Pressure Drop to Inlet of PSV 1

[RJB32482]

Guys,
I am checking a pressure drop calculation for the inlet of a PSV for a reactor jacket system. The heat transfer fluid exits the reactor in 5 places (from the top and bottom of the reactor)and all go into one header. The header then flows vertically to the PSV. Where would I take the beginning point to perform the PSV calculation? The bottom exit of the jacket fluid? The entrance of the fluid to the common header? The entrance of the fluid to the top of the header? Heres a rough sketch (not to scale)


PSV
____________________
____________________
Rx____________________ header (one pipe)
____________________
____________________

 

[StoneCold]

RJB32482
Can you tell us two more things? What is the fluid in the jacket? What type of jacket is it? (dimpled, half pipe, open anulus style)

Thanks
StoneCold

 

[RJB32482]

Jacket fluid is a 30% propylene glycol solution and the jacket is half-pipe.

 

[Montemayor]

RJB32482:

I presume you are refering to API RP 521, Section 5.4.1.2 which deals with the 3% maximum non-recoverable pressure loss in PRV inlet piping. If so, then I also assume you are trying to protect the half-pipe coil welded on the external surface of your reactor - and not the reactor per se.

If that's the case, then I believe you are confronted with locating the PRV directly on the actual coil itself in order to stay within the 3% pressure drop loss because you have to start the pressure drop calculation from the half-coil pipe itself (since that is what you are protecting). The 3% loss is referred to the PRV's set pressure - which normally would be the MAWP of the half pipe coil. Since the pipe coil is no longer a true "pipe", the MAWP of the welded coil has to be rated and identified for setting the PRV. The MAWP is logically going to be far less than that of the original, circular cross-section pipe.

 

[RJB32482]

Thanks for the good help Montemayor,
The reactor is already in service and I do not think that the PSV can be moved. The piping is around 2'6"long going to the header and 1'0" from the top of the common header to the PSV itself. Will performing a pressure drop calculation for each coil itself (I'm assuming that each outlet has its own set of coils) be sufficient? The pressure drop calculation in the file right now only accounts for the 1'0" itself.

Thanks

 

[Latexman]

Mr. Montemayor,

I am very familiar with RJB32482's problem having faced it several times. In a fire scenario a tank-like vessel usually provides entrainment disengagement and very low pressure drop as it's contents boils off, even when the vessel liquid level is near the very bottom. However, a half-pipe jacket is just the opposite. Being "pipe-ular" it provides very little disengagement and the pressure drop is quite high since there is likely to be two phase flow. I'm familiar with vessels that have hundreds of feet of half-pipe in a zone, thus making the pressure drop even higher. And, the dP is quite different near the end of the liquid (nearly all boiled off) with maximum travel compared to when the PSV first pops open. I don't think I have ever seen this problem addressed to my satisfaction and I'd love to hear what others have done. The past practice I am familiar with when we pushed it up the relief device technical ladder was that it was basically pipe and we don't protect pipe from fire scenarios. While that was true, it was still legally a pressure vessel and the mandated solution of vapor only venting with no accounting for pressure drop in the half-pipe was tough to swallow.

Good luck,
Latexman

 

[Montemayor]

RJB32482 & Latexman:

One of the real great things about this forum is the opportunity to share and discuss similar experiences and discover short and accurate routes to a specific resolution of a problem based on another’s experience.

I also have had similar – if not exact – experiences with kettle reactors equipped with half-pipe cooling jackets. They were a pain and a half to deal with. The heat transfer obtained was miserable and the pressure drop through the coil was even worse. When confronted with the fire case as the PRV design basis, I discovered the same thing you guys are reporting: I couldn’t get the hot water (or steam) out through the PRV due to the tremendous pressure drop (& subsequent accumulated pressure) just getting to the relief device. What I did was I welded a “weldolet” coupling on the actual pipe coil and directly installed a rupture disc at that point. That was some years back. Today, I would go with a buckling pin device – which is cheaper, more accurate and better in other ways. If you are cooling with water and heating with steam inside the coils, the buckling pin will operate just fine since you can route the fluid out of the area or through the roof.

On new reactors, I started specifying a 100% cylindrical carbon steel jacket on the shell. These yield a better heat transfer rate and are much easier to maintain and operate. If you are interested, go to thread391-71471 and read the interesting discussion we had on the half-pipe jacket.

I hope this experience helps you out.