Natural Frequencies on machines at rest vs. operating speed 1

[skfboiler]

On a typical motor driven fan mounted on rolling element bearngs, does the natural frequency change when the machine reaches its operating speed?

 

[GregLocock]

No

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[GregLocock]

...or at least, not to within an amount that makes much difference.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[electricpete]

If the motion of the mode of interest involves tilting of a thin disk, there can be a big difference.

When rotating, the gyroscopic effect increases natural frequency (compared to lump mass at location of the disk).

When stationary (for example lateral bump test), the rotational inertia decreases the natural frequency (compared to lumped mass at location of the disk).

=====================================
(2B)+(2B)' ?

 

[electricpete]

Attached is an illustration of what I was discussing. Between bearings machine with disk attached.

In this case there is no difference for 1st mode, but 2nd mode is 162hz during stationary bump test mode vs 244hz 2nd mode while rotating.

Slide 1: Graphical Rotor Geometry (steel)
Slide 2: Numerical Rotor Description (SI Units)
Slide 3: Bump Test while stationary incl tilting disk effects: F2 = 162hz
Slide 4: Treat disk as lumped mass with no disk effect. F2 = 193HZ
Slide 5: Rotating Critical Speed. F2 = 244hz


=====================================
(2B)+(2B)' ?

 

[electricpete]

If the disk-like attachment happens to be overhung beyond the bearings, then the effect on 1st mode can be large.

See roman numeral VII at the bottom here:

http://home.comcast.net/~electricpete/rotosolve/xfmatdemos1.htm

=====================================
(2B)+(2B)' ?

 

[MikeyP]

Hmmm. I think that most fans will have most of their mass concentrated at the hub and will not be rotating *that* fast (I am assuming something like a desk fan here).

M

--
Dr Michael F Platten

 

[electricpete]

A fan with all its mass at the hub wouldn’t push much air to my way of thinking. The gyroscopic effect is certainly is not relevant for all fans, but on the other hand it is certainly very relevant for many industrial fans (particularly if there is high D/L ratio of fan wheel and assymetrically positiong with respect to the bearings).

The 12 Dec 10 20:05 posted between bearings example was done in 5 minutes to show the principle, but if it would make you feel better I can surely change the parameters to lower the frequency of the 2nd critical near 60hz (let me know if you want to see that).

The linked overhung fan 12 Dec 10 20:14 is model of a real industrial fan.

Attached is an article of a well-studied overhung fan. On the bottom of page 39: you’ll see:

Gyroscopic forces and moments acting on an overhung fan wheel can increase the first lateral mode by a big amount—in the authors’ case, by 37.5 percent (by 625 cpm—to 2290 cpm at 2080 rpm from 1668 cpm at rest).

=====================================
(2B)+(2B)' ?

 

[electricpete]

In case it wasn’t clear, my point is: no-one can say without knowing the details of the fan whether the gyroscopic effect is relevant or not. Would like to hear more details on the fan.

=====================================
(2B)+(2B)' ?

 

[FeX32]

That's an interesting point Pete.
I would have thought that the gyroscopic forces played a larger role in increasing the damping coefficient. Moreover, as the gyroscopic effect effectively changes the reaction axis to the mode shape, would not the mode shape itself change?


Fe

 

[electricpete]

I would have thought that the gyroscopic forces played a larger role in increasing the damping coefficient.

Most non-rotating vibration problems involve symmetric K and C matrix. In that case the K matrix represents conservative forces (associated with energy storage) and the C matrix represents non-conservative (damping associated with energy dissipation) forces.

In rotating equipment there are a number of scenario’s where skew-symmetric components are added to the K and C matrix and the role of these skew symmetric terms is reversed (skew symmetric terms in the K matrix represent non-conservative/damping forces and skew symmetric terms in the C matrix represent conservative forces).

An example of skew symmetric terms in the K matrix would be in a fluid film bearing with cross-coupled stiffness term kxy = -kyx. This can add or remove energy from depending on direction of rotation of the mode compared to the machine.

An example of skew symmetric term in the C matrix would be the gyroscopic terms +/-w*Ip. It represents conservative force because grysoscopic effect represents stored kinetic energy only and therefore cannot possibly add or remove energy from the system. Therefore it is not damping.

Moreover, as the gyroscopic effect effectively changes the reaction axis to the mode shape, would not the mode shape itself change

Yes, the mode shape changes. I never said otherwise… in fact showed the changing modeshape in my attachment.


=====================================
(2B)+(2B)' ?

 

[FeX32]

Thanks Pete,
I should have opened the attachment .

Also, I like you explanation. I now see why the damping is not effected.

Gyroscopic coupled vibration would make for some very interesting experiments. Could be used in vibration isolation to actively change the resonant frequencies ect.


Fe

 

[electricpete]

Thanks. My apologies to everyone if my tone was gruff or worse. I have messed around with modeling vibration of machines, but am by no means an expert and can learn a lot from everyone here.

=====================================
(2B)+(2B)' ?

 

[MachineryWatch]

Great question and an eye opening discussion. Thanks for the time you spent on your answers.

http://www.machinerywatch.com