axial vibrations in generator

[rob768]

Hi all,

is it possible that electrical problems in a generator cause high levels of axial vibrations at a frequency of 2x line speed? Any reference to articles or books would be highly appreciated. I am examining a case in which high axial vibratios seem to cause the roller bearing cages to wera out rapidly, introducing failure of the sets without prior warning (other than smoke coming from the seized bearing). (sets are rated 2300 kW, running at 1800 rpm)

 

[Strong]

Rob,

2x line frequency is the dominant frequency from magnetostriction force in electric motors, generators and transformers. It is typically low in the axial direction, so I would suspect a resonant structure that causes amplification. It could be the generator end bell/plate, bearing pedestal, or foundation depending upon the machine design. A natural frequency test could confirm this. Stray current passing through bearings in another cause of rapid bearing failure from electrical discharge machining (EDM). EDM typically shows in the higher frequency range and is not associated with 2xLF.

Walt

 

[rob768]

Thanks Strong, but we already confirmed resonance of the bearing support and a certain load dependance as well. However, we are looking for a source to excite the resonance.Initially, since the first sets we encountered this on were driven by 16 cylinder engines, we assumed a relation with 4th order of the engine, travelling through the steel subframe in some way. However, we now have a similar damage in a pto generator, and no longer a clear relation with an engine orde. So, basiscally, i am looking for a source to excite the resonance.

 

[TPL]

Quote:
I am examining a case in which high axial vibratios seem to cause the roller bearing cages to wear out rapidly

a) Is the axial vibration present at FSNL? If it is, then electrical issues might not be the cause

b) How is the axial position of the gemerator maintained? Most designs of roller bearings cannot accomodate anything other than light axial loads. Have you eliminated the possibility that high static axial loading (e.g thermal gorwth or alignment) is overloading the bearing and causing failure?

 

[rob768]

I don't understand the abbreviation FSNL, but the vibration occurs at some 120 Hz (4x rotational speed). Axial loading should not play a role as there are relatively soft elastic couplings between engine and generator, and the NDE bearing is free to move. Also, the wear path in the bearings is not eccentric indicating proper operation at least most of the time. Quite serious cage wear in all directions is the best indication we have for high vibrations being at the root of the problem. In one bearing, the little grease that was present was gold-colored with bronze particles.

 

[TPL]

Full Speed No Load

 

[rob768]

aha, makes sense now. No, the measured vibrations are load dependant.

 

[BobM3]

The guys in the Electrical Forum: Electric Motors and Motor Controls... might be able to answer if there are electrical sources exciting the vibration. Can you spin it (or hit it with a hammer) with no electrical and just measure the resonanant frequencies? It might be reassuring to know it's not mechanical.

 

[Strong]

Rob768,

You didn't indicate that an engine was driving the generator in your original post. Apparently the vibration is minimal at full speed no load, so the engine can be ruled out as the source. I said previously that 120 Hz is an an excitation force generated within the generator. You have confirmed this by saying vibration is load dependent. That is exciting the resonant structure. Detune the resonant structure by changing mass or stiffness or adding a dynmaic mass absorber. I have worked on several generator s with this type of vibration problem.

Walt

 

[rob768]

To BoBm3: hitting is with a hammer has been tried on previous sessions, but if the generator isn't rotating, the full weight of the rotor leans on the bearings, significantly changing natural frequencies, mode shapes and damping to quite another situation.

To Strong: Vibration isn't absent at low loads, just less. Detuning has been tried by the end user, as he was very pressed for time and wanted to use the sets. Stiffening the bearings shields gave some improvement.

It can still be engine related, but the probablility with two different types and speeds of engines in use is much much lower.
A solution may not be the most difficult part, if only i know what the source is. I am mechanical, and no little about electrickety. Hence my post..

And yes, we have considered tuned mass dampings, but I'd rather solve it completely.

 

[rob768]

ow, and did I mention to thank you for the advices so far? i now do.

 

[seymours2571]

Rob,

You mentioned:

"we now have a similar damage in a pto generator"

Does this mean you are using a universal joint connection between the engine and the generator? If so, what is the angle of universal joint. The steaper the angle of a universal joint the larger the axial thrust loading.

Just a thought,

Steve

Stephen Seymour, PE
Seymour Engineering & Consulting Group

http://www.seymourecg.com

 

[seymours2571]

Sorry I meant "steeper" not "steaper". That is what a get for typing after a big lunch and not using spell check.

Stephen Seymour, PE
Seymour Engineering & Consulting Group

http://www.seymourecg.com

 

[rob768]

No, it is not all that simple. The generator is connected to the PTO output of a large marine gear, and is all in a straight line. There is no evident mechanical source.
Gonna take BobM's advice and will cross-post this in an electrical forum.

 

[JJPellin]

If I am reading this correctly, you are seeing high axial vibration at two times line frequency. You have confirmed that you have a bearing housing that resonates at this frequency but are interested in determining the exciting force for the resonance. A two-times line frequency vibration would be produced by an uneven air gap between the rotor and stator. This normally shows up as radial vibration, but in your case, it could be exciting the axial resonance in the bearing housing. I would still suggest that you de-tune the bearing housing so it does not resonate. Change the mass or stiffness so that it is no longer resonant. If the remaining 2x line frequency vibration is acceptable, then you are done. If not, then investigate the sources of this vibration including uneven air gap. Once the bearing housing resonance is eliminated, the vibration should be dominant in the radial direction.

Johnny Pellin

 

[Yvanhof]

To follow up on JJPellin, in my opinion he is right but you can also consider that if your magnetic center does not correspond to your mechanical center, the rotor will pull axially.

When generators are mounted on sleeve bearings that can be seen and measured. We then take this in consideration when ajusting-aligning the couplings. It acts just as a thermal growth would do.

 

[rob768]

Tanks for your thought, but we actualy measured axial displacement as it was one of our suspect items as well. However, there was a little during engaging of the set, but certainly not periodic.

 

[Yvanhof]

Funny how we can see things differently.

2FL indicates a variable air gap. it can come from a soft foot condition, but it can also come from a stator which is not concentric with the bearing journals or bearings in the end cover of the housing. That condition originates at the manufacture of the generator. The air gap can be measured physically, but depending on how it's built, it can be a hard thing to do.

You say that you have a measurable axial movement at start-up. This force is constantly applied during operation and it refers to the relation mag-center to mech-center. It means that the 2FL signature is always there and since you have amplitudes that are strongest in the axial direction, that's where it shows the most.

Fix that and it will go away. It is a magnetic force it comes from the function of the generator.

There could be a way to check this. On the reverse thinking of an electric motor, where you simply cut the power and witness the disapearance of the phenomenon. You could cut the demand and try to see if it goes away with real time analysis from your instrument.

If the problem is electric, it will go away instantly and you will be able to see it. You must program a sample that takes no time to renew itself. One that takes only a second to sample, but continuously. I think it would be the primary cause to fix.

You may still need to fix the resonance problem, but at least it would be less excited in my opinion.

Let me know what you think.

Yvan

 

[rob768]

it is an item we are going to check next time we perform measurements. the thing is, if the field change causes the rotor to be pulled at start, it is a static force, and with one sliding bearing, it should not cause a problem.

 

[dicer]

I'm curious what was discovered for a cause?

Did this gen set just have plain roller bearings?
And there were no thrust bearings?
I don't understand.