Cyclic radial vibration in a vertical motor driving a pump

[edison123]

A 1.5 MW, 500 RPM motor drives a centrifugal pump delivering sea water for generator cooling. The motor sits on a 2.5 meter high stool, which is anchored to a steel base, which is embedded in concrete and the pump is directly connected to the motor. The pump has 3 water cooled guide rubber bearings, which are not visible being underground.

In decoupled condition, the motor radial vibration at both top and bottom is less than 1.0 mm/sec rms. When the pump is coupled, the motor radial vibration is cyclic going from 1.2 mm/sec to 3.6 mm/sec rms in east-west direction (with 1XRPM predominant) but only 0.5-0.6 mm/sec rms in north-south direction. The tangential vibration is also high at around 4.5 mm/sec rms. The axial vibration is less than 0.5-0.7 mm/sec rms in both decoupled and coupled conditions.

The GA sketch of the motor-pump setup is attached.

I look forward to your ideas about why the coupled radial vibration is cyclic and that too why only in one direction. Also, is tangential vibration such an important measure in vertical machines?

Thanks.

Muthu
www.edison.co.in

 

[Tmoose]

Is this a new installation?
Or, has the high vibration developed over time or recently?
Does the steel base plate have anchor bolts and jack/leveling bolts.
Are there any pictures of the installation, to show how the steel plate was leveled before grouting?

Vertical pumps frequently have first bending mode resonant frequencies in the operating range. Hydraulic institute used to call it the "reed" frequency. With the piping and asymmetrical structure there usually wide variations in the resonance in various directions.

"2.5 meter high stool, which is anchored to a steel base, which is embedded in concrete".

It was common for pump manufacturers to calculate reed frequencies based on "rigid" boundary conditions, and tuck a disclaimer in the specs or contract saying they could not be responsible for resonance at operating speed due to installation. That is understandable if only to caution against poor workmanship in that critical area. But even well done installations are far from rigid, so the installed resonant frequencies are going to be a bit lower numerically than calculated.

What is the phase of the 1X vibration in the good and bad directions? Does it vary or swing when the vibration is cycling?
It // should // be ~90 degrees. If more like 0 or 180 resonance is suspected.

Is there any perceptible motion (fingertip, and finger nail) at any of those locations? Is there a difference across any bolted or "embedded" interface? For instance, the steel base plate and any of its anchor bolts?

I'd measure and record the 1X vibration amplitude and phase of the stool mounting feet, those anchor bolts, the steel base plate near the stool attachment points, the perimeter of the steel base plate, the steel base plate near its anchor bolts, the base plate anchor bolts themselves, and the concrete surrounding the base plate. See attached image for locatoins.


I'd do a "bump test" on the stopped pump in both directions to look for resonant frequencies. I take a back ground measurements to subtract from the "bump test" result.