Estimation of Leakage Flow Through A Labrynth Gas Seal 4

[TPL]

Can anyone point me to a resource where I can find out how to estimate the gas leakage rate across a labrynth seal. I have the gas pressures on either side of the labrynth, the number of teeth, the inner diameter and the clearance?

Thankyou

 

[turboengineer]

I run into labyrinth seal leakages often in the design of liquid fuel rocket enginer turbopumps and other varieties of turbomachinery, and at our company we always refer to an oldie but a goodie called "The Leakage of Steam Through Labyrinth Seals" by Adolf Egli, ASME June 19-21, 1935, but my copy is a copy of a copy of a copy (so you can see how important this paper is to us here). While the paper was written about leakage rates for steam in particular, the emperical data, charts, etc, are non-dimentionalized, and the equations have variables (viscosity, density, etc) so that you can use other fluids as well. I've used Egli for LOX, RP-1, nitrogen (liquid and gas) and numerous other fluids, and I have checked my results in the field in order to confirm my analysis. Egli is perhaps the most difinitive guide on the subject. If you cannot find the paper anyplace else I can send or fax you a copy of what I have. It's readable enough for you to figure it out. I also have a spreadsheet using the formulas from the paper to calculate laby leakage rates. Simply reply to me through this site and I'll get back to you.

Good Luck,

Tim, TurboEngineer

 

[TPL]

Thanks Tim, that info would be great.

I have a centrifugal compressor (pumping natural gas) fitted with tandem dry gas seals. I know the leakage rates of natural gas from the secondary dry gas seal. To stop contamination of the secondary seal with lube oil, an outboard labrynth tertiary seal is fitted between the dry gas seal and the sleeve bearing - this labrynth is ported, with N2 being pumped through it to stop oil migration across it. The resulting N2 and natural gas mix, in the void between the dry gas seal and the labrynth, is vented to atmosphere.
The need for N2 has been challenged (cost cutting) with a proposal to replace its buffer function with air. Functionally I don't have a problem with this, but from a safety perspective, an air and natural gas mix is not ideal.

Knowing the labrynth parameters, I am looking to estimate air flow across the labrynth to determine the likely mix ratios of air and natural gas to see where we might be relative to explosive limits. Anything that might help put a perspective on this is welcomed

thanks

Terry

 

[turboengineer]

Terry,

Whoever said you should use air instead of GN2 in a purge gas seal with natural gas in it needs to get their head examined! Due to the tight clearances in laby seals, they rub the shaft. This could generate enough heat that, with an oxidizer present, fire is a possibility. Even if there is not enough heat generated to ignite the gas, all substances have an ignition pressure in the presence of an oxidizer. LOX is certainly a better oxidizer then air, but if I can light a piece of wood on fire by throwing a rock at it while it is covered with LOX, then perhaps before your people go to air they should do a little research into ignition pressures. All it takes is one good rub to ruin your day!

I use a similar sealing arrangement in rocket engine turbopump interpropellant seals. My seals are similar to yours except there are two outboard laby galleries on both sides of the purge gas gallery. I have never used anything other then helium or nitrogen gas as a purge gas because it will not support combustion. I have actually worked on the design of centrifugal natural gas compressors at work, but I am not familiar with their sealing arrangement. Take a look at

http://www.barber-nichols.com

if you’d like to see what I’m up to here.

Good Luck!

 

[TPL]

I cannot disagree with your sentiment, however, 2 major manufacturers of centrifugal compressors have stated that newer designs can use air instead of N2 as a buffer gas 'under certain conditions' - they are both bidding for a couple of new trains that we will install over the next couple of years. These newer designs use a carbon ring contact seal (instead of labrynths) as the tertiary seal and use air as a buffer - the reasoning here is that the carbon ring buffer seal passes insufficent air to reach the EL of air/gas mix (i.e. the vent contains insufficent air to explode). Management know that the existing N2 generator membranes are failing (the O2 content is way up) and are happy not to replace them (cost savings??). Plenty of subjective opinions abound but until I can produce some objective reasoning as to why we should replace the N2 membranes then the buffer gas O2 content will increase until it really is just air.
So any methods of calculating labrynth leakage are appreciated
thanks

Terry

 

[1969grad]

TPL, never fear this is not rocket science.

The use of air for a buffer gas is quite common for compressors in natural gas (transmission)service. In fact, air is used far more than N2 and I cannot remember seeing N2 used for any recently installed units in gas transmission lines.

The reason is simple. Air is available at most gas transmission stations and N2 bottles are a hassle given that most compressor stations now operate unattended. If you run out of N2 during the night then the unit shuts down. Now the maintenance guy on the call out is perhaps upset and Gas Control is upset because throughput is reduced. And I hope you have spare bottles available.

Furthermore, the amount of seal gas getting through the second seal is nearly zero so I doubt that you would ever be able to get above the LEL.

If there was a problem using air then it would be public knowledge by both the users and the equipment manufacturers.

The vendor should be able to tell you how much buffer air you need if you need to know now. Otherwise, the quantity of buffer air is shown on the P and IDs submitted by the vendor during the drawing review stage.

BTW, the downstream seal in a tandem arrangement is used to provide a safe and orderly shutdown in the event that the first seal fails. It's purpose is not to keep the unit running as a backup.

You should plan on having an air tank in the compressor room in the event of an ESD for the roll down process assuming that the air compressors are not running.

Final thought. You need to use instrument quality air.

 

[israelkk]

Try

http://www.flexwareinc.com/index.html

they have a free software for labyrinth leak calculations.

 

[unclesyd]

turboengineer,

It look like israelik has found Egli for you at his recommended website.

“Laby Flex is a Windows program based on A. Egli's equation for calculating
Labyrinth Seal leakage for straight through or for interlocking (stepped) labyrinths”

 

[TPL]

Thanks Israelkk and Unclesyd.

The reason why air can be used in recent installations of dry gas seals is that they use contacting tertiary seals, with low leakage rates.

The leakage from our dry gas seal cartridges (if fitted with a contacting tertiary seal) is 70sl/min expected and 268sl/min guaranteed (manufacturers figures). We have labrynth tertiaries at the moment and expect a much higher leakage rate (up to 10X in fact) - whatever, the assertion that leakage is nearly zero is wrong and the potential for an explosive mix (air/nat gas) in the secondary vent does exist.

Our N2 buffer gas comes from an N2 skid (running whenever the lube oil pumps are on) where air is pumped through membranes which separate O2 from air: the remainder is used as the buffer gas. The membranes are quite expensive and we would like to go to air but..the use of N2 was HAZOP driven and undoing a HAZOP is not easy.

Going to contacting tertiaries is also expensive and we want to move forward from a position of understanding, where the risks are fully understood.

 

[airplaneman]

Hi. Here are other sources for different information about seals behaviour. If you dont find it, I can email it.

1) "Steam and gas turbines" by Sodola - volume 1. You can find there a formula of a flow through different seals. This formula is approved by tests for sure...

2) "Gas turbine engineering handbook", M. Boyce.

3) " Jet engine construction", Chronin. - This book is written in russian, however it has the most accurate data. It represents a results of continued work done by Stodola.

 

[sailoday28]

airplaneman (Aeronautics)
WOW- It's interesting to see that someone has the classic stodola.