Skip to main content

Dear All

We are facing problem of repeated Dry Gas Seal failure on an Integrally Geared Centrifugal Compressor. The Compressor consists of Bull Gear Driving 03 Pinions with each Pinion carrying 02 Impellers i.e. 01 Impeller at each end.

Each Pinion is axially located by 02 Thrust Collars that are running against Bull Gear Face. Bull Gear is Axially located by Thrust Bearing. Axial Probes are installed only on Bull Gear Shaft.

The Dry Gas Seal manufacturer has attributed the failure to Axial Vibrations. I am trying to understand the source of Axial Vibrations.

Is it possible that Axial Runout on Pinion Thrust Collar Face is causing To & For Axial Motion of Pinion in each revolution e.g. if Axial Runout on Pinion Thrust Collar Face is 0.012 mm TIR (API limit) and Pinion RPM are 21945, resulting axial movement will be 0.012 mm/revolution which implies Axial Vibration at rate of 4.389 mm/sec

Is it possible similar effect will be created by Face Runout of Bull Gear resulting either from machining tolerance or Bull Gear Shaft position within its Bearings (Say extreme right on Non Drive End Bearing and extreme left in Drive End Bearing) e.g. for 0.25 mm Bearing Clearance, Gear Dia to Bearing Span ratio of 1.5 and Gear RPM 1490 resulting movement will be 9.3 mm/sec

Please refer to the attached pdf for schematic of what I am thinking of

Attachments

Original Post

Replies sorted oldest to newest

Both of what you describe are possible but in my experience I have never seen either, and I have worked on dozens of this type machine and well as worked for a major gear manufacturer that made them.

Some questions, the collars on the pinions are removable.  When they are factory installed, there is a very tight tolerance on the axial runout of each collar.  Have they ever been removed in your instance?

The surface on the gear also has a very tight axial runout tolerance as the gear comes from the factory.  Is it still "original" or has it been machined/repaired in any way due to some previous problem?

What type of coupling is used between the prime mover and the gear?

If the prime mover is a motor, is it induction or synchronous?  In either case how many starts per day?

For those not familiar with the arrangement, see Figure 23 in the paper at https://core.ac.uk/download/pdf/84815308.pdf







Last edited by Registered Member

Dear John

Thanks for the reply. The collars on pinion are factory installed and never removed/machined by us. the Gear is also original and no rework or machining has ever been done.

The coupling is Diaphragm Coupling and Motor probably is Induction Motor though I will reconfirm. Its a continuous running machine unless there is some problem. Average number of starts may be less than 6 per year.

I have one question. You have never seen either. As per API the allowable runout on Thrust Collar is 0.0125 mm and for a pinion turning at 21945 RPM vibration don't you think it will surely cause Axial Vibration at 4.57 mm/sec and if you think no, how it can be prevented

In my understanding this will not be the case with conventional machines where Thrust Collar has 360 Deg Contact with Bearing and any Runout on Thrust Collar will probably cause accelerated Bearing Wear due to excessive stress resulting from pont contact due to collar high spot running against bearing ?

Further to the information in above message, please note that we have 03 such Compressors and though overall reliability of Dry Gas Seals on this Compressor had always been a problem, on one of the unit the problem has become much more Serious with Dry Gas Seal on one of the 06 Stages (the Stage with highest Operating Temperature) of this unit failing approximately every 06 months.

The Seals on problematic Compressor had material change of Elastomer which was implemented around 10 months prior to first failure and can be one cause. we have decided to revert back to original Elastomer Material.

Also around 06 to 07 month prior to failure Compressor Motor was replaced with new one which is identical to original and realignment was done. Uptil now I was never thinking it as a contributing factor but based on reply from John it seems it can be a possible cause

Seal manufacturer has put strong emphasis on Axial Vibration and highlighted factors such as possible Bull Gear "swash", mis match of Axial Probe Readings on Bull Gear (though can be a probe setting issue) significant difference in shaft Vibration and Non Drive End & Drive End of Bull Gear Shaft etc.

Initially I was unable to interpret how aforementioned factors highlighted by Seal Manufacturer will cause Axial Vibrations or impact the Seal. Latter I found article on internet pointing towards the possibility of Pinion Axial Vibrations due to Bull Gear Face Runout.

Latter upon checking the Gap Voltages of Bull Gear shaft Probes, I found that  the unit having Seal Failure Issue has much higher difference in Bull Gear Shaft Gap Voltages when Comparing Drive End with Non Drive End.

I am still not convinced if Axial Vibration is cause of Seal Failure as Identical Seal on the opposite end of same pinion is giving much higher service life of around 17 months

Dear John

Thanks for your confirmation regarding my interpretation for possibility of Axial Vibrations. I have limited experience with such Integrally Geared Compressors Initially I was not able to understand how Axial Vibrations can be caused in this Compressor and kept disputing Seal Manufacturer claim but latter realized this may be possible based on case history

Though still not sure of this is really the Case but is there any other way to cross check this ? May be we can expect fluctuations in Motor Amperes as pinion will be forced against normal acting thrust during Rotation ?

Now in order to move forward I am trying to evaluate what can be maximum possible Axial Vibration with all Gear Box Components within Tolerance Limit so that it can be compared with Seal Limit.

I have already asked this question to Compressor manufacturer but still didn't get any answer

You are correct in that API 617 states that the the allowable runout on the thrust collar is 0.0125 mm (0.5 mil).  In my experience however what is typically is realized by the manufacturer is more on the order of 0.00254 mm (0.1 mil).  That is in fact the tolerance at the gear manufacturer where I was employed for many years.

Can you elaborate on the term you or the gear manufacturer use, i. e., "Bull gear swash"?  I've never heard that term.

Two other comments you've made may be deserving of clarification.

"Latter upon checking the Gap Voltages of Bull Gear shaft Probes, I found that  the unit having Seal Failure Issue has much higher difference in Bull Gear Shaft Gap Voltages when Comparing Drive End with Non Drive End."

How much difference are we talking about?  Since this gearbox is likely single helical oil is known to move across the mesh at very high velocity and when impinging on the casing can cause a casing temperature differential side to side.  That might be the cause of the gap differences.   But simply put, looking at probe average gap voltage changes on opposite ends of the same rotor almost certainly has to occur due to some reason other than axial vibration. 

I am still not convinced if Axial Vibration is cause of Seal Failure as Identical Seal on the opposite end of same pinion is giving much higher service life of around 17 months.

This may be a key observation.  I mentioned earlier a temperature differential that can occur due to oil moving through the mesh.  Due you have end to end temperature data on this same pinion?  Can you provide a photograph with directions of rotation and gearing hand of cut so that direction of oil can be determined?

There is an abundance of literature out there discussing this type machine.  Have you researched any of the information that discusses axial motion and dry gas seal failures.  One such article, actually a PP Slide compilation, can be found at https://oaktrust.library.tamu....ce=1&isAllowed=y

A more general article can be found at https://core.ac.uk/download/pdf/84815308.pdf























   

Dear John

Pictures of Bull Gear and Pinion are attached. Direction of Rotation is Clock Wise viewing from Bull Gear Drive End

Temperature of Gas at the Pinion End where Seal is failing is 164 Deg Celsius and at the Opposite End where we are getting more service life is 112 Deg Celsius

The Difference in Gap Voltage across both ends of Bull Gear Shaft is app 1 Volt. While mentioning this difference in Gap Volts, I didn't mean to say that this difference is due to Axial Vibrations. In fact What I am thinking is this difference can cause Pinion Axial Shuttling i.e. Axial Vibrations at the rate of 0.125* 1.5 (Bull Gear Dia to its Shaft Bearing Span Ratio) at around 4.7 mm/sec

Bull Gear swash is the term used by Seal Manufacturer and he is referring to Axial Face Runout of Bull Gear

I also got the clue of Bull Gear inducing Axial Vibrations form the article you shared. Can there be any other possible source of Axial Vibrations ? May be due to Single Helical Gear Teeth meshing ?

Attachments

Images (2)
  • Bull Gear Picture
  • Pinion Picture

Taking your items one at a time...

Temperature of Gas at the Pinion End where Seal is failing is 164 Deg Celsius and at the Opposite End where we are getting more service life is 112 Deg Celsius

This could be a coincidence but this would be a concern to me so you might approach the gear manufacturer about some form of baffle design.

The Difference in Gap Voltage across both ends of Bull Gear Shaft is app 1 Volt. While mentioning this difference in Gap Volts, I didn't mean to say that this difference is due to Axial Vibrations. In fact What I am thinking is this difference can cause Pinion Axial Shuttling i.e. Axial Vibrations at the rate of 0.125* 1.5 (Bull Gear Dia to its Shaft Bearing Span Ratio) at around 4.7 mm/sec

I think I mistakenly inferred by your words that the change in gap voltage differed from end to end.  It is perfectly normal to see a difference in the gap voltage (not the change) from end to end.  There is quite a range of voltage over which the probe output is linear.  For most Bently Nevada probes something around -10 vDC is about the center of the linear range.  The linear range extends from about -4 vDC to about -18 vDC.    I could gap one probe at -9 vDC and the other at -11 vdc and they would work fine.  If the gear moved toward one probe, say the one gapped at -11Vdc by 5 mils (sorry about being non-metric) the gap voltage would go more positive to -10 VDC.  At the same time the other probe would go more negative since the target moved away from the probe.  It's voltage in a normal arrangement would then be -10 vDC, a change from the original -9 vDC.

Bull Gear swash is the term used by Seal Manufacturer and he is referring to Axial Face Runout of Bull Gear

I would assume they mean the area where sliding contact exists between the gear and the pinion collars.  That too is an area where the runout should be quite small, on the order of what I had previously stated.  That area on the gear should be a  ground finish and should be inspected visually and if you have the tools can be swept for runout.

By the way, the area on the gear, is that also the target for the probes observing axial position?

I also got the clue of Bull Gear inducing Axial Vibrations form the article you shared. Can there be any other possible source of Axial Vibrations ? May be due to Single Helical Gear Teeth meshing ?

There could be many sources; distress on the sliding surfaces, a nick in a tooth from mishandling, so surfaces have to be carefully examined.  But keep in mind these would explain axial sliding as a potential cause on both seals, not one seal.

Add Reply

Post
×
×
×
×
Link copied to your clipboard.
×