The attached trend clearly shows appearance of 6xLF in motor vibration. The motor is powered accross the line. Although not verified directly by line voltage measurement, the timing of appearance coincides with converting some of the nearby motors on the grid to VFD and possibly polluting the grid.
I just wonder if someone has information or personal experience as far as how damaging is this vibration on the motor mechanically and also effect of line voltage distortion on motor operating parameters ( possibly additional energy losses, loss of torque, etc. )?
1 – Heating effects. Harmonics on the voltage input causes harmonics in the current which cause I^2*R heating but contribute no useful work. Also, there will be increaseed heating from several loss (heating) components that increase with frequency such as stray loss and core loss. If the motor is not operating near full load, likely the heating is not a big problem.
2 – Winding vibration. I would put this in the same category as 2*LF. It is relatively benign as a source of vibration (as compared to unbalance or worse misalignment). Yes, vibration can have several adverse effects on windings, but I don't think it can easily be predicted by bearing housing vibration monitoring. If I read right, you have < 0.03 ips. And even though I know you are reading the housing remote from where the winding is, I think it would be not much to worry about (my opinion).
Now, before we get too far, let's back up with another question. You are assuming the 6*LF is created on the input of the electronics and transferred to other machines fed from the same ac that feeds the electronics. But can 6*LF be created on the input side of electronics? My answer would be probably not. 6*LF occurs on dc output and on vfd output. But these electronic devices generally create 5*LF, 7*LF, 11*LF, 13*LF on their ac input. I would guess the 6*LF is likely getting to your accelerometer some other way as we have wondered about in other threads. (nearby electronic-fed equipment creates magnetic field interacts with mounting magnet, magnetic field causes EMI, or perhaps magnetic field that causes actual vibration of your motor).
At least that's my opinion. Maybe others can chime in.
This motor shows 21600 CPM and harmonics. This is typically seen in DC motors and results from rectification of 3 phase AC into ripple DC. Is this an AC or DC motor?. If it is DC, then there may be an issue in one of the control cards that has caused the growth in the 21600 vibration.
Posts: 165 | Location: US | Registered: 26 May 2005
... Winding vibration. If I read right, you have < 0.03 ips. And even though I know you are reading the housing remote from where the winding is, I think it would be not much to worry about.
.... You are assuming the 6*LF is created on the input of the electronics and transferred to other machines fed from the same ac that feeds the electronics. But can 6*LF be created on the input side of electronics? ...(nearby electronic-fed equipment creates magnetic field interacts with mounting magnet, magnetic field causes EMI...
Good points, but here are few comments.
Regarding mechanical effect on the motor: this is a very rigidly installed AC motor. As one can see, other vibrations are much smaller then that at 21600 cpm, so 0.03 ips should not be interpreted as low and not damaging as we accustomed to vibration numbers in general. This is vibration of an offensive component transmitted to the bearing housing or just due to EM force acting upon the rotor. What are the actual force values at play remains unknown.
Regarding 6xLF on the AC bus: I do not know the mechanism as to how 6xLF penetrates from vfd output to the input 60 hz bus, but it apparently finds its way. The detected 6xLF vibration is not EMI induced which has penetrated accelerometer circuit. The reason I claim that 21600 cpm is true vibration occured due to distorted SIN voltage waveshape of the grid comes from the additional data I have supplied in the attached file. It shows that vibration taken on the driven equipment bearing ( Figure 2) located 3 feet away from the motor shows about 15 folds less vibration at 21600 cpm. Stray magnetic field would've induced about the same amplitude.
If opportunity arises I'll capture spectrum of the 60 HZ grid.
Just to recap my opinion: 6*LF shows up on the output of electronic power supplies while the 5*LF and 7*LF show up on the ac input of electronic devices and propogates to other equipment fed from the same ac supplies.
Good idea to measure the ac power... would fill in another piece of the puzzle (and I believe it would confirm what I said although I can't particularly explain where your 6*LF vib comes from). Current would be easier to measure. Of course the current spectrum will have some influence from the motor itself which can draw 5*LF and 7*LF current even when connected to pure 60hz ac sindusoidal voltage. But if there were 6*LF on the input voltage it should show up on the current spectrum also (and my prediction is you will see no 6*LF in the current). Voltage measurement would be a more revealing measurement, but sometimes a little more difficult to get that measurement safely.
Here is a voltage and power spectrum from input side of a chiller drive running at 60 hz. The curser is set at the sixth harmonic. There is virtually nothing there. This is typical of drive operation. Typically, there should be no even numbered harmonics in the voltage or current. As Pete indicates, voltage and current spectra from the bus feeding the motor would shed a lot of light on the issue.
I have acquired voltage and current spectrum and as predicted the 6xLF component is quite low in both of them. The 6xLF peak is dominant in the vibration spectrum though.
I took additional vibration data ( with higher resolution ) accross the stator from IB to OB bearing location and it shows about twice as much vibration in the middle at 6xLF. It means that the stator body is flexing ( similar in nature to 2xLF vibration ).
A bump test showed that 21600 cpm, which is 3rd harmonic of 2xLF=7200cpm ( a range of small amplitude 2xLF harmonics is present as usual), is probably amplified due to resonance occuring at 23615 cpm ( 9% away ). Resonance peak is more pronounced in the middle of the stator. Prior to this occurance the resonance data has not been acquired.
Posts: 1423 | Location: Texas | Registered: 22 February 2005
I don't think it is harmonics of 2*LF such as might have been caused by looseness excited by 120hz magnetic force. The reason is that you have harmonics of 21000cpm. So 21000cpm is the forcing frequency.
I still think this is the result of nearby electronic-powerred equipment such as your vfd driven motors. Either vibration transmitted from the equipment or fields transmitted from the equipment.
An interesting experiment to get dig further would be to check the vib when the machine isn't running. If 21,600cpm is still there, you know it comes from the other equipment.
Or else a survey of this machine when nearby equipment is not running.
Originally posted by electricpete: The reason is that you have harmonics of 21600cpm. So 21600cpm is the forcing frequency.
Either vibration transmitted from the equipment or fields transmitted from the equipment.
Actually there are harmonics of 7200 cpm present, not 21600 cpm. They are clearly visible in log scale. (I could not upload the file. Will try later). The only harmonic sticking out is the 21600 cpm one. A bump test revealed a closely situated resonance ( 9% away). So, the puzzle almost seems to come together. The only possible rub may be that at 9% away from the resonance the amplification factor (AF) is not that high to produce what we are seeing. I''ll calculate AF from the bump test.
In regards to stray magnetic field which could induce voltage in the accelerometer I can say that a measurement taken on the driven equipment just across coupling did not show 21600 cpm.
Posts: 1423 | Location: Texas | Registered: 22 February 2005
I was referring to the waterfall plot in your first post,. It looks like you have 1x 21600, 2x 21,600 and 4*21,600. But I can't read the frequencies exactly. Are those peaks out near 2x 21,600 and 4*21,600 actually some other frequency?
Attached are results from a bump test. It does not appear from it that 21600 cpm is close to resonance.
Taking readings when this machine is off is not feasable so far. In regards to induced voltage or transfered background vibration, note that there is almost no peak at 21600 cpm on the bearing just accross the coupling ( slide 5)
