- #1

- 1,496

- 2

OK, in bearings (particularly roller) often times there will be a "spring ring" around it. The ring acts to soften the support structure for the bearing. Ring is typically a thin ring with "pads" on both the inside and outside. They are staggered such to put the thin ring into bending when load is applied at a pad.

OK. So, when analyzing rotordynamics, the stiffness of this ring is crucial. We have been depending on a 50 year old study that we did, which gave us a correlation based on ring thickness, diameter and axial length. It is analagous to an equation which is printed in Roark, which looks something like:

[tex]

k = CEh\left(\frac{t}{D}\right)^3

[/tex]

Where C is a constant, E is the modulus of elasticity, h is the axial length, t is the thickness and d is the nominal diameter.

The Roark solution is for a pure radial load on the pads, whereas in a bearing situation, the load is applied in a global direction (i.e. x or y) rather than simultaneously radial. Our study basically modified that constant.

I was hoping someone else (I imagine you, Fred) have some experience with these things. We have some new experimental data that completely disagrees with our old experimental curve fit. Furthermore, I simply cannot get a numerical run to converge using standard contact (only bonded which severely over-predicts stiffness).

I cannot find a journal article, a design memo, anything else besides this old study that we have. I really really hope someone can shed some light on these things.

p.s. Attached is a numerical representation of the ring if you're having trouble imagining it.

OK. So, when analyzing rotordynamics, the stiffness of this ring is crucial. We have been depending on a 50 year old study that we did, which gave us a correlation based on ring thickness, diameter and axial length. It is analagous to an equation which is printed in Roark, which looks something like:

[tex]

k = CEh\left(\frac{t}{D}\right)^3

[/tex]

Where C is a constant, E is the modulus of elasticity, h is the axial length, t is the thickness and d is the nominal diameter.

The Roark solution is for a pure radial load on the pads, whereas in a bearing situation, the load is applied in a global direction (i.e. x or y) rather than simultaneously radial. Our study basically modified that constant.

I was hoping someone else (I imagine you, Fred) have some experience with these things. We have some new experimental data that completely disagrees with our old experimental curve fit. Furthermore, I simply cannot get a numerical run to converge using standard contact (only bonded which severely over-predicts stiffness).

I cannot find a journal article, a design memo, anything else besides this old study that we have. I really really hope someone can shed some light on these things.

p.s. Attached is a numerical representation of the ring if you're having trouble imagining it.