I am analysing a diesel powered UPS system which works as follows and would like some assistance.

A 350 KW 4 pole synchronous machine is connected to the mains at all times and drives the “ 4 pole rotating stator” section of an induction coupling at 25 Hz. The inner part of the coupling is a squirrel cage rotor which is connected to a sprag (one way) clutch, which is connected to a stationary diesel motor.

The inner rotor runs with just the drag of the slipping sprag clutch at 49.898 Hz relative to the foundation, or 24.898 Hz relative to the rotating stator.

In the event of a power failure the inner rotor acts as an inertia flywheel and via a dc injection to the “rotating stator”, provides energy to the synchronous machine for a few seconds while the diesel starts and reaches a speed equal to the slowing high speed rotor, when the sprag clutch locks and takes over the load. The diesel runs at 26 Hz and again via a dc voltage to the “rotating stator” allows the necessary slip to maintain the synchronous machine at 25 Hz and provide power to the plant.

After a diesel engine bearing failure which shook the induction coupling badly, was repaired, a “beat” appeared with an average value of around 3 mm/s rms overall, and remained constant for about 6 months. The period was around 7.6 seconds which corresponded with (50 – 49.898) Hz = 0.132 Hz = slip frequency. Or, relative to the rotating stator (25 – 24.898)Hz = 0.132 Hz slip.

The induction coupling was replaced with a unit repaired 15 years ago, which was again dismantled, check balanced, Megger tested, with new bearings and then installed. No one saw anything unusual about the squirrel cage rotor of either machine while apart. Amazingly the vibration characteristic was the same but 1.5 times more severe. The speed of this rotor is slightly slower at 49.33 Hz and the vibration period is 1.5 seconds.

In both cases the vibration modulation peaks are at + /- 1x “slip speed”. The upper one is at 50 Hz almost as high as the 1x peak and the lower one is very small. The asymmetric signature looks like amplitude and frequency modulation mixed.

I interpret “slip speed” as the difference between synchronous speed and running speed, i.e. a motor with a full load speed or 24 Hz has a slip speed of 1 Hz at full load.

Pdma operational testing on the present machine showed a “yellow” rotor fault on a 4 pole squirrel cage motor with classic side bands of 2x slip frequency around 50 Hz line frequency. The static RIC test showed a “notch” on each phase in the same position on each wave, about 10% of the wave size.

Question
Why does the mechanical vibration appear as 1x slip speed sidebands around shaft speed (relative to the foundation) and the electrical variation occur at 2x slip speed around the line frequency ?

Urgent replies really appreciated.

Cheers

John

John,
As a former Heemaf / Holec design engineer it is not difficult for me to recognise an Heemaf diesel-no-break model that you are referring to.
Just a few remarks:
1) The high speed rotor is a massive iron eddy current rotor, not a squirrel cage rotor. So no rotor bars in there and no pole pass frequency considerations applicable.
2) This model has a structural torsional resonance frequency at about 45 .. 48 hz. In no load operation the high speed rotor may excite this resonance depending on actual rotor slip and actual 2-plane unbalance level. If bearings and clutch have more friction than the slip will increase, speed will drop and vibration level at 1*rpm increases as you are approaching the resonant condition. That is the reason why your second test was more severe. The induction coupling excitation current displayed on the panel will also have shown more ac current. (This ac current is a good indication of high speed rotor bearings and clutch bearings condition, an early electrical variant of condition monitoring!).
Here comes the trick: if you loosen one out of four mounting bolts of the induction coupling or synchronous machine you might be able to tune the torsional resonance to a lower value and as a result reduce 1*rpm vibration level at normal operation mode.
3) I am also puzzling as to the reason why there is a 2* slip side band around 50 hz induction coupling excitation current, not normal for a 4-pole machine, isn't. This modulation has never drawn my attention in the past, so never investigated either.
IMO there is no direct relationship between the 'mechanical' and 'electrical' sidebands. Interesting discussion currently going on another thread!
Regards,
Arie Mol
Rotating Equipment Consultant
NL

quote:

IMO there is no direct relationship between the 'mechanical' and 'electrical' sidebands. Interesting discussion currently going on another thread!
Regards,
Arie Mol
Rotating Equipment Consultant
NL

You have recently chastised me for bringing up unrelated stuff into other threads. Why don’t you post your objections to the thread it is relevant to? And be specific, please. If you say that the electrical sidebands are not related to the mechanical sidebands, please provide an explanation.
It is my opinion that the position of the mechanical sidebands(if they exist) and electrical sidebands (if they exist) can be calculated from the slip. The relation is right there. Large electrical sidebands are frequently accompanied by the mechanical sidebands. Please, correct me if I am wrong (in the right thread).
jank

What is diesel-powered UPS? Is this emergency diesel engine generator?

John,

A non-uniform magnetic field would cause the Slip Speed vibration. I would check rotor-stator air gap, rotor eccentricity (or shaft bow), and verify that bearings are shrink-fit (tight) on shaft. Also verify the axial position of the slip rotor is correctly centered.

I recently worked on a 500 hp 3578 rpm motor-pump that had electrical (rotor fault) symptoms (2-poles x Slip), but the actual cause was a mechanical fault with a loose inboard motor bearing. The bearing was not shrunk on the shaft, because shaft journal was worn.

My advise is to keep things simple by checking the important mechanical and electrical issues.

Walt

Walt,

Sorry, my question is not directly related to the original one.

Loose bearing on the shaft can explain 2xSlip SBs since it caused dynamic eccentricity . I am curious whether or not you have seen any other symptoms of a loose bearing, such as 1x orders, etc.?

David

David,

The motor in question had high (significant increase) vibration at 1xSS with side-bands at 2xSlip. I originally thought it was a rotor electrical fault (rotor bar), but the motor current was not fluctuating like the vibration and audible sound. Ultrasound measurements pinpointed the inboard bearing.

I arrived at the plant to do the shaft alignment. The motor was being assembled and they had just put the bearings (heated) on the shaft. By luck I checked the tightness of each bearing by hand, and I found the new inboard bearing to be loose and easily slid off the journal. The shaft was machined with a sleeve shrunk on and then machined for proper bearing interference fit.

Walt

John, I hope you don't mind that I respond to one of the postings here.
Jan:
a) I apologise, I never meant to punish or chastise with my words. I appreciate your valuable contributions to this forum very much.
b) My statement that there is no direct relationship between mech/elec sidebands is only valid for this particular case. In general a direct relationship may exist or maybe not. The interesting discussion on the other thread may shed some light on this issue.
c)My statement was not an objection towards the statements raised in this other thread. If it was then I agree I should respond in the other thread. Indeed, cross-coupling threads is not a good idea and should be avoided.
d) To those interested: above mentioned 'other' thread:
http://maintenanceforums.com/eve/forums/a/tpc/f/7161085912/m/2411076923
Regards,
Arie Mol

Thank you all for replying.

The Pdma test showed assymetric side bands around 50 Hz of 2x and 4x slip frequencies with the 4x peak being the highest, and just above the "yellow" limit. I wonder what this means given it's not a squirrel cage rotor.

Nice to hear from you again Ari, and thanks for the information on the rotor design and torsional resonance. Can you be a bit more specific about the design, for a mech.eng. please.
I also measured the lateral natural frequencies on the overhung sprag clutch with 50 Hz vertically and 48 Hz horizontally being small but definite peaks, but the big one is around 67 Hz so this is likely to be the shaft bending mode,with the others being likely structural.

There is definite beat at 1x slip speed audible at around 3 t 4 mm/s on the IC casing and up to 18 mm/s vertically on the sprag clutch. The unusual ability to measure on the clutch give an idea of the shaft motion compared to the case motion. Horizontal on the clutch is less as it is less resonant.

Cheers

John

John, (and Josh)
In the attachment some more design data if this lovely elec./mech. rotating equipment, made in NL.
For more info on newer designs: www.hitecups.com
Regards,
Arie Mol

Uh, I am afraid something went wrong transferring the attachment.
Arie

One more try:

rotary_ups_330_-_500_kVA.doc (4,100 Kb, 14 downloads)

John,
a) The rotor is a solid steel mass rotor producing torque like an eddy current rotor. I think one may consider the outer shell surface of this rotor to have an infinite number of imperfect rotor bars all adding to an angular speed variation and (real) current variation with a 4* slip modulation in this 4-pole design stator. I would not be too concerned about this phenomena because there can be no broken rotor bars simply because they are not in there. And a solid mass rotor has eternal life.
b) The clutch is positioned in the middle of a cardan-like shaft between diesel engine flywheel and induction coupling high speed rotor. When the two bearings of the clutch are properly aligned having zero ax/rad clearance then this cardan shaft is rigid. If not, then there is a hinge in the middle of the cardan shaft resulting in elevated vibration level. Vertical level much more than horizontal level may indicate an outer ring bore being oval. This needs re-machining of outer bore housing.
Your 67 hz is too low, should be approx. 90 .. 110 hz for a rigid cardan shaft.
Regards,
Arie Mol

Thank you Ari,

I've had this article for 20 years and it describes 1x modulation as the result of 2 faults occuring together, imbalance and off center positioning of the rotor in the stator.

If anyone can get a .pdf copy and post it, everyone will benifit, and it has really good diagrams which the text refers too.

Proceedings from the Nineteenth Turbo machinery Symposium, Houston Texas.
Understanding the Vibration forces in Induction Motors, Michael J Costello, Magnetic Products Inc.

The most common electromechanical forces are due to broken rotor bars have a slip frequency of (2x slip frequency x number of pole pairs), which for this 4 pole motor creates modulation at 4x slip frequency.

The second most common electromechanical forces have a frequency of 1x slip speed which appears in all our vibration measurements.

For this to occur, two asymmetries must occur simultaneously between the rotor and stator.

One example is a rotor which is not adequately centered radially in the stator and which also exhibits excessive unbalance.

Assuming this example is of a 2 pole motor, when a magnetic pole lines up with the point of minimum air gap, and the mechanical unbalance is 180 degrees from this point, the unbalance magnetic pull with tend to balance the rotor.

The resulting force will therefore be negligible during one half cycle of slip. During the other half cycle the magnetic pull will line up with the unbalance, amplifying the unbalanced magnetic pull force. Therefore the resulting modulating force will occur once in one cycle of slip.

While it is also possible for this to occur on slower speed motors, a 1x slip frequency force is much more difficult to produce. For a 4 pole motor a pronounced force occurs when the unbalanced force lines up with the unbalanced magnetic pull force created by the point of minimum air gap. In addition a force is created each time the unbalance force comes under the influence of a magnetic pole except when the pole is 180 degrees from the point of minimum air gap.

This in effect also creates an unbalanced magnetic pull force but it is not as great as when the unbalance lines up directly with the pole. Because of this the unbalanced magnetic pull force will have a tendency to modulate a 1x slip frequency.


The nature of the proposed faults fit well with the scenario of events leading to them.

Fault after diesel engine repair.
The most likely cause of the excessive vibration with a 1x modulation is the combination of the high speed rotor being eccentric with the rotating stator since manufacture, with the addition of imbalance.

The imbalance was added due to severe diesel vibration moving the sprag clutch carrier “off square” on it’s press fit on the high speed rotor shaft. This caused the sprag clutch to whirl in a conical mode and create mass imbalance of the high speed assembly. We know this because previously the assembly vibration was very low.

Fault after installing the IC previously repaired 15 years ago.
The cause is the same, but in this case the imbalance was added due to fitting the sprag clutch carrier “off square” on it’s press fit on the high speed rotor shaft.
Alternatively, there may also some residual imbalance in the high speed rotor from poor balancing at XXX plus the above imbalance. All we know is the assembly must be out of balance.

The news that it's not a squirrel cage rotor does not change this conclusion.

The improved alignment will be done on Tuesday and I'll let you know how it goes.

The low lateral natural frequency is a worry though, maybe the shaft is cracked ? But because of the similarity of the faults it would have to be cracked on both IC's, which is not likely.

Once again, thanks to all who replied.

Cheers

John

John,

Your comments about the 1x Slip match my example that I mentioned earlier with motor having eccentric mass unbalance caused by a worn bearing journal.

Walt

Hi All,

Yes Walt, your advice is the same as in the paper and I'm thankfull for it, as it made me a bit more confident of my previously written report.

We have now done more tests.

To look for eccentricity between the high speed rotor and the "rotating stator" we turned the rotating stator and measured the run out of the sprag clutch with a dial indicator and the max variation was 0.01 mm which is low for this cantileverd position.

The Pdma test did not show significant eccentricity.

Before the report was written we did some lift tests and it was "tight" a both ends, no slop at all, just increasing deflection with increasing force on the lever. The the low speed bearings are ok and the sprag clutch end of the HS rotor is ok as well.

The blind C3 "parallel roller bearing" always causes 2x HS rotor speed vibration which was not there on any machine when this bearing was a ball bearing, until about 1993.

So given that the sprag clutch and the drive shaft to the diesel are around 200 Kg and the rotor weight is 750 Kg this blind bearing is not highly loaded and on all machines normally shows some looseness in operation. BUT this is the only machine ever to do this severe vibration and modulation.

So with no significant eccentricity between rotors, the mechanism in the paper does not apply.

We have improved the concenticity of the sprag clutch to the HS rotor from 0.06 mm to 0.01mm p-p at the diesel end of it, and subjuctively the vibration has reduced and the modulation has reduced. This will have improved the balance quite a bit, BUT there is still a problem.

The balancing people got the rotating stator balance quite good but we don't really know about the HS rotor.

Even though the sprag clutch is now better centered the HS rotor could be incorrectly balanced and causing the remaining part of the problem. It's a very big issue to have the whole thing pulled apart again.

I'll be doing some measurements tomorrow morning and I'll give an update.

NB, I'm using our remote monitoring system from Melbourne to diagnose this problem in Perth.
Plus phone and email so it's good to also get help from good people on the net. The future working today !

Once again, thank you all for contributing.

Cheers

John

John,

quote:

So with no significant eccentricity between rotors, the mechanism in the paper does not apply.

In the past I have done coast down tests on induction motors to verify the effect of magnetic pull on unbalance and critical speed.
Two tests I did:
a) coast down with magnetic field, volt/hertz on terminals remained constant (using an external genset)
b) coast down without magnetic field (power switched off)
Not much differences recorded for 4-pole and 2-pole motors.
BTW: Maurice Dooley still there in Perth?
Regards,
Arie Mol

Hi Ari,
Thanks for your help.

I've done some run down tests and with no power on during the 2 sec measurement window there is a very clear beat between HS rotor at slip frequency, see fig 16. Also see fig 15 for a beat at 44 Hz and 50 Hz which is 2x rotating stator speed which at that instant must be mechanical. There is a small structural resonance at 44 Hz.

Just a thought about my 67 Hz major resonance in fig 19 and your thought that it should be around 95 Hz. My measurements were done with the diesel drive shaft conected, and it's quite heavy. Were your measurements done like this or just the sprag clutch ?

Yes, Maruice is still there and getting pretty worried about all this.

Cheers

John

BWest1d.zip (76 Kb, 7 downloads)

John,

quote:

My measurements were done with the diesel drive shaft conected, and it's quite heavy. Were your measurements done like this or just the sprag clutch ?

Yes, done like that = completely assembled clutch.
Will respond later to other info.
(Send my sincere regards to Maurice and his family)
Arie Mol

John,
1)
My description of the clutch, point b) / 6 October, needs correction. This description applies to 400 .. 500 kVA designs. I thought it was such a design because in your very first posting you mentioned 350 kW which refers to 400 kVA or more. Now I know we are dealing with a 330 kVA design that has an overhanging clutch directly coupled to the high speed rotor (Mercedes diesel, not an MTU, right?). For such a design I have to correct: the resonance frequency is not 90 .. 110 hz but lower, your 67 hz is probably a normal value.
2)
About the magnetic pull and unbalance relationship discussed earlier in this thread: It should be noted that the magnetic pull is very small for this particular type of machine. So there can hardly be any interaction between mechanical unbalance and magnetic pull. The reason why the 4-pole magnetic field is very low is this: when the unit is initially started-up the IC should not load the mains power supply too much with high inrush current. As you can observe: the starting current is low and the run-up time is very long for this size of machine. It is designed to be a soft machine towards the mains power supply.
3)
If I make the correct interpretation of your recent coast-down data then it becomes obvious to me: there is an 1*rpm high speed rotor mechanical excitation involved together with a 2*rpm rotating stator mechanical excitation (the frequency is 2* 25.00 = 50 .00 hz because the rotating stator is directly connected to the synchronous motor/generator running 1500.00 rpm). These two frequency components are producing the audible beat and time waveform modulation. It is certainly not an electrical fault. This 50.0000 hz is, although with many zero’s after the dot, mechanical not electrical.
Likely not a clutch flaw either.
What makes this situation unusual is the relatively high contribution of the 2*rpm rotating stator component.
This points towards:
a) A mechanical flaw in the rotating stator radial/axial alignment (alignment of rotating stator in it’s own stationary housing, not alignment towards synchronous machine, assuming the beat is not observable on the synchronous machine)
b) A mechanical flaw inside the 1500 rpm bearing arrangement, something like an oval bore that makes the highspeed rotor and the directly connected clutch running out or ‘jumping’ twice a revolution of the rotating stator.
To investigate the origine of this 2*rpm it is important to observe what happens during the time that rotating stator changes speed. In other words during running-up after the unit is initially started-up and during coast down after taking the unit out of operation (k41 opened, diesel start disabled). This could reveal influence of resonance. I remember a case of a resonant condition of stationary housing endshield having an axial resonance frequency too close to 50 hz (normally ++ 60 hz). Also, there is the test option to have the rotating stator, including synchronous machine, to run up to approx. 1420 rpm with highspeed rotor blocked (and diesel start and other functions disabled).
The practice of (local?) revamping of IC units may have led to a rotating stator bearing problem such as Walt has descibed earlier.
Arie Mol
NL