2 times line frequency in vibration reading 5

[ravindranathan]

What exactly is meant by "2xLine frequency"in motor vibration readings?
Why is it more prevalent in 2 pole motors?

 

[electricpete]

no problem. I kind of blew by the way you suggested, comparing signals using a scope. You're dating yourself. That is old school! (which is a compliment on my book).

=====================================
(2B)+(2B)' ?

 

[edison123]

Pete - Your toon shows the varying air gap correctly which is kind of air-gap eccentricity, I think.

Bill - The simplest way to check for electrical or mechanical bugs is to watch the vibration the instant power is shut off. Electrical vibration will die immediately.

Muthu

http://www.edison.co.in

 

[electricpete]

edison - You can call it eccentricity if you like, but it is not what is traditionally called static eccentricity nor dynamic eccentricity.

Further, that terminology would imo blur the line between the two items discussed as potential 2*LF causes which are fundamentally different:

1 – asymmetric airgap (static eccentricity)
2 – deformation of the stator core caused by the rotating fundamental mmf wave

(1) results in unbalanced magnetic pull on the rotor which loads the bearing, (2) does not.

(1) represents a condition that could be corrected through centering or truing up, (2) does not. Specifically for (2), when examining a machine in the shop the stator bore could be perfectly round, rotor could be perfectly round and centered in the airgap... there's nothing apparent to fix (It is only when power is applied that the stator deforms as shown in my cartoon).



=====================================
(2B)+(2B)' ?

 

[edison123]

When you are that deep into the weeds, anything is possible dude.

Muthu

http://www.edison.co.in

 

[electricpete]

I'll admit to being in the weeds earlier in this thread, but not in my last few posts.

If you're disputing the distinction between (1) and (2) above, then I guess we'll have to agree to disagree. I consider them completely different.
If you're just disputing the terminology, that doesn't matter to me (call it whatever you want).

=====================================
(2B)+(2B)' ?

 

[edison123]

pete. For me, any spatial air gap variation, either on-center or off-center, static or dynamic, is still going to create 2X vibs and much is not much distinguishable from one another. So, I guess, we disagree.

Muthu

http://www.edison.co.in

 

[electricpete]

Before this thread falls away down the threadminder, I wanted to update my summary for posterity with two more things (3,4) that were mentioned above but not included in the previous summary.

Possible causes of 2*LF vib measured on bearing housings
1 – asymmetric airgap / static eccentricity
2 – deformation of the stator core caused by the rotating fundamental mmf wave (transmitted from stator core to frame to bearing housings)
3 – Soft foot (causes either 1 or 2, I vote it affects transmission path in 2).
4 – Unbalanced voltage.

Also for completeness I wanted to add another important assumption in the math "proof" that static eccentricity / asymmetric airgap does not cause 2*LF... which if violated would explain why 2*LF might occur from asymmetric airgap. That additional assumption in the math analysis was that the rotor axis was offset parallel to the stator axis (not tilted). But if a 2-pole rotor had a pure tilt such that its axis intersects the stator bore axis in the axial center, then when the N pole is close to the rotor in one axial half, the S pole is close to the rotor in the other axial half and the no-homopolar-flux constraint (which otherwise limits 2*LF) does not come into play in this situation and elevated 2*LF vib could be expected (acting opposite direction on each bearing… might be identifiable by 2*LF phase comparison between bearings). That's assuming that iron is a short circuit to flux flowing axially in the core which is not really correct... the laminations are not in the correct orientation to prevent circulating currents that would reduce the flow of axial flux. It's a similar consideration that also affects reluctance of the homopolar path. As a first approximation the flux only flows on the outside of the iron, in to a depth of the skin depth. The skin depth of iron / structural steel to 60hz is about 0.009". For silicon steel used in core it's probably a little higher. So it's definitely not zero reluctance, but maybe low enough to allow this 2*LF mechanism, who knows. And of course most airgap deviations are probably not pure parallel nor pure tilt, but instead mixed somwhere between... again more difficult to predict.


=====================================
(2B)+(2B)' ?

 

[electricpete]

Here is a followup on something I said I would followup on in 2nd quote below. First quote below gives background on what I was following up on:

https://www.academia.edu/35527368/T...Practical_Experience_Analysis_and_Diagnostics

Later, in practical vibration analysis, an idea was introduced that the major source of motor vibration at twice line frequency is a static eccentric air gap [13]–[16]. Modern training courses and diagnostic charts associate vibration at twice line frequency in induction motors with an eccentric air gap in the motor.

However, a number of special researches performed on induction motors tailored for the artificial change of the air-gap eccentricity [17]–[20] are not confirming this statement. Usually, vibration levels at twice line frequency depend very slightly on the amount of air-gap eccentricity [20].

I'm going to see if I can find those references 11 and 13-20 when I have time to see if there is anything in there that provides any more insight.

I found three of the articles are available for free. Two in the first batch that mention 2*LF vibration occurs as a result of asymmetric airgap:
[ul]
[li][14] G. H. Bate, “Vibration diagnostics for industrial electric motor drives,” Application Notes, Bruel & Kjaer, Nacrum, Denmark, 1987.[/li]
[li][15] W. R. Finley, M. M. Hodowanec, and W. G. Hotler, “An analytical approach to solving motor vibration problems,” in Proc. 46th Annu. Meeting Petroleum Chem. Ind. Conf., 1999, pp. 217–232.[/li]
[/ul]
Neither of these offer any testing. They may be based on authors’ experience, but I don’t think that was stated, certainly there were no case studies. What was offered to support these claims was a very simplistic explanation, nothing resembling a proof.

The third article is one of those cited as showing asymmetric airgap has no significant effect on 2*LF vibrations.
[ul]
[li][20] R. Supangat, “On-line condition monitoring and detection of stator and rotor faults in induction motors,” Ph.D. dissertation, School Elect. Electron. Eng., Univ. Adelaide, Adelaide SA, Australia, 2008.[/li]
[/ul]
That third reference [20] is available at the link above in three parts. It involves a lot of testing on 2*LF vib and other motor vib / current parameters. Below I tried to outline the relevant parts of [20] regarding 2*LF:
[ul]
[li]Part 2 page 141 section 4.1 describes the approach for testing. They took a 2.2KW induction motor and varied the static eccentricity from -0.2mm to +0.2mm off center. That corresponds to-50% to +50% of full airgap (apparently the airgap is pretty small). They also re-aligned the machine every time they adjusted the airgap. At each eccentricity step, they checked 2*LF vibration 0 to full load.[/li]

[li]Part 2 page 154 section 6.3.5 summarizes the results as follows: "Table 6.10 and Appendix B.5 present the experimental results of analysing the magnitude of the twice supply frequency component in the motor vibration signal. As can be seen in the results, the magnitude of this component does not show a consistent (useful) variation pattern between the healthy and the faulty motors. Table 6.10 - Characteristics of the twice supply frequency component"[/li]

[li]Part 3 page 29 Appendix B5 (screenshot linked here) gives full results of this testing in the form of side by side charts. The left chart gives 2LF vib vs load, with a separate curve for each eccentricity level. The right chart gives 2LF vib vs eccentricity, with a separate curve for each load level. If you look at that graph, I'm pretty sure you'll agree that there is no obvious trend for increasing 2*LF vib with increasing eccentricity within this particular data.[/li]

[li]Aside - the unusual labeling of the graphs in B5 deserves some discussion: The caption definitely tells us that the vertical axis definitely represents the 2*LF component of the vibration. However it also has an unusual label F(f) * F(f) in db. That is the same labeling used for most of the graphs in this study, including current. We know what db is, we just don’t normally use it for vibration. I gather the F(f) * F(f) might indicate that it is a plot of power spectral density of vibration.[/li]​

[/ul]

One more thing to add. The winding configuration (series or parallel) apparently can also play a role in whether or not asymmetric airgap causes 2*LF vibration, based on Measurement and Calculation of Unbalanced Magnetic Pull in Wound Rotor Induction Machine by Dorrell and Kayani. You can scroll through the link and look at the pictures of the rig they used to do the testing. A relevant passage on page 2:

The UMP [Unbalanced Magnetic Pull] quoted in this paper is a steady pull which dominates the force. This is because the winding is four pole and series wound. If parallel windings were used then it is possible get a pulsating or rotating force vector at twice supply frequency [7]

=====================================
(2B)+(2B)' ?

 

[electricpete]

One more thing to add. The winding configuration (series or parallel) apparently can also play a role in whether or not asymmetric airgap causes 2*LF vibration, based on Measurement and Calculation of Unbalanced Magnetic Pull in Wound Rotor Induction Machine by Dorrell and Kayani. You can scroll through the link and look at the pictures of the rig they used to do the testing. A relevant passage on page 2:

The UMP [Unbalanced Magnetic Pull] quoted in this paper is a steady pull which dominates the force. This is because the winding is four pole and series wound. If parallel windings were used then it is possible get a pulsating or rotating force vector at twice supply frequency [7]

I was able to get hold of [7], and I think I understand what’s going on.  I'm not sure his conclusions about eccentricity causing 2*LF in parallel connected motors are represetnative of the parallel motor connections people may see in the real world (comment after you've read below if I'm mistaken).
He took a 10-pole motor with coils brought out to a patch-board so it could be reconfigured many ways.  The parallel configurations where he saw 2*LF component of unbalanced magnetic pull associated with static eccentricity were:
[ul]
[li]* [figure 5] 10 pole, with every single pole group connected directly to the power supply (10 circuit wye).[/li]
[li]* [Fig.7] 10 pole with five parallel circuits per phase and two physically adjacent groups within each circuit[/li]
[li]* [Fig 8] 6 pole with three parallel circuits per phase and two physically adjacent groups within each circuit[/li]
[/ul]
In these configurations (especially the first one), it’s easy to see that the current will not be equal in all the parallels… a coil located at the narrow gap will carry less exciting current than a coil located at the wide gap.   This makes the current distribution around the periphery LESSS sinusoidal and the flux density distribution around the periphery MORE sinudoidal.  The less sinusoidal current distrubution around the circumference destroys the assumptions of that earlier math analysis I described which had concluded there is no 2*LF component of unbalanced magnetic pull from eccentricity ( So it’s not suprising this particular motor would not match the results of that earlier paper.). The more sinusoidal flux density distribution around the circumference results in lower unbalanced magnetic pull (which is what he was trying to demonstrate... winding designs to reduce unbalanced magnetic pull).

The most extreme example is the 10 circuits per phase. It is not realistic at all. Why would someone want to construct or test a motor with something like 10 parallel groups per phase? The answer in the paper: it greatly reduces the steady component of the unbalanced magnetic pull (the 10-pole parallel has a factor of 10 decrease in steady unbalanced magnetic pull for a given eccentricity compared to the 10-pole series). It comes at the expense of creating a new 2*LF unbalanced magnetic pull in the parallel, but that is much smaller than the original large steady pull in the series connection.

Likewise, I think what he's doing even with the second and third configurations listed is different than how a typical parallel motor would be wound (although I'm open to comment). When he has multiple parallel circuits that each have two groups in series, he selects the two groups which are physically adjacent. As far as I have seen, most winders in this situation would put the groups that are physically 180-degrees-opposite in series. Their thought process is to keep the exciting current roughly the same in both parallels so that one doesn't have higher exciting current than the other as discussed by EASA. So ironically, it seems that that EASA approach of putting 180-opposite groups in series comes at the expense of higher UMP (dc) than the Dorrell approach of by putting physically-adjacent groups in series (not to mention that it may be tougher to run the jumpers around the endwinding to connect 180-opposite groups in series). Although there may be additional benefits of the EASA approach besides reducing current imbalance that were not mentioned in the tech note. Again please correct me if I'm wrong to assume that most winders would choose the 180-opposite groups to be in series as described by EASA.


=====================================
(2B)+(2B)' ?