This is a 13.2kv 4000hp 1200 rpm vertical motor rigidly coupled to a multistage mixed flow centrifugal pump. (Motor top bearing is Kingsbury thrust bearing plus tilting pad radial bearing, motor bottom bearing is large greased deep groove ball bearing, pump has several water-lub'd sleeve bearings).

We replaced the motor with a rewound motor. Uncoupled run was fine (lower than previous readings... motor was better balanced).

When we started the maching after coupling, , we heard nasty noises coming from the area of the pump bearings. We took a set of vibration readings and secured the pump.

The vibration readings are the most recent readings in attached waterfalls (which are shown a little darker than the others so easy to pick out). You can see there is nothing to indicate misalignment. The highest overall at any point is 0.08 ips. Axial didn't jump up and harmonics didn't jump up (perhaps there is some nonsync stuff that jumped up a little on the pump... but low levels).

We went in and checked alignment and found the pump 0.015 mils offset (not TIR) horizontally from the center of the stuffing box.

One moral of the story: vibration cannot be relied upon to show misalignment (once again). (if you consider off-center shaft to be a form of misalignment)

Now the story of how we got to be misaligned. We had aligned the motor to the position where the pump had been dropped when we previously uncoupled it. We assumed the packing would keep it reasonably centered. That was apparently a bad assumption. Apparenlty the packing was weak and when the pump dropped to the bottom, the loose packing allowed it to tilt toward the side of the stuffing box. Then we aligned the motor to the off-center pump.

Now my question: what is the standard way to ensure the pump is centered within the stuffing box?
A - by replacing packing and tightening down the packing
or
B - by performing measurements between the stuffing box and the shaft
or ?

CD_Misalignment_small.ppt (754 Kb, 47 downloads)

Send picture.

4000HP not one to have to replace cuz u & me @$$ume- as they say.

Also sure you know IRT is better for misalignment. Haven't done it much on verts though. Would be a good one for study. May need to try phase on it.

It appeared to me that the spectral profile data did indicate misalignment. However, the rub in the bearing can easily set a sinusodial pattern and a snake-like pattern that can mimic faults. Generally on verticals you can plot the vibration and through the fulcrum should be much less. The farther you go from the fixed point there should be a linear plot line from center if imbalance or more parallel if misaligned.

As oju requested: send photos of the type of setup and pump to motor coupling - rigid via flange hub or????????

Two things to note about these waterfalls:

1 - The top of the scale is 0.06 ips
2 - The misalignment was only introduced when the motor was installed and coupled - the very last set of data in the group.

So it doesn't scream misalignment to me. Unless you're suggesting there was misalignment present previously (and in fact I can extend the waterfall back to 1989 with similar pattern even though we re-aligned several times during that time).

I'll look around for some photo's.

Attached are some photo's.

From the first slide, the dark colored piece on top is the upper coupling hub (which mounts on the motor shaft)

The lighter colored pieces below that are the adjusting nut (which threads on the pump shaft - used to adjust lift), and below that the lower coupling hub (which mounts to the pump shaft).

This particular machine photographed had the upper coupling hub replaced - so it is a smaller upper coupling hub than our other machines - it has a smaller rim diameter than the adjusting nut / lower hub. On the remaining sister machines we have a larger upper hub than this, and the upper coupling hub has the same rim diameter as the adjusting nut / lower hub.

CD_Pump_Small.ppt (348 Kb, 23 downloads)

Pete; sometimes on a machine such as this I take care that the impeller doesn't kick the pump shaft over. New packing and with mag bases and indictors I loosen the coupling bolts and let the pump settle. Monitor with indicators and move 90 degrees; still zero? When you have enough room, you can put 2 indicators on one shaft and take two radial or rim readings spaced at 6" apart. The difference is the angle and the top indicator I tread as a parallel reading.

If you have a new pump (new brgs & packing) you may find the pump's shaft always maintains in concentric center. But with this pump I would make sure the shaft is in center and then align the motor to it. Plus, make sure you use a master level to ensure plumb as this will increase the life of the pump.

E-Pete,

We have several pumps with a similar set up as you do. Our mechanics have created a device that will allow them to plum and center the pump shaft and lock it in place. Then they align the motor to the pump prior to coupling them together. We have noticed a significant reduction in vibration levels and pump life since.

I few months a go we also have the same problem. Our mechanics claim the alignment was good, but vibration spectrum wasn't good and shows different pattern than it should be. Further investigation reveal that during the alignment the pump shaft was concentric to stuffing box but not free. they push the shaft to get concentric so when they release it, the shaft move from the concentric position ant it ruin the alignment. This was caused by bad pump column dimension and alignment.

regards
jasmine