Gas filter differential pressure

[CostasV]

We had a natural gas filter differential pressure reading that we can not explain. Maybe someone could help.

The equipment is a filter OD 1320mm horizontal pressure vessel, design pressure 80 bar. It contains 37 cartridges D192mm X d90mm X 878mm. Design flow 500.000 Nm3/h at line pressure 40 bar. Flow direction in the cartridge is from outside to inside.
Normal reading of differential pressure was about 100 mbar at 400.000 Nm3/h.
After an upset that lasted about 10-20 sec (the upset was 1 or 2 cycles of increase of flow to about 600.000 Nm3/h, followed by a sudden decrease to about zero flow, and finally a stabilization to the original flow of 400.000) the dp reading dropped to about 25 mbar. During the upset, maximum dp exceeded 400 mbar but remained below 500 mbar.
We suspected that some cartridges were damaged and we opened the filter. Unexpectedly, none of the cartridges was damaged, and everything seemed ok. We replaced the cartridges. After putting the new filter in operation, the dp reading was 40 mbar with 400.000 Nm3/h. The "old" type of cartridge is identical to the "new". We also checked the accuracy of the pressure transmitter and found it very reliable.

We checked 6 of the 37 old cartridges in detail. All were found to be similar. They had an "expected" dust on the surface of the pleated paper. The filtering media layers under the paper were intact, without any sign of letting the gas pass unfiltered. No deformation or sign of gas passing through the end caps.

At first, we thought that the upset could remove some of the "blocking" dust of the filter, leaving it with much lower dp than initially, but the new filters give a higher dp than the old ones. We really can not explain that.

Any ideas?

 

[CWB1]

What does the filter manufacturer spec as the allowable dP and max momentary pressure? Given the pressure drop, I'd say you either damaged the filter or a PRD somewhere in the system.

 

[CostasV]

Collapse dp is 2,2 bar according the data sheet. Based on this collapse pressure we set the max allowable dp at 500 mbar, (the 500 mbar dp switch did not activated). The data sheet has no information about max momentary pressure.
Pressure relief devices are much higher, over 80 bar, while the pressure was about 40 bar at the time of upset and later as well.

 

[EngineerTex]

Definitely strange. If you're sure the differential pressure reading is correct, then all of the following have to be investigated:

1. Are you sure that you are passing as much flow as you think you are?
2. Is there a leak path bypassing the filters in a way that is not expected?
3. Is there a leak path to atmosphere?
4. Is one out of each batch of the 37 new and 37 old filters damaged? Does swapping out half of the new ones with the most-recently used filters change the dp by an unexpected amount, either up or down? As a second step, swap the half of "new" ones just removed for the "new" ones still in there and see if there is a different dp and if it is an expected change. Does that isolate where the problem may lie?
5. The damaged PRD seems like a pretty good possibility, as CWB1 says.

Or did you already determine the problem?

Engineering is not the science behind building. It is the science behind not building.

 

[CostasV]

Hi EngineerTex

Thanks for your answer. No , I have not found any explanation to the problem yet.

1. Yes
2. The only path is the 20mm thickness steel plate that separates the "dirty" gas area from the "clean" gas area. Even if there was a leak by passing the filters, this leak would still had the same effect (which is to lower the dp). But we see that with the "new" filters, the dp has become more than the "old" dp.
3. No, no leak to atmosphere.
4. Not one of each batch, new or old is damaged.
Your idea (swapping the half "new" with the "old") is great. But it takes a lot of time to test this.
5. No. The pressure relief device is intact. But, if it was damaged, how could a leak create a dp=25 mbar with dirty filters, and dp=40 mbar with clean filters?

 

[LittleInch]

I think your original suspicion that some of the cartridges were damaged still holds true.

The damage might well be difficult to identify visually and only by doing discrete DP tests on an "old" filter cartridge and "new" filter cartridge will you be able to really check this.

Whether it is worth doing that I can't say, but if you have a design of the filter cartridge you can share and any relevant photos it might be possible to suggest some other cause?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[CostasV]

LittleInch

I don't have a design but here are 2 photos.

 

[LittleInch]

Unfortunately in examining the filter this way you've probably destroyed any damage from the actual event.

The damage would probably not be visible to the naked eye.

differential flowing pressure tests on a cartridge is your only way forward here I think.

Or send one back to the vendor and ask him to test it / examine it.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[CostasV]

Asking the vendor to test/examine it, is a good idea, thanks.