In the attached plot it is obvious that this machine is operating above critical speed. It is actually operating in close proximity to the next critical speed of 1350 RPM based on the impact test but this is not the point I'd like to make.

Here is the question in this regard.

Does operation above a critical speed necessarily indicate that a rotor takes certain shape, such as a bow-shape or S-shape depending on bearing/mounting stiffness vs. shaft stiffness and number of critical speeds it passed?

What if the speed at which diameter of the polar plot circle is at maximum (which we call "critical speed" ) is associated not with the shape of a rotor but rather to another component in the system ( even stationary ) coming into a resonance ?

Thanks,

David

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

Nyquist_plot.doc (48 Kb, 73 downloads)

quote:

Does operation above a critical speed necessarily indicates the fact that the rotor takes shape of a bow shape or S-shape depending on bearing/mounting stiffness and number of critical speeds it passed?

My two cents fwiw: As a simplified generalization, I believe the operating deflection will resemble the modeshape of the closest critical speed. The first critical might be a flex rotor or rigid rotor critical. If you were just above the first critical which was a rigid rotor critical and no other criticals in sight, there would be very little bow of the shaft. If you were just above the first critical which was a flexible rotor critical and no other criticals in sight, the rotor would be bowed. If you were near the second flexible rotor critical, you can expect the S-shaped rotor deformation.

A friend of mine has explained this situation in the following way:

"If the measured vibration is from a structure other than the rotor, then the rotor may not have passed through a critical. There is a difference between critical speed and resonance. The rotor has to resonate at a certain speed in order for it to be a critical speed. However, vibration due to rotor speed can cause other structures to resonate. Most of the time when a machine has a resonance problem, it is not the rotor but another part of the structure that is being resonated by rotor speed. If this is the case, the rotor shape will not change. The Nyquist plot could reflect the resonance of a machine component other than the rotor. Even though there is a phase shift and resonance, the other structure has not passed through a critical because it isn't rotating. It is just resonating and can take on any bending mode allowed by its geometry."

I have to agree with the above and add that in this case I have acquired seismic vibration data on a bearing secured to the frame of a hammer mill. Obviously every structural component will affect the readings.

I guess a conclusion to make is as follows: don't use seismic coast down data in order to assess where a rotor is running as far as critical speeds or mode shapes are concerned.

Sorry if I just have reinvented the wheel.

David

Dave,
You mentioned that this is a hammermill. They usually have very stiff shafts, so I would tend to believe that you are encountering support modes. If you want to really know what these modes represent, get yourself some rotor dynamics modeling software, put in the shaft dimensions and material properties, lump on the weights and inertias of the hammers and bolts, and viola!!!
Seriously, usually there are two modes you will encounter between stop and running speed. These are rigid body modes, meaning the rotor is NOT bending, just bouncing in it's bearings. The first rigid body mode has the bearings in phase and the second rigid body mode has the ends out-of-phase. Again, since these are not bending modes, the only thing that affects them is the stiffness from the bearing, to the concrete that supports the machine.

David,
The nyquist plot shows that you have a structural resonance and the rotor operates under the first critical. The imbalance stays within 40 degrees from start-up.

The impact data needs to be improved. You have a 2 second waveform with an impact event of 0.2 seconds. You need to lower the resolution or increase the FMAX. Try 800Hz Fmax with 400 lines of resolution, that will get you one second of data. Also use a dead blow hammer and lightly tap the rotor. That should help resolve the spectrum.

Pete, Ron, Erik,

I agree with you all. The only thing I would like to say is that if somebody looked at this Nyquist plot without knowning as to how data has been acquired, he would've concluded that this machine is operating above 1st critical of the rotor with associated rotor modeshape, regardless of mode shape type.

In reality though, IMO, the circle in the plot represents just a resonance somewhere in the structure and has nothing to do with the changing relationship between the heavy and high spot occuring normally in a rotor going through a critical speed.

Erik, I agree, the data collection setup for the bump test could have been improved although I think it wouldn't have changed the spectral pattern to a degree that that may reveal any additional information.

David

The support bearings and the shaft are 3 springs connected in series. Effective system stiffness Keff is calculated as:

1 / Keff = 1 / (2*Kbrg) + 1 / Kshaft

System resonance and rotor speed when it occurs will be defined by system effective stiffness Keff and system mass M.
“Before resonance / at resonance/ after resonance” angle relationship between the heavy and high spots changes from 0 to 90 to 180 deg. If bearings are much stiffer then that of the shaft then the rotor will acquire a bow-shape at this so called 1st critical speed, otherwise – shaft’s mode shape is a straight line ( rigid shaft ) experiencing translational motion.

So, in this particular case (assuming the loop represents system a resonance) according to the theory, the machine is running above the 1st critical with heavy-high spot angle difference of 180 deg.
In fact the loop in Nyquist plot shows an almost 180 deg angle change.

Dave

To what do we attribute the dramatic steady increase in the bump test response (acceleration) between 4000 and 6000 cpm in the top plot of Figure 2? It looks like it increases by at least a factor of 4 in this range.

quote:

Originally posted by electricpete:
To what do we attribute the dramatic steady increase in the bump test response (acceleration) between 4000 and 6000 cpm in the top plot of Figure 2?

Same impact plot in velocity units does not look that dramatic.

Also, how one would know as to what particular component to attribute it to if it is structure related? Could it be that this is due to the noise as a result of improper set up parameters in the TWF Erik was talking about?

Nyquist_plot2.doc (38 Kb, 12 downloads)

quote:

I guess a conclusion to make is as follows: don't use seismic coast down data in order to assess where a rotor is running as far as critical speeds or mode shapes are concerned.

What I get out of this data is there is something resonant in the system that may give someone trouble at some point. Is that what you are saying?

Regards

Mick,
I think that in general what Dave_man is saying is correct. But if there is a component other then bearing mount in a machine which resonates at certain machine speed, thus creating a false impression of bearing-shaft system resonance in the Nyquist or Bode plot, then a wrong conclusion could be made for instance as far as heavy-high spot relation on the rotor is concerned. Some other vital conclusions will also be wrong.

David