How to Prevent Rotor from Bowing

P

Thread Starter

PRANAB ROY

Myself Pranab Roy working in Alkamil Power, Oman. We have three GE frame 9E (125MW ) m/c., Recently we have faced a problem, the problem was: We have to shutdown the GT-1C, came in to turning gear at 99rpm but after 45min of running in turning gear jacking oil pump tripped due to single phasing (one of the phase got burnt). We have stopped the turning gear by making the module off, m/c came in to zero speed, in that case there is chance of rotor bowing . In GE frame 9E m/c , m/c remains at turning gear for 24hrs.

I want to know as a control room engineer in GE frame 9E m/c what we can do if Jacking oil pump tripped within 24hrs to prevent from rotor bowing.

Plz let me know i shall be thanfull if u do the needfull ,to know what are the steps we can do to prevent bowing.

Regards
Pranab Roy
 
This has been covered in other posts on control.com. The search feature here takes a little getting used to, but it finds results very quickly. Search for terms like "cooldown" and "slowroll" or "slow roll" (I think you need to use the syntax '+cooldown' or '+slowroll' or '+slow +roll', without the quotation marks.)

When something like this happens, it's extremely important to keep the Aux. L.O. Pump running, to keep L.O. flowing to the bearings to cool the bearings. The heat from the shaft can damage the bearings if the L.O. flow is shut off.

The next most important thing is something that everyone lacks the first time something like this happens, and that is: be patient.

You need to wait approximately 24 hours after the unit reaches zero speed before trying to put the unit back on turning gear or slowroll or cooldown or whatever it's called at your site. The compressor rotor (the longest and heaviest component of the shaft) will bow under its own weight when it's hot and as it cools it will actually straighten *almost* completely. If you try to get the unit turning with extraordinary measures (hydraulic rams, starting motor, etc.), you can catastrophically damage the compressor rotor.

If you wait until the unit has cooled (approx. 24 hours), the shaft will return to nearly straight. At this point, the unit can be CRANKed for a while to help straighten the shaft, and I like to select FIRE mode and leave the unit on FIRE mode for 10-20 minutes or so (*not* too long because the cranking motor is pulling a *LOT* of current at this speed) to help get a little heat into the shaft and help straighten it.

Then just select AUTO mode and the unit should continue accelerating to FSNL. You may see some high vibrations during this acceleration, but they should be well below trip setpoints, and once the unit reaches FSNL, it should be fine.

Patience is the key. Managers and operators just freak out and want to do everything they can to get that shaft turning. They get hydraulic rams out and try to jack under coupling bolts and all manner of schemes. But, if the rotor was hot when the unit reached zero speed, it's very likely that the compressor will have bowed under its own weight so much that the rotor blades will be touching the compressor case at the six o'clock position and any attempt to turn the rotor may damage the blades.

It's the characteristic of metal and of the rotor design, that it will straighten itself out, not completely, but enough that the unit can be started and rotated to remove the last of the bow.

Just keep the L.O. running to keep the bearings cool, and just wait. Fix the jacking oil pump motor, or make the preparations to do so. But, in the meantime, everybody should just remain cool.

This method has been pooh-poohed and rejected and criticized. But, just stop and think about it. There are literally thousands of GE-design heavy duty gas turbines, and some percentage of them experience this at least once, sometimes more (depending on the configuration of the plant and the reliability of the AC supply to the plant). And, with the exception of those who damage the compressor rotors trying to use extraordinary methods to turn it when it shouldn't be turned all of those units continue to run just fine with no ill effects. At first, during acceleration and sometimes during initial loading, there might be some higher than previously experienced vibration levels, but those dissipate quickly with heat, time, and rotation.

Units don't get disassembled and inspected, unless the L.O. flow is shut off and the bearing metal temperature sensors indicate more than about 150 deg C, or so (which is the temperature above which the bearing metal begins to become very soft and can actually melt or become deformed because of the weight of the rotor). (Wheelspace temperatures are the best indicator of turbine and compressor rotor temperature. Compressor discharge temperature is useful as well, but for two-bearing machines and most other units wheelspace temp's are the about the best indicator there is available.)

So, this or something like this should actually be worked out and written as part of the standard operating procedures to be observed during an event. That way, it's already been decided and everyone knows how to respond and what to do.

For those unit with hydraulic ratchet mechanisms, I recommend forcing the ratchet sequencing off while still maintaining L.O. flow to the bearings. Otherwise, some sequencing will continually try to ratchet and damage the ratchet pump and/or its motor by trying to turn the rotor continually for hours instead of cycling as it normally does.

But the keys are to keep the L.O. pumps running and to just be patient (the latter is much easier said than done). When the highest wheelspace temperature drops to near ambient temperature or just slightly above, the rotor has probably cooled sufficiently to try to turn it.

When the highest wheelspace temperature is below approximately 90 deg C, it's probably safe to shut off the L.O. But, keep an eye on the bearing metal temperatures, and if they start to increase too much after shutting off the L.O. pumps, then the pumps need to be re-started and at least cycled to cool the bearings.

24 hours is usually a good safe amount of time, in general, for a turbine rotor which has been running at Base Load and which went to zero speed fairly quickly after trip or shutdwon and couldn't be put back on cooldown or CRANKed within about 20 minutes of reaching zero speed.

If the unit was on cooldown for six hours or so and then went to zero speed and couldn't be put back on cooldown, then the time could be less, but it all depends on internal temperatures.

Hope this helps.
 
Top