TurboGenerator Journal Bearing Shaft Surface Finish - Too fine? 3

[jt2001]

Hi All,

We have completed a grind to cleanup damage on a turbogenerator journal shaft.
The current surface finish is around 0.4 - 0.6 RA, Could this be too fine for reliable operation?

Regards,

JT

 

[edison123]

We routinely grind polish the shaft journals to mirror finish at those levels for both anti-friction and white metal bearings. No issues so far.

Muthu

http://www.edison.co.in

 

[FacEngrPE]

I would not expect finely polished journal surfaces to cause trouble in a hydrodynamic journal bearing. The only time metal to metal contact occurs is at the beginning of a startup and at the end of a stop. The rest of the time the shaft is "flying", and the oil flow in the journal is laminar.

Search for more info

https://duckduckgo.com/?t=lm&q=hydrodynamic+journal+bearing+design&ia=web

 

[electricpete]

You're asking in an electric motor forum and I believe the responses you got are correct for electric motors and small generators

I would not assume they are correct for turbines and generators of all types, especially larger ones which can be highly engineered machines. OEM should be consulted for journal surface finish requirements (which may include both minimum and maximum roughness).

There is lore handed down at our plant where decades ago high speed feed pump turbines were handled on teflon rollers. The teflon transferred enough to make the journal "too smooth" and the bearings later had problems immediately upon installation. During investigation it was noted that OEM repair specifications for these journals carry both a minimum and maximum roughness spec (which was one datapoint supporting the "too smooth" explanation for the event). It is non-intuitive to most people (including me) but our knowleedgeable turbine engineers have been unequivocal on this point.

A link which may or may not be directly relevant to journals, but illustrates that smoother surfaces on rotating sliding surfaces are not always better

[ul]
[li]* Kingsbury Universe Brochure - COLLAR/RUNNER/JOURNAL SURFACE - The most commonly overlooked bearing component is the collar. It is the single most important part of the bearing. Collar rotation draws oil into the region between the collar and shoe surfaces. Oil adheres to the collar and is pulled into pressurized oil wedges. This occurs due to the collar surface finish. If the collar finish is too smooth (better than 12 RMS), it will not move an adequate supply of oil; too rough, and the bearing shoes will be damaged. Ideally the finish should be between 12 - 16 RM[/li]
[/ul]

 

[FacEngrPE]

electricpete
I had not considered ring lubricated bearings, friction (traction?) is required to drive the ring, and as the contact surface is going to be lubricated - rather low friction at best, I can see where this could be important.
However if this is a pumped lubrication arrangement, no oil rings will be present, so the friction tradeoffs could be different.
Details Details

And then there are the kingbury bearings


I just noticed the same oil ring / shaft finish question occurs in this thread. thread407-520600

 

[electricpete]

I agree too smooth might be a consideration for driving oil rings. But my post had nothing to do with oil rings (The Kingsbury quote referenced had nothing to do with oil rings, the turbine at our plant has journal bearings with forced lube, no oil rings).

I will link a thread from another forum:

https://assetmanagementprofessionals.org/discussion/acceptable-journal-surface-finish

[ul]
[li]In that thread, the op happens to quote the exact same words that I quoted above (which doesn't mean much, probably copied from same original source). What is noteable imo is that 3 responders (William Foiles, Mike O'Neal, and JohnFromPA) all agreed that too smooth can be problematic for sliding bearings.[/li]
[/ul]

I don't have an explanation for this (as I said it's not intuitive for most people including me). I searched and found an article as one more datapoint:

Effects of shaft surface texture on journal bearing pressure distribution

The primary reason for measuring surface roughness is predict the performance of the component. A bearing surface, for instance, requires a level of surface texture that allows lubricant to be retained in small pockets as well as allowing the bearing to roll with minimum friction. If the surface is too rough, wear can quickly develop; however, if the surface is too smooth, inadequate lubrication and seizure might occur [1], [2].

There's a lot more available at that link, including links to its references. I can't say it makes sense to me but I haven't dug into it that far.

The strongest data point in my book is the comments from our knowledgeable turbine engineer, which are more than enough to convince me that it would be wise to check with the OEM on this point when dealing with large turbogenerator equipment.

EDIT - I didn't give much attention to the specific values: 0.4 - 0.6 Ra. Is that microinches or micrometers?

 

[DAVIDSTECKER]

There are a few examples of where having a "rougher" surface is desirable.
One is the crosshatch pattern in an engine cylinder to hold a layer of oil and help seal the piston rings.
I belief that the oil impregnated bushings recommend a 32rms to hold a layer of lub.
There is a marine paint called shark skin that is rough to hold a layer of water for improved performance on racing boats.

 

[GregLocock]

The problem with using pistons as an example is that the velocity drops to 0 twice per rev, so you don't have hydrodynamic lubrication. Perhaps more relevantly our thrust bearings, which are hydrodynamic, started failing, which was wipiping out the crank sensor. A rather urgent investigation revealed that during maintenance the lapping machine's rotation had been reversed.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[FacEngrPE]

R[sub]a[/sub] is the arithmetic average of the roughness profile, can be presented in either micrometers (μm) or microinches (μin.).
I am guessing here, as the OP did not state units 0.4 μm ~ 16 μin. ~ N5

Reference

https://en.wikipedia.org/wiki/Surface_roughness

I suspect in the case of a pressure fed lubrication journal that surface roughness may most important at startup and very low speed. It is likely the OEM has adjusted his design based on service experience, so the advice above regarding getting the OEM's advice has my Vote.

Literature can be found on this subject, which seems to indicate roughness does impact the hydrodynamics of the oil film, it is more complicated than just measuring R[sub]a[/sub], as the literature describes patterns.

 

[electricpete]

Thanks, LPS for that literature link (the first result was relevant). Clearly there's a lot more to the subject than can be gleaned from a standard textbook on hydrodynamic bearings (which would tell us the velocity of the lubricant matches the velocity of the solid at their boundary irrespective of surface roughness, and the only role of surface roughness is disadvantageous if it is not kept a small fraction of film thickness)

As a practical matter in the motor world, I have never heard of any sliding bearing problem attributable to "too smooth", nor seen any minimum smoothness requirement in a motor repair spec.

I believe something must be different about turbines bearings that this aspect can have a real world impact. Maybe it has something to do with the startup and shutdown operating modes as you suggested. If I had to guess not turning gear as long as oil lift is applied, but maybe extended operation at lower speed. Or maybe they are designed with less margin and therefore more sensitive to these smaller effects that don't get noticed in applications that have a lot more margin (lower psi loading)