Compressor selection - Reciprocating vs Rotary Screw 3

[edta]

We have a booster compression package for offshore platform. The natural gas to be compressed with the process data as below:

+ Flow rate : 9 - 25 mmscfd (1065 - 1900 acfm)

+ MW: 21.8 - 22.2

+ Suction pressure: 550 kPag (min flow) - 850 kPag (max flow)

+ Discharge pressure: 1700 kPag (min flow) - 2300 kPag (max flow)

+ Compression ratio: 2.52 - 2.77

From the preliminary screening, reciprocating and rotary screw (oil free and oil flooded) compressors seem adequate for above application. IMHO,the reciprocating is not preferred due to below reasons:

+ Large space

+ High vibration problem

+ Bulky and Heavy

+ Lower availability

+ much more wearing parts

The oil free screw compressor seems to have the same problems with noise and vibration. Hence, I prefer the oil flooded screw for my project due to its higher availability. But this type still have problem with expensive lube oil, which may be saturated with natural hydrocarbon gas. However, other guys in my team said that we may have the limit manufacturer and after sale supportif we go with the screw type (the platform is located in southeast Asia).

Any advice on the compressor selection for my project is greatly appreciated.

Thanks

 

[zdas04]

The only machine I know of that handles varying suction pressure very well is oil-flooded screws. Recips want suction pressure to be ±5% of design pressure. Dry screws can handle nearly ±8%, but a small increase in compression ratios tends to break the timing chain.

Your ratios are very very low for flooded screws in hydrocarbon service. [You mixed your units so much that I'm not sure which ones you are more comfortable in so I'll just guess.] The difference between success and failure in an oil-flooded screw is managing the condensible vapors. Both water vapor and hydrocarbon vapor. Hydrocarbon vapor is kind of a go/no-go. I haven't had much luck cooking it out of the screw oil after absorbtion, so you have to pick an oil that does not act as an NGL plant and suck the heavier hydrocarbon vapors out of the gas stream. That tends to push you towards synthetic oil. Expensive, but not as expensive as damaged oil wrecking a compressor.

The other side of this is water vapor. Screw oil is very hydrophillic. It really wants to extract the water vapor from the gas stream. If you do nothing then after a week or so you will begin losing your brown fluffy oil down the line. What you have to do is manage the temperature of the oil out of the screw. That really is the only control feature that you have to get absolutely right. The rest is pretty small beans. I want the temperature out of the screw to be 205-215°F (with synthetic you can go higher without damaging the oil, but it is going to be hard enough to get to 205 with those compression ratios). Your only tool to do that is heat of compression. With low ratios I insist on a secondary oil cooler (NOT a cooler bypass "constant temperature valve"). Build a glycol loop that goes to the compressor cooler and put a VFD on that pump. Also put an automated bypass around the oil cooler on the oil side to eliminate all cooling during startup.

If you get that right, everything else becomes simple. There are a number of excellent packagers in Asia, but getting a good packager that understands the issues and is willing to address them is the key. I really don't care which manufacturer's screw someone picks, they are all pretty darn good. I care a lot which packager they pick, most are absolutely incompetent on screw packaging for field use (the ones that say "we've packaged 500 units for plants, we KNOW what we're doing" are the worst in the world at variable conditions).

I always want an oil pump on a field machine, at your compression ratios, it is not optional.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. ùGalileo Galilei, Italian Physicist

 

[edta]

Hi zdas04,

Thanks for your kind advice.

I did the process simulation for this compressor. The discharge pressure varies from 106 - 114 deg C (218 - 238 F).

We have checked with the packagers in Singapore and they tended to propose the reciprocating compressors. It might be their suppliers/manufacturers want to push that kind of machine to the offshore user, or they have limit experience in the screw paackage.

 

[zdas04]

I see that a lot. I've had clients send me preliminary drawings from several of the big guys in Singapore and mostly they looked really lame.

I don't know what kind of temperature control you put in your simulation, but if you used a 3-way valve to selectively bypass your cooler, then your simulation is wrong. Every run I've ever seen has been incredibly optimistic about the performance of those stupid valves. Mostly people set the them at 180F (sometimes 140F), which is the right inlet oil temperature for a very limited number of possible combinations of process temperature, inlet pressure, outlet pressure, and gas flow rate. Very limited. Hysys (for example) seems to be fine with it and simulations move all the parameters in some odd orchestration. Never seen a field machine behave like the simulation predicted.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. ùGalileo Galilei, Italian Physicist

 

[edta]

We are the early state of the project, hence I've just made a simple steady state HYSYS simulation. The calculated discharge temperature at this stage does not take into account the effect of the injected oil. WE need more detail information from manufacturer if we decide to go with this type of compressor.

 

[zdas04]

You've "simulated" an oil flooded screw without the oil??????????????? And you are willing to quote numbers from it?????????? I would say that you really need help. A lot of help. So basically you used HySys to calculate the adiabatic heat of compression for your less than 3 ratios? The multi-phase actual process is a tiny bit more complex than that. I think we've discussed how complex in this forum a few times.

I posted the method I use and Art Montemayor found a small glitch and described how to fix it, I can't put my finger on that thread right now (if I ever find it I'm going to put it in my archive, I've needed it a couple of times). It might be worth your time to try to find it. There is a lengthy Q&A at thread1036-183642 that you might find helpful as well.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. ùGalileo Galilei, Italian Physicist

 

[edta]

As said above, the type of the compressor is not decided yet, and we are in the very early stage of the project (still not in the FEED stage) and still need a lot of work and help to do.

 

[edta]

And thank you for helpful reference

 

[edta]

@zdas

If the discharge cooler is not required by our process (ie: actually we need hot gas of the downstream of compressor) , what is your advice on the temperature control of injected oil. Thanks

 

[zdas04]

Discharge cooler not needed? Remember that the oil is a mostly closed loop, so as you add heat of compression it has to go somewhere.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. ùGalileo Galilei, Italian Physicist

 

[edta]

I mean discharge cooler for the gas stream only. We definitely need a cooler for closed loop oil

 

[dcasto]

We use a synthetic oil on 1500 btu casenhead gas that is a dew point at suction and 4 ratios we have no issues with the oil. The discharge temp is under 180 degrees

Screws do have issues with high frequency vibration mostly related to the gear box and some with the rotors

Recips with pockets can take mire than a plus minus 5% on suction pressure

But all of those issues, you are still are better off with a screw than the issues with recips

 

[zdas04]

dcasto,
Let me make sure I understand. You are saying that you have raw field gas with 1500 BTU/SCF and 180F discharge and have "no issues with the oil". I must say that that is outside of my experience. I have never seen a flooded screw with the discharge pressure that low that didn't: (1) lose a lot of oil; (2) have bearing failures, and (3) controlled capacity properly. Never. Not once in hundreds of field machines. Yours must be magical. If you don't reject the water and liquid hydrocarbons that disolve in the oil, bad things always happen. At 180 F out of the screw you are not rejecting any of these bad things. Did you mean 180F into the screw? That is a pretty common setting for the bullett on the Amont, and usually results in outlet temps over 195F that will often reject light hydrocarbons (but not much water).

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. ùGalileo Galilei, Italian Physicist

 

[zdas04]

Dang, I wish this site had edit capabilities. In the last post I meant to say "discharge temperature that low", and I wanted to finish my thought.

When you say "at dew point at the suction", lets assume that suction conditions are 14.5 psia and 140°F. For methane that means you have 9400 lbm/MMSCF in the gas. Now you add 4 ratios to 58 psia and at 180°F the dew point is 7,100 lbm/MMSCF--where did the missing 275 gal/MMSCF go? That is kind of a lot of water to just vanish.

If you had allowed the process temperature to drift to 206°F then the dew point would have been 10,800 lbm/MMSCF and everything that had come in would go out (with the gas at 87% RH) and be able to drop out in mechanical separators after the cooler.

I have not had the high-frequency vibration issue, but that is likely because I won't deploy a gear box with a screw package. The one time I gave in and accepted a speed changer gear box I got the biggest maintenance headache of my career. Oil leaks. Oil foaming. Running out of oil. Broken gears and bearings. It was the only gearbox in the fleet and I was never comfortable that the operators were paying enough attention to it, but that piece of kit was a serious headache and I was really happy when I connected the well to a different gathering system and no longer needed the higher discharge pressure and could send that machine home.

As to the ±5% claim, that comes from an unpublished study done at one of the major recip manufacturers. The decision not to publish came from marketing and when I talked to the engineer that wrote it he was not happy with the decision. His study found that when suction pressure was more than 5% from design conditions you were starting to see valve stiffness concerns that caused the suction valves to open early or late. A few degrees of crank angle late can increase the discharge temp by tens of degrees. A few degrees early can run you out of hp. His statistical data showed a sharp decrease in the mean time to failure with a definite direction change after about 5%. Will a machine designed for 100 psia suction eat itself if you give it 90 psia? No. Will it break earlier than it would have if you'd kept suction pressure above 95 psia? His data said yes. There are many thousands of recips operating outside this design limitation all over the world. This engineer says that machines that should last 50,000 hours without failure are having a higher frequency of failure during their 5 year overhaul cycle. Is it one in a thousand or one in a dozen? I don't know, I didn't read the report I only got a tirade about it over too many beers. Since that conversation I've been way more careful to operate recips as close to their design conditions as I could pull off and I think (I wish I had been keeping statistics) I have had fewer failures than I had before I started paying attention to design conditions.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. ùGalileo Galilei, Italian Physicist

 

[dcasto]

The suction is 30 psig at 45 oF the discharge is 160 to 220 psig (min back pressure of 100 psig to make the a100 synthetic oil run around). Heck we have flooded our slug catchers and ingested condensate into the screws. let the sit a while and drain enough oil out to get back on then let then hang in min flow till the condensate kicks out and put the oil back. I was like you, I didn't believe it till I saw it. We do like it better as we get the ratios up and the temp gets to 200 F. Of coarse, water is not a concern with the suction conditions.

http://news.compasscompression.com/2012/11/how-to-avoid-oil-dilution-with-high-dew_7.html

 

[zdas04]

Dcasto,
I was fine with the link until he said

Control lube oil temperature supplied to the compressor with a 3-way temperature control valve.

then I wanted to strangle him. For raw field gas, those stinking leaking valves cause more problems than any other single thing on a flooded screw. I hate them. They give you a good result for such a limited combination of flow rate, suction/discharge pressure, and suction temp that they do more harm than good in an environment where all of those parameters change dramatically from second to second.

Looking at temperature into the screw worked OK in plants with constant suction/discharge pressure/temperature and very consistent gas make up. In blood, guts, and feathers you are looking at the wrong parameter. You need to be managing the outlet temp, not assuming an outlet temp from a managed inlet temp.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. ùGalileo Galilei, Italian Physicist

 

[racookpe1978]

i'm wondering when the original poster is going to start talking to the pump/compressor manufacturers' rather than to an internet forum. It's too complex a problem to solve with a closed book and some other company's pipeline/oil field.

 

[Longinthetooth]

zdas04-
Thank you for an interesting series of posts on this topic. I get involved in screw and recip compressors now and again, and your posts included some eye openers. I'll be taking another look at some of our machines now.

 

[zdas04]

Glad I could help. I seem to have better luck with fab shops that have never built a flooded screw package than with shops that have been building plant machines for years. The "experienced" shops always fight me on temperature control, the need for an oil pump, and the location of suction/discharge flanges (to say nothing of speed control). Shops that are new to screws only have to be un-taught the stuff in the screw manufacturers packaging guides (which universally lead to horrible skids).

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[edta]

I did meet some packagers from Singapore. Most of them are salesmen and have strong desired to push the recip. machines to our project. The others in my project team (with no back ground in process or mechanical, only project management in general) likely to go with the recip.

I am tired of fighting with those men and let them decide what they want.