Dear All,

We have ID FAN with the specification Double-Stage Axial Flow Fan, Speed 750 RPM no variable speed drive. Each stage has 20 Blades, and each blade can rotate to control flow.

This is newly installed ID FAN, Before comissioning we did balancing, the 1xRPM reduce from 2.6 mm/s to 0.1 mm/s (Horz NDE, Opening Blade 0%) and when commissioning the vibration was fluctuative depends on opening blade pitch (See Attached File 1).

When there is Load the Vibration also very fluctuative (See Attached File 2), We see it as problem because the other ID FAN with same specification have stable vibration even the load change (opening blade change).

We suspect structure problem or flow problem, and the we try to take measurement on the casing and open the insulation. then suddenly the vibration dropped from 2.3 mm/s to 1.3 mm/s and it become stable even the load change. After we install back the insulation the vibration increase and it became unstable again (See Attached File 3).

Because there is no change on parameter process and the other id fan also stable,  we don’t suspect flow problem anymore.

So what do you guys think? Is it Resonance?

Best Regards



Original Post

I would not expect much of change in fan case natural frequency with removal of thermal lagging (mineral wool and aluminum jacket) from mass change; especially near 12.5 Hz. A simple impact test could be conducted by removing small areas of lagging for access to casing. The impulse-response measurements could be made with fan off or running (negative averaging method).

Another explanation for the change in vibrations with that large area of lagging removed is that there was a temperature effect. Bolted joint stiffness can change with temperature. Bearing support stiffness can change with temperature. Are any temperatures (bearings, bearings, oil, etc.) measured that showed a change?

If you balanced fan rotor when it was essentially cold, then you can expect balance might change when hot. Did you follow the 1xSS amplitude and phase change at each bearing from cold to hot and no load to full load?


We would try to find natural freq when there is opportunity to shut-off the FAN. Today we tried to use negative averaging method with CSI2140 with just small rubber hammer and the result was 12.4 Hz(??), I don't know if the result valid or not.

We tried bump test on smaller equipment when running and off and the result quite valid. but for big structure like ID FAN i dont know how much impact do we need.

There is no significant change on any temperature, the average was 50-60 C

For now we tried to remove thermal lagging layer by layer (there is 3 layer mineral wool) and see vibration trending.

Thanks Walt and fburgos,


Images (1)

The DE phase angle was about 340-deg. throughout the blade angle stroke, while the amplitude increased from about 0.5 to 0.9 mm/sec.

The NDE phase angle had a lot of variation (was all over the compass), while the amplitude increased from about 0.1 to 1.2 mm/sec. as blade pitch opened.

This could indicate looseness in 2nd stage hub fit or blades or pitch control that affected rotor unbalance. It could indicate poor/loose fit of NDE bearing. Also the no-load spectra shows sidebands of 1xSS around the blade pass frequency indicating eccentric motion at 1xSS or a single blade that does not have the same pitch as the others.

I assume by no load that you mean a discharge damper was closed. If this was the case, then as the blade pitch was being opened the rotor was probably heating up and increasing rotor deflection and vibration at the heavy spot.

The correlation between vibration levels and with/without lagging remains a mystery. If the thermal lagging was also designed for sound reduction, then typically a septum of sheet lead (or material with equal weight) is placed between the mineral wool layers or bonded to the aluminum jacket. This could add 1 to 2 lbs. per square foot of area and make the total weight more interesting.

Get a bigger impact hammer or at least 5-lb head with rubber impact tip!


Last edited by Walt Strong

Add Reply

Link copied to your clipboard.