PSV inlet pressure drop from Control Valve with small diameter piping 4

[chemter]

Trying to perform an inlet pressure drop for a steam relief valve (on a utility header). The system is a 3” pressure reducing control valve, which expands to a 6” pipe (for about 1 – 2 ft) which then expands to a 12” steam line. A PSV is located off this 12” steam line via an 8” tee-branch (8” PSV inlet). All protected equipment is located downstream of the PSV. The problem is, when performing the inlet pressure drop, there is a short length (1 – 2 ft) of 6” pipe located just downstream of the 3” control valve. It would be a major piping change to try and expand the pipe immediately downstream of the control valve, however, I don’t see how you can ignore this section of 6” pipe. I was hoping to find some documentation in API that would allow you to take credit for a pressure recovery in the 12” line and use that point as the start of your inlet pressure drops (similar to a pipe entering a “vessel”), but I haven’t seen anything like this. Curious how other engineers approach this case. Make the recommendation to replace the 6” pipe back to the control valve? Thanks.

 

[don1980]

How high are the inlet losses, based on your calculation which includes all the piping from the control valve to the PSV? I assume they're higher than 3%, otherwise you wouldn't be asking this question. But, how much higher are they, above 3%.

I think most engineers would regard the inlet piping as just the 8" segment between the 12" header and the PSV inlet flange. I not saying that's technically correct. But, there is an argument for doing that when the supply line has a high capacitance.

If you had, say a 2" line and a 2" PSV located 50 ft from the let-down valve, then there's a clear argument for saying the inlet line is everything from the let-down valve to the PSV. If that inlet line was 12", rather than 2", then I'd disregard everything except for piping between the header and the PSV. That's because the system has enough capacitance that it can effectively be regarded as a vessel.

In your case, with an 8T10 PSV (I'm assuming that's what you have) then it's harder to make the argument that the 12" line sufficient capacitance to effectively act as a vessel, but that argument can be made.

The bottom line is that there would have to be a very strong case for change, before I'd consider replacing that 6" segment of pipe.

 

[Chance17]

I would not view the 6" pipe as part of the safety valve inlet.
If you can not live with that view, you could replace the conventional style safety valve with a modulating pilot.
The safety valve would not "pop" fully open so that inlet pipe PDrop is evaluated at "required rate" and not "actual valve capacity".

 

[chemter]

don1980 and chance17,
Thanks for the response. I tend to agree with you both. If I use any portion of the 6" line, I'd be well over 3%.
I've tried to come up with a calculation for cases like this, but it's fairly difficult. Generally speaking (for installed systems like similar to this), I could calculate the velocity of the relieving fluid in the larger diameter pipe compared to the smaller diameter pipe and possibly show a dramatic decrease in velocity. The problem is, how much decrease is enough? The next step (I think) would then be to show you have a large capacity available, as don1980 suggested, but then another question is how much capacity does one need? In the past, I've tried to estimate the volume and time it would take to deplete a given system volume in an effort to see if rapid opening and closing would occur at the relief device, but again, what amount of time is considered acceptable. I've not been able to come up with a good number for any of my questions.
Thanks again for your responses.

 

[Latexman]

I don't think Code says how rapid it can be, so it's up to Engineering judgement. For me it's easy to see that 1 second is too rapid and 1 hour is very comfortably long enough. In my mind, the grey area is roughly around the 1 minute frequency. If the event lasted two hours before it could be safed off, the PSV would have to cycle 120 times. A good PSV should survive that.

Good luck,
Latexman

 

[TD2K]

The 3% inlet line loss criteria is driven by the desire to avoid getting into a relief valve chatter situation (and also any inlet line loss reduces the capacity through the PSV if you have choked flow).

Do you overpressure any component in the 6" section of piping? If not, I'd stop there. Since the PSV is located upstream of any downstream equipment (and assuming the PSV setpoint is equal to or less than any MAWPs), I'd consider your system protected.