Does anybody know how in terms of increasing and decreasing the load of, let say a gas turbine an effect on the vibration if the gas turbine is slightly unbalanced? In terms of speed, I guess the vibration tends to increase with increasing speed. But how does load affect unbalance in terms of vibration?

Probably you have thermal vectors present. check verify whether the amplitude and phsae are changing at increase or decrease load.

You are not likely to get an answer to your question here as every machine is different so it is difficult (and dangerous) to generalize. Also, what is "slightly unbalanced"? A GT that is running 1.5 mils pp might respond differently to load changes than it would when running 5 mils, due primarily to the underlying cause of the imbalance.

Also, even though you used as an example a GT, be cautious when looking at different machine types. A gearbox with some residual imbalance will often go down in amplitude as it is loaded.

John from PA

ianski,
Could you explain more on the thermal vectors effect on unbalance?

John,
Well, I have a GE turbine driving a generator. Initial run ups from idle to full speed no load, there was no significant vibration. After 3 hours of running full load however, the vibration starts to move up from 2 mils p-p to 7 mils p-p (only on 1XTS). I am seeing somehow thermal influences here. And all these readings are taken from an accelerometer integrated twice to mils p-p.

If you care to look at thermal transients look at polar plots. The description you have about waiting 3 hours seems unlikely for a thermal transient, but the picture may be different from what you describe if you track the entire start and loading on polar plots (1X).

Also, you should track load changes.

John is correct, giving a simple answer to a complicated problem isn't worth trying. I find it difficult to believe that nothing happened prior to 3 hours. The turbine rotor should be approaching or have reached thermal equilibrium by that time without load changes.

wookp,

You also don't mention the type of gas turbine; e. g., aeroderivative (typical rolling element bearings) or industrial (fluid film bearings). Typically these types of machines will have significantly different coupling methods and also exhibit significantly different behavior due to alignment changes.

It is entirely possible that your behavior over a three hour period is the coupled machines undergoing alignment changes, and finally reaching thermal equilibrium.

Contrary to the "charts" etc. that say 1X is imbalance, it may very well be a changing alignment situation causing the 1X change.

John from PA

William,
Do you have any information on how to read polar plots? I could get some polar charts but I have no idea on how to diagnose a condition based on a polar chart.
The increasing vibration only starts after 3 hours of running the turbine at full load which is at 17MW. During the first 3 hours, nothing actually happened (well, that's what at least i believe).

John,
Sorry for the lack of information. It's aeroderivative type of turbine from GE with rolling element bearings. It is a dual shaft design with a GG and a PT. The PT drives the generator directly.

But coming to my main question, i guess load does affect unbalance. But what is the theoretical explanation behind it? I could understand that speed play a part in unbalance, but load?

Wookp,

Let me use a slightly different scenario to explain a polar plot, which is simply, in your instance the 1X amplitude and phase vs. time.

Let's say you have a generator that due to some electrical fault has a thermal sensitivity to load. This is not that unusual and is caused by localized heating in the generator rotating element. Say we synch the machine and sit at minimal load for an hour and the 1X vector remains constant at 2 mils pp at 90 deg lag angle. We slowly ramp the load up, allowing it to stabilize at discrete increments and observe that at full load the 1X vector is at 10 mils, still at an angle of 90 deg lag angle. This means the generator goes thru a thermal transient of 8 mils from a no load to a full load condition. More importantly, lets say it pretty much plots a straight line from the 2 mil/90 deg point at minimal load out to the 10 mil/90 deg point at full load. You don't want to operate at 10 mils pp for any extended period, so lets call a balance shot that takes the unit from 4 mils at 270 deg lag at no load to 4 mils at 90 deg lag at full load. The unit still has undergone the 8 mil thermal transient but starts and finishes at a tolerable amplitude. Something like this may buy you a few months operating time until you can schedule an outage to properly fix the generator.

On the 2nd point, what type coupling is on your GE aeroderivative?

John from PA

John,
Thanks for the input. If I am not mistaken, the coupling is a spacer rigid type coupling. Until now, I have oly been able to collect one set of data from only on vertical accelerometer probe on the PT.

One thing I am confused of this probe configuration is that, GE actually installed accelerometers and on the vibration panel, it is integrated twice to get displacement readings. Anyone encounter this before?

Does the power turbine have rolling element bearings or fluid film bearings?

It sounds like this area is completely new for you. Have you considered getting a consultant or calling the OEM?

William,
The power turbine has rolling element bearings. Well, we contacted a consultant and they said that the generator is unbalanced thus causing the power turbine to have high 1XTS vibrations, just based on spectral readings. However, the vibrations on the generator are very low. That is why we need to further reconfirm this before any maintenance is done. And no, we have not called the OEM.

Well, it seems that the high vibration phenomenon on my turbine has disappeared. No maintenance has been done so far. Just shutting it down and startin it back up. Any comments to how this could happen? Anybody had a similar experience?

You may be relieved for now, but I wouldn't be too comfortable about this. Without data this is but a guessing game.

Vibration that suddenly changes may return. On the up side, it could be nothing more than a sensor. I don't know if there are any limits on the down side as to severity.

That is true and i do believe that the same high vibration condition may resurface. I have checked the sensor and the sensor is working as it should be.