API 520 Multiple PSV Clarification 1

[sandstars]

I have 2 PSVs that protect the same vessel. One is 2J3, the other is 6R8. The vessel in question has a MAWP of 75 psig and both valves are set at 75 psig. We are running up against the relief flow rate (calculated using API 520 and 521 at 116% overpressure) and the lift pressure of the valves so my problem is twofold.

Without going in to a long, detailed discussion of the system setup, trust me when I tell you this: Due to a modulating valve, increasing the pressure set point on the larger valve will give us additional room for pressure whereas increasing the pressure on the small PSV gets us nothing from a pressure standpoint. We are currently running around 60 psig but plan to increase flow rate, which will subsequently increase the pressure of the system.

So, while parusing API 520, I ran across this little tidbit (paraphrased): "In multiple relief systems, the first device must be set at the MAWP. However, the second device may be set 5% higher." To kill two birds with one stone, I would like to increase the spring set point on the larger valve. API doesn't mention anything about different sized valves that I could find and I did not bother to dig through the ASME code yet that API references for this section. The feedback I've gotten internal to my company is that it's not the "conservative" approach and increasing the smaller one by 5% would be preferable even though that gets me nothing from a pressure standpoint and I would still have to add PSVs to the system anyway.

However, to their point, what's to stop someone from sizing a really tiny relief, calling that the primary, then sizing the larger one at 5% higher? Is this still acceptable per code? Does API assume both valves are sized equally for the 5% case to be true?

Thanks!

 

[TD2K]

To answer your last question, nothing. Once you have multiple PSVs, code allows you on pressure vessels to go to 116% of MAWP. So, if I was just slightly capacity limited, I could add a 3/4" by 1" D orifice new PSV in parallel to an existing T orifice to get the additional capacity I need even though virtually all of the capacity increase is due to the T orifice being calculated now at 116% accumulation versus 110% and very little from the "new" PSV.

Which one to set lower? Typically if I'm analyzing a system there is the sizing case for your PSV and several other cases that 'could' result in a relief, some of them much smaller than the sizing case. Setting the smaller PSV at MAWP and the second one at 105% of MAWP allows the smaller PSV to handle those other cases if they occur. Oversizing PSVs will result in chatter if there is insufficient gas flow to keep the PSV open which is exactly the case you can have with those other possible relief cases. The other solution is to install several smaller valves with staggered setpoints but that is not typical in my experience.

"Without going in to a long, detailed discussion of the system setup, trust me when I tell you this: Due to a modulating valve, increasing the pressure set point on the larger valve will give us additional room for pressure whereas increasing the pressure on the small PSV gets us nothing from a pressure standpoint." I'm not following you on this but you did say trust me.

 

[zdas04]

I'd like to add that API 520/521 says what must be accomplished and mostly leaves it to engineering judgment to accomplish that goal. What must be accomplished is preventing the vessel from exceeding MAWP plus allowable accumulation. How you do that is your business. I've set PSV on 600 psig MAWP vessels at 250 psig, calculated the flow at that pressure and sized the orifice to stay below 275 psig. I had sound engineering reasons for doing that and was absolutely confident that I had complied with both the letter and the intent of API 520.

As TD2K says, there are any number of reasons for setting the smaller valve at MAWP and the larger one at 105% of MAWP. The most important one is that most transients that cause over pressure are smaller than the maximum credible scenario and you will often get away with only the smaller one opening. When the transient that the system was designed for happens, the combined flow keeps everything safe. The guy that said that setting the larger one at the lower pressure is "more conservative" is smoking something that is not legal anywhere in the U.S. Doing that does not add ANYTHING to process safety and actually increases the risk of valve chatter.

One aside if I may, right now your two valves are set at the same pressure. Really they are not. Small differences in spring fatigue, calibration events, etc. make is nearly impossible for two PSV to actually lift at the same time on a vessel. What that implies is that whichever one is less likely to reseat will be the one that lifts (a variant of Murphy's Law). Parallel PSV set at "the same" value is the worst of all possible worlds and when I find it I generally raise enough hell to get some rational thought put into sizing and set point.

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.

 

[sandstars]

Thanks to everyone for the feedback! You certainly brought to light a few things I hadn't thought of (sometimes, I get so focused on the pressure and flow ratings, I forget about things like chatter).

So maybe explaining the system would help. Then again, maybe it'll just be more confusing. Either way, here's my best shot:

In a blocked flow case (controlling case), the pump pushes product through a heat exchanger that vaporizes the product, past the larger PSV, through a forward pressure control valve, and in to the vessel in question (which has the smaller PSV attached). The control valve (Fail Open) is currently running almost always at 100%. As we've continually ramped up rates, the pressure drop from the pump to the heater has increased significantly and taken away the pressure needed where the PSVs are.

What I'm driving at is to limit the pump's VFD to some predetermined pressure and flow set points so that we don't exceed either per PSV calcs but want to maximize this so we're not limited on production later down the road.

Hope that clarifies any confusion.

 

[sandstars]

Not sure how to edit posts but forgot to mention we're upgrading the pump's impeller to compensate for the pressure losses at the higher flow rates.

 

[SNORGY]

I might be speaking a bit out of turn and out of my element here.

It appears to me that if it is the intent to take advantage of the wording in the API guideline so that you preferentially open the smaller PSV and let the larger one be set higher than MAWP, then,in essence, this is suggesting that it is the system dynamics (as opposed to steady state relieving) that drive the rationale for overpressure protection. If so, then what is implied (to me) is that you actually only need a "BURP" valve as opposed to a "prolonged puke" valve.

I would suggest that you might run HYSYS in dynamic mode to see if the system behaves the way you want it to in a relief scenario. My initial suspicion is that if the governing case is pump blocked flow, it implies a dynamic event of longer duration than can be handled by a "BURP" valve.

If that is true, have you actually gained enough?