Suction pressure of compressors 1

[engr2GW]

Hi all,

1. the suction pressure of a reciprocating compressor is not set right? it pretty much is what it is (adjusts) at any given time based on teh working pressure of the pipe or vessel it is pulling from right?

2. There are three pressure vessels (separators), one is operating at 100psig, the second at 30psig, and the third at 5psig. the gas outlets are feeding to a common pipe to run to a 2 or 3 stage recip. compressor suction which is expected to discharge at pipeline pressure of 300. Is that set up going to work? I saw this set up and thought the lower pressure vessels may not be able to push into the common line heading for the compressor suction, instead there'll be back pressure on the lower pressure vessel leading to the gas going to a different route or leaving through the liquid dump line.

Please advice, I don't know if the compressor can adjust to make this possible.





As much as possible, do it right the first time...

 

[MikeHalloran]

The setup as you describe it sounds a little funky. ... BUT a lot depends on what sort of valves exist between the separators and the manifold.

Has this setup ever run? If so, you may have missed a detail in your observation.

Is there a schematic that describes how the system is supposed to work? E.g. a P&ID?

Absent any such documentation, folks here might be able to make sense of some photographs.



Mike Halloran
Pembroke Pines, FL, USA

 

[zdas04]

The suction valves on a recip are just spring-loaded check valves. The have to have a dP to operate. If the suction pressure gets really low, then the piston will reach the BDC point without them ever opening. On the upstroke (on the head end cylinder) you are going to recompress the gas that was in the clearance when the discharge closed. The discharge will not reopen and you'll just generate windage losses until the machine melts.

Compressor manufacturers use +/-5% of design pressure as the appropriate suction pressure. Outside of that range asks for problems. If suction drops and the suction valve opens late, then the actual compression ratios can go up dramatically (until the heat of compression takes the machine down). If the suction increases then the valves open to early and put too much gas into the cylinder, which often takes the machine down on low rpm as the engine tries to do the required extra work).

The scenario that you proposed will not work. You don't even have to look at the compressor, the piping will not let it work. If there are check valves then the 100 psig stream will keep the other two shut. If there are no check valves then the 100 psig stream will flow into the compressor and into the 30 psig vessel and into the 5 psig vessel until everything equalizes.

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.

 

[TD2K]

I did some work in an oil plant that had separators running very close to your pressures. What we did was take the gas off the 5 psig separator and compress it up to 30 psig in an recip (call it the LP separator), some gas was recycled to maintain suction pressure. The gas then combined with the gas coming off the 30 psig separator (I think the separator was under pressure control maintaining the separator pressure) and went to the next compressor, basically an MP compressor. The LP and MP compressors were driven by a common motor.

The 100 psig separator was maintained on pressure control. The gas from the 100 psig separator combined with the gas discharged from the MP compressor and went to the suction of the main compressor which took it to the final pressure.

 

[engr2GW]

@TD2K, that's the ideal thing to do, but the gas rate has to be high enough and gas price good enough to justify 3 reciprocating compressors for one production well pad right?
But you're absolutely right, I actually do this on some sites. Thanks for the input.

As much as possible, do it right the first time...

 

[zdas04]

TD2K,
The only other way to do it is a pressure regulator off the 2 higher pressure vessels and dump that energy and design the compressor for a 2 psig suction. Probably lower CapEx, much higher OpEx.

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.

 

[engr2GW]

zdas04,
thanks for the reply.
do you mean a back pressure regulator to let the two higher pressure vessels only release at 2psig? if so, will there not be too much backpressure on the two vessels, to where the gas may push the liquid and itself down the liquid line?
If not, please explain.
Thanks a lot.

As much as possible, do it right the first time...

 

[zdas04]

You have a tricky problem here. I started off thinking backpressure regulator to hold 100 and 30 psig respectively, but that does not manage downstream pressure, it just looks at upstream pressure. So depending on the inflow to the two vessels the outflow from a BP regulator could easily be well over your 5 psig stream. A pressure regulator would look at downstream pressure and make sure that all three streams could flow, but you have no control over the actual vessel pressures.

I could design a 3-stage compressor and bring the 5 psig stream into the first stage (discharging at 28 psig), merge the 30 psig stream into the suction of the second stage (discharging at 95 psig). Merge the 100 psig stream into third stage suction and discharge the whole thing at 400 psig. If you need more pressure than that you would need a fourth stage for the full stream. Getting cylinder sizes and clearance right for this mess would really be a challenge, but it could be done. That scenario wastes the least energy and minimizes CapEx and OpEx, but any small change in flow rate or operating pressure would cause major problems quickly.

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.

 

[racookpe1978]

Your problem starts with, and ends with for that matter, with three very different pressure chambers trying to "naturally" exhaust to a single very low pressure pipe that is the single inlet to a single compressor.

1) Use three different pipes from the three different chamber exhausting into three different compressors. Most expensive from a equipment standpoint. Most workable, most successful, most flexible when conditions change from the "best case" optimum-everything perfect solution.

2) Exhaust the lowest pressure (least energy) system into the condenser, then run the next two via a pressure reducer from the HP collector to the medium pressure collector, then run a separate pipe from that MP collector to the pump. You will lose the energy in the LP collector, but the other two will work.

 

[zdas04]

I had another idea this morning. You never said what the relative flow rates of the three streams was. If the lowest pressure stream is the lowest flow then you could reduce your overall complexity by using an ejector with the 100 psig stream to suck on the LP stream exhausting at 30 psig so your compressor suction accommodates the 30 psig stream. Depending on flow rates you could also do it with a 2-stage ejector (use the 100 psig stream to bump the 5 psig up to 30 psig and then in another stream use the 100 psig to suck on the combined 30 psig stream and boost it to around 43 psig--would require that the 100 psig stream have a mass flow rate equal to twice the sum of the 5 psig stream times 3 plus the 30 psig stream times two.

Even if the relative flows are wrong (which would not be surprising), you can use ejectors to homogenize the streams and reduce the hp required by the compressor some (e.g., you could use an ejector with the 30 psig stream as power gas to raise the 5 psig stream to 8 psig, then use the 100 psig stream to raise the combined stream to 15 psig)

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.

 

[engr2GW]

zdas04,
that's not a bad idea. I have never used an ejector before or seen it in our kind of facility, but I looked it up and it appears promising. Thanks.

As much as possible, do it right the first time...

 

[engr2GW]

does the ejector work when the fluid flow is not constant, i.e. if it fluctuates a bit. because the flashing of gas from this vessels increases when the liquid dumps and stays fairly constant thereafter. Also, will the ejector have to be resized/redesigned if the flow rate from all 3 vessels drop due to production drop?

Thanks.

As much as possible, do it right the first time...

 

[zdas04]

Ejectors can handle pretty wide swings, but not for free. Keep in mind my "rule of twos": (1) Power gas pressure must be greater than twice (in absolute units) exhaust pressure: (2) Power gas mass flow rate must be in the neighborhood of twice suction mass flow rate; and (3) Exhaust pressure can be no more than twice suction pressure. None of these rules are hard and fast (I designed an ejector for 6 CR in one step, but that beast was so touchy that a 5% change in mass flow rate would drop suction flow by 70%, but with constant power gas flow rate it is awesome, the "trick" was to get 1.5 Mach on the power fluid instead of the more common 1.1-1.3 Mach). A unit designed for the rule or twos will tolerate pretty wide swings in all the parameters.

One thing that they don't like much is widely changing power gas pressure (they start with a choked flow nozzle followed by a divergent nozzle). In a choked flow nozzle, changing upstream pressure results in a constant velocity but varying mass flow rate. Varying mass flow rate in a super-sonic divergent nozzle causes boundary layer separation and rapid loss of lift. My clients are always saying "if that gadget works this well at 80 psig power gas it should really kick ass at 160 psig". Then they call me and say "this stupid thing isn't working". My answer is always "what did you raise the power gas pressure to?". The answer is generally "that doesn't matter, it stopped working". Small changes of a rule-of-twos ejector don't hurt much. Big changes do. Generally small changes in a designed-for-purpose ejectors are disastrous.

The is a section on ejectors on my web page (under "samples", "CBM ans Shale Gas Upstream Facilities Fall 2010", starting on page 248 or 249 depending on whether you are looking at acrobat number or page number). It has some theory, some cases, and some experiences. Might be worth looking at.

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.

 

[MikeHalloran]

Can we get a photo of this installation?
I can't stop wondering if the 'common pipe' is actually multiple intercoolers in a shared shell.





Mike Halloran
Pembroke Pines, FL, USA