Recip compressor skid design best practices & flare header requirements

[gwalkerb]

Typical reciprocating compressor skid design necessitates keep all piping in the package as low as possible, to decrease the magnitude of vibrational moments, and prevent the skid from shaking itself apart. My company's standard package designs follow these best practices, and they've worked well for us up until now.

However, we've been seeing more and more customers request that our on-skid flare header both has no low points / liquid traps, and exits the skid high up in the package (generally somewhere between 5 and 25 feet). This is becoming a design issue for us, as we are having difficultly reconciling these two opposing design criteria. We run into trouble sometimes trying to explain to our customers why it's a really bad idea to have 8" pipe 10 feet in the air in a recip compressor, they just don't seem to get it. Mostly this seems to be a problem with EPCs who have very little experience with compressors, and are primarily familiar with process systems only.

Does anyone have any ideas of how to best handle both of these problems? Firstly, how best to explain to people that recips really don't do well with lots of piping high up in the air. Secondly, are there any good package designs out there that meet both criteria of trap-free flare headers that exit high in the skid, while keeping process piping, and especially PSVs, low in the skid?




As an aside, does anyone know if there's a way to change your listed discipline? I selected the wrong one when I first signed up, but I can't find any way to change it.

 

[51stovebolt]

Having spent my career in compressor packaging, I can say that you face a common dilemma. I agree with your comment about EPCs and customers not grasping the concept. Sadly, some packagers do not either.

Explaining the situation to others can become complex because the fundamental concepts are vibration theory, a college course which many math and science oriented engineers still have nightmares about. As you appear to understand, good engineering practice is to keep mass to a minimum and stiffness at its maximum in order to push the resonant frequency of the piping system up above the running frequency of the compressor package. Keeping pipes low comes from the part about maximizing stiffness. Short pipe supports (in other words close to the skid) maximizes stiffness. I suppose you could come up with some sort of "show and tell" type example where you can change the mass and stiffness and show how that impacts the resonant frequency. But even that is boring to most or still to complicated. Generally, I stick with explaining the best engineering practices and leave understanding the concepts to the individual to study on their owny.

Regarding routing/supporting of the flare header piping, here are a few thoughts:
1.) All flare headers I've ever seen are isolated from the main process system. In other words, they are not a conduit for the high-pressure, pulsing flow that can couple with pipe elbows and fittings to cause pulsation-induced linear vibrations. As such, support of flare headers is not quite as difficult and critical as process piping.

2.) Secondly, support the piping from another tall and substantial structure, a cooler structure perhaps (assuming that is on skid).

3.) In the event you do not have a cooler structure on skid and are trying to connect with pipes in overhead pipe racks, could you run the header low along on the skid then exit at the skid edge, where vibration should be at a minimum? At this point perhaps you could provide or recommend a flanged bellows for vibration isolation, and support a vertical pipe run from the pipe rack, or some other structure that is not vibrating.

Good luck.
Will

 

[bicycledisciple]

gwalkerb :

I've seen the PSV from compressor discharge bottle teeing into the vent header via a separate pipe spool taken off the compressor discharge pipe. This PSV is typically at a higher elevation, with an elbow at the inlet and outlet sides and satisfies the criteria for free draining. The vent header on the other hand didn't have any slope or trap.

If your system differs and you have to succumb to placing the vent header pipe high, try I-beam type supports with pipe clamps.

The popping of the PSV and the discharge through the opening are dynamic events since the gas will expand, and perhaps in a non-ideal condition you may even have some chattering all having the potential to cause variation in discharge over time. Using API RP520 it is possible to calculate steady state reaction forces from a conventional type PSV. Best if you get this from your vendor. A study of transient forces and pipe stress on the header where the PSV tees in could be tried out. Ceasar II appears to have a way to calculate stress due to a PSV popping condition. This maybe one strategy to check bending loads and rule out the higher vent header.

Ron
-------------------------------

http://www.georgeron.com