Gas Coalescing Filter Failure

Contaminated gas, despite significant investment in gas coalescing systems, is very common. This paper investigates the reasons that often lead to these disappointing results in the deployment of state-of the-art gas/liquid coalescing systems.

Gas coalescing filters are used to protect compressors, capture oil carryover, capture amine carryover, low NOx burners, catalysts and molecular sieve beds, and even gas turbines running on pipeline grade natural gas. The number of poorly performing gas coalescing systems currently in use is considerable. The results of poor performance can not only reduce system uptime but can also be very costly in terms of maintenance and repairs.

Current gas/liquid coalescing technology is capable of removing most hydrocarbon and aqueous aerosols found in gaseous fuels. In most industrial process applications, the industry standard is removal of 0.3 micron (um) aerosols or larger at an efficiency of 99+ percent. There are newer systems that can achieve 0.1 micron (um) removal at 99.9 percent. There are less efficient systems available as well. There is a cost/performance relationship. More effective coalescing systems usually involve higher cost vessels as well as higher cost coalescing filter elements.

Coalescing filters are not the solution if the contamination is in a vapor state. Gas coalescing filters will not remove a vapor. The vapor must reach dewpoint (become an aerosol) before a gas coalescing filter can be effective. In some systems, the change from vapor to aerosol only occurs downstream of a coalescing filter. This is a common occurrence in refinery fuel gas systems where ammonium chloride is formed by vapors that travel with fuel gas. These vapors (NH3 & HCl) can pass through a coalescing filter but later form to deposit salt ammonium chloride salts (NH4Cl) on hot burner tip orifices in the form of a white solid. This situation often exits where gases, exiting high pressure compressors, carry hot oil vapors. Coalescing filters are not a solution when the contamination is in vapor form.

The failure of gas coalescing systems is normally the result of poor design or lack of maintenance.

There are Four Rules that apply to gas coalescing system design – violating any one of these rules will likely result in reduced performance. These rules assume that the quality of equipment purchased is a high-performance filtration product to begin with.

Rule 1 – Don’t use a gas coalescing filter to remove large amounts of liquid.

A gas coalescing filter will only remove what would be considered a mist or fog. As vapor condenses en route to the coalescer, large amounts of liquid can accumulate. This liquid needs to be drained or carried away before it reaches the coalescing system. Gross liquids reaching a gas coalescing element will often block gas flow (slugging) completely until the liquid attempts to drain down. A knock-out drum located upstream is typically employed to take care of this situation. In some facilities, there is a knock-out drum in the gas loop but its performance is sub-par due to several reasons:

  • The knock-out drum is undersized which allows excess liquid to carryover
  • The knock-out drum does not have a mist-eliminator pad or vanes which could reduce excess liquid carryover
  • The knock-out drum’s mist eliminator has corroded away
  • The knock-out drum is too far upstream to be effective – condensation forms downstream due to piping distance and temperature drop

Rule 2 – Follow the manufacturer’s recommendation regarding flow rates and differential pressure.

Not only is there a maximum flow rate for a gas coalescing system, there is a maximum turn-down rate. Excessive flow rates cause gas velocity to propel liquids that should be draining into the sump to become re-entrained into the gas flow ending up downstream. If the flow rate becomes very low, due to process conditions, the coalescing media may lose efficiency allowing some liquid to pass through. Exceeding the maximum differential pressure rating has the same result as exceeding the maximum flow rating. As the coalescing filter becomes clogged with solid particulate, the gas must flow through less media which increases velocity and the likelihood of liquid re-entrainment downstream.

Rule 3 – Cool the gas as much as possible upstream of the coalescing system and maintain downstream temperature.

By cooling the gas, you encourage the formation of aerosols from vapor when it reaches dew point. The more vapor that becomes an aerosol before it reaches the coalescing system, the less vapor goes downstream. Gas exiting a coalescing filter will still contain vapor (hydrocarbon based and/or aqueous based) and that vapor may be close to dewpoint. Anything downstream of the coalescing system that will cause the temperature to drop may bring the remaining vapor to dewpoint causing liquid to form downstream where it will cause damage. Low piping skin temperature is an obvious cause for temperature drop. At a minimum, the piping downstream of a gas/liquid coalescing system should be insulated, even in parts of the world with moderate temperatures (ex. a rainstorm in Monterrey, Mexico will produce 50ᵒF water on the surface of gas piping). In colder climates, the downstream piping should be heated and insulated. Anything that causes a pressure drop will also cause a temperature drop (due to the Joule-Thomson Effect). In a methane system, a temperature drop of 1 degree F will occur for every 15 psi drop in pressure. Pressure drop is the major reason for liquid formation downstream of a coalescing system. The following items, located downstream of a gas coalescing system, can each cause a pressure/temperature drop sufficient to cause coalesced gas to reach dew point:

  • Pressure regulators / control valves
  • Orifice flanges
  • Restricted orifice valves
  • Distribution piping with cross sectional area that increases

Rule 4 – Don’t generate corrosion products downstream of a coalescing filter.

This can be a very difficult rule to keep if the gas contains chemicals that are corrosive to the piping materials. In most refinery fuel
gas systems, the make-up of the gas is often corrosive to the piping. Acid gases can include H2S and CO2 and there are usually some chlorides included. Any and all of these will attack carbon steel and, to some extent, stainless steel. Stainless steel is about 70% iron. Iron is the fuel of corrosion formation. In a fuel gas system, there is no oxygen so there is no oxide formation of a passive layer on the ID of piping. The lack of a passive layer (oxide) on piping allows corrosion to continue occurring. The reactions can take several forms but here are some examples:

H2S (20-30 ppmv) and Iron
H2S + Fe = FeS + H2 (a form of black powder)

H2S and Iron and water vapor
2 H2S + Fe = FeS2 &2H2 (another form of black powder)

The resulting ongoing corrosion of the piping materials causes particulate contamination that makes its way downstream to the point of use. The contamination will eventually foul instruments and heater burner tips and most other downstream components. The fact that these compounds cannot be removed from the gas makes the solution more complicated.

Conclusion

The performance of gas/liquid coalescing systems can be significantly impacted by both poor system design and lax operational practices. The operational practices are easily remedied by adequate monitoring and disciplined maintenance procedures. The system design issues are much easier to remedy during new construction or major system upgrades when piping and vessels are updated or replaced. Overcoming system design issues in an operating environment is much more difficult and expensive to remedy. There are some things that can be done within existing systems to overcome poorly configured coalescing systems. These are outlined in our paper “How to Salvage a Poorly Performing Gas/Liquid Coalescing System”.

About the Author
Bill Burns has worked with coalescing filtration systems since 2001. He is the Technology Manager for Filtration Technologies, LLC.

DOES ANY OF THIS SOUND LIKE YOUR GAS COALESCING SYSTEM? If so, please check out our paper that describes how to address these issues.