I performed analysis today on a 100hp direct driven pump (VRD powered) that shows high 1x rotation vibration (.26 ips) on both the A & B bearings in the horizontal direction at 100%. All other directions are low and the C & D bearings remain low no matter what the operating parameters are. No other harmonics or defects are present in the spectra.

Phase analysis accross the coupling is ok, no noticeable phase shift to indicate alignment problems or 2x rotational speed.

One thing I did notice was the mechanic used a half key on the motor side of the coupling. When the machine was worked on during the last shutdown the motor was run un-coupled and ran rough. Someone told him to run the un-coupled motor with a half key which smoothed out the vibration. This is why the half key remained on the unit during re-assembly.

I know there is an unbalance condition in this machine. The problem is when the VFD is at 90% the vibration at A & B in the horizontal direction jumps to .67 ips. Lower than 90% vibration drops with every reduction in VFD frequency.

As speed decreases should the amplitude be more linear and drop with speed? Could the half key cause this response at certain operating speeds? Could this be a resonance that we are encountering?

I only recorded phase at 100%. I plan on taking more phase readings at different speeds. In the horizontal direction will the phase remain the same at different speeds? If there is a phase shift can we lean more towards a resonance issue?

I also plan on taking bump tests to try and identify naturals.

Any sugestions will be apprecaited. Thanks in advance.

J Gorman

I guess the A & B bearings are the motor bearings? And the vibration peaked at 90% and was lower at <90% and >90%? Sounds like resonance at 90% motor speed. I don't think the half-key has anything to do with it. I would think the half-key would fill the void better. And as far as getting a linear, or anything close to linear, response, a lot depends on the system. Not too many things are linear.

In theory, when changing speed far below resonance, the velocity response to a constant unbalance decreases as speed^3. i.e. when decrease speed by a factor of 1/2 the inches per second should decrease by a facgtor of 1/8. If the upper speed was below but somewhere close to resonance the ratio would be higher. I agree with Patrick that machines don't always read the books.

I'm not sure I understood the discussion on half key. I would be surprised if the same key used during uncoupled run could be used when coupled. Stepped key would make sense but not half key. Also as has been discussed before another common approach (instead of stepped key) is that a full-depth key should normally extend approx 1/2 the length of the exposed keyway.

This message has been edited. Last edited by: electricpete,

Lawrencep yes the A & B are the motor bearings. Again the horizontal direction is the highest. I am leaning more towards the resonance problem now. Tomorrow I will try and identify if the motor or base has a natural frequency that is causing the vibartion at the 1605 cpm. From there I might try and calculate dynamic absorber. Has anyone had luck in reducing vibration with one? Seen this vibatrion is so directional (horizontal) would you mount the absorber horizontal outward from the motor base?

Why consider a dynamic absorber for a variable speed machine, since it would only work at one speed and potentially introduce two new resonant frequencies in the speed range? I suggest a natural frequency test by impact hammer or an ODS test at the 90% speed. If you map the structural deflection, then it is possible to determine if the motion is shear or rocking in the horizontal direction. I suggest adding stiffness (instead of mass) to increase the natural frequency from 90% to over 110% service speed.

Thanks Walt! I like the idea of stiffness verses mass. I think adding stiffness might be an easier option. Also I did not consider or know correcting one resonance could possibly induce 2 new ones.

Thanks

[QUOTE]Originally posted by electricpete:
In theory, when changing speed far below resonance, the velocity response to a constant unbalance decreases as speed^3. i.e. when decrease speed by a factor of 1/2 the inches per second should decrease by a facgtor of 1/8. If the upper speed was below but somewhere close to resonance the ratio would be higher. I agree with Patrick that machines don't always read the books.QUOTE]

Pete - Why is the response proportional to speed^3? I always thought it was speed^2.

I was working on the pump this morning and discovered a natural frequency (using a bump test) of the motor plate. This frequency is at 1590 and the pump when opertaing near or at this frequency vibrates bad. The further away from 1590 up or down the the vibration smooths out.

I tried stiffening the motor plate with angle iron and t-bar in every way imaginable with no success on moving or raising this natural frequency.

I beleive my next step is to add weight and try to move the frequency hopefully downward out of the operating range. This would be an attempt before a total redesign of the motor mount.

Has anyone been successful with a vibration absorber?

quote:

In theory, when changing speed far below resonance, the velocity response to a constant unbalance decreases as speed^3...

Pete - Why is the response proportional to speed^3? I always thought it was speed^2.

The unbalance force is proportional to speed squared. The velocity transfer function is proportional to speed to the first when below resonance. Velocity response is product of force times velocity transfer function ~ speed^2 * speed^1 ~ speed ^3

Above was talking about velocity. The more familiar transfer function is displacement.
H(s) = (1/m)* (1 /[s^2 + s c/m + k/m])
Far below resonance s << sqrt(k/m) the term k/m dominates the denominator and as frequency s=jw changes the tranfer function is approximatley constant with speed. So if we were talking about displacement far below resonance the transfer function is constant and the displacement response to unbalance would be proportional to speed^2.

Since velocity is derivative of diplacement, |v| ~ f*|d| and transfer function for velocity is power of 1 times speed higher than transfer function for displacement. i.e. transfer function for velocity proportional to speed^1 (when below resonance). When we multiply that by unbalance force~speed^2 we get velocity~speed^3.

This message has been edited. Last edited by: electricpete,