From "Transformers for the Electric Power Industry" by Westinghouse Electric:
"Magnetostriction:
When a strip of iron is magnetized, it contracts ever so slightly. This very small change in dimension, occurring 120 times per second is a vibration which creates an audible noise."
It goes on to say that this magnetostriction is not directly proportional to flux (there is a curve), so harmonics of 120 Hz are generated as well. There is also a relationship between strain in the steel and magnetostriction and that bending the laminations increase the noise level and should be avoided.
Thanks. I am beginning to think that magnetostriction is a big effect in transformers but not in motors. In a motor, we could picture a flux wave only rotating around the stator. So even though high-flux-density portion of core is trying to expand at a given point in time, the low density portion is trying to contract... total length of the core does not change. In contract each leg of a core of transformer carries a uniform flux throughout and would be trying to expand or contract at any moment in time. The three legs will be expanding/contracting differently.
I didn't think of that myself... read it in a 1932 IEEE article by some GE old-timer.
Maybe it has something to do with the fact that the transformer core probably has more joints.
Of course, the motor is making a lot of other noise, so maybe the iron noise gets swamped out.
I've noticed that motors on adjustable frequency drives get extremely noisy at low frequency. I don't know if that is from the core or torsional oscillations.
I found this post looking for an answer to a vibration technician's finding of a 120Hz vibration in a 4 pole induction motor, >200HP. This machine had been subject to excessive vibration earlier due to coupling mis-alignment, also has a rather severe duty cycle for an industrial application, maybe 5,000 starts per year. From your work earlier, I wonder if the stator laminations might be loose?
Further to the previous, (checking my facts) that is a 6 pole motor, 500hp, and is operated in (unusual for this app) continuous mode, with Multilin reporting 232 starts in 7824 hours, voltage on bus balanced within ~0.2%, one leg current 2% higher, ~matching one voltage leg reporting low. Other motors on bus report the same one high leg. Still could be damage from alignment induced coupling failure.
Pete,
I'm replying from your previous thread to tell you that I'll have to lookup some info at my workplace to provide my best answer which may not be till early next week. For now, I'm inclined to agree with you that magnetostriction is not a big contributer to 2E in induction motors ; however, current unbalance, introduced by Mr. Williams above, may be the biggest player of all. Other possibilities have been raised though by some knowledgeable EE associates who have been doing some heavy duty electromagnetic analysis and testing in recent years. I need to refresh on their most recent findings. Incidentially, you apparently haven't seen the thread I started in the near-dormant Electrical Acoustics forum which is aimed at all combinations of line frequency peaks we have seen in our pumpmotor spectra. It was an effort to stir up some interest over there and I hoped to find someone to confirm our present speculative understanding of what we are seeing. So far a forlorn hope!
Regarding my comments about 2E sidebands around Rotor Slot Passing, that is definitely linked to rotor free-floating and heavily loaded eccentricity which can vary continuously with fluid film radial bearings particularly the pivoted-pad kind that allow rotor eccentricities up to 200% of nominal bearing clearance and let the rotor precess slowly between pad gaps. Whenever we see really high level 2E sidebands around RSP harmonics we worry about what the rotor was doing to our poor, defenseless bearings at the time the data were acquired. More to come. Regards, vanstoja
Further to the current unbalance comment, one thing I haven't gotten a good grip on is how negative sequence rotor currents can cause a mechanical event observable by vibration instrumentation. If the rotor has suffered damage, I would think you would see much more activity around bar passing or slip frequencies? If there was no damage, shouldn't we see a net zero of mechanical activity?
Electromechanical Pete,
I'm a bit late with this but our UMP engagement intervened. Here's a direct quote from an "authoritative source" from 1994;
"2E noise is generated by many sources within an electric motor. The primary sources are:
1. The fundamental magnetic field which develops the steady state torque to turn the rotor. This source of 2E is primarily a radial bending of the stator core and is best minimized by increasing the backiron thickness of the motor.
2. Negative sequence fundamental voltage/current which is caused by the unbalanced line voltage. This source of 2E generates an oscillating torque and , according to analysis, could be reduced by [active control measures]. However, an alternate way to reduce this source would be to present a better line voltage balance to the motor.
3. Rotor/stator offset (rotor not centered in stator bore) or rotor and/or stator out-of-round (elliptical shaped rotor or stator bore). These sources of 2E are typically radial in magnetic force direction and would not likely be good candidates for application of [active control measures].
4. Magnetostriction, or contraction and/or expansion of the steel punchings caused by conduction of magnetic flux, generates a 2E signal in electric motors. While this source is not thought to be a strong effect in our motors, no quantitative analysis of this source has yet been done."
Some testing reportedly found that:
2E is reduced with less flux density since radial magnetic forces are proportional to air gap flux squared.
2E increases as voltage imbalance increases. The knee of the curves were 0.5% voltage imbalance for a 4-pole machine and 0.25% for an 8-pole machine.
2E is affected by changes in pressure at the stator bore. It remains to determine if this is a source change or a transmission path effect.
A 1967 report by a major motor manufacturer on motor noise implicated space harmonics, saturation harmonics and combinations of both in the generation of n2E sidebands around Rotor Slot Passing frequency.
Good research on vanstoja's part. I'm thinking information overload, at this point. Looking at this from a practical predictive maintenance standpoint, the 0.2% voltage unbalance is better than many places I've seen. Rotor out of round may cause 1 per revolution fundamental acceleration that is masked by other 1/rev noise from the driven machinery, so the 120Hz noise shows up on our radar?
Levels of vibration observed have actually decreased, machine may be somewhat warmer...
One thing I will throw out for rwilliams' benefit since it sounds like he is troubleshooting a twice LF problem in the field:
Try loosening one foot at a time and checking vibration levels (only one foot loose at a time). Soft foot is known to cause twice line frequency. Also mounting resonance or frame resonance may be involved.
Also, if unbalanced / negative sequence plays a role in the twice line frequency vibration, I think it will be an effect on the stator, not the rotor. Negative sequence also has a heating effect which is severe on the rotor... but that's an issue unrelated to the twice line frequency vibration.
Soft foot & 120Hz was a new one for me, we'll check it out. Probably you have nailed it, since our techs just noted we never check for soft foot, & this machine was just aligned.
Thanks much!
