Radial forces due to uneven air gap 1

[tmoritz]

We have a 100MW pumped-storge unit that is running with an uneven air gap. Nominal air gap is 0.700 inch. Eccentricity is 0.065 inch. The winding is going on 30 years and is showing weakness. We are in the process of procuring a new winding and big question is, should we risk moving the stator?

I'm a mechanical engineer and have forgotten most of my electromagnetics from freshman physics. I have often heard the phrase, "It'll find its electrical center." I know this is true from an axial perspective but is it true radially? Is there a way to estimate the radial forces resulting from an uneven air gap?

So far I have identified the following issues associated with uneven air gap:
1) increased radial forces on the guide bearings
2) increased vibration
3) uneven stator heating resulting in thermal distortion of the stator

What other issues are at hand?

I look forward to reading the responses to this thread.

Thank you,

Tom Moritz
Mechanical Engineer
US Bureau of Reclamation

 

[electricpete]

I have often heard the phrase, "It'll find its electrical center." I know this is true from an axial perspective but is it true radially?

No. It is the exact opposite radially. The magnetic force tends to pull the rotor off-center.

One concern beyond what you mentioned is the possibility of rotor to stator rub, particularly during starting conditions of a motor (not relevant for hydro) and particularly for rotors that are flexible, and only when eccentricity is severe, which I don't think is the case for you. By the way is 0.065" the max minus minimum, or the max deviation from average?

Your other questions about estimating forces - can be estimated theoretically and also there are some correlations... I'll dig around when I have some more time.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Attached is a theoretical approach:
# Compute force vs time for static eccentricity
# airgap flux density 59,000 lines per square inch
# radius to airgap = 12"
# length of airgap = 12"
# eccentricity parameter 0.1

The results are shown in plots at the end (units are Newtons). You can see for 2-pole there is both 2*LF component and dc component. As number of poles increases (same size motor), the 2*LF goes away but the dc remains constant.

You can scale these results proportional to airgap area (proportional to radius times length) and proportinal to flux density squared. The relationship for eccentricity is not as simple.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[esee135]

I stator rubbed a medium sized hydro machine.

Just as you identified, you should be nervous if you are going to reshape the stator. It really depends on how long you want to stretch the life.

In my case the results post move were as follows:
- core vibration (stretching and relaxation of the core steel)
- new core hot spots
- risks of stator coil/bar failure

It will depend on the insulation system you have intact. We found that our stator frame and core moved together very closely, so you have to judge whether your stator coil will crack under movement. Our case was asphalt-mica, which still had some flexibility remaining.

Your eccentricity is not ideal, however, there are many machines with worth cores. Our eccentricity was on the order of .125" on a design air gap of .380", plus cloverleafing etc, with an average gap reduced to .250".

The other effects is you will get increased circulating currents, your heating can cause permanent core damage, and you can easily set into positive feedback failure modes. The core heat was enough that the asphalt-mica ground wall began to slump.I had no access to the air gap side, but the back iron measurement was unusually hot showing heat eminating from the air gap. I wouldn't doubt added heating to the pole face from linking more flux. Your level of core damage is easily shown at speed no load voltage on, the no load losses will be apparent on the core if you got that far in the damage stages.

If you have an epoxy-mica or polyester-mica winding, you shouldn't be so quick to reshape the stator. We found that the stator frame and stator core moved together very closely, so in your case to correct eccentricity, your stator coil/bar may crack and your done. If you wait, and you have a new core and coil assembly on the way, the unit will probably last.

Heat is bad, but you can live with it short term. Bearings get worn, but if the vibration is not significant (ie: you're not changing guide bearings quickly), you can stretch your way to the core replacement.

I can just imagine the added core heat at a .700" air gap with your eccentricity. At .700" your field current is significantly higher than a .400" unit, you would be playing the square of the distance game with flux.

 

[tmoritz]

Electricpete,

Thanks for confirming my suspicion that the radial forces are attractive not repulsive. I can't tell you how many engineers here at USBR believe in the concept of magnetic center. As I indicated in my original e-mail, I think they recall hearing something about axial restoring forces and have transferred that concept to the radial domain.

As for the air gap, the values are basically max-min. We took multiple (20) air gap readings as we indexed the rotor. I then fit the data to a sinusoidal curve (DFT). The amplitude is what we are calling the eccentricity.

Tom Moritz
Mechanical Engineer
US Bureau of Reclamation

 

[tmoritz]

esee135,

Our objective is to center the stator but keep distortion to a minimum.

We do have an epoxy/mica or polyester/mica insulation on these windings so brittleness is an issue and the primary concern of those in the "DON'T MOVE THE STATOR" camp. I'm trying to get my mind around all the advantages and disadvantages associated with moving the stator.

I'm leaning toward running this unit without moving the stator. Seems like the mechanical risk is less than the electrical risk. We've been running this unit for several years in this condition. I just want to get my mind around the dynamics involved so I can determine what and how to monitor.


Tom Moritz
Mechanical Engineer
US Bureau of Reclamation

 

[wolf39]

Tom:

The radial forces due to rotor eccentricity are dependent of the following technical data:

1. rotor eccentricity
2. air gap flux density
3. axial pole length
4. pole shoe width
5. number of poles

It is very unlikely, however, that the air gap under every single pole is identical in case the rotor eccentricity is zero. The magnetic pulling force of each pole depends of its individual air gap.

For a stator rewind it is normal to pull the rotor. It then is possible to re-adjust the field pole air gaps and to re-align the stator.

For a unit 30 years old it may be even economical to replace the complete stator, in case the loss evaluation is high enough. Also, outage is then reduced to a minimum.

Regards

Wolf

http://www.hydropower-consult.com

 

[Clyde38]

To All:
Not really any of my business, I'm strictly curious. I really don't have any experience in very large motors. This sounds like a very large motor and I am curious about the sheer weight of the rotor (what does this rotor weigh?). How much influence will electromagnetic forces have over the mass of the rotating rotor, especially if this motor is in a horizontal position?

electricpete:
It appears that the force on the graphs is in Newtons, is this correct?




[link

http://www.linkedin.com/in/clydehancock

]

[/url]

 

[tmoritz]

Clyde38,

The rotor weight is 575,000 lbs as it sits on the pedestal. This is a reversable vertical synchronous machine, 136MW pump, 106MW generator.

The eccentricity may seem like a small issue but it is a contributing factor in an extended outage on this unit. In November of 2008 I found two cracks in the spider arms. We identified fatigue as the root cause. The cracks initiated in an area of stress concentration. In addition to the elevated stress we were also exposing the structure to cyclical magnetic stress, once per revolution, generated by the varying air gap. Smaller air gap higher torque, and evidently higher radial force. The third problem we found was insufficient preload on the rim. This allowed the rim to "float" at synchronous speed and resulted in unbalanced torque distribution over the eight spider arms.

Our repairs will replace the fatigued material and reduce the magnitude of the stress concentration. We will be replacing all eight compression springs used to preload the rim hoop stress. This will ensure even load transfer to the spider arms. We were proposing centering the stator to eliminate the cyclical stress but several people are arguing against it. Therefore my inquiry regarding the dynamics of a rotor running inside an eccentric stator.

Tom Moritz
Mechanical Engineer
US Bureau of Reclamation

 

[electricpete]

It appears that the force on the graphs is in Newtons, is this correct?

Correct.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[esee135]

Clyde38 --- Typical weight for a hydro ~20-50MVA is on the order of 165T with poles attached.

Or for a 100MVA unit, rotor spider 165T + rotor rim + rotor poles + fan blades etc.

Typical pole weight of a 50MVA machine is ~ 2000lbs

 

[esee135]

Tom: I agree with your theory. In evaluating some machines with bypassed turns, it does affect the orbit of the shaft centerline. Similar to your eccentricity of the stator, the tight air gap section should induce a larger radial synchronizing torque. Also the tight section should result in more flux coupling creating heating as was my case.

Without deep analysis the rotor would be pulled to the stator, and the rotating field would result in a 1x rotating torque as you described with respect to a reference key phasor, which I would imagine would affect the shaft orbit, and wear the generator guide bearing in the appropriate orbit shape.

Does your stator frame sit on radial freedom keys, or are you grouted to the concrete, or do you have radial dowels? I've ran into similar constraints where it is evident that the upper bracket constricts growth, and the core waves follow a predictable pattern. Was your stator originally pile stacked, or has core splits?

 

[edison123]

When you say eccentricity of 0.065", is it in stator or in rotor ? How did you measure it ? If it is only the difference in air gaps across all the poles with rotor sitting in one position, it is not eccentricity.

In vertical hydro generators, I have come across 10% air-gap variation without any major issues of vibration or excessive core/winding heating.

As long as you are planning on a rewind, it would be better to bring the roundness both in stator and rotor. With such an aged machine, replacing the core is also a good idea to ensure lower losses and longer life. This also gives you an opportunity to eliminate core splits, which are usually sources of winding failures and noise.

Muthu

http://www.edison.co.in

 

[tmoritz]

esee135 - The stator frame sits on twelve sole plates with radial dowels. The stator is comprised of three sections bolted together.

I am concerned the uneven heating will result in the uneven stator frame expansion. This could cause the stator frame to bind on the radial dowels resulting in stator distortion further aggrevating the situation.

I appreciate your feedback.

Thanks,

Tom Moritz
Mechanical Engineer
US Bureau of Reclamation

 

[tmoritz]

edison123 - We took a total of 80 air gap readings, 40 at the top, 40 at the bottom. At each elevation we took 20 readings from a fixed point on the stator and 20 readings from a fixed point on the rotor. The readings from the fixed point on the stator were used to compute the eccentricity and roundness of the rotor. The readings from the fixed point on the rotor were used to compute the eccentricity and roundness of the stator. Both are relative to the axis of rotation.

Here are the results:
Stator top eccentricity: 65 mils @ 169°
Stator bottom eccentricity: 52 mils @ 185°
Rotor top eccentricity: 14 mils @ 360°
Rotor bottom eccentricity: 12 mils @ 350°

The phase angle similarity top to bottom in both cases indicates the misalignment is mostly parallel, not angular.

In regards to eliminating the split lines, the USBR tends to make core replacement an optional item in the bid schedule. Once the old winding is removed we will do a loop test to check for core hot spots. The decision to exercise the option is then made. Of course this often extends the unit outage. Someday I need to work through the economics of this process. We could probably show the cost of the core replacement is less than the lost revenue due to the extended outage.


Tom Moritz
Mechanical Engineer
US Bureau of Reclamation

 

[edison123]

Tom

From the readings, I agree it looks a like a parallel shift. Is it too difficult to move the whole stator frame ?

I agree that replacement of core + winding is less of a time driver than, removal of winding + test core + core replacement. Along with the lost generation cost, you could also add any cost savings due upgrading to low loss core steel.

Is uneven heating now an observed fact or an assumption ?

Muthu

http://www.edison.co.in

 

[tmoritz]

We can move the stator. It's just there's a camp of experts telling me not to move it. This is a 30 year old winding on a pump-generator. It's past it's prime and a recent PD test is showing weakness. I'm trying to learn all I can about the implications of not moving the stator.

Regards,



Tom Moritz
Mechanical Engineer
US Bureau of Reclamation

 

[wolf39]

Tom:

According to the USBR power performance data publications your Mt. Elbert units produce energy at $0.025 per kWh. With a plant factor of 18% this is equivalent to a daily revenue of about $11,000. In case the outcome of the core loop test requires a new stator core you can expect an outage of one year or even more. It would make more sense to order a new stator winding plus a new stator (core and frame) designed to state-of-the-art and assembled in one piece in the erection bay at site. Lifting the old stator and mounting the new one in the generator pit can be done in two months at the most (see Grand Coulee G22-G24). The substantial efficiency increase plus reduced outage plus less future maintenance does in most cases offset the extra expenditure for the stator core and frame. An extra stator lifting device may be required but the scrap value of the old stator should be sufficient to cover its cost.

Regards

Wolf

http://www.hydropower-consult.com

 

[edison123]

If you have not observed any excessive vibration/uneven heat/ excessive bearing temps in 30 years, then I guess the problem is not so severe. But moving the stator should not be problem too since there is no relative motion between the core and the winding to damage the winding. I have centered two machines - one during the rewind and the other during overhaul - without any problem.

Muthu

http://www.edison.co.in

 

[esee135]

Tom: Depending on the head room designed in the core, you should be able to rehabilitate the core if the shape is reasonable. Our utility has a history of operating cores well past 70 years with few problems. This is acceptable so long as the due diligence engineering is performed such as the loop test.

However, playing the statistics of generator rewinds, you will eventually require replacement of the core. I have found recently that total stator frame replacement is prevailing over straight rewinds due to the ability to minimize the outage. In your case at >100MW, with good power sales pricing, you may fall into this case. Nevertheless, stator cores are very forgiving even when abused.