Screw Compressor Driver 2

[EmmanuelTop]

We are in the selection process for depletion/booster compressor, and very likely an oil-flooded screw machine will be a winner from process point of view.

There might be some issues with electrical power supply on site (frequent voltage dips and power loss), so various alternatives need to be considered for the compressor driver.

We are talking about LP screw compressors, pulling the gas (essentially pure Methane saturated with water) from 3 barg to 15 barg. Compressor power varies from 500 kW to 1,500 kW depending on the number of machines operating in parallel, i.e. if the configuration will be 1x100%, 2x50% or 3x33%.

What sort of reliable drivers is out there in the market, and what can we consider as an adequate replacement for electric motor?

Thank you in advance.
Best regards,

http://antwrp.gsfc.nasa.gov/apod/astropix.html

 

[zdas04]

3 barg suction is higher than the maximum allowable for several air-derivative machines, but it is reasonable for several of the process derivative machines. I'd verify max suction pressure with the manufacturer before I went very far.

As to drivers, most of the nearly 100 flooded screw packages I've acquired over the years have had natural-gas engine drivers. About half of them have been from Cat and about half from Waukesha. Both engines have done very well. There are other engines out there (Cummins and Perkins jump to mind), but they are not as common in Oil & Gas and I don't see them as often.

With most low pressure screw applications you have to be really careful not to over run the process needs (i.e., pull suction lower than you intend to), and you really want the suction controller to be fully open at steady state conditions (i.e., use a suction controller for start-up and for upset conditions, in other words the suction controller is not part of the process control scheme and should not be connected to the PLC). The control scheme that one manufacture recommends (and that I've used with outstanding results) is:
1. Change driver speed in response to changing load
2. Operate the unloader (turn valve or poppets on air-derivative, slide valve on process machines) when you reach min or max rpm
3. Throttle a suction controller
4. Activate a bypass

That sequence maximizes process efficiency while minimizing energy use.

For gas-engine drivers there are speed changers that can be controlled by the PLC, but you have to be really careful with the programming to keep the engine from hunting (i.e., the physical speed changer is far slower than the PLC can "think"). For electric drive I always specify a VFD (for process reasons, not for cost savings). I drive the unloader with oil pump discharge, not with pneumatics (I found that pneumatic control tends to be too springy for precise positioning and the unloader will tend to hunt). For suction controllers I use pressure regulator valves controlled by local process without any tie to the PLC or the instrument air/gas system.

The key to success on these machines is the discharge temperature control. Oil temp inlet and oil flow rate need to be controlled to keep the oil temp out of the screw at 205-215[°]F to cook that water vapor out of the oil. The real key to success is managing that temperature. Everything else is a second-order function.

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.

 

[EmmanuelTop]

Thank you for such elaborated and informative reply. Highly appreciated.

I have limited experience with oil-flooded screws. Can the same machine be designed for variable flow and variable suction pressure at the same time? The suction pressure range is 3-5 barg.

Thanks.

http://antwrp.gsfc.nasa.gov/apod/astropix.html

 

[zdas04]

Variability is one of the strongest features of a flooded screw, that is why they are always my first choice for wellsite use (where nothing is very constant for very long).

You need to pay attention to the manufacturer's maximum suction pressure (which is an odd parameter for someone to worry about, but they all do).

Going from 4 bar(a) to 16 bar(a) is about 4 compression ratios. The adiabatic efficiency vs. compression ratios curve would probably point you toward a VI of around 3.6. This gives you a maximum efficiency of about 75%, and a range of compression ratios (assuming that you are only willing to give up 10% efficiency) from 2.2 to 8. This means that for 16 bar(a) discharge pressure you get better than 65% efficiency with suction pressure ranging from 2-7 bar(a), and with 4 bar(a) suction you are better than 65% efficiency with discharge pressure ranging from 8.8 to 32 bar(a) (if your compressor discharge is up to that, some are but packagers use ANSI 150 flanges on the discharge which limits you to about 19 bar(a)).

Flow rates changing can be a problem with maintaining temperature out of the screw. The best machine I ever designed had a VFD on the cooler fan, a secondary coolant (glycol) in a plate and frame heat exchanger (the glycol went to the fin-fan cooler) with a centrifugal pump and a VFD, and a VFD on the oil pump. The PLC programming was a touch complex, but basically we watched temp out of the screw and if it was less than 205[°]F we would slow the cooler fan to a minimum, then we'd slow glycol flow toward a minimum, then we'd slow the oil pump to a minimum, finally we'd throttle the compressor discharge to get more compression ratios (last resort, but necessary in very cold weather). For temperatures above 215[°]F we'd speed up the oil pump, speed up the glycol pump, and finally speed up the cooler fan. That temperature control scheme has worked from -45[°]F ambient up to 110[°]F ambient with flow rates from 0.8 MMSCF/day to 3 MMSCF/day, suction pressure all over the map, and discharge pressure all over the map. This silly machine just runs like the Energiser Bunny.

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.