Originally Posted By: rubberchicken
Thanks, I think I have read something similar happening in other applications, but never in this large of a scale. Also, I do not think you ever answered the question on the oil whip final fix: was the 10 c change in oil inlet temp an increase, or a decrease ? How did you effect the temp change ?
We had fitted a refurbished generator rotor, that was exceptionally well balanced to the machine...thereafter, it had wild vibrations (over 130um) on bearing 9 appearing sporadically and randomly...reducing load sometimes helped, othertimes didn't. Looking at the FFF of the bearing vibes, the 1x was really low, in the single digits, while there was an element at 0.25x that was pretty well even with it, but moving up and down at scan interval...like I said, it was well balanced, so the 1x was small, and the 0.25x was about the same.
1x was pretty well rock solid, and the variable 0.25x would peak at 6um, drop, 12um, drop, then go berserk...classic whip.
So the bearing characteristic number is the Sommerfeld Number, it's a dimensionless unit (r/c)^2 x uN/P
r= shaft radius
c= radial clearance
u= viscosity
N= Revs/second
P= applied pressure, the load on the bearing divided by the projected area (length x dia) of the bearing.
So with a running synchronous machine, there was little that could be done to try to ameliorate the problem...so we started messing with u...indirectly by changing the main oil cooler outlet temperature setpoint...lower, more viscosity, more instability, higher less viscosity, less vibes. This decreased the "u" component, dropping the So number into a more stable place)
It was inserted into the SOP to correct abherent conditions until the other parameters could be addressed (the reference to jacking oil is the shaft lift oil used during starting...it disrupted the behaviour)...but also affected every other bearing, by dropping the Minimum Oil Film Thickness...not ideal.
I tried changing bearing oil pressures, through the manipulation of the Booster baffler and bypass baffler valves, but it was not as effective, and upset other locations more.
When we had a suitable shutdown window arranged, we moved the generator casing, both up (to increase the vertical load), and across (to push the bearing against the shaft in the direction of the Minimum oil film thickenss)... increasing the "P" component of the So number, but locally...could have increased the clearance, but that would also increase the flow ratio, and therefore keep the viscosity higher by having higher replenishment rates, fighiting against itself...IIRC, we lifted the bearing 0.001", and slid the generator sideways 0.005", moving a couple hundred tonnes of generator to affect the moves. Design and implementation via the OEM.
Back and running with standard oil pressures, standard oil temperatures, the 0.25x was still a little wonky, although not at the point that it caused a runaway, so the OEM designed another fix...to reduce the flow rate to the bearing, and push the viscosity back down for the bearing 9 itself.
The bearings themselves have the oil feed on the vacuum side of the bearing. The main oil pressure is 10s of psi, and there is an orifice/strainer between the supply and the suction side of the bearing...the suction side when running IS suction, pulling a partial vacuum as it pushes the oil out the side clearance.
When the ducks were aligned, we pulled the unit out, red hot (HP and IP temps 1005F), and dropped to barring. When stable, we dropped off barring, took the unit off oil, hung a lock out tag out, opened the oils system, changed the strainer orifice, removed the LOTO, got the oil back on, and barred the shaft to 180 degrees inverted...in 15 minutes...Looked like an F1 pit-stop, and walking the floor, I've never sweat so much in my life.
That fixed it...