Dew Point Management in Oil Flooded Screw Compressors

[zdas04]

I keep having conversations with process engineers where they say they are doing "dew point control" and don't need to manage the discharge temperature on oil flooded screw compressors. When I ask them what "dew point control" means I get the "you are NOT supposed to ask that, everyone knows the answer" look and typically no answer.

I'm just a MechE not an exahalted ChemE, but I didn't sleep through ALL of my thermo classes. "Dew point" is "the temperature (varying according to pressure and humidity) below which water droplets begin to condense and dew can form". In other words at a given temperature and pressure, the 100% relative humidity water vapor content is the "dew point". If I have 7 lbm/MMSCF [112 mg/SCF] gas at 1000 psia [6.9 MPa] and 32°F [0C] then I am at dew point or 100% RH. When I put that gas into a pipeline and lose 200 psi [1.3 MPa] to friction and heat the gas to a ground temp of 66°F [18.9C] then the 100% RH point goes to 33 lbm/MMSCF [528 mg/SCF] so the actual RH is 21%.

Now lets look at what happens in a compressor. Lets say that suction conditions are:
[ul]
[li]Suction Pressure ==> 14.5 psia [100 kPaa][/li]
[li]Suction Temperature ==> 140°F [60C][/li]
[li]Suction Water Content ==> 9483 lbm/MMSCF [150,200 mg/SCF][/li]
[li]Gas flow rate ==> 1 MMSCF/day [28 kSCM/day][/li]
[/ul]

If the machine does 10 compression ratios and we are managing the temperature out of the screw then we see:
[ul]
[li]Discharge Pressure ==> 145 psia [1 MPaa][/li]
[li]Discharge Temperature ==> 206°F [97C][/li]
[li]Dew Point water content ==> 4384 lbm/MMSCF [70,180 mg/SCM][/li]
[/ul]

My question is "What dew point are you controlling in your dew point control scheme and how are you accomplishing this 'control'?" I really want to know. I had a packager tell me last month that the skid he was designing for a client had dew point control so using a cooler bypass 3-way valve with a 140°F [60C] bullet in it would be fine. I asked him to send me drawings of his dew point control device that would let him ignore the thermodynamics of gas compression and phase change (sounded like an AMAZING device). He blew me off.

In the scenario above, 206°F is not high enough for all of the incoming water vapor to leave as water vapor. For that to happen I would need to heat the oil to 248°F ]120C] (which has the potential to create other problems) or cool the inlet to 110°F [43C] (which requires an inlet gas cooler). It seems like something called dew point control would allow a PLC to compare the inlet 100% RH point to the outlet 100% RH point and make pressure/temperature adjustments to make sure that the water content at 100% RH is higher on the discharge than the incoming stream so the gas leaves the compressor with all the water that came in and still be sub-saturated. I've never seen any of this kind of equipment or real-time calculations on screw packages pulling on raw gas. Has anyone else?

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

 

[rotw]

Zdas,

I am not experienced/knowledgeable on screw compressor so just guessing.

Would that make sense if the condensated liquid are monitored via the drain system for instance using some sort of flowmeter, of course supposing these drains happen to exist on a screw compressor...

In this case the flow signal would be brought to controller; but then the next question is what to control ?


"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[dcasto]

well, if you know you cant get there, then why try? In your example you cant win. So, I'd use a two stage screw from mycom so you can win the issue.

now for your question, the algorithm for water vapor or dew point is simple enough to program into a PLC. It can try to adjust the oil temperature to get it hot enough, but as you said, 250 oil is a problem.

 

[zdas04]

Rotw,
The oil is hydrophilllic so it basically acts as a dehydrator to suck the water vapor out of the gas into the oil. You never really see continuous phase water anywhere in the loop. There is no drain system.

dcasto,
Programming is the easy part. Doing something is really difficult. In the example above, you either need very flexible oil temp control (and oil that can tolerate high temps, and bearings that can tolerate high temps, and clearances that don't seize up with expansion, etc.) or a heat exchanger on the inlet. The temperature I used is not out of line for any of the big new shale gas projects. We are going to be dealing with these kinds of suction temperatures on wellhead compressors by early in the 2020s since the reservoir is fundamentally isothermal with drawdown. So if I have really hot gas coming in at 100% RH I'm going to really have to understand managing water vapor much better than screw packagers do today.

I know what we are going to have to do, but the chance of me being involved is nearly zero (not many 80 year olds holding an active role in equipment selection). The first thing someone is going to have to do is get the fabricators to stop making crap up. They keep saying they are doing dew point control when in fact they ran a simulation and with ideal conditions water doesn't accumulate so they put a 140F bullet on a 3-way valve and call it solved. I call it a broken compressor.

In the other thread you said that your compressor was running at 30 psig and 45 F on the suction side. I'll assume you are still in the Rockies and the suction pressure is 43 psia so your water content is 171 lbm/MMSCF on the suction. If I call your discharge pressure 200 psia at 180F I get a discharge dew point of 1300 lbm/MMSCF so the gas leaves at 13% RH. No problems at all. If that suction temp had been 110F (like the Barnett, or Yuma County, CO in the summer) then the water content would have been 1400 lbm/MMSCF and at 180F out of the screw you would be leaving about 12 gallons/MMSCF in the oil.

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

 

[jbattershell]

there isn't a way to control water content or "dewpoint" to any meaningful way in this screw compressor unless you are cooling and reheating the discharge air.

even if the discharge temperature of the compressor were kept high enough to be above the dewpoint, you still have a bunch of water in the air just waiting to come out as condensate. the piping and reservoir in the plants system will all be at temperatures below the 250F or higher, therefore, water will condense out of the air and into the piping somewhere and then flow into the equipment.

there is no magic to removing water without doing something. a dryer of some sort must be installed on the discharge to remove water vapor in the air if you truly don't want it coming out somewhere in the air system. many dryers are available and vary in cost from low to high based on how efficient they are at removing water and how much energy they consume to do that.
John

 

[zdas04]

First, sorry for the confusion, but I'm not talking about air, I'm talking about field natural gas. Second, I'd rather have the water in the gathering system than in the oil (I can pig the gathering system to remove it). What happens after the compressor is really not the concern of this discussion. Typically from well site compressors, we gather the gas to a central delivery point where we can jack the pressure high enough to be economical to dehydrate the gas. At less than 150 psig, the economics make no sense (it takes too much fuel to cook the water out of the glycol).

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

 

[snips13]

When we package rotary screw gas compressors, we take the gas analysis and run a dew point check to see what is the dew point temp of the gas. We run a thermostatic valve in mixing mode in order to keep the oil temp warm. Rule of thumb is to keep the oil temp 20F degrees above whatever the dew point temp is of the gas that you are compressing. What you are trying to avoid is, condensate build up that will contaminate your oil and cause catastrophic compressor failure.

http://www.compressorforums.com

 

[zdas04]

snips13,
That is the traditional way it is done, and has worked well in plants. Move the machine to a well site and you cannot promise a consistent mix of reservoir fluids or operating conditions from one millisecond to the next. I've seen samples 5 minutes apart that had a BTU content over 500 BTU/SCF different one from the next. We also see suction pressure change 30 psi in 10 seconds, suction temperature change 40F in an hour, flow rate vary 30% from one minute to the next. With a fixed oil flow rate and fixed inlet temperature you have the right dew point control several minutes out of every day.

I find the "we take the gas analysis ..." discussion usually mean "when it breaks it is the client's fault for giving the machine the wrong inlet/outlet fluid/pressure/temperature". The reservoir is going to give you what it is going to give you when it is ready to give you anything at all. Zero control of the inlet stream, minimal control of the outlet stream. A fixed bullet in a 3-way valve is not a good well site answer.

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

 

[cosh]

providing you know composition (for example in realtime from DCS, local GC etc.)
you may code an accurate model of compressor in your PLC,
if that is not possible,
there are soft sensors which can model a screw compressor (or a different type) as polytropic stage, see
'

http://www.prode.com/en/opcgashydrocarbon.htm'

of course there are several areas of application (phase equilibria, thermophysical properties, simulation etc.)

 

[zdas04]

The models are all expecting reasonably steady state. The only way that a reservoir gives you steady state is if you define "chaos" as a state. Also I've never seen an online GC last more than a few weeks on a wellsite. GC's tend to be pretty delicate stuff and they don't stand up to the range of fluids and solids that wells produce.

What I'm trying to say is that a successful wellsite screw compressor needs to have been designed with an understanding that every possible parameter has a range of values and that the combination of changes can be in any direction and they often are not related one to another. A simple solution of a 3-way valve results in broken machines far too often for it to be a coincidence.

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