Lifetime of an oil filter compressor

[LotteVV]

Hi,I'm doing an internship about mechanical maintenance. I have come to a problem that is a bit over my head. The problem was also discussed by a few process engineers but they did not come with a good solution.

So, the oil filter is part of a biogas compressor. The biogas comes from the anaerobic waste water treatment, it is then compressed and used as fuel. The compressor is oil lubricated so right after the compressor the oil is being filtered out by a filter. Normally the lifetime of the filter is 3000-4000 working hours. However, the filter now needs to be replaced in less then a month.

Can someone explain in what direction I should look for answers? I don't think temperature is a problem and there is no visible pollution on the filter. The decision of replacing the filter depends on the differential manometer.
There was a study once that mercaptans where found in the gas but I don't see why that would cause any harm to the filter.

Thanks for helping

 

[ChemAir]

I don't think temperature is a problem and there is no visible pollution on the filter. The decision of replacing the filter depends on the differential manometer.

Have you used a second instrument to verify the filter pressure drop is correct (impulse lines clear, etc.)? When things look normal, but a reading is unusual, sometimes it is wrong.

Can you tell us about the filter itself? Is it a cartridge filter, bag filter, or what kind?

Is the concern downtime due to filter changeouts?, cost, etc?

 

[LotteVV]

Have you used a second instrument to verify the filter pressure drop is correct (impulse lines clear, etc.)? When things look normal, but a reading is unusual, sometimes it is wrong.

Can you tell us about the filter itself? Is it a cartridge filter, bag filter, or what kind?

Is the concern downtime due to filter changeouts?, cost, etc?

The manometer has been looked at and is not the problem.

The filter is a fiber demister pad. I can't find a link to the original filter but this link shows the filter best( first picture): https://www.demisterpads.com/demister-pad/fiber-demister-pad.html

The concern is partly the cost of the demister but also the loss of biogas that has to be flared when the compressor is down.

 

[Asymptotic]

My experience is limited to air compressors, but the filter you describe functions similarly to what in that application is called a coalescing filter.

http://teamtechnical.com/html/Parker Basics of Coalescing.pdf

Details differ depending on type (screw vs scroll vs recip ...), but a common cause of increasing filter replacement rate is a progressively more worn compressor. As internal clearances open up, more and more oil passes through, and ends up in the outlet air. Perhaps you are experiencing something similar.

Another thought - can you test spent filters for active bacterial and fungal contamination? Biofilms are known to clog filters rather rapidly and thoroughly in process water systems without an effective biocide treatment program. Perhaps the same thing can occur in a biogas application.

 

[Dullard]

At right angles to your question:
It is not uncommon to install parallel filters (with isolation valves) in situations where service is too painful to the process. You probably still need to figure out why the service interval has shortened, but you can reduce the pain in the interim.

The suggestions above are all good. I'd add:
there is a pretty good chance (I'm guessing) that your supply gas is 100% humid (water). That water is going to condense when the gas is compressed, and condense more as the gas cools. Exactly where that happens could make a difference to your demister. If you were getting some free 'refrigerated drying' from the plumbing (Winter), it may not be working so well now.

 

[LotteVV]

My experience is limited to air compressors, but the filter you describe functions similarly to what in that application is called a coalescing filter.

http://teamtechnical.com/html/Parker Basics of Coalescing.pdf

Details differ depending on type (screw vs scroll vs recip ...), but a common cause of increasing filter replacement rate is a progressively more worn compressor. As internal clearances open up, more and more oil passes through, and ends up in the outlet air. Perhaps you are experiencing something similar.

Another thought - can you test spent filters for active bacterial and fungal contamination? Biofilms are known to clog filters rather rapidly and thoroughly in process water systems without an effective biocide treatment program. Perhaps the same thing can occur in a biogas application.

In the oil analysis it does sometimes show that small steel particles are present. But I don't see clear connection between the two. Sometimes the amount of particles is normal but the demister still fails.

Is it common for filters to not show any visible contamination when bacteria or fungi grow on them? I'm not very educated on the topic of bacteria and filters. Unfortunately, I don't think I'll have enough time for a proper test.

At right angles to your question:
It is not uncommon to install parallel filters (with isolation valves) in situations where service is too painful to the process. You probably still need to figure out why the service interval has shortened, but you can reduce the pain in the interim.

The suggestions above are all good. I'd add:
there is a pretty good chance (I'm guessing) that your supply gas is 100% humid (water). That water is going to condense when the gas is compressed, and condense more as the gas cools. Exactly where that happens could make a difference to your demister. If you were getting some free 'refrigerated drying' from the plumbing (Winter), it may not be working so well now.

The water gets filtered out with a knock-out drum. So the correct cycle is actually:
Anaerobic reactor -> knock-out drum -> Biogas cyclone separator -> compressor -> oil supply tank with the demister.
Will the water still have such a big effect with this cycle?

 

[Dullard]

In the oil analysis it does sometimes show that small steel particles are present. But I don't see clear connection between the two. Sometimes the amount of particles is normal but the demister still fails.

Is it common for filters to not show any visible contamination when bacteria or fungi grow on them? I'm not very educated on the topic of bacteria and filters. Unfortunately, I don't think I'll have enough time for a proper test.

The water gets filtered out with a knock-out drum. So the correct cycle is actually:
Anaerobic reactor -> knock-out drum -> Biogas cyclone separator -> compressor -> oil supply tank with the demister.
Will the water still have such a big effect with this cycle?

As I understand it, both the knockout drum and the cyclone separator would remove only particulate (condensed) water - I was assuming that you were doing that.

 

[Asymptotic]

In the oil analysis it does sometimes show that small steel particles are present. But I don't see clear connection between the two. Sometimes the amount of particles is normal but the demister still fails.

Has the demister filter always had a high failure rate? If no, did it come about gradually, or abruptly?
Clearances open up gradually as wear progresses, and permit more oil (and whatever is in the gaseous phase) to blow through. If the compressor is already worn you won't necessarily see a lot more steel in the oil sump.

Is it common for filters to not show any visible contamination when bacteria or fungi grow on them? I'm not very educated on the topic of bacteria and filters.

Don't know about your application, but in water systems biofilm can be clear and colorless, depending on filtration quality. For instance, if filtering at one micron a heavy biofilm build-up looks like clear snot, but at 50 microns it'll take on a reddish, iron oxide hue.

Excerpted from "A Review of Biogas Utilisation, Purification and Upgrading Technologies" PDF.
https://hal.archives-ouvertes.fr/hal-01619254/document

Removal of Water
The biogas that leaves digesters is always saturated with
water and the absolute water content depends on the temperature
(at 35 °C, the water content of the biogas is usually
5%). Generally, the lower the temperature, the lower the
water content in the raw biogas. This water must be dried
if the biogas is to be used for grid injection or vehicle fuel,
and even gas turbines and combined heat and power (CHP)
(Tables 1, 2). Water can be removed by physical separation
(condensation) and chemical drying (adsorption).

The physical drying methods by condensation are
demisters (liquid particles are separated by wired mesh,
with microspores 0.5–2 nm that can attain dew point of
2–20 °C), cyclone separators (utilizing centrifugal force
to separate water droplets), moisture traps (by expansion,
causing low temperature to condense water), water traps
(design with biogas pipe to collect and remove water) [3, 4,
7]. However, they are less efficient in separating water since
it can only decrease the biomethane dew-point to 0.5 °C,

I'm wondering if upstream moisture removal (knock-out and cyclone) isn't working as well as it should be, and placing more burden on the demister.

 

[LotteVV]

Has the demister filter always had a high failure rate? If no, did it come about gradually, or abruptly?
Clearances open up gradually as wear progresses, and permit more oil (and whatever is in the gaseous phase) to blow through. If the compressor is already worn you won't necessarily see a lot more steel in the oil sump.Don't know about your application, but in water systems biofilm can be clear and colorless, depending on filtration quality. For instance, if filtering at one micron a heavy biofilm build-up looks like clear snot, but at 50 microns it'll take on a reddish, iron oxide hue.

Excerpted from "A Review of Biogas Utilisation, Purification and Upgrading Technologies" PDF.
https://hal.archives-ouvertes.fr/hal-01619254/documentI'm wondering if upstream moisture removal (knock-out and cyclone) isn't working as well as it should be, and placing more burden on the demister.

So, apparently there is some foaming on the oil when they change it. Which could indicate contamination from water.

Also, the problems seems worse in the summer. This could perhaps also point in the direction of bacteria. They like it warm, right?
But if the problem is water. Wouldn't the amount of water particles tell us what the problem is? In the oil spem the ppm of the water is given. Sometimes this amount is classified as normal but the demister still fails. Or does the amount of water not give the right perspective? Could just a small amount of water with bacteria have a lot of effect?

Unfortunately I don't have enough data to see when the problem started. The data of SAP goes back to 1998, at that time the problem was already quite severe (approx. changing of one week)

 

[Asymptotic]

So, apparently there is some foaming on the oil when they change it. Which could indicate contamination from water

If by foaming you mean an emulsion, then yes, and it suggests quite a bit of water contamination.

How much moisture an oil can hold without providing visual "cloudiness" clues varies - as low as 100 ppm for highly refined mineral oils (such as transformer cooling oil) up to 3000 ppm for certain synthetics - but once passing that saturation threshold you'll end up forming an emulsion, free water, or often, both.

These articles should prove informative.
https://www.machinerylubrication.com/Read/1084/water-contamination-lube
https://www.machinerylubrication.com/Read/301/visual-crackle-oil-test

In the oil spem the ppm of the water is given. Sometimes this amount is classified as normal but the demister still fails. Or does the amount of water not give the right perspective?

This concerns me. Run-of-the-mill oil testing usually employ the 'crackle' test (which ideally detects down to 500 ppm, but is closer to 1000 ppm in actual practice) rather than the more expensive and capable Karl-Fisher test.

If it only 'sometimes' passes this test I'd believe there is too much water in the oil.

Could just a small amount of water with bacteria have a lot of effect?

Yes. For example, water-contaminated kerosene is a breeding ground for microbes that feed on hydrocarbons, and is a big problem when it comes to long term storage. Runaway bacterial growth isn't possible without water, and won't occur if moisture levels are kept low enough. To put it in perspective, although fuel line freezing is a more immediate concern than infestation, maximum allowable water content in kerosene-based jet fuel is 30 ppm.

 

[Asymptotic]

Also, the problems seems worse in the summer. This could perhaps also point in the direction of bacteria. They like it warm, right?

Yes, bacteria like it warm, but I'd still be looking more towards water. Warm air holds more moisture than cool air.

 

[rbelli1]

The data of SAP goes back to 1998, at that time the problem was already quite severe (approx. changing of one week)

So this problem has been going on for over twenty years?

BoB

 

[LotteVV]

So this problem has been going on for over twenty years?

BoB

Yes, it has.
They looked over the problem several times but didn't find a solution. But I do think they perhaps didn't do a thorough investigation.

 

[LotteVV]

If by foaming you mean an emulsion, then yes, and it suggests quite a bit of water contamination.

How much moisture an oil can hold without providing visual "cloudiness" clues varies - as low as 100 ppm for highly refined mineral oils (such as transformer cooling oil) up to 3000 ppm for certain synthetics - but once passing that saturation threshold you'll end up forming an emulsion, free water, or often, both.

These articles should prove informative.
https://www.machinerylubrication.com/Read/1084/water-contamination-lube
https://www.machinerylubrication.com/Read/301/visual-crackle-oil-testThis concerns me. Run-of-the-mill oil testing usually employ the 'crackle' test (which ideally detects down to 500 ppm, but is closer to 1000 ppm in actual practice) rather than the more expensive and capable Karl-Fisher test.

If it only 'sometimes' passes this test I'd believe there is too much water in the oil.Yes. For example, water-contaminated kerosene is a breeding ground for microbes that feed on hydrocarbons, and is a big problem when it comes to long term storage. Runaway bacterial growth isn't possible without water, and won't occur if moisture levels are kept low enough. To put it in perspective, although fuel line freezing is a more immediate concern than infestation, maximum allowable water content in kerosene-based jet fuel is 30 ppm.

I found that they use the Karl Fischer method (ASTM-D6304). The values I have indicate the following: (DDMMYY)
632 ppm on 16/09/16
351 ppm on 27/01/17
640 ppm on 20/07/17
584 ppm on 25/10/17
834 ppm on 4/05/18

 

[Asymptotic]

I found that they use the Karl Fischer method (ASTM-D6304). The values I have indicate the following: (DDMMYY)
632 ppm on 16/09/16
351 ppm on 27/01/17
640 ppm on 20/07/17
584 ppm on 25/10/17
834 ppm on 4/05/18

Is the oil basestock mineral (petroleum) or one of the synthetics? I'm guessing mineral oil if you are seeing emulsification at these water concentrations, although (in air compressors, at least) typically a synthetic is preferred. Can you post what specific oil is used, for instance, Mobil Rarus 400 series, etc.?

How often is the oil changed? Was it changed at least once since the first (Sept 16, 2016) water ppm reading provided?

Time was, water content under 1000 ppm was considered OK (at least, in gearbox applications - what the oil is used for makes a big difference), but rule-of-thumb had dropped to 500 ppm or less as the deleterious effects of water contamination have become better understood. Less water >> greater equipment reliability.

If this were an unpressurized gearbox, a cheap and dirty way to reduce oil water content would be to drain the oil, thoroughly clean the gearbox, refill with new oil, and replace the breather cap with an indicating desiccant breather. In normal service, the color indicator should change gradually from the ambient, outside air side toward the gearbox. However, if the water source is on the gearbox side (typically, a leaky oil-to-water heat exchanger) then the color change will concentrate on that side, and indicate immediate service is required.

This isn't much use for you because, in this application, the oil sump is likely to be pressurized.

Consult with your equipment and oil vendors (processing combustible gases undoubtedly adds factors I'm unaware of), but I'm wondering whether an oil-drying 'kidney' loop would help. Something along the lines of this company's offerings.

http://www.precisionfiltration.com/products/phoenix_membrane_purifier.asp

 

[LotteVV]

Apparently, there are two dryers after the compressor( a knock-out drum is in between the compressor and dryers). Would a solution to get the water out properly be that the dryer is placed before the compressor?

 

[LotteVV]

Is the oil basestock mineral (petroleum) or one of the synthetics? I'm guessing mineral oil if you are seeing emulsification at these water concentrations, although (in air compressors, at least) typically a synthetic is preferred. Can you post what specific oil is used, for instance, Mobil Rarus 400 series, etc.?

How often is the oil changed? Was it changed at least once since the first (Sept 16, 2016) water ppm reading provided?

View attachment 243548

Time was, water content under 1000 ppm was considered OK (at least, in gearbox applications - what the oil is used for makes a big difference), but rule-of-thumb had dropped to 500 ppm or less as the deleterious effects of water contamination have become better understood. Less water >> greater equipment reliability.

If this were an unpressurized gearbox, a cheap and dirty way to reduce oil water content would be to drain the oil, thoroughly clean the gearbox, refill with new oil, and replace the breather cap with an indicating desiccant breather. In normal service, the color indicator should change gradually from the ambient, outside air side toward the gearbox. However, if the water source is on the gearbox side (typically, a leaky oil-to-water heat exchanger) then the color change will concentrate on that side, and indicate immediate service is required.

This isn't much use for you because, in this application, the oil sump is likely to be pressurized.

Consult with your equipment and oil vendors (processing combustible gases undoubtedly adds factors I'm unaware of), but I'm wondering whether an oil-drying 'kidney' loop would help. Something along the lines of this company's offerings.

http://www.precisionfiltration.com/products/phoenix_membrane_purifier.asp

I managed to fix a test of the oil for contamination of bacteria. So, I will hopefully soon know if bacteria are the problem.

 

[Asymptotic]

Apparently, there are two dryers after the compressor( a knock-out drum is in between the compressor and dryers). Would a solution to get the water out properly be that the dryer is placed before the compressor?

I don't know enough about biogas applications to hazard a guess. This is a good question to bring up with the engineers who designed your system, and/or vendors that supply its components.

Good job in setting up the bacterial contamination test. Even if the answer is negative, it'll be one more check marked off the list of things that aren't involved.