Bearing lubrication flowrate effect on lubrication regime

[ktgottfr]

Hi Folks!
I have a question here that I hope to use to improve my intuition of how rolling bearing lubrication works, and to settle a disagreement with a colleague if possible

Basically, what we would like to know is whether or not the particular oil flowrate to the bearing has much of an effect on the lubrication regime beyond keeping the bearing itself cool.

My understanding is that beyond a small nominal amount of oil/grease being present, adding more oil will not really change the lubrication state (mixed, full film etc).

I'll explain a bit of the background of this question. We have a bearing that normally receives 6 liters per minute of iso32 oil at 80C. However, sometimes this bearing will experience a cold start at -40C where the oil flow is estimated to be more like 0.25 liters per minute.

My feeling is that the bearing lubrication at -40C will actually be better than the lubrication at 80C because the oil viscosity will be way way higher at low temperatures. It should not matter that the oil flowrate is reduced to 0.25 lpm becuase this is still a lot of oil for lubrication. I imagine that even just a few ml/min of this high viscosity oil would be enough to maintain full film lubrication.

What do you think? Do the bearings need a very high oil flowrate even at super low temperatures/high viscosity operation? Or is a very low oil flowrate fine when the viscosity is high?

The bearing in question is an NK 16/20 XL needle bearing.

Thanks!
-Kristjan

 

[geesaman.d]

For a roller bearing, I would not expect the flow rate to matter at low temps because the rollers will trap oil between the race and roller surfaces. If the d*n is high enough for the viscosity, it will establish EHD separation of the metal surfaces. At low temps this is very easy to do. SKF has a lubrication calculator IIRC that would estimate the film thickness.

-40F though is quite cold - you're below the pour point of most oils - if the oil is viscous like a paste, and the bearing is moving fast enough, and there is a place for lubricant to bypass the rollers, it could reject the lubricant flowing into it and run the rollers dry (and seize). But that's only a risk in my mind if the oil flow can bypass the rollers.

 

[electricpete]

Basically, what we would like to know is whether or not the particular oil flowrate to the bearing has much of an effect on the lubrication regime beyond keeping the bearing itself cool.

As pointed out EHDL is a special regime which should not be confused with the regimes discussed for sliding bearings (boundary/mixed/film regimes per stribeck curve).

My feeling is that the bearing lubrication at -40C will actually be better than the lubrication at 80C because the oil viscosity will be way way higher at low temperatures.

If the oil supply temperature is lower then the viscosity is higher and there will be more heat from friction. It may make the machine will be less efficient

More heat from friction means higher temperature rise across the bearing at a given flow. I'm not sure how your decrease in flow relates to your decrease in supply temperature, but if you are decreasing oil flow rate then that's another factor that would cause temperature rise across the bearing to go up. So with 2 factors causing temperature rise accross the bearing to go up it's not clear whether oil exit temperature goes up nor what's going on in the rest of the oil loop.

I'm no lube expert but from my viewpoint there are rarely simple answers in lubrication. Most questions demand a broad examination of the application details to decide what is "better"

 

[ktgottfr]

Thanks for the input!

Yeah -40 is very cold but the oil is being supplied under pressure and will likely still flow directly into the bearing from the side at a rate of perhaps 100ml/min. This cold start condition is very temporary, with the whole system getting up to full temperature after around 1 minute.

The concern was that damage might be done to the bearing during these few cold seconds as the oil flowrate will be much lower than the 6 liters/min that the bearing sees at 80C oil temperature. However, my reading of the bearing literature seemed to say that lubrication gets better as the viscosity goes up, so that only a few drops of oil per minute would be needed at -40.

It sounds like there is some agreement on this.

Yes, the very high viscosity would cause higher friction losses and faster heating, but for a cold start this is ok, as we want the thing to get up to normal operating temperature quickly.

Thanks again for your responses and feel free to comment more!

-Kristjan

 

[geesaman.d]

will likely still flow directly into the bearing from the side

This would worry me. Roller bearings push lubricant out. I've seen roller bearings fail from lack of lubrication in cold temperatures. It's not that the lubricant wasn't present, it just was getting pushed out of the roller path and was viscous enough that it didn't flow back in. The heat that seized the bearing was under the rollers and developed too rapidly to warm the lubricant. The longer and narrower the roller, the more risk.

The NK 16/20 XL bearing seems to have a lubrication slot and hole that feeds to the center of the rollers. I would want oil injected directly in there. It would take precious little flow to work.

David

 

[ktgottfr]

Hi David,
Ok, in our case the oil is being forced thru the bearings by around 100psi of pressure. There is no where else for the oil to go but thru the bearings from one side to the other. So I am guessing we will be ok for those few seconds before the oil heats up and the flow increases.

Thanks again for your comments!
-Kristjan

 

[TugboatEng]

Rolling bearings don't need any flow per se. They are suitable for grease lubrication as a result. Provided the bearing is wet with oil it should tolerate a momentary lack of flow during startup.

Highly viscous oils are not better for lubrication of rolling bearings. As stated above, they can be pushed out and not flow back in. There are also other problems at the higher speed end of the equation related to high oil viscosity. Your temperature range is quite broad. I suggest identifying an oil with a higher viscosity index to maintain the correct viscosity over the entire operating temperature range.

 

[electricpete]

> I imagine that even just a few ml/min of this high viscosity oil would be enough to maintain full film lubrication.

Film lubrication is a term that applies to sliding bearings, it is not a term that applies to rolling bearings. Rolling bearings operate in an elastohydrodynamic lubrication regime (EHDL). Example article link discusses boundary / mixed / film for sliding bearings and EHDL for rolling bearings.


> Yes, the very high viscosity would cause higher friction losses and faster heating, but for a cold start this is ok, as we want the thing to get up to normal operating temperature quickly.

The bearing has a temperature distribution. Rapid temperature change generally means higher temperature gradients and more potential for differential expansion among theouter ring, rolling elements, and shaft or inner ring. That can cause a variety of problems, mostly skidding if internal clearance increases and lockup if internal clearance disappears. It's not obvious exactly how your bearing will expand (at least you're not cooling the housing which would more obviously risk lockup) but it's still something to think about whenever you have a rapid temperature change.

If cage is a different material than the rolling elements that's another differential expansion part of the puzzle. Also viscosity can affect skidding directly and if higher viscosity increases friction between cage and rolling elements then that increases tendency for skidding. Skidding is typically a longer term consideration for bearing reliability as it degrades the surfaces during each start, but if it's severe then who knows... maybe the increased cage stresses break the cage which can be an immediate catastrophic failure.

I'm not saying any of these scenarios will cause a problem, only these are some of the things to think about.

 

[electricpete]

in our case the oil is being forced thru the bearings by around 100psi of pressure. There is no where else for the oil to go but thru the bearings from one side to the other. So I am guessing we will be ok for those few seconds before the oil heats up and the flow increases.

That introduces another potential differential expansion pattern coldest at the oil inlet end and hottest at the oil outlet end which could introduce a taper to the rollers or the outer ring or the cage. It may be fine but again something to think about.

The NK 16/20 XL bearing seems to have a lubrication slot and hole that feeds to the center of the rollers. I would want oil injected directly in there. It would take precious little flow to work.

That would help the end to end differential expansion taper concern.

 

[electricpete]

in our case the oil is being forced thru the bearings by around 100psi of pressure

Just curious, is that pressure roughly constant with temperature. Some oil systems struggle to control pressure during wide temperature variations.

 

[Tmoose]

FAG and SKF have charts and calculators to figure a parameter called "kappa." Viscosity ratio or somesuch.

To start, maybe the viscosity of the cold lube could be used to determine if kappa remains in a useful rang

Viscosity ratio The viscosity ratio  is an indication of the quality of lubricant film
formation:
k – Viscosity ratio = v/v1
v mm2s–1
Kinematic viscosity of the lubricant at operating temperature
v1 mm2s–1
Reference viscosity of the lubricant at operating temperature.
The reference viscosity 1 is determined from the mean bearing diameter dM and the operating speed n.
​

At values of k = 4 and above, full lubrication is present,
i.e. the partners are not in contact.

 

[electricpete]

To refresh my memory on that, I googled it, and it's described here:

https://www.skf.com/us/products/rol...e/lubrication-condition-the-viscosity-ratio-k

Just to describe it in my own words (after having just read it): Diagram 2 gives "rated" viscosity (I think of it as minimum required viscosity) as a function of Dm and N. As Dm or N increases, the rated viscosity decreases (in a similar way that we can get away with lower viscosity as the linear speed increases on sliding bearings). Kappa is the ratio of viscosity at operating temperature to that "rated" viscosity. They talk about the desirable range of kappa and mention the 4.

I shouldn't have rejected the term "film" in this context, since skf uses it.

If the application is working fine at normal temperature there is no problem meeting any minimum viscosity at -40C. But there are a lot of other concerns discussed above.