Big Air Compressor

[zdas04]

I'm working on a project where I need to take 35 MMSCF/d of atmospheric air from 11.5 psia (atmospheric pressure at 6,800 ft) to 200 psig. The air then runs through a separation process (losing about 30 psig) and 17 MMSCF of nitrogen is further compressed to 2,000 psig for a tertiary-recovery application.

The separation process is pretty unforgiving of compressor oil so I was looking at doing at least the air-compressor part with centrifiguls. If I put both the air-compressor (4 stages) and the nitrogen compressor (3 stages) on the same shaft, then it is a 10,000 hp application. Since 10,000 hp is a mid-range gas-turbine application it looked like a fit to me from a hp viewpoint, I just don't know if it is reasonable from a fabrication viewpoint.

I could also do it with a centrifugal air-compressor and a 3-stage recip for the nitrogen.

Any ideas on which approach (or some other approach) would be more effective?

David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem

 

[pmover]

a process operation perspective . . .

i do understand the compressor oil carryover matter - not good for the end product and equipment. as i understand, there are two different process (air and n2)compressors driven by the same driver. a process upset condition anywhere in the system impacts the entire process.

any requirements for controlling or varying process flow?

how will compressor operation be at varying flows - if needed? basically, any turndown requirements.

think about start-up and shutdown activities. seems straight forward, but a robust and simple control system is needed.

what is the primary process variable (n2 outlet pressure?).

surge control is certainly another consideration, if opting for centrifugals . . . rather than recycle, perhaps vent to atmosphere, which is a cost & operational consideration.

certainly speed of each compressor is another matter, any gearboxes?

a fun project i must say. perhaps the folks from the air separation facilities can offer some insight.

good luck David and i will monitor this posting and gladly will offer further help.

btw, i lived and professionally worked there in nw nm for about 3 years in late 80's and early 90's - gas industry.
-pmover

 

[sms]

I have worked a number of air seperation plant projects, and the configuration typically ends up as a 1200 or 1800 rpm integral geared compressor driven by a 10,000+HP electric motor. An integrally geared machine gives some flexibility advantage in mixing and matching air vs N2 wheels on the same machine. There are also some nice oil free designs where the bearing is completely outside the volute so there is no way for oil to get into the process. There are pretty established designs for this from Demag, Mann/GHH/Borsig, and others. The machines I have purchased have either been Demag or Mann.

Do you have electricity? Or does it need to be a gas turbine? I don't see any great reason why a GT drive would not work, I just haven't seen it done before.

 

[zdas04]

pmover,
The injection rate has been given as "constant". With as many injection wells as are proposed, maybe that is a good assumption (e.g., one well might only be able to take 1.5 MMCF/d while two others can take 2.25 MMCF/d each).

The primary process variable is nitrogen discharge. The air-compressor discharge can vary a fair bit without upsetting the extraction process.

I've had pretty miserable luck with gearboxes because in small frames they are always under sized and under lubricated for the hp they're pushing. The reduction gears we had in the Navy were incredibly effective, but they were pretty delicate too. In oil & gas I just haven't had a lot of luck deploying stuff that navy nuc's took in stride. Bottom line, since the suction pressure, demand volume, and discharge pressure are so rock-steady I want either internal gears or no gears.

SMS,
It is unlikely that there will be enough power anywhere in the Rockies for this to be run on purchased power - you just don't find gas close enough to the electric infrastructure.

It is good to hear that having the two compressors on the same shaft is a reasonably normal idea. My initial contact with Solar made me wonder. I'll look at Demag and Mann's web sites.

David

 

[sms]

So I guess it is not right there in Farmington, spitting distance from the 4 corners power plant? ;-)

 

[zdas04]

Even spitting distance doesn't assure access. All of that power is going to Southern Arizona (so that the "clean" electricity can go to California at a premium price). It is really hard to get a 100 kW drop in this area.

David

 

[pmover]

it is a definite reality to have multiple compressors with one driver - referred to "train unit" or just "train". usually a LP unit and a HP unit. these train units are installed in refrigeration systems, gas/air processing, etc. throughout the world. i do stress the importance of a robust and proven control system. since n2 discharge is primary process variable, that makes the control system more clearly defined. it is just a matter of sizing and selecting the control elements, actuators, etc. to have an effective control system for the equipment furnished.

something to consider is have a packager of these type units put them together (toromount process systems, edti, enerflex(?), etc.). basically, you furnish/define the process conditions (fluid, fluid quality, p, t, and flow) at the skid's edge and they do all engineering details/design. i've first hand knowledge of these outfits and they are reputable outfits. i can furnish you contact data upon request. basically, you receive a skid package ready to set on the pad, make pipe connections, and make any electrical/instrumentation connections. very simple and effective concept. these outfits have hysis and hysim to conduct simulations, if needed. you certainly should be involved with the design review process and have the rotating equipment tested at mfg shop before shipment. quality is of importance.

just something else to ponder . . .

good luck David! definitely a fun little project!
-pmover

 

[zdas04]

pmover,
This ain't my first rodeo, and the advice you gave about letting the packager do their part of the job is vital to long-term success and needs to be highlighted to any new guys that read this thread in the future.

I have one client that wanted to specify their (recip) package down to the brand of needle valves on gauge lines. I can't talk them out of it, so I politely declined further participation in the project. They'll never learn--an earlier package that was over-specified to that point had major operational problems and the packager said "show us where we deviated from your specifications and we'll cover it". The company couldn't and had to eat significant repair costs.

I'm kinda insistant (to the point of being a real jerk sometimes) on what the fittings at the skid edge look like (flange sizes, flange pressure ratings, exact locations, etc.), but the internals of the skid ain't my area. I will question why someone selected a Cat 3412 LE when a Waukesha H-24 LE is a better fit or why they chose a non-adjustable temp control valve--but my design choices are limited to (1) I buy the package they designed; or (2) I don't buy it. I don't have time to do my job and theirs.

Any help you can give me with contacts would be greatfully appreciated. You can click on my web-site link below and go to the "contact us" tab for my e-mail address.

Thanks a lot.

David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem

 

[cryotechnic]

It’s a very interesting project I think.

The oil is not the biggest problem. I mean, there are many very reliable oil-free compressors available. (Sulzer, Demag, MAN)

As a process-operator in an air-separation plant I think about process problems when I read your post. I assume it’s a cryogenic air-separation process.

I don’t know if following problems can be eliminated in the design of the compressor, but it’s worth to take it into consideration.

There are situations that you have to regenerate your molsieves with air by running the molsieve in a short circuit. (normally you do that with waste gas from your process)
At this point there is no air going to the process which means no nitrogen product.
Is this a problem for the compressor??

In other words:
Is it a problem that the air side of the compressor is loaded and the nitrogen side is not?

What happens when that product (GAN) gets impure? Normally in our air separation plants, we vent the product to the atmosphere when it gets impure. This means, the compressor will switch off due to low suction pressure. (or we turn them off on time if possible)

Point is:
What happens with the “air/gan” compressor if the GAN supply is suddenly switched off?

Cryotechnic

"Math is the ruler of your potential succes...."