I have a vibration problem with an industrial overhung fan. The vibration values measured when mounted in silent blocks is unacceptable while when it is attached by bolts directly to the floor, the 1X peak become acceptable.

The phase readings taken suggest unbalance as cross phase between horizontal and vertical is 98deg. All the components had been shop balanced. I haven't tried to balance in-situ because the rotation speed is 3800 rpm (58.4 Hz)and the trial mass didn't remain fixed after trial run. The rotor is operating above its first critical.

The FRF comparison between fixed and mounted in silent blocks also doesn't justify the vibration value differences.

Please see the attached file and give your opinion.
Thanks

PO.doc (246 Kb, 99 downloads)

I take it you want it to run on the "silent blocks", huh? I guess a dumb question on my part is why?

What happens when the power is cut while monitoring the coastdown on the "silent blocks"? Does the amplitude gradually decrease or does it reach a frequency where the amplitude drops drastically, like maybe the slient blocks are resonanting it?

Rubipako,
In a brief reveiw of your data, I see what one should expect to see. When mounted on the blocks, the vertical amplitudes increase. this is normal. When you add in the fact that your modal analysis shows a resonance around turning speed when using the blocks, then your issue is even more evident. The easiest course of action at this point, assuming you cannot use an alternate isolation block with a differnt natural frequency, would be an in-situ balance. If you reduce the driving force of your resonance, then your going to reduce the effect of the resonance on the system. Keep in mind though that you will be incurring a 90* phase lag in your system due to operation at a resonant frequency.

Best of Luck,
Nicholas A. Martin

I tried to balance in one plane because the width of the rotor is small compared with the diameter. I have reduced the vibration level in one bearing to 1mm/s, but the other bearing increased a lot up to 22 mm/s. I suppose I need to balance in two planes. The problem is that the width is so small that a trial mass in one plane affects also to the other plane.

What can I do?

quote:

The rotor is operating above its first critical.

While on the slient blocks or bolted to the floor?

How far above critical? Where is the second critical?

Seems like if it is OK bolted down to the floor, then the silent pads are resonant at running speed, Difficult to balance and keep in balance when running at ANY critical, not necessarily the rotor's critical or shaft critical but even the silent pad's critical. Sounds like the pads are making the entire system resonant at running speed. Have you done an impact test while on the pads?

This message has been edited. Last edited by: Ralph Stewart,

If you are operating close to a bending mode of the shaft, I highly recommend trim balancing, at speed, in the housing, but do not use a balancing program with phase information. The problem with that is the phase changes you are expecting are influenced by the fact that you are dealing with a 180 degree phase shift due to the critical.
Use the four run (or as Arn would say, three run, save one run) balancing method. Same weight at three locations separated by 120 degrees and one original. The vectors will accurately pinpoint where your trial weight will go.
As far as keeping the weight on the fan, you have two choices. You should be able to get a spring steel clip balance weight on the inside edge of the fan blades going through the inlet. If you can't, and the wheel is steel, then tack weld on a small washer as a trial weight, pry it off, move it and go again.
At this speed, you shouldn't need much weight. I just balanced a 5 foot diameter fan wheel at 3585 rpm and 60 grams brought it down from 2.5 mils to 0.2 mils.
Good luck!

quote:

Disposition with 7 silent block is made in order to support each mount the same load taking into account the center of gravity. 12 silent mount disposition is uniform distribution along the structure.

The plots and data don't suggest that you have obtained isolation. The blocks (number of blocks affects this too) may be too stiff.

As to the rotor bending, that doesn't look like the major motion (resolution not great). Also, since you have the rotor, I have to assume you did this stationary; the overhung fan forward resonance will increase while spinning.

The force most likely is comming from imbalance (the sheaves can also cause vibration on the fan). Balancing may help (check the hubs for runout, too).

Don't balance without phase; this is a bad idea, as has been discussed here previously. The minor resonances around running speed may not be important - this is what peaks are in the FRF, unless the modeshape corresponds to one of these? I've balanced these with the mounts and with blocking the mounts - both methods work (if the vibration is high enough to bottom out the mounts, then that could cause trouble during a balance, but not using phase doesn't help).

How/why did you decide to install these isolators? Did you start by looking at the weight of the equipment compared to what the vendor advise is? How much sag on the isolators did you achieve?

After the bouncing mode is taken care of by considering weight, there are other modes to cosider, like rocking, which this seems to be doing.

I've never seen exactly this type of mount, but my first question would be why so many? The fan doesn't look heavy enough to need so many.

I also don't see the need to balance without phase. If you take the data properly and input the correct information to your balancing software, it doesn't matter what the "phase shift" is (i.e., where you are on the resonance curve). As far as balancing, I would first balance using only the outboard (fan end) bearing, placing the weight on the rotor as close as I could get to the bearing. When that bearing is at a low value, I would then balance the sheave using data taken at the inboard (sheave end) bearing. The easiest way to do this is with a two-part, steel "stick" putty, assuming there is an adequate "rim" on the sheave for the putty to lay against (centrifugal force will hold it in place). You just cut off a length of the putty and press it into place on the rim of the sheave, without mixing it. If it is in the wrong position, you can lift it off with a putty knife and add or subtract putty as needed. When you have the final amount and position right, take a marker and mark the outline of the putty on the sheave (it will flatten out as you run). Then mix it up (knead it, best with rubber gloves) and put it back in place being sure to work it back to the line you marked. The sheave balance will have very little effect on the outboard (fan end) bearing.
This is really not dangerous if done properly. You of course want to stay away from the sheave/belt side of the unit if you leave the guard off while balancing the sheave. You can of course put the guard back on before each run if you want to.

But I still think there is a problem with the isolator design.

Hi,
I have added new information to attached file. I see that most of you suggest that the problem is related to the isolators. These are only for isolate vibrations from the machine to the floor. I don´t see that the blocks are resonating at running speed. The natural frequency of the pads is 8.5Hz.

As I posted before, trying to balance decreased vibration in on bearing but increased in the other. As 'rustythevibeguy' said balancing in the sheave to decrease the other value may help. I haven't tried for now.

Bill, I don't understand when you say that other modes like rocking may be taken place. And, of course, the modal analysis was carried out stationary. If the frequency increases when spinning, then the 54.4Hz mode may be closer to the running speed(58.4Hz) and that could be one of the reasons of the growing of vibration. The same mode shape when bolted to the floor is at 49Hz. This is far from spindle speed so that doesn't affect so much and the vibration level become acceptable.

The number of blocks and its stiffness is the recommended by the vendor taking into account the total weight and the position of its center of gravity.

I am going to take the posibity to vary the speed of the motor, so that I would be able to change the rotor's turning speed. I would see then if the vibration level increases quadratically with the speed or not and could determine the exact frequency of the resonance.

thank you everybody for your comments.

PO(2).doc (252 Kb, 17 downloads)

It´s not uncommon for the isolator supplier to only calculate one direction of freedom, the vertical, leaving the other rocking modes etc. free to end up close to running speed, giving us poor consultants a healthy, steady and good source of income. So is the mount complete with machine on the isolators having a natural frequency in all 6 directions of freedom at 8.5Hz? I would not think so. Brutal way to verify this is to remove or change a couple of the isolators. Olov

Rubipako,

I wonder whether 54.4 HZ is really a resonance frequency confirmed in your modal analysis by phase in FRF function. Please post it.

If it is not a resonance, then existing vibration with the isolators active may be normal and could be reduced by balancing. You just have to balance out couple unbalance causing high vibration on the sheave bearing using known methods.

David

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

Yes, the 54.4 Hz resonance is confirmed by phase in the FRF.