Strange vibration problem

[geesamand]

I'm working with two similar machines showing massively excessive vibration (50mils vs. 2.7mil limit). It's a VFD-reduced double-reduction gear drive mounted on a tank.

Regardless of motor speed, the vibration is showing very strongly at 1x motor speed. 2x, 3x, 4x are measureable but small. There are no resonances near the motor speeds that show max amplitude. We realigned coupling alignment with a laser alignment tool. Motor ran smoothly when decoupled (and unloaded). We improved the stiffness of the motor-gearbox connection.

I have heard of cases where flimsy mounting will cause normal excitation sources to result in excessive vibration, but I'm puzzled why 1x motor speed is the *only* vibration showing and why it's so intense. While I can see a stiffer mount reducing amplitude, intuitively it seems like this is not going after the source of the vibration energy - true?

Has anyone else seen such intense response at motor rpm due to insufficient mounting stiffness? Insights appreciated.

 

[Tmoose]

Are the motor or gearbox on isolation mounts (generally yikes)? Is the motor on a "scoop" type mount, so it is cantilevered in space? Is the vibration similar ampitude at both ends, and in "vertical/horizontal" directions, with appropriately different phases?

What type of coupling? The rubber element types are inherently bent, forming an offset crank between the driving and driven machines.
I'd rig a dial indicator up comparing "ground" to the motor motion and turn the motor by hand. If it's cranking 50 MILS statically, it's not going to get any better at speed.

Disc couplings tightened up in the misaligned state cause a similar cranking action. Escess coupling capacity makes it worse.
If the mounting is flexible the 1X amplitude (in MILS or mm) will be pretty constant from 1 rpm on up. Add some resonance, and some speeds will have tasty 10 X higher vibration.

"Tank mounted" sounds suspiciously flexible to me at the start.

 

[geesamand]

Tmoose, thanks for the snappy response.

The motor is a scoop mount. No isolation mounts right now. I've done some reading and the spectral plots are similar to those of soft foot, but I'm not convinced that a scoop would have the rigidity to pull a motor frame into misalignment. (We'll check for soft foot anyway)

The vibration is similar on various parts of the motor and motor scoop. It has a vertical support, and adding more bracing hasn't done much. We didn't check phase - I don't have experience yet with dual probes. (vib unit is a CSI 21/20, if it even has that capability)

Coupling is a new Lovejoy C276 Hytrel jaw type. It was laser aligned to .004" parallel, but it wasn't verified using another method.

I do suspect that the tank mounting is part of the issue. The tank feet are load sensors at the bottom of the tank. Tank stands maybe 10-20' tall. Even if the tank flexibility plays a major role, I would prefer to eliminate any/all possible sources first. Could the tank flexibility alone cause this kind of vibration at 1x?

 

[GregLocock]

Can you plot 1x vibration vs motor speed?

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[geesamand]

The vibration is always at motor speed. The amplitude is worst at 40% turndown, but all speed show the primary vibration energy equals motor speed.

Dave

 

[GregLocock]

Can you post your plot of 1x vibration vs motor speed?

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[TPL]

What transducers are you using to measure vibration displacement?

If you are using proximity probes, you could be looking at glitch or runout rather than true vibration

 

[geesamand]

I don't have a plot handy. The guy with the vibration meter software is out of the office right now.

The probe is a magnetic accelerometer. I don't have the details handy but we've used this probe extensively on many machines. And it's quite obvious the unit is operating with destructive levels of vibration.

Our last set of data looks like this, at the NDE of the motor (worst case):
Motor speed Vibration
(rpm) (pk-pk, mils)
270 1.38
865 12.0
1077 50.12
1494 13.9

Prior to adding some supports to the system:
427 4.82
851 70.2
1486 16.49

Are you suggesting we take readings at more motor speeds, or do a capture during coast-down?

Thanks everyone

 

[GregLocock]

Looks to me as though you've got a mounting resonance, which you've successfully modified quite a bit with your bracing, excited by out of balance in the input shaft of the gearbox or more likely runout or misalignment.

Yes, I'd do a coastdown waterfall plot. I suggets you do it with and without the gearbox coupled, just to make sure that it isn't a problem with the motor

Increasing the stiffness by a further factor of two or so might well be a challenge, the trouble is you are going to make things worse before they get better, ie you'll be pushng the resonance up in speed where there is more energy.

Sorting the forcing out is likely to be a much more satisfactory approach.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[geesamand]

We searched exhaustively for resonances. None came up around 800-1000cpm. The highest freq resonances were at 565 and 675cpm. That's why I have looked away from resonance. Could the ring test have missed something?

 

[GregLocock]

You had a peak at 851 that you moved up to 1077 with structural mods, so unless you are really unlucky you've found the resonance. a 20% shift is quite significant.

Admittedly that theory doesn't square with your observation that the vibration is the same (magnitude?) at all locations and directions.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.