What do ya'll think about this orbit. It is off of a turbine bearing running at 3600 rpm. This is my first time posting an attachment, so if it doesn't show up, I might need some help posting it.
The sensitivity is 200 mv/mil. So each gridline is 1 mil.

In saying that, which do you think would be better using a shaft rider or prox.probes on turbine bearings. We have both and the prox. probes are showing lower vibration readings. Which one would you think would be correct.

Thanks in advance.

Rafael

orbit.doc (46 Kb, 120 downloads) Orbit

The orbit looks fairly typical of what I see on turbine generators.

Normally shaft riders and prox probes are measuring different parameters and this is why they don't always agree. The shaft rider measures absolute shaft vibration (shaft vibration relative to earth) while the prox probe measures shaft vibration relative to its mounting location which could be the bearing or the housing. In many cases, a prox probe output is combined with a seismic transducer output to give a shaft absolute reading which would be more comparable to the shaft rider reading. However, before you could make a direct comparison, you would have to make sure that the shaft rider and the proximity probe system were mounted at the same radial angle.

Shaft riders have some nice features, but they are dated technology and should be replaced by modern prox probe systems. For plants that replace shaft riders with prox probe systems, I recommend, at a minimum, that access be provided to the old shaft rider locations. This will help to evaluate prox probe runout/glitch and give a consistent location for taking data to use when balancing the unit while new data is developed using the new probes. Many times, old balance data is perfectly reliable, but is based on shaft rider data which is not directly comparable to the new prox probe system -- as you are finding out.

Another consideration is that the manufacturer's alarm recommendations are typically based on shaft rider data which is not really directly comparable to prox probe data -- I know, the point of your post. Most plants with which I am familiar ignore this subtle point and just use the original alarm values anyway.

Since you now have both shaft riders and prox probes, you have the best of the old and the new, and I would maintain that situation. I would limit the information available to the operators to one system or the other and make sure that what they saw was reliable.

Michael Titone

Rafael,

Not certain that the orbit is telling you much beyond amplitudes. Can't judge precession as it like a keyphasor mark and seems to be multiple revolutions. If orientated properly it looks like some slight vertical preload but probably not abnormal considering gravity.

One thing I see is a lot of "inward" and "outward" orientated spikes on the orbit. This could be something as simple as surface finish but you should examine your time domain for random positive and negative going spikes which could support electrostatic discharge.

I don't know much about it but I have a few questions for my own benefit.

1 - Aren't the bearing clearance needed to help judge the magnitudes? Is the distance 8 mils from top to bottom of the plot supposed to represent the the bearing diamterifcal clearance
2 - Is it rotating CCW?
3 - It seems like a heavy preload to me. If I draw a circle 4 mils long centered at the origin, this orbit goes through it.

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

One thing that I like to review is the orbit orientation. I assume that the transducers are at 45 degrees either side of vertical(?). If this orbit is similar to a scope view of an orbit and not corrected for probe orientation, then the orbit's long axis would be situated at an angle to the horizontal, instead of appearing horizontal as in the picture.

Depending upon the direction of rotation from an assumed viewing direction, this slant to the orbit would be normal or potentially indicative of a side force (alignment?). To know this, you need to know the probe orientation as well as rotation of the shaft.

To find the slant for orthogonal probes, rotate the picture to line up the horizontal and vertical planes with the actual location of the probes in their physical orientation. Based on how the data was taken this also asignes a viewing direction implicitly. Is the orientation normal for a fluid film bearing? What type of bearing? Elliptical? Tilt pads have a different shaft orientation under normal gravity load.

Fixed arc bearings have a natural stiffness asymmetry. So, the orbit would generally be elongated, particularly with some elliptical bearings with vastly different horizontal and vertical clearances.

This message has been edited. Last edited by: William_C._Foiles,

Have you filtered(1x)the signal? Looks like a lot of trash riding the orbit, possibly from surfaces errors or something else like that discussed earlier.