Rolling Element Bearing Stiffness Matrix

Determination

Yi Guo Robert G. Parker

National Wind Technology Center Department of Mechanical
National Renewable Energy Engineering
Laboratory Virginia Tech

NREL/PR-5000-60639

NREL is a national laboratory of the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.
Motivation
Limited work on stiffness matrix in the literature
o Diagonal matrix approximation typically used
Elastic deformation of race causes nondiagonal terms
Coupling between the radial
k xx k xy k xz k x x k x y 0 and rotational displacement

k yy k yz k y x k y y 0
Coupling between the axial
k zz k z x k z y 0 and rotational displacement
k x x k x y 0

Symmetric k y y 0

0
Coupling between the axial
and radial displacement
2
Finite Element/Contact Mechanics Model
Three-dimensional finite element model includes
microgeometry
Analyze contact between rolling elements and races
o Contact searched at every time instant as bearing rotates

Radial Ball Bearing Cylindrical Bearing Contact Grid

3
Accuracy Order of Finite Element Analysis
Numerical Jacobian used to compute K
Order of Jacobian approximation formula should be comparable
to the accuracy order of finite element analysis (FEA)
Method to obtain the accuracy order of FEA

V c1h p1 + c2 h p2 + ...
Vh =+
Vh : finite element solution
eh V : exact solution
eh3 eh2 Vh3 Vh2 h2p1 h3p1 h : finite element size
= p1 p1 : order of accuracy
eh2 eh1 Vh2 Vh1 h1 h2p1
h2
log
Vh3 Vh1 h3 Accuracy Order FEA FEA/Contact
log{( ) }
Vh3 Vh2 log h1 p1 1.11 1.94
p1 = h3
h
log ( 1 )
h2

4
Comparison Against Published Experiments
Calculated stiffnesses by FEA agree with experiments
(Royston et al. 1998) (Kraus 1987)

FEA Radial Stiffness Experiment

Stiffness (kN/mm)
Radial Stiffness
Stiffness (kN/mm)

FEA
Experiment

Experiment

Axial Stiffness FEA Axial Stiffness

FEA
Experiment

Radial and Axial Load (N) Axial Load (N)

5
Comparison Against Theoretical Models
FEA stiffness agrees with the theoretical model
o Only with unrealistic races that match Harriss assumptions
Theoretical models predict higher stiffness with design
dimensions
FE, thick race

Harris, Theoretical
FE, actual race

Harris, Theoretical
FE, thick race

FE, actual race

6
Bearing Stiffness Is Time-Varying
Number of rollers in contact changes periodically
Can excite gearbox vibration

Ball pass period

Five Rollers in Contact

Four Rollers in Contact

7
Off-Diagonal Stiffnesses Affect Gear Vibration
Gear dynamics with off-diagonal stiffnesses differ
from gear dynamics with a diagonal stiffness matrix
o Need to include off-diagonal stiffnesses

Full matrix
Diagonal matrix

Photo by Yi Guo, NREL

8
Conclusions
A method developed to determine bearing stiffness matrices

Method validated by experiments

Comparison against theoretical models expose their

limitations

Bearing contact is nonlinear and time-varying

Bearing microgeometry affects stiffness

Off-diagonal bearing stiffnesses affect gear dynamics

9
Thank You!
This study was supported by the National Aeronautics and Space
Administration Glenn Research Center. The authors would also like to
thank the National Renewable Energy Laboratory for follow-on support.
NRELs contribution to this work was funded by the U.S. Department of
Energy under Contract No. DE-AC36-08GO28308.

yi.guo@nrel.gov
303-384-7187

10