9 replies to this topic

[Ahmed M]

Greetings to all,

 

Hope this finds you all doing fine and staying safe.

 

We have been analyzing the performance of a particular filter coalescer vessel in our plant which is very critical to the operation of our plant. The coalescer is designed to remove upto 99% of 1 micron and above free and entrained liquid droplets from a natural feed gas stream. We have conducted a site survey and have observed that this coalescer have been passing free liquid. Its efficiency is below 28% which is way below the design. 

 

Upon reviewing the historical trend (normalized dP) we have observed a strange pattern whereby several spikes in delta pressure have been observed preceded by a positive slop dP pressure increment before it spikes and returns back to normal.

 

However the historical trend never showed the dP increasing to above 3 PSI. At first glance we suspected that the increasing dP across the elements caused mechanical damage to it upon spiking after which dP reduces. But the mechanical collapse delta pressure of these filters according to the data sheet is around 45 psi.

 

Please find the attached.

 

Appreciate your input and opinion on the possible causes behind the dP spikes and the drop in the performance of the coalescers.

Attached Files

•  dP Normalized.png   85.6KB   4 downloads

Edited by Yusuf 245, 26 January 2021 - 02:22 PM.

[Bobby Strain]

I don't believe that the choice to correlate the variable has any meaning physically. Seems that the filter is damaged and you need to replace it. What is it protecting? I assume that it is in a natural gas stream.

 

Bobby

[Ahmed M]

Thanks Bobby for your reply.

 

I am trying to normalize the dP across the filter elements against gas flow in order to eliminate the effect of the changing flow rates on changes in dP. I rearranged the simplified Ergun's equation for dP across packed beds. I think this is applicable here unless you have a differing opinion.

 

As for your question, the filters are meant to be the last line of defense U/S our dehydrators which are basically molecular sieve beds. Service fluid is natural gas.

 

We have already replaced them and are under observation. However we are trying to explain these spikes in dP where our initial assumption is mechanical damage. But we are doubting this explanation because according to the data sheet filter elements are designed to handle much higher dP.

[Bobby Strain]

You should display actual DP and normalized DP. But not per flow unit. That has no meaning and distorts the data using the scale you have chosen. You should confirm the filter manufacturer agrees your use of the Ergun equation is correct. You should recognize, too, that liquid loading will affect the filter performance and pressure drop.

 

What did you find by examining the filters?

 

Bobby

[Ahmed M]

Hello Bob

 

Well we managed to replace the filters. Visual inspection didn't reveal any mechanical damage. However the outer layers of the filters were covered with black solid substances (awaiting for analysis but we suspect corrosion product). This is quite unusual for this application, usually the solid contaminants get trapped in the inner layers while the outer layers remain clear. The strange thing  is that even though the filters appear to be clogged, the dP trend did not reflect anything.

 

The bigger problem right now is that upon replacing with new ffilters, we still have found free entrained water D/S the coalscer.

[Bobby Strain]

You likely have flow bypassing the element or level control malfunction. I would talk to the equipment supplier. Also, the liquid loading could be to much for the filter. I usually install a separator ahead of such equipment if the liquid loading warrants.

 

Bobby

Edited by Bobby Strain, 31 January 2021 - 01:38 PM.

[tx123]

Hello,

 

Interesting discussion, I realise it is more than a year ago but did you perhaps find out what caused the liquid carryover?

 

I am also looking for ballpark sizing rules of gas-liquid coalescers in ranges of 20-40bar,  Shell DEP states that radial outflow velocity shall be < 0.1 m/s but not much more info on this.  I understand that it obviously depends on filter type, liquid load and type of liquid and gas properties.   Are there any suitable design guidelines to get a first indication of vessel size required for a specific application  or do users normally just request filter vendors to come up with a design?

 

Rgds, Craig.

[Bobby Strain]

Give specs to filter vendors and ask for information.

 

Bobby

[breizh]

Hi tx123,

Better to start a new post!

Anyway, I've attached some literature on the subject.

Breizh 

[tx123]

Hi Bobby and Breizh,

 

Thank you for your comments, appreciated and useful information!

 

 Craig