acceptable, and conservative. Just because the limits are greater than MAWP one shouldn't conclude that they're risky. Quite the opposite is true.

 

[MechMad1]

Thank you! I found this very interesting.
Problem solved!