ROTARY SCREW COMPRESSOR 4

[aberta]

Can anyone tell me the how the typical efficiencies change of such compressors with speed. we have multiple units, all running at different speeds (1800, 2400,2900,3600). They don't provide the design flows, and are not off loading. All compressors are the same make and type.

 

[pmover]

first off, to obtain the appropriate aidience, this posting is best placed in heat transfer and thermodynamics forum...

the screw compressor is a positive displacement machine. clarify whether or not these compressors are oil-flooded.

also, clarify what efficiencies are you referring to?

compression efficiency is a function of the gas being compressed (i.e. gas properties) and pressure ratio of compressor or by using robust industrial strength software to determine enthalpy states (i.e. inlet, outlet, & isentropic) of the gas.

machine efficiency is another term and i recommend you obtaining this information from mfg. in fact, forwarding your question to the mfg will likely get you the results you requested.

good luck!
-pmover

 

[4842]

Several things which you do not mention.
- are the units all the same size (and just at different speeds)?
- what gas are they handling? MW can impact efficiency
- what was original design basis as a unit running at reduced speed (i.e. away from optimum) may be very inefficent?

 

[aberta]

I have been on vacation for two weeks. Hence the delay in responding. My apologies. I also apologise (pmover & 4842) for the vague nature of my question.
These are air compressors, flooded screw type. All compressors are the same type and size (but I am going to reconfirm this by their S/Nos.).
The original "experts?" thought running at slower speeds would mean compressors will last longer. They have since retired. The compressors speeds were increased ( two steam driven units) as needed to meet the demand. These compressors have been tuned several times by the manufacturers. The question of the speed of operation was never entertained.
Because of lack of supply, a 500 HP mobile compressor(#5) was justified and purchased. It still does not meet the demand as other units back off on pressure. We are going to do temperature profile photographs of the piping to determine pipeline blockages. I will update you when we get that done.
For our operation, I feel two compressors operating at 3600 rpm (steam driven units) are adequate for 80% (guess) of the time. By adjustin pressure supplies to different areas, we can compensate and rely on two steam driven units only and use the electric (3600 rpm unit) as stand by. For us steam is much cheaper.
What is bothering me is that these compressors are positive displacemnet units. But are they also positive suction units?
Do they have suction efficiencies?
As these compressors were designed for 3600 rpm, that speed would very likely be the most efficient.
I have talked to the local rep but they want studies done on flow characteristics of each individual unit. This study is quite costly and we do not want to go that route. We have ability to do flow studies with units in operation and we are doing this ourselves.
This I hope describes the problem.
What I would like to know is the basic characteristics of a flooded screw type compressor. We would like to run the compressures based on pressure after the receivers. At present the receivers pressure drops first and later the compressors reat to it. We have resolved the immediate problem of low pressure alarms by using different pressures in different areas. We should be using two compressors but are using three.
I would appreciate any comments.
Regarsd.

 

[pmover]

1) obtain original design criteria (i.e. data sheets or technical write-up) for compressors from original equipment mfg (if possible) or from other reliable resources. this perhaps will be difficult.

2) oil-flooded rotary screw compressors are postive displacement compressors. i'm not familiar with the terms "positive suction units" and "suction effieicncies". these units are design to flood the compression process with oil, in which the oil absorbs a portion of the heat of compression. adequate cooling of compressed gas and lubricating oil is essential.

now as far as efficiency, that will be a little difficult since a portion of the heat of compression is absorbed by the oil. the best solution would be to obtain gas composition, measure P, T, & speed at compressor inlet/outlet and obtain a flow measurement reading. lacking a flow measurement device, flow can reasonably be extrapulated from knowing driver data (i.e. steam consumption, electricity consumption, etc.) and accounting for mechanical/thermal losses.

obtain the same data measurements for the oil flow. with this information, you ought to reasonably determine the efficiency of the compressor. short of that, contact the compressor mfg and obtain recommendations from them.

as far as adding 500hp unit and still having a capacity shortfall, perhaps it is time to conduct a thorough plant/process analysis (get the big picture).

not having flow measurement devices will certainly hinder diagnosis of machinery operation. while i'm not familiar with portable flow measurement devices, perhaps there are devices that can be rented and used for diagnostics purposes and then justifying the the cost for permanently installed flow devices.

a long posting, but i trust it is helpful.

good luck!
-pmover

 

[aberta]

pmover
I hope you did not misunderstand me.
The compressor reps say that since it is a positive displacement screw compressor, the output should be proportional to speed. To do this, the suction must also be positive(mass) suction. I donot believe one can get positive(mass) suction. This is what I question. The suction will probably (I am guessing) have an efficiency curve which depends on speed, like that of a fan eficiency. Can you please comment on this. I need to resolve this issue in my mind.
Your comments about oil and gas heat exchange are very much appreciated and we will look into that.
Can you please comment on moving the compressor controls to after all recievers and dryers. These are at present before these vessels.
Thank you again.

 

[pmover]

the compressor reps statement is accurate provided that inlet conditions are steady (i.e. not transient). changing inlet conditions will vary capacity or amount of gas entering compression process. again, i fail to understand your 2nd and 3rd sentences.

thinking about them though....
perhaps you are referring to the compressor performance based curves which are valid on certain inlet conditions. changing inlet conditions will simply cause a shift (up/down or left/right) in performance curves.

another analogy...
i am certain you've used a hand operated compressor (not a tire pump!) to increase tire pressure in your bike, etc. depending upon make/model (i.e. single-action or double-action), air is drawn into the cylinder and then compressed or squeezed during the compression stroke. although this applies to oscillating/reciprocating compressors, the same applies to rotary units. a fixed amount of gas enter the compression chamber and is squeezed to a higher pressure per stroke. well, changing atmospheric conditions (like an increase in elevation) will change the capacity of these type compressors - with certain assumptions be applied. so, it will take a few more strokes to achieve the desired result. who cares, we all need the exercise...

regarding control system, define the main process variable (i.e. pressure, temperature, or flow) or what parameter is desired to be maintained and define its setpoint. main process variable control is to be achieved by what means (i.e. speed, other control action, etc.)? a change in main process variable results in a change in speed or some other control action. as far as relocating controls, i again recommend investigating the overall plant process (i.e. "getting the big picture&quot and not just the dilema your currently experiencing.

my last, long-winded post.

good luck!
-pmover

 

[zdas04]

Aberta,
If the mass flow rate into the compressor doesn't equal the mass flow rate out of the compressor then you've violated the law of continuity and mother nature will get angry. As you try to move more air, if it is not available at the suction then the suction pressure will decrease. The thing that will always swing is the pressure.

On a flooded screw the discharge pressure (in absolute units) = suction pressure (absolute) * VI^k. VI is the ratio of the inlet chamber volume to the outlet chamber volume. "k" is the ratio of specific heats. Any of the screw manufacturers will have a sizing program that should tell you the energy required for compression (add your auxillaries to get to your motor or turbine load). The output volume flow rate is proportional to speed as long as there is enough gas on the suction to be compressed.

In varying-suction-pressure service, I've had very good luck moving the sensing line for suction pressure to upstream of the suction controller. That way the PLC sees the pressure you're trying to control instead of the controlled pressure. This scheme is much more stable. I am a bit confused though - most of the time air compressors pull air from the atmosphere at a pretty constant pressure (suction controllers are rare on screws in air service). If your suction is varying, are these screws downstream of a "booster" compressor?

If you're not getting the flow you expect, have you checked the operation of the turn-valve unloader? Sometimes those things stick or jump a cog and the don't close or open the way the PLC expects.

David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

 

[aberta]

pmover & zdas04:

It appears I am pretty good at confusing people. I have done a fair amount of research and am a bit wiser. We are working very closely with our maintenance people as this has been a longterm problem. To explain a bit more clearly:
1. The suction side of the valve has a butterfly valve which is controlled pneumatically to adjust compressor loading based on discharge pressure. The compressor air mass output I believe is inversely proportional to the compression ratio: outlet pressue abs/inlet pressure abs. Therfore it controls capacity of each compressor.
2. Our inlet pressure is controlled by a butterfly valve. The butterfly valve is designed to reduce mass flow from the compressor if the discharge pressure increases above design set point.This was giving us much reduced output.
3.The above technology is about 20 years old at least.
4.Since we have four compressures on line serving one manifold, compressors controls start to hunt. To prevent hunting, we adjust the linkage on the inlet valve controller, thus reducing output.
5. To make things worse our 3600 rpm compressor is actually running at 2460 rpm.It has taken us only 20 years to discover. We strobe tested all compressors for speed. Apparently some 20 years ago, someone changed gearbox ratio without changing flow diagrams, bills of material or anything.
6.The three compressors are designed for 3600 rpm operation but are run at reduced speeds. On another part of our plant, they run really good at 3600 rpm and are a variable speed drive.
Now for my question again:
1. My aim is to run these compressors(3) at 3600 rpm with the 4th. stanby at 1800 rpm. Mechanically I have no problems doing this.
2. I want to eliminate the "off loading" inlet valves and associated controls. Can you please comment.
3.I would like to relocate the control point(one) for all the compressors to after the receivers and dryers and on the distribution pipe to plant. This way I can sense demand change as soon as it happens and not be influenced by the accumulation volume in the vessels.
4.The next problem is how to run the compressors (2 steam and 2 electric). We are contacting OEM with respect to this. At present I feel that as our base load is high, run
the 2 steamers full output with two electric on VFD. When demand reduces, off load surplus compressed air or discharge into a slightly lower pressure cross tie (about 5 psi lower)which is run by (2) VFD compressors. Note that demand is always high except one down day every six weeks. Can you comment please on your experience and what would you do.
Regards.

 

[pmover]

assuming process fluid is atmospheric air and pressure is primary process variable...

summarizing, simply install pressure transmitter at location to control pressure. the signal from transmitter will be used to modulate/control either the compressor inlet valve or speed of compressor driver (electric motor via vfd).

i've not heard of using a butterfly valve for controlling capacity of a rotary compressor; hence, my reservation in providing a response. are you certain the compressors are rotary, in type?

your description of equipment & process operation is practical and achievable.

frankly stating, i'd highly recommend contacting a reputable engineering firm to perform the services you desire. particularly a firm that has expertise in rotating equipment design, construction, & installation of associated control systems.

good luck!
-pmover

 

[zdas04]

I'll second pmover's comments and add one caution. Rotary compressors in parallel are never exactly equal to each other. Trying to maintain the same setpoint on all of them at the same time will always cause hunting. I've had good luck with stepwise control.

Stepwise control is an easy concept that is hard to explain. Let's say for illustration that your target system pressure is 100 psig, the system MAOP is 230 psig, and the discharge-pressure instrument dead band is 2 psig. I would set up machine #1 to try to maintian 118 psig. Machine #2 to try to maintain 112. Machine #3 to try to maintain 106. Machine #4 should try to maintain 100 psig. Since #1 and #2 will never have enough capacity to raise system pressure to their set points, they will run all the time fully loaded. #3 will also run fully loaded most of the time. #4 will be running with a moderate load and may swing some. With an extra dead band between any two machines, they shouldn't hunt, and you can rotate set points periodically to balance the long-term loading.

I've done this with suction-pressure control and it really smooths out parallel-machine unloader cycling.


David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

 

[aberta]

zdas04

Thankyou for your valuable comments. We are planning to do exactly what you stated. We have a four compressor operation at our hot strip mill, and we use the stepper control. In our case all four machines run from a pressure switch & we can select the lead compressor. The rest follow the sequence. All these compressors are on/off operation.
At the moment we are looking at pressure losses in the pipework. I will keep you updated.
As mentioned earlier the design speed for 3 of the compressors is 3600 rpm. Our first objective is to run one at 3600 for evaluation. The rest will follow afterwards.
I have saved your website address for possible future reference on other projects.
Many thenks.

 

[COPELANDJIM]

Regarding "eliminate the "off loading" inlet valves and associated controls."

Although you only need one of the 4 air compressors to vary its capacity to match the load, you may wish to retain the "off loading" inlet valves and controls in two of the machines, in case one fails.

I used an HVAC PLC digital controller to properly load and stage a pair of refrigeration compressors. I used the "built in" SEQUENCE software and it worked beautifully. I used the DIGITAL OUTPUTS to activate a total of 6 unloaders as well as the START & STOP of a second compressor. The software includes adjustable time delays that are used to avoid short cycling STARTS & STOPS during light or moderate load conditions. When this software was configured to match the compressors and loads, it did a wonderful job of hold near its setpoint, without excessively starting and stopping compressors.

Those compressors did not have variable inlet valves, so we did not use an analog output signals. The controllers are capable of varying analog signal to actuators to control inlet valves.

James Copeland, P.E.

 

[aberta]

AN UPDATE ON OUR 4-AIR COMPRESSORS.
We checked the compressors for deficiencies and corrected them. This was mainly oil cooler fouling on the water side, thermostatic valves replaced, pneumatics cleaned and adjusted etc.
We then increased the speeds of the two steamers to 3170 rpm in steps. We can now run on 3 compressors instaed of 4 or five. The inlet valve controllers(pneumatic) do not work too well and at times when the demand is low, we blow sump safety valves. We then shut off the electric air compressors.
We are very reluctant to make more than one change at a time. Our first objective to to reach 3600 rpm, but we may stop at 3400 rpm to give us a safety margin.
After that we will try to re-tune are pneumatic controls on our inlet valves. We donot have any conclusions about them yet, but it is quite likely, we will install new type controls.

Many thanks for your contributions. I will keep you updated.
Thanks.

 

[aberta]

Second UPDATE ON OUR 4-COMPRESSORS.
We have increased the speed of our two turbine compressors to 3350 RPM. Now we export the excess copreesed air capacity to adjacent mill. Our BOSC runs fine on just 3 compressors. We have no problems with low pressure alarms on our Instrument Air systems and also on our utility compressed air system. We have now tuned the compressors (this is that inlet valve control again) as best as we can. Our next step is to increase the speeds of the turbine driven compressors to their original design speed of 3600 RPM. We will then close our export valve (to be used in an emergency only) and run on two steamers. The next step would be to monitor and come up with a new logic for sequencing compressors, as mentioned in your earlier responses.
I appreciate your help, and I will keep you updated.
thanks.