"The fouling/coking of the oil labyrinth is suspected as origin." I doubt this hypothesis is correct. Analyze and correct the vibrations first, and the oil coking from high temperature may go away.

Walt

I have never heard of such, are you talking about with the unit in service. I would love to see some pics and more info on the process.

Most oil seals are designed so the drainage is back into the bearing cavity, so the dissolved muck will containinat the oil.

Just thinking about such a process, If I had the nerve to do it;
1) pressurize the oil tank air space so that the bearing pedistals are positive pressure
2) be sure the spray temperature matches the shaft temp
3) be prepared for high vibration and trip out

Sorry but it does exist, I already met this phenomenon in machines( hot gas expander, steam turbine) and others people also (GE, Bently Nevada).
Cleaning would be in service, ok for spray temp, too costly to shut down the plant for this light rub that is why we imagine this scenario.
thank for your replies

I acknowledge "coking" of an oil deflector happens and causes vibration problems.
ensure proper air flow in the area is something I tried to ensure during outage to prevent. Usually needed to install a shield that would hold thermal insulation back.

a "far fetched" ideal....could you drill into the oil deflector cavity so that cleaning fluid could be injected that would dessolve the burnt oil. Then an air purge could be used to midigate the reoccurance.

edit;
I looked at a cross section of an unit I though I could have drill to the oil deflector and imediately realized the access was too limited. then looking at drawing, considered the pressure washing spray and thought about the gland steam would also suck in the wash. and it hit me!!

how about if the gland exhuaster was secured and allow shaft steam to soften and maybe wash out the burnt oil

byrdj: you're right, I will try to inject air (not easy but feasible) and protect the oil deflector with a shield.

this image is an unit I had experiance with where thermal isulation would block the oil deflector's air gap and I would as a SS plate to the end of the HP casing to defind the needed clearance

in my case the brg housing and oil deflector are hot due to steam leakage.
1. What I'd like to understand is the difference of vibration rub pattern between rotor rubbing on steel strip (light rub on steam laby supported by springs) and coked oil deflector which is normally a soft pad.
2. I read, in coked oil deflector case, the rotor phase sometimes change heavily sometimes almost no change. Why there is a difference?
3. I don't understand the hot spot principle in this situation and more so close to the journal brg.
Normally the coked oil is found like a ring inside the deflector and so around the rotor but not with the same hardness/stiffness in all directions . The rotor displacement is limited in some directions and vibration may happen. I have seen deep groove in rotor due to accumulated particules inside oil deflector.
4. In the coked oil deflector case is it possible to have a reverse precession?

I do not have a great insight into rotor dynamics and resulting vibrations so I can't provide answers.

I made a hasty response to the OP, so I will offer some links that may be of help:
"Turbine oil seal coking"
http://www.maintenance.org/topic/abnormal-vibratio...

http://www.maintenance.org/topic/light-rub-in-comp...

http://www.maintenance.org/topic/intermittent-vibr...

http://turbolab.tamu.edu/proc/turboproc/T42/CST09....

https://www.orbit-magazine.com/2014/11/19/steam-tu...

http://www.slideshare.net/GE_MC_India/vibration-mo...

http://www.miataturbo.net/diy-turbo-discussion-14/...

https://books.google.com/books?id=f-MDAAAAMBAJ&...

https://www.google.com/search?sourceid=navclient&a...

Getting water in the lube oil may not be a problem, but what about fugitive coke particles that get into oil? An air stream for cooling was mentioned as a short-term measure to minimize rubs. The GE paper mentions a long-term solution to minimize coking. Some of the papers discuss the difficulty of removing coking that is common to steam and gas turbines, compressors and turbochargers. A rub that causes phase change is sometimes referred to as Spiral Vibration when vibration vector amplitude and phase change. This may occur when a shaft critical speed (balance resonance) gets excited by the rub.

Walt

Thank you Walt
I knew all of them (except miataturbo).
Nicolas