'_. (NA._A-CR-1953 bS) RF.5 [AECB 5E_ti, z : Ui_' $ N8Z-3 1909 ..

_
5HABEETH. 5_' ",SY STAIE A_3 'IRA_SIEN= . :
; TdERMAL ANAL31S O_" A Sh_:[ _[At_ING 351_S Jnclas
i
RESEARCH REPORT - USER' S MANUAL : -:
FOR
COMPUTER PROGRAM AT81Y003 SHABERTH
2
! -
-2
_ _ MAY 1981 ':_-.
| -'
_- Ste .gy State -and Transient Thermal --
iW Analysis ,, a Sha_ Bearing System
W
Including Ball, Cylindrical and
Tapered Roller Bearings
CONTRIBUTORS:
|
G. B. HADDEN
R.J.KLEC_
|
_ M.A. RAG_-N
_ L. SHEYNIN
SKF REPORT NO. ArS1D040
!
I "SUBMITTED TO:
NATIONAL AERONALrTIC5 _ SPACE ADMINISTRATION
|
LEWIS CENTER
21000 BROOKPARK ROAD
CLEVELAND, OH 44135
I UNDER CONTRACT NO. NAS3-22690
KFTECHNOLOGYSERVICES
]982024093
i
ii
_ AT81D040
PROPRIETARY NOTICE
FURNISHED UNDER U.S. GOVERNMENT CONTRACT NO. NAS3-22690. SHALL
NOT BE EITHER RELEASED OUTSIDE THE GOVERNMENT, OR USED, DUPLICATED
OR DISCLOSED IN WHOLE OR IN PART FOR MANUFACTURE OR PROCUREMENT,
WITHOUT THE WRITTEN PERMISSION OF THE SECURITY CLASSIFICATION
OFFICER, NASA-L_IS RESEARCH CENTER, CLEVELAND, OH 44135.
1982024093-002
LI
fJ
RESEARCH REPORT - USER'S MANUAL it
f
I FOR i
,- COMPUTER PROGRAM AT81Y003 SHABERTH
i j
" MAY 1981
Steady State and Transient Thermal
Analysis of a Shaft Bearing System
Including Ball, Cylindrical and
Tapered Roller Bearings
! "
CONTRIBUTORS:
r- G.B. HADDEN
R. J. KLECKNER
M. A. RAGEN
L. SHEYNIN
APPROVED: __ S_FREPORTNO. AT81D040
RELEASED:
SUBMITTED TO :
NATIONAL AERONAUTICS & SPACE ADMINISTRATION
LEWIS CENTER
_" 21000 BROOKPARK ROAD
' CLEVELAND, OH 44135
"' UNDER CONTRACT NO. NAS3-22690
i
I__ SUBMITTED BY:
_ SKFTECHNOLOGYSERVICES
SKFINDUSTi'#IES. INC
KINGOFPRUSSIA. PA
: L"
1982024093-00
j t .z ,_
i,,+,+ - ,++_+_
i ++:i
1. fqee_ No, I 2. Gov_ i Ne. "t tqemji_'u CNUq NO,
l NASA CR. NO. 165365 +.:.
I
, "nee _ --,,_m nero- 0_u ,
Steady State and Transient Thermal Analyni8 of a . Nay, 1981 .,. _,+'_+
Shaft Bearing System Including Ball, Cylindrical L _ Oquvu_nCMn +++_
and Tapered Roller Bearings +' ;'+
ii
7..,,.._ m. pe,+e..,_8o,_,_nero.e. %+"_(
+ Hadden, G. B.; Kleckner, R; J.; IUigen, N.A.; ATOID040
Sheynin, L. I0. Work _ No. -+" _'-,_
SKF Industries, Inc. 11, _ w O_ NO+
Technology Services Division l-
ll00 Print Avenue NAS3-22690
King Prussia I PA 19406 13. Tvge_ mlu, u,t _ _ .
of CmmW
_1 Semum_ *4m_ mum me Admm Final
NASA Lewis Research Center s_ _ amsw Cma
_; 2100 Brookpazk Road
I Cleveland, OH 44135
-J i_ S.emmmm_ Nem
IlL Abmrl_
The material presented in this manual is structured to guide the user in the
_ practical and correct of the SBABERTH
_pl_en_tion computer program.
SHABC_PH is capable of s4mulating the _hez_mmechanical performance of a load
support system consisting of a flexible shaft supported by up to 5 rolling
element bearings. Any combination of bell, cylindrical, and tapered roller
bearings can be used to support the shaft.
In this version of SHABERT_, the _ser can select either SKF or NASA models
in calculating lubricant film thickness and traction forces. Also, the
formulation of the cage pocket/rolling element interaction model was revised
- to improve solution numerical convergence characteristics.
!
L]
[
_.+ 17. Keywo_ ISummuduv Aum_l_) [tlL O_ Smm_m
"For Sill bytltl Nlt,On/I T_hn,cal Informt'on S_n_e,$0rzn_ittd, Vitlmzi 22111
NA_A-C-I_(Rev.10-75)
1982024093-004
k
TABLE OF CONTENTS _I
L ;-__<_{
PAGE I,
I -<%>>3
TABLE OF CONTENTS ...................................... i i _c
LIST OF TABLES vii -
.......................................... I i,_ _,
LIST OF FIGURES ......................................... viii
FOREWORD................................................ 1 \
1 0 INTRODUCTION 3 tI_6Oee6QeOQIeeeQOe6QO OeQQe_t_OOeOQ
2.0 PROBLEM FORMULATION AND SOLUTION ................... 4 _:_:{:
2.1 Temperature Calculations ...................... 5
2.i. 1 Steady State Temperature Map .......... 5 _l _ :_
2.1.2 Transient Temperatures ................ 6
2.1.3 Calculation of Heat Transfer Rate ..... 8 I
2.1.3.1 Generated Heat ............... 8
2.1.3.2 Conduction ................... 9
2.1.3.3 Free Convection .............. 9
2.1.3.4 Forced ConvectJ on ............ 10
2.1.3.5 Radiation .................... ii
2.1.3.6 Fluid Flow ................... 12
2.1.3.7 Total Heat Transferred ....... 13
2.1.4 Conduction Through a Bearing .......... 14
2.1.4.1 Thermal Resistance ........... 14
" 2 i g Di sio 1 Ch ge Analysis 17
, .2 Bear n men na an ...........
J
2.3 Bearing Inner Ring Equilibrium ................ 18
?I
2 4 Bearing Quasi-Dynamic Solution 20
2.4.1 Cage Degrees of Freedom ............... 23
3.0 PROGRAM INPUT ...................................... 28
3.1 Types of Input Data ........................... 28
3 2 Data Set I - Title Cards 29
3.2.I Title Card 1.......................... 29
3.2.2 Title Card 2 .......................... 29
i
! SKFTECHNC)L(X> ' SE VI<:ES
.i Sk_INI)(ISTk_{S;N,"
i
......... l
1982024093-005
TABLE OF CONTENTS (continued} fl
H
PAGE
3.3 Data Set II- Bearing Data .................... 32 II
3.3.1 Card Type I - Bearing Type and Material
Designations .......................... 33 [I
3.3.2 Card Type B2 - Bearing Geometry and
t_
Outer Ring Speed ...................... 34
3.3.2.1 Ball Bearing Geometry ........ 34 W
3.3.2.2 Tapered Roller Geometry ...... 40
3.3.2.3 Cylindrical Roller Bearing [|
Geometry ..................... 40
3.3.3 Card Type B3 - Rolling Element
Geometry .............................. 42 II
3.3.3.1 Ball Geometry ................ 42
3.3.3.2 Tapered Roller Geometry ...... 42 _I
3.3.3.3 Cylindrical Roller Geometry.. 44 IJ
3.3.4 Card Type B4 - Rolling Element i|
Ring Geometry ......................... 44
II
3.3.4.1 Ball Bearing ................. 44
3.3.4.2 Tapered Roller Bearing |I
Contact Geometry ............. 44 [I
3.3.4.3 Cylindrical Roller Contact
Geometry ..................... 46 []
I!
3.3.4.3.1 Card Type B4A - Roller
Raceway Geometry ........... 46
3.3.4.3.2 Card Type B4B - Roller
H
Flange Geometry for Cylin-
drical Roller Bearings ..... 46
3.3.5 Roller-Raceway Non-Uniform Profile
Definitions ........................... 48
3.3.5.1 Card Type B5 - Outer Raceway U
Roller Contact ............... 48
3.3.6 Card Type B6 - Inner Raceway Roller fl
Contact .............................. 48
3.3.7 Ring - Rolling Element Surface Data.. 48
U
i
i
l
SKFTECHNL)L_k_YSE_VI_'[S
U
1982024093-006
ORiGiNALPAGEIS
OF pOOR(_JALITY
]
7 TABLE OF CONTENTS (continued)
J PAGE
! 3.3.7.1 Card Type B7A - Raceway
.. Rolling Element Surface
Data ........................ 48
- 3.3.7.2 Card Type B7B - Flange
ul Roller End Surface Data ..... 50
3.3.8 Card Type B8 - Cage Data ............. 50
I
I
- 3.3.9 Card Type B9 - 6haft and Housing Fit
Dimensions ........................... 50
[
_ 3.3.10 Card Type BI0 - Shaft & Housing Fit
Dimensions ........................... 51
J
[ 3.3.11 Card Type BII - Elastic-Moduli ....... 51
3.3.12 Card Type BI2 - Poisson's Ratio ...... 51
[-i
3.3.13 Card Type BI3 - Density .............. 51
[ 3.3.14 Card Type BI4 - Coefficient of
i
L Thermal Expansion .................... 51
r 3.3.15 Card Type BI5 - Lubrication and
i Friction Data ........................ 51
L
3.3.16 Card Type BI6 ........................ 52
3.4 Data Set III- Thermal Model Data ............ 53
3.4.1 Card Type T1 ......................... 53
3.4.2 Card Type T2 ......................... 58 i
i
3.4.3 Card Type T3 ......................... 58 i
3.4.4 Card Type T4 ......................... 58
i 3.4.5 Card Type T5 & T5A ................... 58
[- 3.4.6 Card Type T6 ......................... 59
_ 3.4.7 Card Type T7 ......................... 59
I'_ 3.4.8 Card Type T8 ......................... 60
3.4.9 Card Type T9 ......................... 61
r"
; iii
Jo
1982024093-007
!
TABLE OF CONTENTS (continued) }I
LJ
PAGE
3.5 Data Set IV - Loading Data .................... 61 II
'| \
3.6 Data Set V - Shaft Input Data ................. 61
i1
3.6.1 Card Type S1 .......................... 62
3.6.2 Card Type S2 ........................... 63 ,_
it
3.6.3 Card Type S3 .......................... 63
4.0 COMPUTER PROGRAM OUTPUT ............................ 64 ii
iJ
4.1 Introduction .................................. 64
!
4.2 Bearing Output ................................ 64
4.2.1.1 Linear and Anqular -; ,
Deflections .................. 64 i
4.2.1.2 Reaction Forces and Moments.. 65
4.2.2 Fatigue Life Data ..................... 65 , i
4.2.2.1 h/sigm_ ...................... 65 'i !
4.2.2.2 Life Multipliers ............. 66 .!
4.2.2.2.1 Lubrication ............... 66
4.2.2.2.2 Material .................. 66 "i
!
4.2.3 Temperatures Relevant to Bearinq '"
Performance ........................... 66
I 4.2.4 Frictional Heat Generation Rate and
Bearinq Friction Torque ............... 66
I
4.2.4.1 Vrictional Heat Genelation ;i
Rate ......................... 66
4.2.4.2 Torque ....................... 67
4.2.5 EHD Film and Heat Transfer Data ....... 67 "|
4.2.5.1 EHD Film Thickness ........... 67 i !
4.2.5.2 Starvation Reduction Factor.. 67 ._
4.2.5.3 Thermal Reduction Factor ..... 67
!
4.2.5.4 Meniscus Distance ............ 67
8 4.2.5.5 Raceway-Rclling Element !I i
" Conductivity ...... 68 '''''''-',
iv '_
[]
1982024093-008
I .
TABLE OF CONTENTS (continued}
PAGE
f
_ 4.2.6 Fit and Dimensional Change Data....... 68
4.2.6.1 Fit Pressures ................ 68 \
i 4.2.6.2 Speed Giving Zero Fit Pressure
(Between the Shaft and Inner
Ring} ........................ 68
4.2.6.3 Clearances ................... 68
4.2.7 Lubricant Temperatures and Physical
Properties ............................ 69
4.2.8 Cage Data ............................. 69
4.2.8.1 Cage-Land Interface .......... 69
; 4.2.8.2 Cage Speed Data .............. 69
I 4.3 Rolling Element Output ........................ 69
i 4.3.1 Rolling Element Kinematics ............ 69
[.
4.3.1.1 Rolling Element Speeds ....... 70
4.3.1.2 Speed Vector Angles .......... 70
r- 4.3.2 Rolling Element Raceway Loading ....... 70
! i
4.3.2.1 Normal Forces ................ 70
4.3.2.2 Hertz Stress ................. 70
I 4.3.2.3 Load Ratio Qasp/Qtot ......... 71
4.3.2.4 Contact Angles ............... 72
i
4.3.3 Roller End-Flange Contact Data ........ 72
4.3.3.1 Normal Force ................. 72
l
4.3.3.2 Hertz Stress ................. 72
4.3.3.3 EHD Film Thickness ........... 72
4.3.3.4 Sliding Velocitl'. ............ 72
|" 4.3.3.5 Rolling Velocity ............. 73
I.
4.3.3.6 Contact Ellipse Semiminor
Axis 7" ,,,,.,,,,,,,,,,,,,,,,,,,,
[ 4.4 Thermal Data .......................... 13
4.5 Shaft Data ............. ...." ..... .ii 3
4.6 Program Error Messages. / 73
:t
5kf T[< F_N%.I<\ -_o[ ,,\ I< I ._
|
I
1982024093-009
TABLE OF CONTENTS (continued}
PAGE I
4.6.1 From Subroutine ALLT .................. 75 I
J
4.6.2 From Subroutine SHABE ................. 75
4.6.3 From Subroutine SOLVXX ................ 74 i
4.6.4 From Subroutine INTFIT ................ 75
i
5.0 GUIDES TO PROGRAM USE .............................. 76
6.0 LIST OF REFERENCES ................................. 78
APPENDIX A - SKF Computer Program AT81Y003 SHABERTH
Hierarchical Flow Chart .................... A:I
APPENDIX B - Heat Transfer Computation Notes ............ B:I
APPENDIX C - SKF Computer Program SHABERTH
Input Data Forms ........................... C:I
SKF Computer Program SHABERTH
APPENDIX D -
Sample Output .............................. D:I i
APP_ 3I., E - Calculation of Cage Pocket and Cage Land
Forces E:I eoeosoeoaeeee.eeeeoeooooooeoeoeeoeeoo
L
APPENDIX F - SKF and NASA Versions of Film Thickness and
Traction Force Calculations ................ F:I !!
'I
71
J
!I
vi
Ii
I
I
] 982024093-0] 0
r
!
ii!
LIST OF TABLES
] i
NO. TITLE PAGE
! -
I
{ 1 Lubricant Properties of Four Oils 54
!
1
!
, o
i.
r
i
i.
'" vii
1982024093-011
il
_Is, o,,,00_s !I
:I
NO. TITLE PAGE
2.1 Convective Heat Transfer 12 !J ,\
2.2 Divided Fluid Flow From Node i 13 []
2.3 Contact Geometry and Temperatures 16
2.4 Bearing Inertial (XYZ)and Rolling i_
Element (xyz) Coordinate Systems 22
2.5 Inner Ring-Cage Land Contact Geometry 24
2.6 Outer Ring-Cage Land Contact Geometry 25
3.1 Angular Contact Ball Bearing Geometry 36 -'!i
3.2 Split Inner Ring Ball Bearing Geometry 39
3.3 Tapered Roller Bearing Geometry 41 ..
3.4 Tapered Roller and Roller Raceway
Geometry 43 I
3.5 Cylindrical Roller Bearing Geometry 45
3.6 Roller-Raceway Lamination Showing _l
Relative Approach (6u) and Crown Drop
(_c) 47 !]
3.7 Cylindrical Roller Bearing Flange
Inversions 49 _
!t
a,
- ]
viii
I
] 982024093-0 ] 2
I
'" ATSIDO40
i
i FOREWORD
"I
The SHABERTH computer program was originally developed by Kell-
strom [1] under U.S. Army Contract DAAD05-73-C-00_sponsored by
the Ballistic Research Laboratory (BRL), Aberdeen Proving Ground,
Maryland, to simulate the thermo-mechanical performance of load
,. support systems consisting of a shaft supported by up to 5 roiling-
element bearings. The program ha_ ;ince undergone extensive de-
velopment to add new capabilities and improve its execution per-
formance.
This user's manual describes the use of a version of the
SHABERTH computer code developed under NASA-Lewis Research Center
Contract NAS3-22690 with Mr. H. Coe as Technical Monitor. The
revisions made to the program include:
I. Modification of the cage module to calculate
cage pocket and cage land forces i" ball, cy-
lindrical and tapered roller bearings, using
the models originally developed for NASA Com-
" puter Program CYBEAN [2]. An option has been
provided to allow the specification of single
or multiple degrees of freedom cage simulation
by input data.
2. Addition of an option which permits the program
to analyze a single ball or roller bearing
without the specification cf shaf _ geometry.
3. Combination of two versions of code within the
program: The SKF version and the NASA version.
The differences between the two versions re-
side in the calculation of the elastohydro-
dynamic (EHD) film tl.ickness and traction
forces which develop between rolling element-
raceway and rolling element-cage concentrated
contacts. The original film thickness models
(Archard-Cowking [3] and Dowson-Higginson [4])
and the Tallian traction model [5] are used
in the SKF version, while the NASA version
uses the Loewenthal model [6] to calculate
film thickness and the Allen model [7] to de-
_" t_rmine traction forces.
Additionally, a new subroutine, FLMFAC (replacing LRHS), is
used to determine the lubzicant life factor as a function o _ A
(film thickness/sufface roughness). The values of the lube
'" life f'_tor produced by FLMFAC adhere closely %o the curve recom-
mended _y the ASME [8. 9].
%
1
1982024093-013
ATSID040
A new section has been added, Appendix E, detailing the cal- i]
culation of cage pocket and cage land forces in ball and roller
bearings. Another new section, Appendix F, describes the dif-
ferences between the SKF and NASA methods of calculating film
thickness and traction forces, and explains the differences in i
executing each version of the code. "'
i-i
'i
I
:i
ii
1982024093-014
!-
I:
OR |NALPA-.IS
[ o,,,oo, ATS1D040 .
l
I. INTRODUCTION "
SHABERTH performs a thermo-mechanical simulation a
of load
support system consistin_ of a flexible shaft supported by up to '\
five (5) rolling element bearings. ]he shaft can be hollow or ' .i
} solid and of arbitrary geometry. Any combination of ball,
{- cylindrical, or tapered roller bearings can be used to support I
the rhaft. The cylindrical roller bearing analysis permits i
- :hrust load to be carried by inner and outer ring flanges. The . applied loading can consist of point or distributed moments, _
poznt or distributed forces, and shaft misalignments. {
[-
{. Concentrated contact EHD traction models [5,10] are included ,
in the prog,'am. Hydrodynamic rolling and shear forces in the
inlet zone of the lubricated contacts are accounted for [10].
f -
The effects of surface roughness [5 ] heating of the lubricant
.- in the contact inlet [11], and lubricant starvation [12] are
considered. Bearing operatin_ clearance is determined as a
Ii function of shaft and housing fits, component temperatures, and
rotational speedy13,14]. A cage model simulates contact between
the cage and rolling element as well as the cage and the piloting
,- land.
{
A lumped mass thermal model allows calculation of steady
state or time transient system termperatures considering free t
i and forced cenvection, conduction, radiation, and mass
transport
heat transfer[15,16,17].A maximum of one hundred (100) temperature
nodes can be used to describe the thermal system.
f-
I. The SHABSRTH program consists of the followinf major
9ubprogram3:
T"
{. I. Bearin_ Analysis These programs are largely
based upon te_ethods of Harris [18,19].
! 2. Three Dimensional Shaft Deflection -
Analysis
_eveloped by Norlan_er and Fri_d:-ichson.
methods of 7imoshehko[131 and adapted to the shaft-
bearing-housing system by Crecelius [14].
4. Generali:ed Steady State and Transient Temperature
_'_vr=n: _nd Heat Dis_zpation Analyses based on
_'.C -:T "''_
_._,_e of Harris [15] Fernlund [16] and
(:
-;,,.:'eason [171.
t
' 3
F"
1982024093-015
@
w-
ORIGINALpAGE iS ^T8 I D040
OF pOOR QUALITY _.
'i t
2. PROBLEH FOmgJLATION AND SOLUTION !i
The purpose of the program is to provide a tool with which "'
the shaft bearing system performance characteristics can be
determined as functions of system temperatures. These system []
temperatures may be a function of steady state operation or a ,-
function of time variant conditions brought on by a change
_n the system steady state condition. Such a change would be [|
the termination of lubricant supply to the bearings and _ther iJ
lubricated mechanical elements.
- !
The program is structured with four nested, calculation il
schemes as follows: '_
i. Thermal, steady state or transient temperatur: calcu- ]
lations which predict system temperat_rcs st a g_ven ]
)
operating state. J
2. Bearing dimensional equfl!br_u_ which uses the bearing
temperatures predicted by zhe temFer_ture mawping
subprograms and the rolling e_ement raceway load distri-
bution, predicted by the bearing subprograms, to calcu- I
late bearing diametral clearance at a given operating
state.
3. Shaft-bearing system load equilibrium whic,, calculates --'
bearing inner ring positions re_a_i'_-_to.the respective
: outer rings such that the external loading applied to the
sha_" is equilibrated by the rolling element loads i!
whi:h deve2op at each bearing inner ring at a given
state. _i
4. BearinK rollin element and cae lqad equi!ib;ium
which calculates the rolling element and cage equili-
brium positions and rotational speeds based upon the |
relative inner-outer ring positions, inertia effects
J
and friction conditions, which if lubricated, are
temperature dep,ndent. 11
The above proram structure allows complete mathematical
simulation o_ :he real thysical system. The program
has been coded te allow various levels of program execution
which prove useful and economical in bearing design studies,
These levels of execution are explained fully in Sections
3, 4, and S.
, I
I
I
1982024093-016
I
#
_. AT81D040
i
{. The structure of the program and the nesting of the solu-
tion loops noted above can be seen clearly in the Program Flow
- [" Chart which is discussed in Appendix A.
The sections below present the systems of field equations
( which are solved in each of the nested calculation schemes.
A more detailed discussion is contained in [I,i0 and 20].
2.1 Temperatur e Calculations
Subsequent to each calculation of bearing generated heat
rates, either the stead)" state or transient temperature mapping
i solution scheme may be executed. This set of sequential calcu-
', lations is terminated as follows:
I. For the steady state case; when each system temperature
is within EPI Centrigrade of its previously pre-
dicted value,(EPI is specified by the user). If it is
_! zero or left blank, a default value of 1 Centigrade is
I used. This criter_n implies that the steady state
' equilibrium condition has been reached.
[
, 2. The transient calculation terminates when the user
specified time up is reached or when one of the system
' temperatures exceeds 600C.
2.1.] Steady State Tempera%ure Map
The mechanical structure to be analyzed is thought of as
divided into a number of elements or nodes, each represented by
a temperature. The net heat flow to node i from the surrounding
nodes j, plus the heat generated at node i, must numerically
eoual zero. This is true fol each node i, i goin_ from 1 _o n,
n bein_ the number of unknown temperatures.
After each calculation of bearing generated heat, which
results from a solution of the shaft-bearing system portion
of the program, a set of system temperatures is determined
f
I which sat=sly the system of equations:
qi = qoi + qgi - 0 for all temperature nodes i (2. i)
I" where qoi is the heat flow from all neighboring nodes to
node i
i qgi is the heat generated at node i. These values
may be input or calculated by the shaft bearing
program as bearing frictional heat
,,[
5
[
1982024093-017
f_
m T.1DO40
_ i,oolI qum.rn'
" t}
This scheme is solved with a modified Newton-Raphson |]
e'hod which successfully terminates when either o two lj
c:_ditions are met:
ti
where: at represents the Newton-Raphson correction to the !]
temperature t at a given iteration such that,
tN+l = tN + At and N + I, and N, refer to the _-
next and current iteration respectively, kl
EP2 is a user speczfied constant. If EP2 is left blank rl
or set to zero f0) a default v&lue of 0.001 is used.
A second convergence criterion dependent upon EP2 is also
us_:d. In the system of equations, qo: q-i = 0 for all nodes !I
i, absolute convergence would be obtained _f the right hand
side (EQ) in fact reduced to zero (0). Usually a small residue
remains at each node, such that Cqoi "" qgi ) = CEQ)i. [I
Q
The second convergence criterion is satisfied if"
[ CEQ)i]< loo x EP2 C2. )
where n ; number of equations in thermal solution _J
numoer of unknown temperatures
2._..Z Transient Temperatures ':"]
6.
In the transient case the net heat qi transferred to a
nod._ i heats the element. It is thus necessary for heat balance ;|
at rode i t]_s_ the following equations are satisfied.
dt
I -. qi ;t
_i Cp i Vi _ C2.4)
,j
6
I
1982024093-018
ORIGINALPA_ m
J OFPOOR_AU._ AT81D040
where 0 = density
'- Cp- specific heat
V- volume of the element
t- temperature
T= time
The temperatures, t__, at the time of initiation T - T s are \
assumed to be known_that is
(
& .
ti(Ts) = toi i = I, 2, ..., n (2.5)
[
The problem of calculatin E the transient temperature
distribution in a bearin E arrangement thus becomes a problem
f
! of solving a system of non-linear differential equations of
, the first order with certain initial values given. The
equations are non-linear since the), contain terms of radiation
and free convection, which are non-linear with temperature as
i. will be sho_m later. The simplest and most economical way of
solving these equations is to calculate the rate of temperature
- increase at the time T _ T k from equation 2.4 and then calcu_-te
the temperatures at time Tk from
[ tk = tk dtk _T - tk _T (2.6]
_ rr- 0-rD---
If the time step dT used as program input is chosen
too large, the temperatures will oscillate, and if it is chosen
too small the calculation will be costly. It is therefore
desirable to choose the largest possible time step that does
not give an oscillating solution. The program optionally cal-
culates such a time step. The step is obtained from the
condition, [16]
dti,k \ > o i- z, z, ..., n (2.7
dti,k
I" If this derivative were negative, the implication would
I be that the local temperature at node i has a negative effect
on its future value. This would be tantamount to asserting
[_ that the hotter a region is now, the colder it will be after
an equal time interval. An oscillating solution would result.
o
L
[
[ 7
1982024093-019
OR,O,N ''" il
OF POOR QUALITY AT81D040
: Differentiating equation (2.6) for node i, one obtains [}
I" dti,k < = 1 aTi . dqi.k i " 1,2,...n (2.8) i_t_
ii dti'k oiCpiVi _ i]
! .I
J
; The derivative dqi,k/dt i,k is calculated numerically _i'
dqi, k = qi(ti * Ati) " qi(ti) (2.9) _i
I
-'l,k at i
8
-t
For each node the value of aT_ giving a value of zero to I
the right hand side cf Eqn. (2.8) _s calculated. The smallest
non-zero value of AT obtained in this manner is chosen as the
time step .
I
2.1.3. Calculation of Heat Transfer Rate
,, !
The transfer of heat within a medium or between two media
can occur by conduction, convection, ratiation and fluid flow.
i
All these types of heat transfer occur in a bearing appli-
cation as the following examples show.
!
1., Heat is transferred by conduction between inner tin z _}
and shaft and between outer ring and housing, i
.o
2. Heat is transferred by convection between the surface
of the housing and the surrounding air. { I
3. Heat is transferred by radiation between the shaft
and the housing.
4. When the bearing is lubrzcated and cooled by circulating
oil, heat is transferred by fluid flow.
Therefore, in calculating the net heat flow to a node all the
above mentioned modes of heat transfer will be considered.
2.1.3.1 Generated Heat
There may be a heat source at Pode i giving rise to a heat !_
flow to be added to the heat flowing from the neighboring nodes.
i
I
1982024093-020
I
ORIGINALPAGE IS AT81D040
OF POOR QUALITY
In the case that the heat source is a bearing, it may
-. either be considered to produce known amounts of power, in which
r case constant numbers are entered as input to the program, or
,o the shaft-bearing program may be used to calculate the bearing
generated heat as a function of bearing temperatures.
L 2.1.3.2 Conduction
The heat flow q which is transferred by conduction
f _
from node I to node _,'_s proportional to the difference in
L
temperature (ti - ti) and thecross sectional area A and
= is inversely p_opor_ional to the distance _ between the two
points, thus
= IA (ti- tj) (Z.iO)
I qci,j l--
where _= the thermal conductivity of the medium.
[" 2.1.5.5 Free Convection
! ,, |
Between a solid medium such as a metallic body and a
[- liquid or gas, heat transfer is by free or forced convection.
!
i _ Heat transfer by free convection is caused by the setting in
motion of the liquid or gas as a result of a change in density
"_" arising from a temperature differential in the medium. With
[ free convection between a solid medium and air, the heat
energy q transferred between nodes i and j can be calculated
_- from theV_ation, (2.11)
l
qvi,j = evA (ti'tj) d. SiSN(ti.tj ) (2.11)
[-
: where % - the film coefficient of heat transfer by free
convection
_ A - the surface area of contact between the media
; d = is an exponent, usually - 1.25, but any v?lue
can be specified as input to the program
I
SIGN , (: if ti) tj if ti< tj
I
9
I t"
] :.
i
1982024093-021
"b- ..mm_'
ORIGINALPAGE IS
OF POOR QUALITY AT81D040
The last factor is included to give the expression qvi,j
correct sign.
The value of Sv can be calculated for various cases, see
Jacob anO Hawkins, [21]
2.1.5.4 Force _ Convect%pn
Heat transfer by forced convection takes place when liquid [_
or gas moves around a solid body, for example, when the liquid
II
is forced to flow by means of a pump or when the solid body is
moved through the liquid or gas. The hear flow q_ _ transferrea
by forced convection can be obtained from the fol_0,_ng equation. _|
_2
qwi,j "w A(ti " tj) (2.12) I!
; where ow is the film coefficient of heat transfer during F_
: forced convection. This value is dependent on _)
the actual shape, the surface condition of the
body, the difference in speed, as well as the
properties of the liquid or gas. ,i
In most cases, it is possible to calculate the coefficient
i of forced convection from a general relationship of the form, [I
i
|
I aRebPrC (2.13) i Nu" 1
I
where a, b, and c are constants obtained from handbooks
.]
_ such as [22]. Re and Pr are dimensionless numbers
defined by
Nu - Nusselt number " =w L/_
L - characteristic length
; _ " conductivity of the fluid _l
Re = Reynold's number . ULo/n
U : characteristic speed
densit" of the fluid [i
- dynamic viscosity of the fluid
!
Pr - Prandtl number - nCp/X
Cp - specific heat "I
!
I
1982024093-022
"t
l
GRIGINALPAGE IS AT81D04 0
OF POOR QUALITY :
The user can input a constant value for the convection coefficient.
Alternatively, he can let the program calculate the coefficient
using one of the three option described below. For options 2 and 5,
the coefficient is allowed to vary with system temperatures.
Constant viscosity \
1. Values of the parameters of equation (2.13) are given
. as input and a constant value of % is calculated
by the program.
- DT
j Temperature dependent viscosity
2. The coefficient aw for turbulent flow and heating of
petroleum oils isgiven by
- aw - k9 . n(t)klo (2.14)
where kg and klO are given as input together with viscosity
at two different temperatures
3. Values of the prameters of equation (215) are given
as input. Viscosity is given at two different
temperatures
- 2.1.3.5 Radiation
If two flat parallel, similar surfaces are placed close
together and have the same surface area A, the heat energy
transferred by radiation between nodes i and j representing
those bodies, will be,
- oA (ti-2_5) 4 - Ctj+:75) 4 (2.15)
& ,
qRi,j
] where c is the surface emissivity. The value of the , , coefficient is an input variable and varies
between I for a completely black surface and 0
i for an absolutely clean surface. In addition
o is Stefan-Bolt:mann's _adiation constant
which has the value 5.76 x 10 .8 watts/_2-CK) 4
and _i and tj are the temperatures at points i
and 3.
Heat transfer by radiation under other conditions can also
be calculattS, [21]. The following equation, for instance
applies between two concentric cylindrical surfaces.
f!
I
1982024093-023
OmGINAI. I'AOIel A_,IDO,O il
OFPOORqu_UTY
" !1
qRi,j I"+ (l-c) (Ai/Ae)' (2.16)
where Ai is the area of the inner cylindrical surface [!
Ae is the area of the outer cylindrical surface
2.1.3.6 Fluid Flow i
Between nodes established in fluids, heat is transferred 1
by transport of the fluid itself and the heat it contains.
!
JII1
qoi !
t
/ / / / / / / /
'| ! " ' i
I _,_ J I
I _ I _ _j_, j I .;
! I I i
.2 / / / '/ / / /' _
t
t
FIGURE 2.1 CONVECTIVE HEAT TRANSFER ""
i
Figure 2.1 shows nodes i and j at the midpoints of consecu-
tive segments established in a streau, of flowing fluid. _l
.]
The heat flow qui,j through the boundary between nodes
i and j can be calculated as the sum of the heat flow qfi through .]
the middle of the element i, and half th_ heat flow qoi trans-
l
ferred to node i by other means, such as convection.
The heat carried by mass flow is, ,
qfi " _z Cpi Vi ti " Xiti (Z.17) i
J
where Vi - the volume flow rate through node i
, .
t
12
] 982024093-024
ORIGINALPAGEIS AT81D04 0
OF POOR QUALITY
The heat input to node i is the sum of the heat generated
at node i (if any) and the sum over all other nodes of the
heat transferrea to node i by conduction, radiation, free and
forced convection.
qoi " qG,i " _,(qci,j * qvi,j * qwi,j + qRi,j ) (2.1_) \
The heat flow between the nodes of Fig. 2.1 is then
I qui,j qfi qoi/2 (2.19)
If the flow from node i is dividing between nodes j 6 k,(Fig. 2.2)
then the heat flow is calculated from
qui,j = Kij (qfi qoi/2) (2.20) "w
J
where Kij = the pToportion of the flow at i going to
l no_einput, j' 0 <Kij _<1- Kij is specified at
=Kik.Zl
] K. - K... Ki I
1 , kL
: J ,
FIGURE 2.2
DIVIDED FLUID FLOW FROM NODE i
_t -=.I3.7 To..talHeat Transferred
_'_ The net heat flow rate to node i can be expressed as,
m
[_ qi " qG i + X + + + * (2 21)
, j=_qci,j qui,j qvi,j qwi,j qRi,j) "
I 13
,I
I
i 982024093-025
OM_ pAN m ATalD040 !
OrpOORqUAUTV
The summation should include all nodes j, both with i]
unknown temperatures as well as boundary nodes, at which the r]
temperature is know_ so long as they have a direct heat exchange
with node i. i] k
This expression is a non-linear function of temperatures \
because of the terms qw and q_. Therefore the equations to be _
solved for a steady state solOtion are non-linear. The sub- _i
program SOL_X for solving non-linear simultaneous equations is
used for this purpose.
2.1.4 Conduction Through a Bearing !_ P!
As described in Section Z.1.3.2 the conduction between "t
two nodes is governed by the thermal conductivity parameter _
of :he medium through which conduction takes place. The value
of _ is specified at input.
1
An exception is when one of the nodes represents a
bearing ring and the other a set of rolling elements. In this
case the conduction is separately calculated using the princi-
ples described below. Note that separate calculations are i
: performed for the rolling element raceway contacts and the roll-
ing element-flange contacts. The methods for both calculations
are identical and are performed within the program.
2.1.4.1. Thermal Resistance
m ,m i
I
o
It is assumed that the rolling speeds of the rolling
elements are so high that the bulk temperature of the rolling
elements is the same at both the inner and outer races, except
in a volume close to the surface. The resistance to heat flow '
can then be calculated as the su_ of the resistance across
the surface and the resistance of the material close to the i
surface. ,l
The resistance _ is defined _mplicitly by _
|
t =n.q (2.zz)
where !
at is temperature difference
q is heat flow
The resistance due to conduction through the EHD film is
calculated as
1
n,. h (2.Z3)
VX-
14
I
1982024093-026
q
"_ ATSID040
ORIGINALPAGE IS
OF POORQ_ALITY
]
where h is taken to be the czlculated plateau film thickness
A is the Hcrtzian contact area at the specific rollinb element-ringa_tact under consideration.
A is the conductivity of the oil. \
The geometry is shown in Figure Z 3(a)
So far, a constant difference between the
temperature
surfaces has been assu_ed. But during the time period of
contact, the difference will decrease because of the finite _
] thermaZ diZusivity o the material near the s'_face, Fig. 2.3
(b). .:
To points at a distance from the surface,this phonomenon
j will have the same effect as an additional resistance _2 acting
in series with _I.
_! This resistance was estimated in [23] as, :
n2= _re,i ( ]_
,
' where _re = contact length, or in the case o_ an
elliptical contact area, 08 times the
major axis
L
, _ = heat conductivity
= thermal diffusivity = _/(O.Cp)
' _ = density
Cp = specific heat
b = half the contact width
V rolling speed
The resultant resistance is
ares _i + 'q2 (Z.25)
15
1982024093-027
M_
oRIGINAl- PAGEB
oFpooeQuALrrI ATeZD040 [I
_:VERAGE FILH i_I
THICKNESS h
TE'_IPERATUREti _ i]
J
i}
(a) Schematic Concentrated Contact
r
i[! ,,ioo4o
/
OR!GINALPAGEIS
OF POORQUALITY
'_ There is one such resistance at each rolling element.
They all act in parallel. The equivalent resistance _ eqv \
is thus obtained from
I n
- _ Z (2.26)
2.2 Bearin_ Dimensional Chanse Analysis
The program calculates the changes ir bearing diametral
clearances according to the analysis described in [14], and
expressed in generalized equation form as,
ADCL - f {(FitS)m , ti, _m, (Qr)m }, m o 1,2 for inner (2.27)
and outer
rings respectively
7 i " 1,2,3,4,5 for
shaft, inner ring,
outer ring, housing
and rolling element
:I respectively
17
1982024093-029
A,,,oo,o I}
ORmmAtPA_Km
OFlOOnqU_UTY l)
t]
where: ADCL is the change in bearing diametral clearance
Fits are the cold mounted shaft and housing fits.
a=e the component temperatures [_
refers to the ring rotational speeds
Q_ refers to the radial component of the minimum
tolling element-race normal force [)
!
A bearing clearance change criterion is satisfied when
1
the change in bearing diametral clearance remains within I
a narrow, user specified range, for two successive iterations
as follows:
](ADCL)N - (ADCL)N.II< EPSFIT for all bearings (2.28)
where: N denotes the most recent iteration and
N-I denotes the previous iteration,
D denotes _he ball or r_ller diameter and i
PSFIT is a user specified value,(default value = .O001D)
It should be noted that although ring rotational speeds, /
and initial, i.e. cold, shaft and housing fits are considered
in the clearance change analysis, these two factors are fixed I
at input and remain constant through the entire solution. .j
Although component temperatures may change as a consequence
o_" the thermal solution, temperatures remain constant through I
a complete set of clearance change iterations. As a result, only _j
the change in bearing load distribution affects the change in
bearing clearance within a set of clearance change iterations.
2.5 Bearing_ Inner Ring Equilibrium :I
The bearing inner ring equilibrium solution is obtained [I
by solving the system:
]
(FMb_i (FMs) i - 0 for all bearings, i (2.29) I
where: F_b denotes a vector of bearing loads and
moments resulting from rolling element/race
forces and moments.
18 !
1
1982024093-030
i r , - . ,.
!
' ORIGINAL PAGE iS ATSID040
] OF POORQUALIlrY
Fbxi
"': Fby i Forces
-,. (2.30)
-_ FMbi - Fbz i
Mby i Moments
If the bearing solution considers friction, FM_ is com-
i prised of the ball race friction forces as well as _he
normal forces.
i If the bea_ing solution is, at the user's option,
frictionless, FM_ is comprised only of rolling element/race normal contact f6rces
FM . denotes a similar vector of loads, exerted on the
inner r_g by the shaft.
'I Fsxi
,l -,. Fsy i Forces
FMsi - Fsz i
I, (2.3l)
Msy i Moments
I Mszi
" The variables in t is syst*m of equations are the_bearing
inner ring deflections _. and the shaft displacements As at
all bearin_ locations. The bearing loads may be expressed as
i a function of the inner ring deflections.
-4, -e. ._
FMb - FMb CAb) (2.32)
I
r" z9
i

1982024093-031
OmaHAL m
il
The deflect.ion _bOf a bearing is described by two radial
deflectins _v and 6z tw angular deflectins : and ezandan t)
axial deflection 6 x. The axial deflection is a s_med to be
the same for all bearings and the shaft.
f!
The solution scheme is ended when [J
-4.
6 (_)i _ < EPS1 (frictionless) i] \
- " (2.33)
(_)ij EPS2 (friction)
rl
j = l,S - for the 5 linea_ and tw_
angular deflections at i|
each bearing [1
If for some i or j, (4) ij = O, Eq (2.34) is used in place of
(2.33) .,. I]
I#
(5(4) ij <__EPSI(frictionless) (,2.34)
(0.0Ol x NBRG) ESP2 (friction) t
NBRG denotes the number of bearings in the system,
EPSI or EPS2 is used depending on whether the bearing solu- i
tions are fricitonless or include friction, respectively. If
the bearing deflections are extrem.-ly small, computer-generated I
numerical inaccuracies may prevent convergence according to the I
above criteria although a perfectly good solution has been ob-
tained. To overcome this problem, the iteration is terminated
if all angular deflections are less than 2 x 10.6 radians i
and all linear deflection_ are less than 5 x I0-8 inches. Any
I
one of the above criteria imply that inner ring equilibrium is
satisfied, j
:.4 Bearing Qua_i-Dyna,m.icSolution
1
The bearing quasi-dynamic solution i._obtained through a ,j
two step proc-.-ss :
!) Elastic Solution - considering rolling element centri- ,,
fugal force,
plus the gyroscopic moment for a tapered i|
roller. *
i
2) Elastic and Quasi-dynamic Solution*
rue - ,I
dynamic equilibrium terms containing first derivatives of the
ball rotational speed vectors and the second derivatives of I
rolling element position vectors with respect to time are
!
replaced by numerical expressions which are position rather
than time dependent. }
..J
20
t
I
1982024093-032
ORt(_|NALPAGE IS
OF PoOR QUAtrrl[ ATSID040
The equations which define rolling element quasi-dynamic
force equilibrium takes the form
m - 1-3 refers to the outer
inner and cage rolling element
contacts respectively
m = 1-4 for tapered rollers with
_llrram" " -_ _ one roller end-flange contact.
c%.
m _,herefour roller and flange [2.35)
contacts are possible.
where:Qm is the vector normal load per unit length
, of the contact. See Ref. [I].
fm is the vector of friction force per unit length
of the contact. See Ref. [10].
F is the vector of inertia and drag forces. See Ref.[1]
s is a coordinate along the contact
perpendicular to the direction of rolling (usually
=he major axis)
is half the contact lengzh. See Ref. [1].
Fm is the vector sum of ths hydrodynamic forces
acting on the rolling element at the m-th contact.
For ball-raceway contact see Ref. [i0]. For the
roller-raceway contact, see Ref. [21].
Rolling element moment equil_hrium is defined by:
Z my x )d m x I = 0 [2 36) m m m "
-a m _ _ _
l Qm,fm,Pm, and s are defined above, M is a vector
J am' of inertia moments. For t_e definition of M I
refer to Ref. [i].
rm is a vector from the rolling element
J center to the point of contact.
In the frictionless elastic solution Fm and fm " 0.
Additionally, the only rolling element inertia term considered
in the frictionless solution is centrifugal force, plus the gyro-
scopic moment for tapered rollers. As a consequence only the
axial and radial force equilibrium equations are solved for each
ball. For each roller the radial and axial force equilibrium
and the tilting moment about the z axis of Fig. 2.4 is solved.
A dummy equation for axial force equilibrium is included in the
21
]
I
1982024093-033
!]
ORIGINAL PAGE _l AT81D040
fJ [I
z
/ [I
/ f}
" 1
\ _.,'6) , /I
I,,"
",_OJ x
I
,,,,, ,, Y
\ x
.,' X !j
\ Y
CAGE POCKE" CENTER
i FOR BALL N9. i i]
CENTER BALL NO. I
y
, !t
FIGURE 2.
i
t
BEAP, I_;G INERTIAL (XY:) AND ROLLING ELEHE._;T (xyz),
COORDI:;ATE SYST_15
22 .i
1982024093-034
ATSID040
solution matrix which keeps the roller centered with respect to
the outer race if the cylindrical roller bearing takes no thrust
load.
The friction solution determines ball quasi-dynamic equilib-
rium for six degrees of freedom. The rolling-element variables
in this solution are xl, yl, _x, _y, _z, and _o.
i where the rolling element axial position relative
xlis
t_ the outer race groove curvature center.
YliS the rolling element radial position relative
t_ the outer race groove curvature center. wx, _y, _oz are orthogonal rolling element rotational
speeds relative to the cage speed, about the x, y, and
[_ z axes and We is the rolling element orbital speed.
The variables x_ and Yl are the ball unknowns in the frictionless
solution. The ?ariabl_s in the roller frictionless solution are
[i
xl' YI' and Oz = arctan (wT/_xl
Details of the cage analysis are contained in Appendix E.
Ii one or cage equilibrium equations are
Either three considered, de-
pending upon the number of degrees of freedom given to the cage.
[ The cage equations and cage-rolling element interactions are
I. not considered when the friction forces are omitted from the rol-
ling element equilibrium equations.
2.4.1. Cage Degrees of Freedom
The program has been modified to allow the user to specify the
I number of degrees of freedom (DOF} of the as either 1 or 3.
cage
The single degree of freedom corresponds to a smaller angular rota-
tion about the bearing axis, measured with respect to rolling
I element i. The angular displacement is converted to a linear di- mension by multiplying it by the bearing pitch diameter and is
noted in Fig. 2.4 as z . When a single degree of freedom is in-
put, the sum of moment_ acting on the cage about the bearing X
I axis is required to be zero. This moment equation considers the
cage-rolling element normal and friction forces as well as the
torque generated at the cage-ring surface.
f-
If the user assigns three degrees of freedom to the cage, it
is permitted to move to an eccentric position with respect to the
i land on which it is piloted. The additional degrees of freedom are the cage center of mass radial displacement, e, and the angular
displacement 0 of the center of mass, with respect to the bearing
Y axis. (See _igures 2.5 and 2.6.) The radial friction forces
i as as pressure build-up between the cage and its piloting
well the
surface are considered in the equilibrium equations. The effect of
the cage mass is neglected.
f-
23
f-
t
i
1982024093-035
* i1
ATSID040 ,
* oRiGiNAL P&Glr" IS
OF poOR QuAUTY I}
/
' 1
"1
' i
J INNER RING
' 1 l
J
,J
_! iI
: ' i]
j
FIGURE 2.5 INNER RING-CAGE LAND CONTACT GEOMETRY i
:I
24
3
,j
i
1982024093-036
AT81D040
ORIGINAL PAGEIS
OF POOR QUALITY
"! h" ,,
le " r% _ !
]
OUTER RING -_
F
E
FIGURE 2.6 OUTER RING-CAGE LAND CONTACT GEOMETRY
E
[
1982024093-037
!
II
ATSID040
; Occasionally, the program will determine that an input value [I
I of three degrees of freedom is impractical, in which case it will
] override the input and allow the cage to have only one DOF. This
occurs if the cage will tend to rotate concentrically with respect
i to the ring on which it is riding. Such a condition is determined II
i as a function of the rolling element orbital speed variation and
LJ
prevails with most roller bearings and with ball bearings subjected
only to axial loading. In both cases, orbital speed variation is [i
often inconsequential. Also, when the cage rides on the rolling
tl
elements, it is allowed only one degree of freedom.
i] The ball bearing friction solution is thus obtained by solving
6Z+(I or 3) equations where Z is the number of rolling elements.
The ball bearing frictionless solution is obtained by solving I, I
(Z/2) (Z/2+l) or Z sets of 2 equations, depending upon the number i
of rolling elements in the bearing and the degree of load symmetry
!
which prevails. The various symmetry conditions are explained be-
low. I
The roller bearing friction solution contains 4Z+(I or 3)
equations and the frictionless solution contains Z/2, Z/2+l or Z i
sets of three equations again depending upon the number or rolling I
elements and whether or not load symmetry exists.
The various load symmetry conditions are as follows. Axial }
symmetry is utilized if the load is axial only, then only one set
!
of two or three equations is solved for the frictionless case and
six rolling element and one cage equilibrium equations are solved
when friction is included. All rolling elements are assumed to I
behave identically.
)
Radial load symmetry is utilized if the non-axial shaft load- I
ing is comprised of only radial components parallel to the Y axis J
and moment components parallel to the Z axis and the position of
the first rolling element is utilized. When this symmetry exists, I
only half the rolling elements need be considered if the number
J
of rolling elements is even and one half plus one need be considered
if the number is odd. Because of inertia terms, radial load sym- |
metry can only be utilized in the frictionless calculations.
J
If load symmetry is not present, then Z sets of two (ball
bearing) or Z sets of three (roller bearing) equations must be 'I
solved to obtain the frictionless solution, i
As with the steady state temperature mapping scheme, the
Newton-Raphson scheme in subprogram SOLVI3 is used to solve the
sets of equations for each bearing. The iteration schem_ term-
inates when either:
26
J
1982024093-038
-] AT81D040
E EPSI frictionless
(K-I) i i=l...n \
_EPS2 friction
E or Where K and K-I refer to iteration numbers
EQ < i00" EPSI frictionless
1 (2.38)
r n _EPS2 friction
L l
Experience has shown that the second criterion is usually
responsible for te_inating the solution. However, w_en rolling
r element loads are extremely large, on the order of i0 = Newtons,
_. it becomes difficult to reduce the equation residues to less
than I0 Newtons. In those instances, the first criterion usually
_[" terminates the iteration scheme.
[
I _
i
I .
o
I
I:
[_ 27
i
1982024093-039
AT81D040 I
3. PROGRAM INPUT
3.1 Types of Input Data il
A complete set of input data comprises data of four distinct _I
categories. Within these categories, cards which convey specific {i
kinds of information are referred to as card types. Depending
on the complexity of the problem, the input data set may contain !]
Done, one or several cards of a given type. The categories are
Jl
listed below.
[1
I. Title Cards
: A title card plus a second card which provides the
program control information for the shaft-bearing
solution. I
II. Bearing Data Cards
A set of up to sixteeen (16) card types. Each set 1
describes one bearing in the assembly. All bearings 1
must be so described. The card sets must be input
sequentially in order of increasing distance from a
selected end of the shaft. I
III. Thermal Data Cards
A set of up to nine (9) card types to describe tne i
thermal model of the assembly.
IV. Loading Data Card
One card describing the loading on a single bearing.
This card type is used only when Shaft Data Cards are
omitted.
V. Shaft Data Cards
A set of three (3) card types to describe the shaft
geometry, bearing locations on the shaft and shaft load- i
ing. Used only when Loading Data Card is omitted. _i
If the program is being used to predict the performance of a
beering assembly, cards from sets I, II, III, and V must be in-
cluded in the runstream. If the program is being used to thermally
model a mechanical system wherein no bearing calculations need be
performed, the cards from sets II, IV and V are omitted.
The review of required input information which follows is
broken into the five sets of data categories given above, with
special emphasis on program control data.
The input data instructions are given in Appendix C, and are
for the most part, self-explanatory. They are laid out in the
format of an eighty column data card. A description of the var- _
iables is given in the input instruction forms.
28
J
1982024093-040
ATSID040
E
The units used for input data are as follows"
Linear Dimensions - (mm)
[_ Angles- (degrees)
Surface Roughness (microns)
Bearing Angular Mounting Errors - (radians) \
Rotational Speeds - (RPM)
Force- (Newtons} (N)
Moments- {N-mm)
Pressure, Elastic Modulus o (N/mm2}
Density- (Zm/cm 5)
Kinematic Viscosity - (cs)
Temperature - (degrees centigrade) (C)
Coefficient of Thermal Expansion - (oc-!)
Thermal Conductivity - (Watts/m/C}
I 5.2 Set Title Cards
Data. !-
5.2.1 Title Card 1
This card should contain the computer run title and any
information which might prove u.eful for future identification.
i The full eighty (80) columns are available for this purpose.
The title will appear at the top of each page of Program output.
-- 5.2.2 Title Card 2
This card provides the control information for the shaft
( bearing solutior.
: Item I: Shaft Speed in rpm, GOV (11. All bearings have
the sam--6"-_'h'_t and inner ring speed.
Item Z: Number of Bearings on the shaft (NBRG), a minimum
of zero is p-6rmitted if no bearing solution is bein.E sought. A
maximum of five is permitted. Note that a bearing is defined as
a single row of rollinE elements. Thus a double row bearing is
treated as two separate, single row bearings.
Item 5: Print Flag (NPRINT), NPRINT equal to zero is normal
and will result in no intermediate or debug output. With a value
of one, a low level intermediate print is obt&ined at the end cf
i each shaft bearing iteration. The values of the i_ner ring dis-
placements (DEL), eoua'.icn residues, bearing inner ring residual
loads, and bearin_ partial derivatives are printed for each iteration.
. Th_s level is recommended in cases where proper convergence of the
[ bea-in solution does not occur.
|.
At the end of each bearing iteration, wherein the rolling
i element and cage equilibrium _quations are solved,
an error
pars-
meter _s rrinted which has the value:
i
1982024093-041
!
_- AT81D040
t]
Error Parameter = &XN/ XN. 1
_X N is the change in the variable X specified at i]
_' iteration N.
is the value of the variable specified at XN- 1
the previous iteration.
t
The Hrror Parameter is calculated for each of the bearing "I_
variables, but only the largest one is printed.
Additionally, at the end of each Clearance Change itera- II
tion, the clearance change error parameter is printed. This
,i
error is defined:
t
DCL N - DCLN. I i
Error Parameter = Ro!Ting Element Diameter
!
where DCL N and DCLN. I denote the clearance changes calculated
at the current and previous iterations respectively. 1
!
If NPRINT is set at 2 all of the above information is
printed. Additionally the variable values and residue values
, are printed for each iteration of the rolling element and cage
]
equilibrium solution. This level is not generally recommended be-
= cause of the large volume of output produced. -i
I
Item 4: ITFIT controls the number of iterations allowed to
: satisfy the bearing clearance change iteration scheme. If ,|
ITFIT is set to zero (0), or left blank, the clearance change ]
portion of the program is not executed. If a positive integer
is input, the clearance change scheme is utilized with a maximum
iteration limit of five (5). If a negative integer is input,
the scheme is used with a maximum iteration limit equal to
the absolute value of the negative integer.
[i Item 5: ITMAIN limits the number of iterations attempted
duringt_solution of the shaft and bearing inner ring equili-
brium problems, i.e. establishing the equilibrium of bearing I
reactions and applied shaft loads. If ITMAIN is left blank,
I
set to zero, or to a positive integer, then (15) iterations
are permitted. If ITMAIN is set to a negative integer the I
number of iterations is limited to the absolute value of that _
integer.
t
30 i
i
I
1982024093-042
q
AT81D040
Item 6: GOV(2) or EPSFIT is the convergence criterion for
the diametral c_ance change portion of the analysis. As men-
tioned under item 3 above, this error parameter is defined by Eq.
2.28.
The iteration scheme is terminated when the error parameter \
is less than the input value of EPSFIT. If EPSFIT is left blank
or is set to zero (0), the program default value of 0.0001 times
the rolling element diameter is used.
Items 7 & 8: Main loop accuracy for frictionless elastic (EPSI) and friction solution (EPS2). These accuracy values con-
trol the accuracy of the shaft bearing deflection solution as
well as the quasi-dynamic solution of the component dynamics. If
are or set to zero (0), default values
EPSI and EPS2 left blank
of 0.001 and 0.0001 respectively are used.
Item 9: JUSTBR, column 78, is a flag indicating whether or
L
not asl--{_e ball or roller bearing is to be analyzed. If a
value of 1 is input, data are needed for only one bearing, and
loads are input on Loading Data card L1 following the Thermal
Data cards. Also, Shaft Data cards are omitted. If JUSTBR is
O or blank, up to 5 bearings are analyzed with Shaft Data, and
the Loading Data card is omitted.
:-. Item i0: IMT, if set to i, the material properties for both
bearing rings and the rolling elements are to be input on card
types B ii through B 14. If IMT is zero or blank, the rings and
[: rolling elements are assumed to be 52100 bearing steel. Card
types B ii through B 14 are required if the change in bearing
diametral clearance is to be calculated or if a system component
has properties different from steel.
Item ii: NPASS controls the level of the bearing solution:
F "
I
[ 0 Elastic Contact Forces are calculated. No lubrication
or friction effects are considered.
1 Elastic Contact Forces are calculated. Lubrication
and friction effects are considered usinq raceway con-
trol (ball bearing) or epicyclic (roller bearing)
assumptions to estimate rolling element and cage speeds.
[
31
!
1982024093-043
AT81D040
2 Inner Equilibrium is satisfied considering only II
the Elastic Contact Forces. Using the inner ring 11
positions thus obtained, rolling element and cage
equilibrium are determined considering friction, i_
3 Complete Solution. The inner ring, rolling element
and cage equilibrium is determined considering all l
elastic and friction forces, il Q _
3.3 Data Set II - Bearing Data
'i Most of the input instructicns are self-explanatory. Where I
certain items are deemed to require more explanation than given
in the input data format instructions they are treated on an in-
dividual basis by card type and item number, i
Most of the bearing input data is read into a two dimensional
array named "BD," which has the dimensions (1830, 5). For each
of the _ive bearings permitted on a shaft, a total of 1830 pieces
of data may be stored. Denoting BD(I,J), I represents a specific
piece of bearing data, J represents the bearing number. The
bearing input data of Data Set II occupies the first 106 locations
of the 1830 allotted. On the input data format sheets the des-
ignation BD(I) where I=i...106, denotes the location within the BD
array where each piece of input data is stored.
I
1982024093-044
I_ 5.3.1 C@rd T>'pe 1 - Bearing T_e and Material Desisnations
.i \
[_ Item 1 : Besting tyMp_ , c _ _um_s 1-10 must be specified,
L_
"B" "T" ttC''
left justified, i.e., , or in column i. This format
must be followed since the Program recognition of bearing type,
[ (ball, cylindrical or tapered roller bearing), is derived from
( reading the "B" "C" "T" , , or in the first column of this card.
r- Items 2 6_3: Columns 13-30 and 31-50, "Steel designations",
t inner and outer rings respectively. The alphameric-literal
"_ description of the steel types such as "M-50" or "AISI 52100"
is input. ,-1
Items 4 _ 5: Columns 51-60 and 61-70, the numbers input
for items 4 and 5 are used to account for improved materials
i and multiply the raceway fatigue lives as determined by
! Lundberg-Palmg-en methods. Typical life factor values for modern
steels are in the neighborhood of 2.0 to 3.0. If the AS_fE
Publication l,ife Adjustment Factors for Ball and Roller BeaTings,
is referenceT]-'oy fhe user, the Material Factor D and the
Material Process Factor E should be used multiplicatively
as inputs for items 4 and 3. The program computes a lubricant
life factor based on the value of h (EftD plateau film thickness/
composite RMS surface roughness). The calculated lube life
factor ranges from 0.21 for h < 0.6 to 3.0 for h > 10.0.
Item 6: Columns 71-78, "Orientation angle of the first
rolling _lement". (1) (degrees). Refer to Fig. (2.4). The
:- quasi-dynamic rolling element bearing problem has an infinite
i number of solutions which fall within a narrow envelope having
a periodic shape. The solution obtained is a function of
the rolling element positions relative to the bearing system
' coordinate axes. 1 : 0 places a rolling element on the Y
_ axis and i_ the cholc customarily made. 1 can be desig-
33
l
l
1982024093-045
AT81D040
11
i|
nated as any value 0<_I< 360/Z where Z is the number of rolling (I
elements. For each _i_erent value assigned to I a different
ent, although similar, bearing solution will be obtained. To
computer time savl"njs !I take advantage of beating symmetry and the
which result_ @I must be specified as zero or Lez DAank:
Item: 7 Column 80, a signal, termed the crown drop fl_g, II
which _pecifles for a cylindrical or tapered roller bearing,
t!
whether the roller-race crown drops will be calculated, or read
directly. If item 7 is blank or zero, the crown drops are cal- I_
culated based on the roller-race crown radius, and effective flat I_
length input information. If the crown drop flag is other than
zero or blank the non-uniform separation of the roller and race- i_
way must be specified at the center of each slice into which the
roller-raceway effective contact length is divided. The slice
widths are identical. The number of slices is input as item 7
card type B4. The non-uniform roller-raceway separation is in-
put on card types B5 and B6.
5.5.2 Card Type B2 - Bearin_ Geomet.ry..zndOuter Ring Speed
5.5.2.1. Ball Bearing Geomet.ry
Items I, 2 and 6 are self explanatory. Item 5 pertains
only to a tapered roller bearing, as discussed later. Items 5
and 4 require explanation however.
i
Through the proper specification o the diametral clearance
and contact angle, the Program can properly handle deep groove,
split inner, and angular contact ball bearings. I
The deep groove bail bearing requires the specification I
of the contact angle corresponding to either the operating diame- I
tral clearance Pd or the off the shelf diametral clearance, if the
!
dimensional change analysis is utilized.
The angular contact bearing is fully described through i _
specification of the contact angle which obtains under a gauge,
axial load. However, this method of input does not accurately }
define the system if there is more than one angular contact 1
supporting the shaft and at least one of those bearings has
its grooves offset in the direction opposite to the other
bearings and if the shaft is capable of axial and/or radial I
play. In other words, if, what are known as angular contact
ball bearings, are mounted such that some diametral shaft
play is permitted, an auxilliary angle as well as the diametral iJ
play must be specified at input. The angle input is not the
9
34
1.1
1982024093-046
L
I o.,o,..L ,,, ,, .T.loo.0
- manufacturer's designated contact angle, a, but an auxilliary
I_ angle, ao, the calculation for which shall be demonstrated.
r Refer to Figure 3.i. The manufacturer's contact angle is
i calculated as follcws: \
- -- C3.1)
i where: ro and ri are the outer and inner raceway groove
_" radii respectively
i D is the ball diameter
i
Under a gauge axial load, a is obtained at both inner and
i outer raceways for each ball. Under this condition, the outer
i and inner raceways are axially offset an amount Sa
(3.s)
When angular contact ball bearings are mounted with some
di_._etral play, the grooves are offset an amount 5ao such that
Sao<Sa. The diametral play which obtains at this condition
is 5d. This diametral play is usually known by the engineer
or designer and is usually required to allow some forgiveness
when thermal gradients are encountered. Assuming that the user
has the values for a , ro, ri, D and $ao then:
__t (3.4)
I _ . tan" 1 _ " "_J
r"
i. where: Pd and A may be calculated from Eqs. (3.i) and (5.2).
[
If S_o is unknown, the following equation may be solved
for ao"
[ 3s
1982024093-047
ORIGINALPAGE IS
',F pOOR QUALITY AT81D040 il
FIGURE 3.1 ANGULAR CONTACT BALL BEARING GEOMETRY I!
(a) (b)
Geometry Under a Gauge Axial Geometry with Grooves Aligned
Load Showing the Manufacturer's
Contact Angle.
36
,I
I
1982024093-048
iI
'_ ORIGINAL PAGEIS
[ OF POOR QUALITY
} AT81D040
.!
!
FIGURE 3.1 ANGULAR CONTACT BALL BEARING GEOMETRY
- (CONTINUED)

k
* (c]
' Geometr}' Showing the as Mounted
: Gl'oove Alignment, the Input
Diametral Play (Sd), and the
Input Contact Angle (ao).
I
37
SKF TECHNOLOGY SEI?VICES
Sk, FINDUSTRIES, INC
I
1982024093-049
OF POOR QUALITY
! ATglD040
= a " cs-I _ 2A ] (3.5)
In order that the Program properly handle split inner ring
ball bearings an auxiliary angle and diametral p!ay must be
input. Referring to Figure 3.2, the auxiliary angle a O and [I
diametral play Sd must be determined and input. Typically the
values of D, ro, ri, as and Sd' (assembled bearing diametral play)
are known.
L
)
The unloaded naif of the inner ring must be removed from con-
sideration and the ball moved such that its center lies on the
line connecting the origins of r_ and ro and positioned such
that the auxiliary clearance Sd/4 exists at Uoth the inner and
outer raceways. The auxiliary angle is given by:
r(r_-D/2)sin e_ ]
a = tan-1 _ro-D/2-Sd'/2 + (ri-D/2)cos a s (3.6)
The input bearing diametral play, Sd, can then be calculated
as follows:
Sd = Sd' + (2ri-D)(l-cos a s ) - 2A(l-cos a o) (3.7)
The angle associated with each ball bearing must be
specified with the correct sign. A positive contact angle
allows the bearing to accept a positively directed axial load
transmitted by the shaft.
38
SKFTECHNOLOGY SEPVICES
., SKFINC_ST_'_S.INC
1982024093-050
ORIGINAL PAGEIS
OF POOR QUALITY
AT81DOqO
L
,-, X
I-t
cD
Z
Z
--1
0
[*
39
[ SKFTECHNOLOGYSEI2VICES SKI-INFL)USTI_IESIN_I"
L
] 98202409:3-05 ]
)
ORIGINAL pAal m AT8 ID04 0
OFPOORQUALm'
3.3.2.2 Tapered Roller Geometry
Items 2 and 6 are self explanatory.
I
Item I: Bearin_ Pitch Diameter )
The tapered roller bearing pitch diameter may be calculated I
through specification of the roller measured large end diameter,
1
the roller large end corner radius, one half the cup included
angle, the roller included angle and the roller total length.
This calculation is performed within the program The user there- i
fore need not specify the tapered roller bearing pit'chdlameter.
Item 5: Bearin_ Axial Play
For _ tapered roller bearing, the bearing axial play rather
than diametral clearance must be specified (See Pig. 53) Note
that this end play pertains to only the bearing in question. Por
two identical tapered roller bearings on a shaft, one half the
total shaft axial play should be specified for each bearing Por
two tapered roller bearings of dissimilar size, the total shaft
axial play should be apportioned according to the bearing size
such that the sum of the axial plays specified for the two
bearings equals the total for the shaft.
Item 4: Bearing Contact Angle
for the tapered roller bearing, one half the included cun
toe contact angle. (S_e Fig. 3.3)
angle (%) is input as
This an_le mgst be specified with the correct sign. A positive
angle allows the bearin E to accept a positively directed axial load
transmitted by toe shaft and vice versa for a negative angle.
Item 5: Tapered Roller Bearin_ Flange Angle
The flange angle is shown bya f in Pig. 33. The flange
an_le must always be positive.
5 3 2 3 Cylindrical Roller Bearing Geometry
Items 1, 2, 3 and 6 are self explanatory. Both items 4
and 5 should be left blank
40 I.
J
I
1982024093-052
1
ORIGINALPAGEIS IV1"B1D040
OF POORQUALITY
l
l,_ "'= _ _ _ :
O
e, J
r_
/ "" I
0 /
, _ /
/
[-- /
t
r._
4!
1982024093-053
ORIGINAL PAN m AT81D040
OF pOOR (_I&LITY
3 5 5 Card T'pe 33 - Rolling Element Geometry
"" .m,_m_.,,i-,|
I
5.3.3.I Ball Geometry. I i'_
I
The geometry o a ball is fully defined by its diameter. ]
I
!
3.3.3.2 Tapered Roller,.Geometrz I
I
Item 1 - Rollerl Measured Large End Diamete T
For a tapered roller, the input diameter required is the
largest measureable diameter. (See Fig. 3.4.) Typically
this measurement should be taken where the large end corner
radius becomes tangent to the roller surface profAle. Within
the program a "working" large end diameter is calculated.
This diameter is shown as Dw in Fig. 5.4. All bearing geometrical
relaticnships are calculated based on Dw.
D_ = I) + Z_zo S;_-d
(3.8)
i
where: D is the measured large end diameter, reo is
the distance from the roller end to the
beginning of the roller effective length at
the outer raceway surface, measured parallel
to the roller surface. I ro is the corner i t
radius at the roller large end
!
i
do_'iV (3.9) i i
is one half the roller included angle.
Item 2 thTough 6 are shown in Fig. 5.4 and are self
explanatory. Note that the program can handle a nonzero roller {
flat length, Item 6. Most tapered rollers are, however, fully
crowned
i
]
42
i ,
1982024093-054
43
I ,
1982024093-055
ORIGINAL PAGE iS
OF POOR QUALITY t_
AT81D040
I4
J
3.3.3.3 C_lindrical Roller Geometry
\
Typically, cylindrical rollers are partially crowned as
_hown in Fig. 3.5. The center of the roller is flat. Toward
the ends the roller profile is formed by a crown radius, R .
There are usually rounded corners at the roller ends. The_e
corners reduce the load carrying _urface of the roller such that
if there are no raceway undercuts the roller raceway effective it
length equals the roller _otal length less the two corner radii ii
(see Section 3.3.4.2). Note that a partially crowned roller is
specified through input of a non-zero flat length. If the flat
length is zero, the roller is fully crowned with its profile
defined by the roller crown radius.
The roller end sphere radius, Rs, is defined in Fig. 3.5.
For a cylindrical roller, the roller included angle is zero.
3.3.4 Card Type B-4 - Ro1.1ing Element-Ring Geometry
3.3.4.1 Ball Bearing i
J
v
Items 1 and 2 refer to the outer and inner raceway curva- I
tures respectively where curvature is defined as the cross
groove radius divided by the ball diameter. Typical values
range from 0.515 to 0.57.
q
3.3.4.2 Tapered Roller Bearing Contact Geometry
Items 1 and 2 - Roller Rac.ewa_ Effective Length _ i
The roller-raceway load bearing surface is measured parallel
to the roller surface such that if there were no relief at the ii
roller ends the effective contact length would be: _].
* = L/cos_ (3.10) il
Le
However, since the ro_ler has corners at the large and small
ends, the actual effective length is less than L_. The con-
sideration of raceway undercuts at the inner raceway flanges i_
may result in a raceway effective length less than the roller
effective length in which case the shorter of the two should
"I
be input, i!
j
44
[I
1982024093-056
ORIGINALPAGEIS
OF POOl; QUALITY AT81D040
I i = i i
.... ! \ _--_
af Flange Angles
RC Crown Radii
LE Roller-Raceway Effective Lengths
,, LF Roller Flat Length
RS Roller End Sphere Radius
Ep Raceway End Play
FIG. 3.5 CYLINDRICAL ROLLER BEARING GEOMETRY
45
L
1982024093-057
ATSID040
f
Items 3 and 4 - Raceway Crown Radius
The present analysis permits both the roller and race- \
way to be crowned. If the raceway is crowned, it must be fully
crowned with no flat specified. If the raceways are flat the i
input crown radius may be left blank, in which case a default I
value of I. x i0+ I0 inches is used.
L
Note that the unloaded roller-raceway separation along ]
the roller profile, ($c) Fig. 3.6, calculated at the center of
each roller raceway slice is comprised of the sum of the roller
and raceway crown drops, i
Items 5 and 6 - Roller Large End Corner Relief i;
These data specify the distance from the roller large
end to the point on the roller surface where the roller effective
length begins. For the outer raceway contact this distance
may be calculated using Fq. 3.9. For the inner raceway use Eq.3.9
or the width of the inner raceway undercut at the large end. (Fig. 3.4)
Item 7 - The number of slices into which the roller
raceway contacts are divided.
A maximum value of twenty (20) is permitted. A default
value of eleven (Ii) is used if a blank or zero is read.
3.3.4.5 Cylindrical Roller Contact Geometry
3.3.4.3.1 Card Type B4A Roller Raceway Geometry _!
Items 1 through 4 and 7 have the same definitions as !
they had for the tapered roller bearing.
Items 5 and 6 - Roller End Corner Relief is not re-
,,aired im_ut for the cviindrical roller bearin_ since the roller
raceway effective length is assumed to be centered along the _
roller. This was not the case for the tapered roller bearing, i_
3.3.4.3.2 Card Type B4B Roller Flange Geometry for "i
Cylindrical Roller Bearings
i
1982024093-058
47
1982024093-059
AT81D040
ORIGINAL PAGE IS
OF POOR QUALITY
il
Items 1 - 4 The flange angle and end play definitions }I
can be seen in Fig. 3.5. Note that when a ring has a single
flange such as Inversion No. 3. Fig. 3.7, the end play is the _i
distance between the roller and the flange when the roller is i
centered on the raceway.
Item 5 - The Flan_e Inversion Index - See Fig. 5.7 i]
The number which corresponds to the particular Flange inversion
i!
being examined must be input. Note that since the inversions
greater than eight (8) cannot carry axial loading, the bearing !
carries load only on the raceways, and thus the Program resets
the inversion index to one if the input value is greater than
elght. The inversion index must bc input as a real number, (with i
a ceclmal polnt).
|
5.5.5 Roller-Raceway Non-Uniform Profile Definition ; t
I i i
3.3.5.1 Card Type B5 - Outer Raceway Roller. Contact i
These cards are used to input the separation between the cute:
raceway 6 roller at the center of each slice along the roller profile
with the high points of the roller and race in contact, i.e., with
all clearance between roller and raceway removed. These cards
must be omitted if item 7 of the Bearing Data Title card is zero
or blank. These data are used onl Z when the roller-raceway
profile geometry cannot be defined by card t[pes B5 and B4A.
5.5.6 Card Ty2e B6.- Inner Racewa Z Roller Contact
Same as Card Type B5 for the inner raceway-roller contact. }
3.3.7 Ring-Rolling Element Surface Data i ,
3.3.7.1 Card Type BYA - Raceway.- Rollin_ Element Surface Data
Items I through 6 define the statistical surface micro-
geometry parameters of the rollers and raceways. Items 1
through 5 require the input of center line average CLA surface i
roughness. Within the program CLA values are converted to RMS I
by multiplying by I.ZS.
Items 4 through 6 are RMS values of the slopes measured i
in degrees, of the surface asperities as measured in a traverse !
across the groove for rings, longitudinally for rollers and
in any arbitrary direction for balls. Typical values for race-
way and rolling element surfaces are 1 to 2 degrees. This card
is omitted if the solution level is NPASS = 0.
ii
48
L
] 982024093-060
ORIGIHAL PAGEIS AT81D040
OF POORQUALITY
) FIG. 3.7 CYLINDRICAL ROLLER BEARING
i FLANGE INVERSIONS
-j
:!
.!
I
'I
49
,I
"i
I
1982024093-061
I
ATSID040
3.3.7.2 Card Type B7B - Flange-Roller End Surface Data ii
'the data are identical to the data in Card Type B7A but refer
to the flange and roller end surfaces rather than the raceway and
roller rolling surfaces. ,I
Note that both card types B7A and B7B are omitted if the solu-
tion level is NPASS = 0. il
J
3.3.8 Card Type B8 - Cage Data
This card is omitted if the solution level is NPASS =0. These !i
data are self-explanatory. Note that the cage weight is an input i
item. It is included mainly for future consideration of cage
stability predictions. A nonzero value should be input to avoid i
divide checks.
The number of degrees of freedom of the cage, MCG, is also an I
input item (either 1 or 3). When NPASS=0, cage DOF is defaulted i
to i. See Section 2.4.1, Cage Degrees of Freedom. f
3.3.9 Card Type B9 - Shaft and Housing Fit Dimensions i
These cards are to be included only if the change in bearing
diametral clearance with o_mting conditions is to be calculated,
i.e. if item 4 ITFIT on the Bearing Title Card is non-zero. On
Card Type B9, tight interference fits bear a positive sign and
loose fits, a negative sign.
Item 3 and 6 on Card No. 9 are termed the shaft and housing
effective widths respectively. The value specified for these
effective widths may be as great as twice the ring width.
Use of an effective width is an attempt to account for the
greater radial rigidity of a shaft being longer than the ring
that is pressed on to it, owing to the fact that the shaft deflects
over a distance that extends beyond the ring width. In the program
the calculated internal pressure on the ring due to its inter-
ference fit with the shaft, is distributed over the shaft effective
width and this (lower) pressure is used in computing the shaft
deflection. Using double the actual width as the effective width
is customary.
5O
1982024093-062
l AT81D040
I
l.
f 3.3.10 Card T_.pe BI0 - Shaft_Housin_ Fit Dimensions
These items are self explanatory.
Note: Bearing System Components Material Properties. Card
types BII throuEh BI4.
These card types define the material properties of the
shaft, inner ring, rolling element, outer ring and
housing, data items I through S respectively. This
set of cards is to be included if either the bearing
clearance change analysis is used, i.e. item 4, ITFIT,
Bearing Title Card 2 is non zero, or if the bearing
rings or rolling elements are not steel, i.e. item 10,
IMT, Bearing Title Card 2 is equal to I. If any item
on card types BII through BI4 is left blank, the
program inserts the appropriate value of the steel
property.
_.3.Ii Card Tv2e BII - Elastic-Moduli
This card defines the elastic modulus for the shaft, inner
ring, rolling element outer ring , and h_using respectively.
A default value of 204085 N/mm _ (29.6x10 v PSI] is used.
5.5.12 C@rd Type BI2 - Poisson's Ratio
This card defines the Poisson's ratio for the shaft, inner
ring, rolling element outer ring, and housing respectively.
A default value of 0.50 is used.
3.3.13 Card Tz_e BI3 - Densit Z
This card defines the density for the shaft, inner ring,
rolling element, qule_ rin_. an 4 hqu_in_ respectively.
A _ezault value o ,,._06 gytma L0._8, l_/in_ is used.
3.3.14 Card T}.pe BI4 - Coefficient of Thermal Expansion
This card defines the coefficient of thermal expansion for
the shaft, inner ring, rolling element, outer rin, and
housing respectively. A default value of 12.24xI0 -Q I/C
(6.8 X 10-6 I/F) is used.
3.3.15 Card Type B15 - Lubrication and Friction Dat9
This card is omitted if the solution level is NPASS - 0.
51
L
1982024093-063
AT81D040
Items 1 _nd 2
Items 1 and 2 are the amounts by which the combined thick-
ness of the lubricant film on the rolling track and rolling
element is increased during the time interval between the pas-
sage of successive rolling elements, from whatever replenish-
ment mechanisms are operative. Item 1 applies to the outer and I
Item 2 to the inner race-rolling contacts respectively, if
Item I is zero or blank the mode of friction is assumed to be
i
dry. _t
_L
At the present time the magnitude of the inner and outer
raceway replenishment layers has not been correlated to lubri- i
cant flow rate, lubricant application methods and bearing size 1
and speed factors. At this point then, the user is forced to
establish proper values for the replenishment layer thickness.
As a rough guide the following suggestions are made.
I) To avoid starvation, the replenishment layer thick-
nesses should be one to two _imes the EHD film thlckness
which develops in the rolling element raceway contacts.
2) Because of centrifugal force effects, intuition
suggests that the oater raceway replenishment layer
should be several times thicker than that prescribed
at :he inner raceway.
Item 3, XCAV, describes the percentage of the bearing
cavity, estimated by the user to be occupied by the lubricant.
0<XCAV<I00.
As with the replenishment layer thicknesses, the amount
of free lubricant should be able to be correlated with lubri-
cant flow rate, lubricant application methods and bearing size
and speed factors. At this time such correlations do not exist.
: XCAV values less than five percent are recommended.
: Item 4 is the coefficien" of coulomb friction applicable
for the contact of asperities. If Item 1 and 2 are zero, then
Item 4 serves as the coulomb friction coefficient which prevails !
in all contacts, i
i
Items 5 and 6 are the lubricant replenishment layer thick- i
nesses for the outer and inner ring flanges respectively. These
iterns!_uldi_etR_f_oezlclenzblank for _l_o_f_#ng_fction applicable i
for the asperity interactions at the roller end-flange contacts.
This value should also be left blank for ball bearings. I
5.3.16 Card Type BI6
This card is omitted if h ,c P.,_S title card 2 is zero or
blank or if Item 1 card BlSis zero or blank which implies dry.
friction.
|
52
I
1
1982024093-064
I .
| "
' AT81D040
This card specifies the lubricant type. If Item I, NCODE is
i i, 2, 3, or 4, the Program uses preprogrammed properties
lubricant
as presented in Table i, and no further information is required.
NCODE Lubricant
1 A specific mineral oil
2 A MIL-L-7808G
3 Polyphenyl-Ether
4 A MIL-L-23699
NCODE may also be specified as negative (-i to -4), in which
case the traction characteristics of the respective lubricant
NCODE noted above are used but the actual properties specified by
Items 2 through 9 override the hazd coded data. This option is
most useful in specifying various mineral oils i.e. NCODE = -i.
If items 8 and 9, AKN and FRIC are left unspecified, default
values are set at 50.0 and 0.07, respectively. AKN and FRIC are
only used when the NASA version of the code is exe: ted (see
Appendix F).
3.4 DATA SET III- THERMAL MODEL DATA
Appendix B has been included to aid the user in data prepara-
tion and calculation of heat transfer coefficients required at
input.
3.4.1 Card Type T1
Card type T1 is a control card. If no temperature map is to
be calculated, this card is to be included as a blank card fol-
lowed by a Type T2 card for each bearing on the shaft. Card Type
T1 contains control input for both steady state and transient
thermal analyses. It is not intended, however, that both analyses
be executed with the same run.
Item i: The highest node number (M). The temperature nodes
must be numbered consecutively from one (i) to the highest node
number. The highest node number must not exceed one hundred (100).
Item 2: Node Number of the Highest Unknown Temperature Node
(N). This number should equal the total number of unknown node
temperatures. It is required that all nodes with unknown temper-
atures be assigned the lowest node numbers. The program assumes
that all node numbers greater than N (from N+I to M) represent
known boundary temperatures.
O
Item 3: Common Initial Temperature (TEMP) C: The temperature
solution iteration scheme requires a starting point, i.e., guesses
of the equilibrium temperatures. Card Type T3 allows the user to
input guesses of individual node temperatures. When a node is not ;
given a specific initial temperature, the temperature specified
53
1982024093-065
i _ _- AT81D040
I _ r-_ _-'q_ O
| _,._0 [._'a_ 0 0_
/ _._f_ - ,_ o o o.
-i
/
10i ! i
_0 _
090
-, " c; _ o i
. _ _,
e _
--1
C_Z
i
'i
54 1
i
.J i
1982024093-066
I ATSID040
i as Item 3 of Card Type T1 is assigned.
Item 4: Punch Flag {IPUNCH): If the Punch Flag is not zero
(0} or blank, the system equilibrium temperatures along with the
respective node numbers will be punched according to the format
of Card T3. This option is useful if, for instance, the user
makes a steady state run with lubrication, and then wishes to use
I the resultant temperatures as the initiation point for a transient dry friction run in order to assess the consequence of lubricant
flow termination.
I
[
I
i j
[
q
1982024093-067
ORIGINAL PAQt IS AT81D040
OF POOR QUALITY
Item 5: "Output Flag" (IUB). If the "Out, at Fla_" is not
zero t-_'_e--bearingprogram output and a temperature map will be
printed after each call to the shaft bearing solution scheme, i
This printout will allow the user to observe the flow of the !
solution and to note the interactive effects of system tempera-
tures and bearing heat generation rates. Two levels of bearing "_
cutput are permitted. If IUB is I, the rolling element output is !{
not printed. If IUB is 2, full bearing output is obtained.
" T
Item 6- "Maximum Number of Calls to the Shaft Bearin_ r
Program'" (IT1). I_l is the limit on the number of Thermal-Shaft-
Bearing iterations, i.e., the external temperaturs equilibrium
calculation. The user must input a non-zero inteRer such as 5 or
I0 in order for the Program to iterate :o an equilibrium condition.
If IT1 is left blank or set to zero (0) or i, shaft bearing
performance will be based on the initially guessed temperatures
of the system. The temperatures printed out will be based on the
bearin_ generated heats. It is unlikely that an acceptable
equilibrium condition will be achieved. However, the temperatures
which result may provide better intial guesses, for a subsequ;-t
run, than those specified by the user.
IT1 also serves as a limit on the transient temperature solu-
tion scheme, by limitin_ the number of times the shaft-bearing-
solution scheme is called. Each call to the shaft-bearing scheme
will input a new set of bearing heats to the transient tempera-
ture scheme until a stead}" state condition is approached or until
the transient solution time up limit is reached.
Item 7: "Absolute Accuracy _f Temperatures for the External
Thermal Solution" (EPI). In the stead}, state thermal solution
scheme, each calculation of system temperatures occurs after a
call to the shaft-bearing scheme which produces bearing generated
heats. After the system temperatures have been calculated for
each iteration, using the internal temperature solution scheme,
each node temperature is checked against the nodal temperature "
at the previous iteration.
If {t(N)i- t(N. lji} _ EPI for all nodes i then equilibrium
has been achieved and the iteration process stops.
t(Nji - temperature o ith node at Nth thermal iteration.
t(N.1]i _ temperature of ith r:o-4 2: ',-i "'" thermal iteration.
56
I
1982024093-068
I_ ORI_|NAL p&G IS
OF pOORQUALITY Am81D040
1.
Item 8: "Iteration Limit for the Internal Thermal Solution"
(ITZ).""X'_er each call to the shaft bearing program, the
internal temperature calculation scheme is used to determine
the steady state equilibrium temperatures based on the calculated
Ii set of bearing heat generation rates. If the program is used \
to calculate the temperature distribution of a non bearing
system it is the internal temperature scheme which is employed.
l If IT2 is left blank or set to zero, the number o internal
iterations is limited to twenty (20).
:" Item 9: "Accuracy for Internal T_ermal Solution" (EP2).
1 The u_EP2 is explained in Section2.1.1. If EP2 is left
blank or set to zero (0), a default value of 0.001 is used.
{" Item I0: "Starting Time" (START) is a time Ts at which
_" the t_ansient solution begins; usually set to zero (O).
I Item ii: "Stopping Time" (STOP) is the tim_ in s-conds at
I. which the transient solution terminates,Tf The transient
solution will generate a history of the system performanc,:
I which will encompass a total elapsed time of
"" cTf - Ts) seconds
Item 12: "Calculation Time Step" (STEPIN). The transient
internal _solution sci_eme solves the sTstem of equations
tk+! . tk , qk LT (3.11) i
pCpV i
_T " STEPIN
The user may specify STEPIN. If l_ft blank or set to zero (0), !
the Progra_ calculates an appropriate value fc_ STEPIN using t
I
the procedure described in Section 2.1.2.
_ Item 15: "Time Interval Between Printed Temperature Maps"
,. (TTI_nds. The user must specify the length of time
which will elapse between each printing of the temperature
I" map. The interval will always b_ at least as large as the
_. "calculation timestep" (STEPIN).
Item 14: "Time Interval Between Calls of the Shaft Bearing
{
, Portion of the Program" (BTIME). BTIME will always have a
value larger tha_ or equal to (STFPIN) even if the user
inadvertently inputs a shorter inter',al. Computational time o_
I savings result if BTIME is greater than STEPIN, however,
_- accuracy might be lost.
57
I
I " B
1982024093-069
ORIGINALPAQE |_ I _
,_ _ _OR QUALITY AT81D040 _'
_. 4.2 Card Type _2
Card Type T2 is rea_'_ired,one card for each bearing if I!
no thermal a_alysis is being performed. The temperature data
is used within the _haft-bearing analysis portion of the pro-
gra= to fix _emperature dependent properties of the lubricant !!
in which cas._ the inner race, ouOer race, lubricant bulk "
cavity and r'iange temperatares are used. Note the flange _
nu_h.e._ing-'=heme depicted in Fig. 5.5 for a cylindrical roller _,
oe_._ n _, _or a tapered"roller, bearing,the innerring flangeis considered "
_'lange No. _, Th_ assemblycomponenttemperaturesat"eachbearin_location
are use_ in the analysisw_ch calcu/atesthe changein bearingdiametra] i
c]earra,ce from "offthe shelP'to operatinzconditions.
5.4,3 Card Type T3
In the steady state analysis this card is used to input
initial guesses of individual nodal temperatures fcr unknown
nodes as well as the constant temperatures for known nodes,
such as ambient air and/o_" an oil sump.
In the transient analysis, Card Type T3 is used to input
the nodal temperatures of all nodes at (START) - Ts i.e.
at the initiation ,f _-hetransient solution.
3.4.4 Card Type T4
With this card, node numbers are assigned to the components
of each bearing, one card per bearing. With this information
! the proper system temperatures are carried into each respective
., bearing analysis. The inner race and inner ring node numb-rs
: may or may not be the same at the user's discretion. Similarly
the outer race and outer ring node numbers may or may no_ be
the same. :
.: 3.4.5 _Card Type T_ _ TS,A
,!
:_ Card Type T5 is recruired, one card per besrzng, if a thermal
, analysis is to be performed This card designates two nodes to
share equally each of the various types of bearing generated heat
i calculated internally by the program. For cylindrical and tapered
i roller bearings an additional card, TSA, follows immediately after
card T5 specifying the two nodes which will equally share the heat
i generated at each of zhe flange contacts. For ball bearings card
; TSA nus_ be omittod.
i
58
"! i
1982024093-070
I _
AT81D040
V
3.4.6 Card Type T6
This card specifies the node numbers and the heat genera- ,
tion rate for those nodes where heat is generated at a constant
rate such as at rubbing seals or gear contacts.
5.4.7 Card Type T7
This card type is used to input the numerical values
of the various heat transfer coefficients which appear in the
equations for heat transfer by conductivity, free convection,
forced convection, radiation and fluid flow. Up to ten
coefficients of each type may be used. Separate values of
each type of coefficient are assigned an index number via
card T7. In describing heat flow paths (Card Type T8 below)
it is necessary only to list the index number by which heat
transfers between node pairs.
Indices I-I0 are reserved for the conduction coefficient
_, 11-20 for the free convection parameters, 21-30 for forced
convection, 31-40 for emissivity and 41-50 for fluid flow
(product of specific heat, density and volume flow rate).
As an example, for heat transfer by conduction with
coefficient t of 53.7 wattslMC one could prepare a card t'_e
T7 with the digit 1 punched in column 10 and the value 53.7
punched in the field corresponding to card columns 11-20.
If a conduction coefficient of 46.7 were applicable for
certain other nodes in the system one could punch an additional
card assigning index No. 2 to the value _ - 46.7 by punching
, a "2" in card column I0 and 46.7 _ny_'hsrewithin card columns
11-20.
i Rather than inputting constant forced convection coefficients,
optionally, these coefficients can be calculated by the program
i in one of three ways. If the calculation option is exercised
j a pair of cards is used in place of a single card containing
! a fixed value of a The contents of the pair of cards depends
upon which of the three optional methods are used.
Option 1) a is independent of temperature but is calculated
as a function of the Nusselt number Nu w_th in
turn is a function of the Reynolds number Re_
the Prandtl number Pr as follows, (cf. [22] }
59
1982024093-071
il ORIGINALPAGE IS
AT81D040
OF POORQUALITY
a = Nu I oil/L '
Nu = aRbpC !i
er
where _ oil is the lubricant conductivity, L is a
: characteristic length (with a unit of meters)
and a and b, and c are conCtants. '_
Option 2) a is a function only of fluid dynamic viscosity i)
_ and viscosity is temperature dependent. I!
d
a = Cn _.
\
Option 3) a is a function of the Nusselt, Reynolds and
Prandtl numbers-and viscosity is temperature
dependent.
3.4 .8 Card Type T8
i This card defines the heat flow paths between pairs of
I
nodes Every node must be connected to at least one other #
node, i.e., two or more independent node systems may not be
! solved with a single Program execution.
:!
i The calculation of heat transfer areas is based on lengths, Ll and L2 input using Card Type TS. Additionally, the
j type of surface for which the area is being calculated is indicated
_: by the sign assigned to the heat transfer coefficient index.
If the surface is cylindrical or circular the index should .,
be positive, if the surface is rectangular the index should
be input as a negative integer.
i
In the case of radiation between concentric axially
s>nmmetric bodies, L3 is the radius of the larger body. For
radiation between two parallel flat surfaces or for conduction
between nodes, L5 is the distance between them.
L
Fluid flew heat transfer accounts for the energy which
the fluld transports across a node boundary. Along a fluid
node at which convection is taking place, the temperature
i varies. The nodal temperature which is output is the average
of the fluid temperature at the output and input boundaries.
If the emerging temperature of the fluid is of interest, it
is necessary to have a fluid node at the fluid outlet. At
this auxiliary node only fluid flow heat transfer occu) s and
the fluid temperature would be constant throughout the node.
Thus the true fluid outlet temperature will be obtained.
k
_i 60
!
1982024093-072
Ii AT81D040
[_ Conduction cf heat through a bearing is controlled by index
51. The actual heat transfer coefficient which contains a con-
ductivity, area and a path length term is calculated in the bear-
ing portion of the program. The term is based upon conduction
[- through an average outer race and inner race rolling element con- \
I tact.
_" 3.4.9 Card Type T9
This card inputs data required to calculate the heat capacity
_ of each node in the system. This card type is required only
for a transient analysis and must be omitted for a steady state
' analysis.
i 3.5 DATA SET IV - LOADING DATA
.
If a single ball or cylindrical roller bearing is to be ana-
lyzed, the user need not model the shaft geometry. Ill such a case,
this loading card may replace all Shaft Input Data (Data Set V}.
Data describing a dummy shaft is generated within the program.
ApDlied loads, acting through the center of the bearing, are
input on this card. These loads may have the form of concentrated
radial forces (FY, FZ), concentrated moments (MY, MY) and a
concentrated axial load (FX).
If more than one bearing is to be analyzed r this card is
omittedf and loads are supplied with the Shaft Input Data.
3.6 DATA SET V - SHAFT INPUT DATA
The shaft-bearing analysis requires all Io. ":ng to be applied
to the shaft. The loads applied to each beari ,re a product of
the shaft-bearing solution. There is no need Z_ the user to solve
the statically determinate or indeterminate system for bearing
loads.
In the analysis the housing is assumed to be rigid. Provision
has been allowed to input data for housing radial and angular
spring characte_istics. However, this has been done for future
c_n_ideratlo,_ of an elastic housing and is therefore currently
unavailable.
61
1982024093-073
AT81D040
I-
1
,i
Ti_e si'.aft input data consists of three card types:
1
1) Shaft Geometry and Elastic Modulus Data i i
2) Bearing Position and Mounting Error Data [i
3) Shaft Load Data
3.6.1 Card TTpe S1 ;i,
4
This card type is used to ..ic_cribe shaft geometry at ._
up to twenty locations alc,',_" _._ie shaft. The user must
place his shaft in a carte,':n coordinate system with the
end of the shaft at the ori;:n and with the shaft lying
alo:_? the X-axis.
The shaft is described by specifying two outer and
two inner diameters at each axial location along the
shaft which define the diameters .immediately to the left
and the right of the X-coordinate. In this way stepwise
or linear variations of the shaft can be handled. A
linear variation in shaft diameter is assumed if the
diameter to the left of one axial location is different
than the diameter to the right of the preceding axial
location. Complex shaft geometries may be approximated
with a set of linear diameter variations spaced at close
intervals.
If an Elastic Modulus is not specified at the desigriate_
! input location, the modulus of steel is assumed, 204083N/mm.
I
62
'4" _:
1982024093-074
[
I_ AT81D040
3.6.2 Card Type $2
This card type locates the bearing inner ring on the shaft \
in the X-Y and X-Z planes. For a ball bearing, the X coordinate
specified locates the inner ring center of curvature. For cylin-
drical roller bearings the X coordinate locates the center of the
inner race roller path.
I For tapered roller bearings in the strictest sense the X
coordinate locates the point where a line from the roller center
of gravitM intersects and is perpendicular to the inner raceway,
I with all bearing end play removed. It is sufficiently accurate however to allow the X coordinate to locate the center of the
inner raceway.
In addition to specifying bearing location the Type 2 card
{
is also used to specify housing radial and angular mounting er-
rors. As mentioned previously, space has been reserved for in-
I putting housing radial and angular spring characteristics, how-
ever, these characteristics are not used in the system analysis
currently.
I Two sets of Type 2 cards may be required. The first set
is always required and defines housing alignment errors in the
shaft X-Y plane. The second set defines the housing alignment
errors in the shaft X-Z plane. Type 2 cards for the X-Z plane
are required only for bearings having alignment errors in the
X-Z plane that are different than those specified for the X-Y
plane. The second set of Type 2 cards must be placed after the
blank card following the Type 3 cards for the X-Y plane (see
Appendix C).
The first set of Type 2 cards must contain a card for each
bearing. The second set of Type 2 cards must give the appro-
priate bearing number in column i0.
3.6.3 Card Type S3
k Type 3 cards are used to specify shaft loadings at a given
X coordinate. Loading may be applied in the X-Y and X-Z planes,
thus requiring two distinct sets of Type 3 cards. Applied loads
i may have the form of concentrated radial forces, concentrated
moments, linearly distributed radial forces and concentrated
axial loads which may be eccentrically applied. If an axial
load is eccentrically applied, the moment which results will be
I included automatically if the point of application in the radial
pl_ne (Y or Z coordinate) is specified in columns 71-80. Alter-
natively, the moment generated by the axial load can be specified _
, I in columns 31-40.
!
63
1982024093-075
AT81D040
Variations in distributed radial loads are handled at input
just as shaft linear diameter variations are handled. Jl
LI
Note that each set of Type 3 cards must be followed by a
blank card. []
Also note that in order for symmetry conditions to be con-
sidered the second Type 3 card must be void of any loading data.
lJ
4.0 COMPUTER PROGRAM OUTPUT i.J
4.1 Introduction _
II
The Program Output is intended to provide the engineer or de-
signer with a complete picture of the shaft-bearing system per-
formance.
In addition to the calculated output data, the input data is
listed, thus producing a complete record of the computer run.
Sample output of three bearing-shaft systems is included in
Appendix D. These studies demonstrate the ability of the NASA
version of the program to calculate performance characteristics of
tapered and flanged cylindrical roller bearings, and to demonstrate
the functioning of the new cage simulation algorithms and the
single-bearin_ analysis capability. The three configurations are: i
i. A system in which an input pinion is supported by tapered
roller bearings in a straddle configuration.
2. A system in which an input pinion gear load is supported
by a flanged cylindrical roller bearing operating in con-
junction with two angular contact ball bearings.
3. A single ball bearing system operating under combined rad-
ial and thrust loading.
Key output items are discussed briefly below.
4.2 Bearing Output
4.2.1.1 Linear and Angular Deflections
These deflections refer to the bearing inner ring relative to
the outer ring and are defined in the inertial coordinate system
of Figure 2.4. The bearing deflections are not necessarily equal '
to the shaft displacements since the bearing outer ring radial or
angular mounting errors may be specified as non-zero input.
64
1982024093-076
t. ORIGINALPAGEIS
OF POORQUALITY ATSID040 :
4
(
)
) 4.2.1.2 Reaction Porces and Moments
[
These values reflect bearing reactions to shaft applied
! loading and outer ring mounting errors.
When the bearing inner ring has achieved an equilibrium
: position, the summation of all bearing reaction loads should
numerically equal the shaft applied loading. When the level
of solution indicated by "NPASS" = 2 is employed, as discussed
in Section S, differences between shaft applied and bearing
reaction loads will exist but will typically be less than I0%.
This difference is a consequence of friction forces contributing
to the reaction loads whereas the inner ring equilibrium posi-
tion has been determined considering elastic contact forces only.
4.2.2 Fatigue Life Data
The LI0 fatigue life of the outer and inner raceways as
well as the bearing are presented. The bearing life represents
.. the statistical combination of the two raceway lives. These
lives reflect the combined effects of the lubricant film thick-
ness and material life factors. The lubricant film thickness
life factor is described in detail in Section 3.3.1.
4.2.2.1
The ratio h/o, also referred to as _, is printed for the
most heavily loaded rolling element. The variable h, represents
the EHD plateau film thickness with thermal and starvation
effects considered. The variable a represents the composite
root mean square surface roughness of the rolling element and
the relevant raceway.
r o
65
1982024093-077
,| AT81D040 !
4.2.2.2 Li , Multipli, ii
[!
4.2.22.1 Lubrication - This life multiplier is a function of
h/_ at each concentrated contact. Its value'ranges from _
0 21 for h/@ < 0 6 to 3 0 at h/_ > i0 This subject is
[i
: covered in mote detail in Section-3.3.1.
[1
4.2.2.2.2 Material - This output simply reflects the input !
value, Agamn,_s covered in Section 5.
:-7
4.2.3 Temperatures Relevant to Bearing Performance i
These temperatures fully describe the temperature con-
ditions which affect the performance of a given bearing. If
one of the temperature mapping options is used, the tempera-
tures printed reflect the results of the particular option.
If, neither temperature option was used, the list is simply a
repeat of the input data. Note tha= there are separate tem-
peratures for outer and inner raceways and flanges and ring
temperatures. The raceway and flange temperatures are used to
determine lubricant properties. The rin_ temperatures are
used in the bearing dimension change analysis. The raceway,
flange and ring tempezatures may be the same value.
4.2.4 Frictional Heat Generation Rate and Bearin_ Friction
Torque
4.2.4.1 Frictional Heat Generation Rate
The various sources of frictional heat generated within
the bearings are listed. The values printed for "OUTER RACE,
OUTER RING FLANGES, INNER RACE, INNER RING FLANGES, R.E.DRAG
A_ND R.E. CAGE" represent the sum of the generated heats for all
rolling elements. Additionally, the heats printed for the
outer and inner raceways and flanges, plus the rolling element
cage, reflect the friction developed outside the concentrated
contacts, i.e., the HD friction as well as the EHD friction
developed within the concentrated contacts. The raceway and
flange data also includes any heat generated as a consequence
of asperity contacts when the SKF friction model is used "R.
E. DRAG" shoul_ be interpreted as the heat resultina from lubri-
cant churning as the roll_ng elements plow through the air-oil
mixture
66
i j
1982024093-078
AT81D040 '_
o
" t
4.2.4.2 Torque
The torque value is calculated as a function of the total
generated heat and the sum of the inner and outer ring rotational
t speeds. The intent is to present a realistic value of the torque i
required to drive the bearing. Under.conditions of inner ring
7
rotation the torque value reflects the torque required to drive
,[ the inner ring.
4.2.5 EHD Film and Heat Transzer Data
4.2.5.i EHD Film Thickness
; -
I These values refer to the calculated EHD plateau film
L thickness at both contacts of the most heavily loaded rolling
element and include the effects of the thermal and starvation
T
I reduction factors.
4.2.5.2 Starvation Reduction Factor
ii . ,
These factors give for the inner and outer ring contacts,
the reduction in EHD film thickness ascribable to lubricant
film starvation according to the methods of Chiu, [II].
These factors pertain to the EHD film thickness for both
the inner and outer race contacts of the most heavily loaded
rolling elements, but are applied to the respective inner and
outer race film thickness for each rolling element in the bear-
ing.
4.2.5.5 Thermal Reduction Factor
These factors are calculated according to the methods of
Cheng, [I0] and pertain to the EHD film thickness for both
the inner and outer race contacts of the most heavily loaded
: rolling elements, but are applied to the respective inner and
' outer race film thickness for each rolling ele_nent in the bear-
ing.
F"
[. 4.2.5.4 Menis;us Distance
These values are calculated according to the methods of
Chiu, [II] and pertain to the film
EHD thickness for both
' the inner and outer race contacts of the most heavily loaded
F
I. 67
1982024093-079
rolling elements, but are applied to the respective inner and I
outer race film thickness for each rolling element in the bear- IJ
ing.
4.2.5.5 Raceway-Rolling Element Conductivity II
These data reflect the amount of heat transfer between Fl
rolling element and raceway for each degree centigrade differ- _|
ence between the two components. These data reflect the average L
i
of all outer and inner contacts respectively. (See Sections 2.1.4
2.1.4.1 for a discussion of the calculation procedure.) il
4.2.6 Fit and Dimensional Change Data
[J
4.2.6.1 Fit Pressures
These data refer to the pressures built up as a conse-
quence of interference fits between shaft and inner ring and
housing and outer ring. Pressures are presented both for the
standard cold-static condition (16oc) and at operating
conditions.
4.2.6.2 Speed Giving Zero Fit Pressure (Between the shaft and
inner ring)
This is a calculated value based upon operating conditions
and provides a measure of the adequacy of the initial shaft
fit.
4.2.6.5 Clearances
[ "Original" refers to cold unmounted clearance which is
specified at input if the diametral clearance change analysis
is executed. "Change" refers to the change in diametral
clearance at operating conditions relative to the cold unmounted
condition. A minus sign indicates a decrease in clearance.
"Operating" refers to the clearance at operating conditions.
For all types of ball bearings the decrease in clearance can
be combined with the initial diametra! clearance, and the free
operating contact angle at operating conditions may be
calculated. Note that the change in clearance should be com-
pared against the diametral play of the split inner ring ball
bearing in order to determine if the possibility for three
point contact exists. The Program does not account for three
_oint contacts even though the change In clearance mlgnt suggest
q that three point contact is obtained.
68
_L
1982024093-080
I
.j ATSID040
J
4.2.7 Lubricant Temperatures and Physical Properties
J The lubricant properties, particularly the dynamic vis-
cosity and to a lesser degree the pressure viscosity coefficient,
are heavily temperature dependent. These factors enter the EHD
film thickness calculation and the HD and EHD friction models,
The lubricant is assumed to be at the same temperature as the \
" relevant raceway. As noted elsewhere, these temperatures may
.i be either input directly or calculated by the Program.
_ The physical properties printed are self explanatory.
The units are enumerated.
f
_ 4.2.8 Ca_e Data
-- 4.2.8.1 .Ca_e-Land Interface
- The cage data indicate the performance parameters at the
interface between the cage rail and the ring iand on which the
cage is guided. The torque, heat rate and separating force
-: require no explanation. The eccentricity ratio defines the
degree to which the cage approaches the ring onwhich it is .6
guided at the point of nearest approach. The radial displace-
ment c_ the cage relative to the bearing axis is divided by one
half tne cage-land diametral clearance. An eccentricity ratio
of one indicates cage-land contact. A ratio of zero indicates
that the cage rotation is concentric with the bearing axis.
Only the cage-_.and and rollinz element pocket forces are
considered in determining the cage eccentricity. The cage
weight and centrifugal force which result from the eccentricity
although available, are not considered in the analysis. The
omission of these considerations helps reduce convergence
prob ler,s.
4.Z.8.2 .Cage Speed Data
Cage speed data present the comparison between t e cage
speed calculated based upon the quasidvnamic equilibrium con-
siderations and the speeds ca:-.ulated with raceway control
i theory for ball bearings and t_te epicyclic speeds of the
roller bearing components.
4.3 Rollin_ Elemen_ Q_tDut
4.3.I _Rolling Element Kinematics
F
.[. 69
1982024093-081
ORIGINAL PAGE m ATelD040
OF POOR QUALITY
4.5.1.1 Rollins Element Speeds
All of the rolling element speeds tend to vary from posi-
tion to position when the bearing is subjected to combined
loading, j
The total rolling element speed is with reference to the
cage and represents the vector sum of the three orthogonal I
components.
4.3.1.2 Speed Vector An1es
The rolling element speed vector angles, Arctan _y,4_,x)
and Arctan (Wz/W,x) are presented in order to show a clearer
picture of the predicted ball kinematics. The ball _peed
i vector tends to become parallel with the bearing X axis with
increasing shaft speed and decreasing contact friction.
4.3.2 Rolling Element Raceway Loadin_
4.5.2.1 Normal Forces
The normal forces acting on each rolling element are
printed. The rolling element race normal forces are self
explanatory. The cage force is calculated only when the
friction solution is emplo),ed and is always directed along the
rolling element Z axis. If the rolling element orbital speed
is positive, a positive cage force indicates that the cage is
pushing the rolling element, tending to accelerate it. Cage
force is a function of rolling element position within the
cage pocket. Its magnitude is derived using hydrodynamic lubri-
cation assumptions, when the distance betwcen the rolling element
and cage web is large, and EHD assumption5 when the separation
is of the order of the EHD film thickness or when rolling
element-cage web interference exists (SKF version). The NASA
version Lsee Appendix F) solves for cage forces using hydrodynamic
lubrication assumptions only.
4.5.2.2 Hertz 5tres_
The stress printed represents the maximum normal stress at
the center of each ball race contact or at the most heavily loaded
slice of the roller raceway contact.
t T0
J
1982024093-082
i ?
\
ATSID040
.i
4.3.2.3 Load Ratio tot-
If the EHD film thickness is small compared to the RMS compo-
site rolling element-race surface roughness, the rolling element-
race normal load will be shared by the EHD film and asperity con-
tacts. The load ratio reflects the portion of the total load
carried by the asperities.
When the NASA version of the code is executed (see Appendix F),
Qasp/Qtot=0, since asperities are not accounted for in the Allen
" Traction model [7].
I
71
1982024093-083
] ORIGINAL PAGE IS AT81D040
i OF POOR QUALITY
I
:I
]
4.3.2.4 Contact An1es
A ball bearing, subject to axial loading, misalignment
,_r mounted such that the inner ring is always displaced axially
relative to the outer rings, (i.e. a duplex set of angular j
contact ball bearings) will have non-zero contact angles. At
low ball orbital speeds the inner and outer race angles will be
substantially the same. At high speeds, ball contrifugal force i
will cause the o,teT race contact angle to be less than the
inner race angle.
i
!
i
I
4.3.3 Roller End-Flange Contact Data
: For a tapered roller bearing a single set of roller ; .
end-flange data is printed. For the cylindrical roller bearing,
'i which may have up to four flanges, the program examines the
j data and prints the results for the most heavily loaded outer
and inn r ring flanges. In addition to the data listed 5elcw
for a cylindrical roller, the semimajor contact axis as well as
the concentrated contact and hydrodynamic heat generation rates
are printed for the tapered roller.
4.5.5.1 Normal Force
The interference between the roller end and flange is
det_rmineJ from _he solution for the relative _ing an_rolling
element pcsitions. Hertz theory is used to calculate the load i -_
which results from this interference.
4.3.3.2 Hertz Stress +
The Hertz stress printed is the maximum normal s_ress
4
which occurs in the contact.
' _.3.3.3 EKD Film Thickness _-
i
The plateau film thickness is <alculated using the
Archard-Cowking [5] equation for point contacts or the Dowson-
Higginson[4] equation for line contacts, when the "SKF" version of
the )rogram is executed (see Appendix F). The "NA3_" verslon uses the , _
Loewenthal podel [6] to calculate film thickpess. Fither result is
then modified to account for starvation [12] and thermal [Ii] reduction :;
effects. This modified film thickness is printed.
4.3.3.4 Slidin$ Velocit 2
Ti,_ sliding velocity is defined as the difference between the i "
f!ang_ and roller end _inear velocities at _he center of the contact. , _-
72 ' "_
i
1982024093-084
.
._ ATSID040
r"
I
- 4.3.3.5 Rolling Velocity
4
Rolling velocity is defined as one half the sum of the
roller end and flange linear velocities at the center of the contact.
: 4.3.3.6 Contact Ellipse Semiminor Axis
To help assess the severity of the roller end flange
contact and the possibility for edge loading_ the semiminor
axis of the contact is printed.
4.4 Thermal Data
As in the case for bearing output, all of the input data is
printed. The calculated output data is presented in the form
of a temperature map in which a node number and the respective
node temperature appear. The appearance of the steady state
and transient temperature maps are identical. The transient
temperature map also includes the time (T) at which the
temperature calculations were made.
4.5 Shaft Data
These data simply reflect the input information.
: 4.6 _rogram Error Messages
4.6.1 From Subroutine ALLT
'_teady State Solution with (EP1)degrees accuracy was not
obtain after _ Iterations".
This message pertains to the external temperature itera-
tion scheme in which system temperatures and bearing generated
heats are being solved for an equilibrium condition.
4.6.2 Fron_ Subroutine SHABE
r
. "It was not possible to obtain the change of clearance
with an accuracy of _ times the rolling element diameter
_ in _ITpIT) iterations .
r

73
1982024093-085
AT81D040
' I
_] This message pertains to the bearing diametral clearance _-
I" change iteration scheme. The solution may be converging in
which case the ,umber of iterations (ITFIT) should be increased.
This can be checked with an NPRINT = I intermediate printout.
The intermediate .rint may indicate that the solution is
oscillating. The most likely cause of oscillation is the
alternate predictim of bearing preload with all rolling elements
loaded, and then in the next iteration, only a subset of the
rolling elements loaded. This problem can usually be over-
come by either of two methods.
1) In subroutine FIT remove the GO TO 20 statement.
This will cause the inner ring load distribution
to have no effect on the change in diametral clear-
ance.
t
2) The solution can be damped by redefining the so!ution
damping factor FA, such that it would take on a value
_! 0_FA_-I. FA is presently set to i in subroutine SHABE.
! If this damping technique is used, the number of FIT _
iterations should be increased as the value of FA is
i decreased. An upper limit of 10 iterations is recom-
mended, j
4.6.3 From Subroutine SOLVXX _
I) "SINGULAR SET OF EQUATIONS"
This message might occur when the thermal input data
:_ is not input properly. {
2) "THE LIMIT FOR NUMBER OF ITERATIONS IS REACHED"
7
This message might occur either during a steady state
temperature solution or bearing solution. Before
_ncreasing the number of iterations check the equa- 1
tion residue values. If they are low, the solution
may be good enough. I
i
3) "THIS IS THE BEST WE CAN DO. IT MAY BE USEABLE"
This message reflects the fact that the next iteration ! :
will result in divergence. The iteration procedure is
thus terminated. The equation residue values should i
be checked, if they are low, the solution may be useable, i
i
4 i
74
i
1982024093-086
.
AT81D040 :
- As noted above the occurrence of messages 2) and 5), do
- no_ necessarily mean that _he solution is not good. Generally #
: ] Lhe messages indicate that the solution has not converged as \
t tightly a_ the user has requested.
_ i Note: The XX suffix on SOLV specifies the version of SOLV contained in the user's program. As of this writing the current
version is SOLV13. The suffix is changed each time improve-
_ ments are made which require a change in the SOLVXX calling
sequence.
L
i 4.6.4 From Subroutine INTFIT
"Sj_gular matrix on tight shaft fit"
"Singular matrix on loose shaft fit"
I These messages reflect an erro_ in the input data usually
as a consequence of inconsistent component diameters, such as
: " the shaft inside diameter being greater thsn the outside
dia_eter.
I
I
' 4
I .
[i
t 75
t.
I t,
l
1982024093-087
AT81D040
t:
:
5.0 GUIDES TO PROGRAM USE
The Computer Program is a tool. As with any tool the
results obtained are at least partially dependent upon the
skill of the user. Certainly the economics of the Program
usage are highly dependent upon the user's technical need
and discriminate use of Program options.
Some general guides for efficient use of the Program are
listed below: _
I. Attempt to use the lowest level of solution possible.
For instance if the prime'object of a given run is to
obtain bearing fatigue lives, execute only the elastic
solution (NPASS = 0). If an estimate of bearing
frictional heat is required, execute the low level ,
friction evaluation (NPASS - i). Execute the friction
i solution (NPASS - 2) only if rolling element and cage
kinematics are of interest. Execute the highest level
of solution (NPASS - 3) if kinematics are of interest i
and the bearing reaction loads deviate substantially
from the shaft applied loading, i.e. a deviation i
greater than ten percent. '
2. Attempt to input bearin_ operatinz diametral clearance
rather than calculate it. Or, execute the diametral
clearance change analysis once for a group of similar ,
runm and use the output from the first run as input
to the subsequent runs omitting the clearance change ,_
analysis.
3. Attempt to i__out accurat_ oper.at.in_ temperatures
rather than calculate them.
4. The more non-linear the p_ob_em th_ more computer time i
required to solve it. In the bearing friction solution _,
large coefficients of friction seem to increase the
degree of nonlinearity. In the thermal solt:tions, if
possible, eliminate nonlinearities by omitting
radiation terms and by using constant rather than
temperature dependent frge and forced convection
coefficients.
5. In the transient thermal solution, _Da_e the ca_Is to
the shaft-bearinm solution [BTIME] to as large an i
interval as prudently Dossib!e. _ however,
too lon_ and interval will produce large errors in heat i
rate predictions.
!
_6
I
1982024093-088
i AT81D040
[i 6. In the steady state thermal analysis attem t to esti-
: mate nodal temperatures on a node by _Isr_ _
Nodes which are heat sources should have higher
[ ,
temperatures than the surrounding nodes.
C
The above suggestions are intended to encourage the use
of the Program on a cost effective basis. The intent is not
t
to discourage the use of important program capabilities but
to emphasize how the program should be most effectively used.
_[ It is suggested that the user take a simple, axially
loaded ball bearing problem and execute the program through
i the full range of options beginning with a frictionless solu-
tion proceeding to the three levels of friction solution with
a low (0.01) and high (0.1) friction coefficient. The dia-
metral clearance change analysis and the thermal solutions
i should also be executed on an experimental basis. This exercise
will provide the user with some insight into economics of the
Program usage on his computer as well as the results obtained
-'_[i from various levels of solution of the same problem.
It is also suggested that a constant user of the program
should study the hierarchical Program flow chart Appendix A,
a]ong with the Program listing to gain an appreciation of.the '
program complexity and the flow of the problem ._olution. The
Program is comprised of many small functional subroutines.
I Zn(,wledge of these small elements may allow the user to more _"
ea,;ilypiece together the philosophy of the total problem
so'tution.
-I
[ "
I
t _
t
t "
2
i .
' 77
t ._
1982024093-089
i!
AT81D040
REFERENCES !I
1. Kellstrom, M., "A Computer Program for Elastic and Thermal
Analysis of Shaft Bearing Systems," SKF Report No. AL74P004, _i
submitted to Vulnerability Laboratory, J.S. Army Ballistic !i
Research Laboratory, Aberdeen Proving Ground, MD, under Army
Contract DAAD05-73-C-0011 (1974).
il
2. Kleckner, R.J. and Pirvics, J., "SKF Computer Program CYBEAN - {I
Volume I: Analysis,,, SKF No. AL78PO22, submitted to NASA/Lewis
Research Center under NASA Contract No. NAS3-20068, July 1978. Ii
tJ
3. Archard, J. and Cowking, E., "Elastohydrodynamic Lubrication
at Point Contact," Proc. Inst. Mech. Eng., London, Vol. 180, il
Part 3B 1965-1966, pp. 47-56. II
4. Dowson, D. and Higginson, G., "Theory of Roller Bearing Lub-
" Proc. Inst Mech. Eng., London, Vol rication and Deformation, .
177, 1963, pp. 58-69.
5. Tallian, T.E., "The Theory of Partial Elastohydrodynamic
Contacts," Wear, 21, pp. 49-101 (1972)
" 6. Loewenthal, S.H., Parker, R.J., and Zaretsky, E.V., "Correla-
tion of Elastohydrodynamic Film Thickness Measurements for
Fluorocarbon, Type II Ester and Polyphenyl Ether Lubricants,"
NASA Technical Note D-7825, National Aeronautics and Space
Administration, Washington, D.C., November 19, 1974.
7. Allen, C.W., Townsend, D.P., and Zaretsky, E.V., "New Gener-
alized Rheologica! Model for Lubrication of a Ball Spinning
in a Nonconforminq Groove," NASA Technical Note D-7280,
National Aeronautics and Space Administration, Washington, D.C.,
May 1973.
I
, "Life Adjustment Factors for Ball and } 8. Bamberger, E.N. et al,
Roller Bearings," The American Society of Mechanical Engineers,
New York, 1971, pp. 8-14. 1
I
9. Kleckner, R.J., Dyba, G.J., "Curve Fit for ASME's Lubrication
Life Factor vs. A Chart," SKF Report No. AL79P007L SKF In-
dustries, Inc., King of Prussia, PA, September, 1979.
"In-
10 McCool, J.I., et al, 'Technical Report AFAPL-TR-75-25,
fluence of Elastohydrodynamic Lubrication on the Life and
Operation of Turbine Engine Ball Bearings - Bearing Design
Manual," SKF Report No. AL75POI4 submitted to AFAPL and NAPTC
under AF Contract No. F33615-72-C-1467, Navy MIPR No. M62376-
3-000007, May, 1975.
7_
1982024093-090
AT81D040
ii. Cheng, H.S., "C_lculation of EHD Film Thickness in High Speed
Rolling and Sliding Contacts," MTI Report 67-TR-24 (1967}.
12. Chiu, Y.P., "An Analysis and Prediction of Lubricant Film Star-
vation in Rolling Contact Systems," ASLE Transactions, 17,
pp. 23-35 (19 }.
\
13. Timoshenko, Strenqth of Materials Part II Advanced Theory and
Problems, 3rd Edition, D. Van Nostrand Co., Inc., 1958. _
14. Crecelius, W.J. and Milke, D., "Dynamic and Thermal Analysis
of High Speed Tapered Roller Bearings Under Combined Loading,"
Technical Report NASA CR 121207.
15. Harris, T.A., "How to Predict Bearing Temperature Increases
in Rolling Bearings," Product Engineering, pp. 89-98, 9th
December 1963.
16. Fernlund, I. and Andreason, S., "Bearing Temperatur'_s Calcula-
ted by Computer," The Bali Bearing Journal No. 156, March 1969.
17. Andreason, S., "Computer Calculation of Transient Temperatures,"
The Ball Bearing Journal, No. 160, March 1970.
18. Harris, T., "An Analytical Method to Predict Skidding in High
Speed Roller Bearing," ASLE Transactions, 9, 1966, pp. 229-441.
19. Harris, T.A., "An Analytical Method to Predict Skidding in
Thrust Loaded Angular Contact Ball Bearings," Journal of Lub-
rication [:achnoloqy t Trans. ASME, Series F, Vol. 93, No. I,
1971, pp 17-24.
20. Crecelius, W.J., Heller, S., and Chiu, Y.P., "Improved Flex-
ible Shaft-Bearing Thermal Analysis with NASA Friction Models
and Cage Effects," SKF Report No. AL76P003, February 1976.
" 3rd 21 Jakob, M. and Hawkins, G.A., "Elements of Heat Transfer,
Ed., John Wiley and Sons, Inc., 1957.
22. Kent's Mechanical Engineering Handbook-Power Volume, John
Wiley and Sons, Inc., 12th Ed., 1960, Chapter 3, p. 20. _
23. Burton, R.A., and Staph, H.E., "Thermally Activated Seizure
of Angular Contact Bearings," ASI.E Trans. 10, pp. 408-417
(1967).
24. Kleckner, R.J., and Pirvics, J., "High Speed Cylindrical
Roller Bearing Analysis, SKF Computer Program CYBEAN-Volume I:
Analysis," SKF Report No. AT78Y002, submitted to NASA Lewis
Research Center, Cleveland, Ohio, under Contract No. NAS3-20068,
July, 1978.
79
1982024093-091
ORIGINAL PAGE IS
OF POOR QUALITY
AT81D040
25. McCool, J.I., et al, Technical Report AFAPL-TR-75-25, "Influ- i
ence of Elastohydrodynamic Lubrication _n the Life and Opera- I
tion of Turbine Engine Ball Bearings - Bearing Design Manual,"
SKF Report No. AL75POI4 submitted to AFAPL and NAPTC under AF
Contract No. F33615-72-C-1467, Navy MIPR No. M62376-3-00007,
May 1975.
26. Dowson, D., and Higginson, G.R., E1_ohydrodynamic Lubrication -
SI Edition, Pergamon Press Ltd., Oxford, 1977, p. 41, equa-
tion (3.15).
27. Pinkus, O. and Sternlich_ B., "Theory of Hydrodynamic Lub-
rication," McGraw-Hill Book Company, Inc., New York, N.Y.,
p. 48, 1961.
_0
i.
1982024093-092
[i ':
[ , AT81D040 ..
V
[i '
'\
APPENDIX A .,
I
I.
[:
SKF COMPUTER PROGRAM ATSIYO03 SHABERTH
HIF __RCHICAL FLOW CHART
:I '!
,
[
[
{
L
A:I
1982024093-093
ATSID040 _-
" APPENDIX A
S K F COMPUTER PROGRAm4 AT81YO03 FLOW CHART
{
Flow Chart
m,l
(
[ The hierarchical flow chart presents the program structure,
listing the program elements in the order in which they would
be called to solve the shaft-bearing dynamic, as well as steady
state and transient temperature distribution problems. The
various solution loops are indicated, as well as notes which
indicate the functions of various subroutine groupings.
Each line in the flow chart represents a program element,
subroutine, function or the main program ALWAYS. The call of
one subroutine by another is denoted by indenting the called
subroutine relative to the routine doing the calling. As an
example, subroutine SKF calls subroutines FLAGS, TYPE, PROPST,
LUPROP, LUBCON, DATOT, CNVRT, CONS and SPRING. Subroutine
CONS calls CONBRI, BCON, TCON, CRCON and CONBR2. BCON calls
ABDEL. Both TCON and CRCON call ABDEL and SLICES.
The first mention of a subroutine within the flow chart
includes the entire list of subordinate program elements. At
subsequent calls to that subroutine the list of subordinate
elements is omitted. As an example the first call to subroutine
AXLBOJ is followed by the subordinate elements, JMVIKT, SNITMT,
' NUMLOS, DUBSIM, MEIE, MEIL and SIMQ. After the call of AXLBOJ
from INDEL, the subordinate elements are not listed but are,

nevertheless, employed. The list of subordinate program ele-
ments are omitted in repeated calls of subordinate GUESS, BEAR, _
SOLVXX and DELIV3 as well as AXLBOJ.
As noted earlier, rolling equilibrium is calculated, first
without, then if required, with friction forces included.
Whether or not frictlun is considered is highlighted with the
' words Frictionless or Friction.
If the Program is too large to fit in its entirety on the
= user's computer, segments of the program may be "overlaid".
7o; this purpose the Program is subdivided into ten (10)
modules which can be sequentially "overlaid".
The Program segments SKF, TEMPIN, and GUESS all per-
form initiation functions and with the exception of GUESS, are
called only once per program execution.
The real problem solving portion of the program is e- Jdied
in segment ALLT. Within this segment the shaft bestir solu-
tion is obtained th_nugh the call to SHABE, then the s{eady
s_ate or transient .emperature distributions are obtained.
A: 2 "'I
"I 982024093-094
"2,
i ATS!D040
i
This scheme is repeated until the end objective, s:eady _a:e
thermal equilibrium or time up for the transient scheme, is
realized.
The nonlinear equation solver SOLVI_ is central to the
program and deserves special discussion as related to the flow "
chart. The first call to SOLVI3 is f_ BEAR. 0nly for this
first call are all of the SOLVI3 sub_._Inate subroutines
listed as noted earlier. These include INSOLV, EQS, PARDER,
SIMQ, EQCHEK, ERI_RIT and ERCHEK. In the subsequent call to
SOLVI5 in which the steady state temperatures are being calcu-
lated, the above listed subroutines are amain called but
these calls with the exception of EQS are not listed on the
flow chart.
EQS is the name given by SOLVI3 to a subroutine which sets
up the system of equations to be solved. EQS is 5_ought into
SOLVI5 through the argument list. When the bearing equation3
are being solved, subroutine BRGGEQ is brought into SOLVI3
and within SOLVI3 is referenced by the name EQS. When the heat
transfer equations are bein S solved as a consequence of the
call of SOLVI3 fro_ ALLT, NET is brought into SOLVI3 and is
referenced as EQS.
Since storage and execution costs rise dramatically when the
program is run at levels higher than NPASS=I, it is recommended that
the program be executed at the lowest solution level possible. If
a level 2 or 3 shaft-bearing solution is desired for bearings with
more than 16 rolling elements, a new executable version must be cre-
ated by expanding the blank common used in subroutine BEAR. Increas-
ing the length of the C array to 33489 elements enab" the program
to simulate bearings with up to 30 rolling element_
A: 3
y
1982024093-095
i
._ AT8 ID040
$ H A B E R T H FLOW CHART (N A S A VERSION)
' ALWA_ '_
P_:_dand _,t tmarlngm_ _r_rlr_soluLion
SKr oontr-I data
--, I FIAO_
f I TYPE
i PROPb'r Set mmterial and lut_'i,uant properties and
0 !.Ul"llnf' culculat_u_r_l_ubL.uXbr' ,.c_q_r_t.ur'u Oe|_nclc._y
l LVDCON calculatloom
I DATOT WPite be_nin6 inputrandta_loo_, presetCUlt.* .\
._ I t TITLE
l I FIll)AT -,
I I ROLDAT
ONVRT Oiloulatebm_-in_r11at_ioxmtant_m .
CONS
COMER1
B_ONABDEL _-
t SLI_ ,
rladand write thenmslsml th_r_ml solu_lon
T_PI_ ecnt_l _t_ and oaloulatebeat tr_nser
I _ eoe1cients
IHAP
5HAF_ RNd and write mhart _:ect_, lo_In_ and
, | I ZErO _ln_ poaitlon data
1 I A_
t
I I O_R
I OOTINP
i
O_lc_ulate shaft _tlo_ eoMt_nt_
J_T
I_xt_
I P_i Calcul=te s_t lnflueno* ooefrlclent_
i I pf
GUESS Oueu wlu_esof -olllr_elemanta_d _,
S_ varlables,posi-,o_s and speeds i
I o_,,
I GBALL !
_'rBRG
I GSTPQ5 _ ,
i GS'_D !
VARRDC
"-_ _in t_m _luti_ or the steady mtat_transient
Calculate5h_t-B_nlr_ oulliorlum,in steps,
for the dcslr_N_A3_--molutlo_le_l
FIT Calc_l.te_ir_ dlu_trsl clemran_m
I_WIT
RI PEI_'_ 11_ _C_l CALCULATIO#3 _atabli_hiterationschmmmto matlsry AIL _?.S VALUES. and t_arln_ i_r rlr_equilibrium
I I
I i
i I
I I
ORIGINN. PAGE B
A:q pOORQUALITY. ;
1982024093-096
ORtGtHALpAGE
OF pOOR QUAL_ AT8!D040 ,
1 BEAR Establish ltorstl_ seherm to sttlsfy r_llir4;
I pREp^;( elow_nt equilibrium. .e_lve rollln_ sl_r,a.t
I fq I INITX equllJ _ ,i,m,besrln_Ly bearing,one ele.-,nt
I 1 UNLOAD st a ti_ 4,
11 I b UNI,ODT
_ ! o I M^)TMI N
fl L _OLVI
l _ INSr_v [,_ II,X
{ l N i {)_iP03
O I ERCHI_
llll{#2 _ = BIICED_ (WZTHCtrrFRICTION) Evaluatoballwleeway eo.taetloads, ball
l
#3
B:BALL (BALLBEARING) _ntrilu_l force end bail equilibrium equatlorul T I B S r _.,CTltL
'i H E II
-_ E l B_LIN
E A ^ M I
R R F
EQTAPR (TAPER_ i_IA, ZI_ BEARZNG)Evnlu,ltn tnp_rndroller-rnenl_ly Ond flnn/_
M I T N_ L._L_.,'LlOOL c_n_act loads,centrifu6alfoz'_, anr..lfjrom_Int
, AL N _ T TAPF_.T_IN and r_llerequilibriumequ_tlccm
Z C A I
t, R q l FI.NOIW4
QU I U l I EqCYL (C_LI, =AL ROTOR BEARI_; _luate cylindricalroller-rac_y _md flm_e
- A N L BGC'P(L co.tactleads, centrift_l foPe_and roller
L R G I ROU2J_ equilibrium_quation_
I A . ' 'LN[_EF'
B
N E _ .Eq
': R C q T I TPHORM
I E
U I M I I l-,Jl
M L i I
I PARDER Caloulate;_rtlal._rlvativ_sof r_llipgelmnt
I B I _ BROGEQ equilibriumequationswith respect to rolling
,, R SII_Q element ix_itlonsand ealculatanew positions
M
, E I _ BBaGEQ
! M l ERWRIT
T _ _ rollln_el_nt-L_mr ring forces_n_ mmnts
R l SU_
N LL_E
S _a_culmte _n_,L"J-'_fatlguelife
I B_II_ Add p_rtlaldcrlvotlvesof bearl_w_inner rlrw_
_HAPA force.Ith re._.t to innerrln_ _ispl_nt
N { FILL8 to the shai% _KN{llbriumequation_nd obt,_in
T $]]_ n_w Innexrin_ posltlon_
_' EEA.9 AIlersh_fl equilibriumis l_tlsfl_l,enlculnte +
_':'.r_ bearingreaction forcesat the bearlr6oqulllbrlum
C VLSC'2 ENalu_te_mp_r_ture dependent iubric&nt
Z _rope-tiesand cOnstants.
{ ALPI'L_O
DRAGNO
A STO_N
! ! EVAI_ (IF N?/_SS: I) _(IP Ir l_/k_ 2 OR _pREPAR rolllngEvaluate b_aringperforman_ with _t J.t_delement sp_'_ {
S BBC._E_ (wITH FRICTIOn)
E_BALL Evaluate ball-racew_ycontact loads.
BCEYRL
f BALLSN
.I FMIX/N_.SA Evaluate bmll-r_ce_y _ film thlcU_,
f TINT inlet_r_ eo_eent_ted contact friction
FILM ror_s and heat_ne_t_on rate.
THERFI I-
STARFC
HDFRI i _
{ rnA_
FRINT/NASA
ALLEN
I'RICTN
i CAGES[= [v_lumt_ _ speed, _ll-eqIe norll m_-',
] C.AGEFJ;} frtetlonforces and heat rates plu_ c_
l L_IJ_D _il-_ing land no_l and frtctlonforces
{ I CGDR plus hut _nnen_tionrmtes.
COAT
r.nl_Z
l C3BALL
I CGROLL
5UMCGL
SUCRE
I , =_
.L_ _J I I { I I l A: 5
i
1982024093-097
I , , l I i I
AT81D040
I RRG^X Evaluato tall, Ix_itlon dcpondont, ncc_lorationn,
- I I Oi_X
[ + + LII"_ Calculatebearlng fatiguelife
I FL_AC
HtrfRAC Calculateball-racewayheat transfercoefficients
_KFR Evaluate roller-race_y and roller-flar_e contact
BGCTRL loads
rum
I T_
TP_
I po.._
I pULL
FMIXRINASA Evaluateroller race_y _D film thicknessinlet
SPEEDS and _trated c_ntaet frictionforcesand
FILM heat 8enerationrite
TH.ERFI
STARFC
HRFRIC
SPEEDS
ALLEN
FRICTN
FLMIX/NA_A Ewaluateroller-flangeE_D film thicknessinlet
FILM and e_-_mtratad contactfriction forcesand
THERF_ heatg_neratiunrate
STARFC
HDFRII
I FBAR
ALLNPT
FRICTN
CAGSPD Evaluate tapered roller bearin6cage speed,
CAG_ forces, roller accelerations, tearing fatigue
BnGAX life, roller-raee_y and roller-flam_ heat
LIFE transfereoefficienta
HtTRAC
HtnTtS
EQCYL Evaluatecylindricalroller-raee_ayand flange
l_"_t normal and frictionforces and heatp_ratl_
ROLLIN rates plus EHD film thic_
1 17_0_
', FI.N0m
I FHIX_/NASA
I FI,HIX/NASA
CAGSPD EValuatecylindricalroller_earin6 cage speed,
CG_EQ ca_e fortes,rollera_-_eleration, be_u,in_fatigue
BRGAX life,plus roller-r_ce_ayand roller-flan6eheat
LIFE transfercoefficients
HOT_C
HUTFLN (IF N_KSS= I, _ TO ALLT.)
BEAR (IF N_ASS : 2) S_IF IF NFA_ : 3 Solve t:.e_olllnEelementand ea_ equilibrium
I PREPAR equationswith frictionincluded,usinE the
#q I PX]LVI3 innerrinkpositionsestablishedby t_e
I I E_ = BRGGEQ (WITHFRICTION) frictionlesaSOlution
SON_ (IF NP_ --3) Solveshaft, rollingelementand eaEe equilibrium
BEARC equation_with frictionincl_ed
#3 I BEAR
I I PBEPAR
I#_1 _V13
l I
I I
_lI_3 ISKF
TITL_
RITE
RITE1 Wrlte shaft-bearing output
REOUT3
REFLOT
[ RITE4
RITE2
FL._ O_*C, mVAA. PAGEIS
I I I
' '' Q 2.
I I , I A:6
I
II ! ,
' I
1982024093-098
: ORIGINALPAGEIS
OF POOR QUALITY AT81D040
SoIve ['or the st,eQdy st_t_ tem_z*at_
FILLGT distrlbuti_. Write shat-_arlng output
_3tV|3 _r_ tel_x?rnture _i[_it'd_:_il_2d.
I EQ_ : NET (FORSTE_Y STAT_ _MAL SOLUTION) _:
DEtIV3
Determinean appropriatetime stepand
! I NET (FORTHE TRANSIENTTHERMAL SOLUTION) solve for the tra_slenttemperature
: [ I _ distribution. Write shaft-bearingoutput
DELIV3 and temperatureap i( de3ir_:d.
NET
I NE_Z
TNAP
DELIV3 Writeoutput for the finalsolution
_aP
NP&,._S: 0 SHAFZ_J_DBEARING _ RING_I.IBRI'UM ARE
,_I._qF'_ CONSIDERINGELASTICCONTACTFORCES.
NO LUBRICATIONOR _I_ION EFFECTSARE
_I_.
NPA.g_: I ,._IAi"T AND BEAP, ING INNER RINC EQUILIBRIUMARE
SATISFIEDCO_IDERING EL_IC C_TrA,qTFORCF..S.
I.!n_I|Tf'ATTr_ AND FIITr';FT{_ I-.Fr.'T.CI:" Ahl." (_,_'_;I_:RF_
USI_ _C_^Y CONT_)L (BALLBEArI_) 0._
EPICYCLIC(ROLLERBEARINGS)ASSU_IONS TO
ESTIMATE _OtLINGEI_ AND CAGE SPEEDS.
NP._ -."2 SHAFTANDBEARINGINNERRING E_OILIBRItIM ARE
SATISFIED_IDERI,'q_ _SI'IC CONTACTFOR('_.S.
USING THE INNERRING POSITIONSTHUS OBTAINED,
ROILING_ AN_ CAGE E_OILIBRIUMA_E
CONSIDERINGFRICTION.
NI:_ASS 3 C_LETE. SO{.UTTON._FT AND BEARINGIN_ER
HINGpLUSPOLLING_ _%'DCA_E E_ClILIBRII_
ARE _ CONSIDERINGALL ELASTICAND
F_ICTioNFO6CE.5.
"i
A:7 _!
1982024093-099
ATSID040
\
.L
,; APPENDIX B
f
{
t.
HEAT TRANSFER COMPUTATION NOTES
I
r
: |o
[
B:I i
1882024083-100
APPENDIX B
AT81D040
[ HEAT TRANSFER INFORMATION
1.
B.I BACKGROUND
The temperature portion of Program AT81Y003 is designed
to produce temperature maps for an axisymmetric mechanical system
of any geometrical shape. The mechanical system is first approxi-
mated by an equivalent system comprising a number of elements of
simple geometries. Each element is then represented by a node
point having either a known or an unknown temperature. The environ-
ment surrounding the system is also represented by one or more
nodes. With the node points properly selected, the heat balance
i equations can be set up accordingly for the nodes of unknown
' temperature. These equations become non-linear when there is
convection and/or radiation between two or more of the node points
considered. The problem is therefore reduced to solving a set
of linear and/or non-linear equations for the same number of
unknown nodal temperatures. It is obvious that the success of
the approach depends largely on the physical subdivision of
the system. If the subdivision is too fine, there will be a large
number of equations t_ be solved; on the other hand, if the
subdivision is too crude, the results may not be reliable.
In a system consisting of rolling bearings, for the sake
of simplicity, the elements considered are usually axially
symmetrical, e.g., each of the bearing rings can be taken as an
element of uniform temperature. For an element which is not
axially symmetrical, its temperature is also assumed to be
i uniform and its presence is assumed not to distort the uniformity
1 in temperature of a neighboring element which is axially symmetri-
cal. That is, the non-symmetrical element is represented by
an equivalent axially symmetrical element with approximately
I the same surface area and material volume. This kind of approxi-
._ mation may seem to be somewhat unrealistic, but with properly
_i devised equivalent systems, it can be used to solve complicated
problems with results satisfying some of the important engineer-
ing requirements.
f
'i The computer program can solve the heat-balance equations
r for either the steady state or the transient state conditions
and produce temperature maps for the mechanical system when the
, input data are properly prepared
i B.2 BASIC EQUATIONS
_I _2.1 Heat Conduction
The rate of heat
flwqci,_(W} that is conducted from node
i i to node j may be expressed Qy,"
i
i
ij _ B:2
.
] 982024093-1 O]
1 ORIGINAL PAOli' IS AT81D040
OF POOR Q_JALITY
. Xi) Aij
qci,j Lij (t i - tj)
t i and tj are the temperatures at i and j, respectively, Ai, j
the area normal to the heat flow, (m2)._j the distance
(m) and hj the thermal conductivity between i and j, (W/mOC).
Assuming that the structure between point i and j is
composed of different materials, an equivalent heat conductivity
may be ca1 ulated as follows:
' [_ _)'x_'__'__ Al///_ 'J Fig.At] B-I_.IAI+ _2 A2
I///// /. " " %
Aij = A 1 + A 2
__/_ '\ \
_ij = ii_ I + _21X 2
-- _.! -.-J_l - tij " tl + _2
The calculation of the areas will be discussed in Section B.2.5.
B. 2.2 CONWECTION
The rate of heat flow that is transferred between a solid
structure and air by free convection may be expressed by
1982024093-102
3 'L
t OR|_NAt PAGE IS
OFPOORQuAUTY
AT8 ID040 ,!
qvi j = ai j Ai j Iti " tj] 1.25 . SIGN (ti - ti) i J | S
L
where
, if _ti tj)< 0
in which
_.5 - 10 -2 W/m 2- (degC) TM for hot surfaces facing upward
and ccld surfaces facing downward
aij .4 I0"2 W/m 2- (degC) TM for hot surfaces facing downward
and cold surfaces facing upward
1.8 10-2 W/m 2- (degCJ TM for vertical surfaces
For other special conditions, aij must be estimated by referring
to heat transfer literature.
The rate of heat flow that is transferred between a solid
structure and a fluid by forced convection may be expressed by
qw,j = a.l,j Ai,j (ti - tj)
in which aij is the heat transfer coefficient.
Now, with a = aij, introduce the Nusselt number
aL
Nu= -_[
I R L
the
Reynolds number
u
and the Prandtl number
o _ Cp
Pr
where
L is a characteristic length which is equal to the diameter
in the case of a cylindrical surface and is equal to the
plate length in case of a flat surface (m)
"1982024093-]03
,q
OR_IFIAt., IPAQI[iS ATSID040
O4=POORQUALn"t
I
U is a characteristic velocity which is equal to the difference
between the fluid velocity at some distance from the surface
and the surface velocity_/sec) ,
is the fluid thermal conductivity(W/MO
is the fluid kinematic viscosity(M2/sec)
p is the fluid density(kg/m 3)
Cp is the fluid specific heat(J/kgC)
For given values of Re and Pr the Nusselt number Nu and
thus the heat transfer coefficient may be estimated frSm one
of the following expressions:
Laminar flow along a flat plate: Re < 2300
Laminar 1ow of a liquid in a pipe:
Nu = 1.36 _ Pr(_)
e
where D is the pipe diameter and L the pipe length
Turbulent flow of a liquid in a pipe:
0.8 _p
Nu = 0.027 Re r
i Gas flow inside and outside a tube:
0.57
i Nu = 0.3 Re
Liquid flow outside a tube:
0.5 0.31
Nu = 0.6 Re " Pr
Forced free convection from the outer surface of a
rotating shaft
Nu - 0.Ii [0.5 Re 2 p_0.35
where the Reynolds number Re is developed by the shaft
rotation.
_ D2
Re -
B:S
1982024093-104
r
ORIGINALPAGE IS AT81D040
OF POORQUALITY _
in which w is the angular velocity (rev/sec)
D is the shaft diameter (m)
The average coefficient of forced convection to zhe lubri-
i cating oil within a rolling contact bearing may be approximated
c by,
using for outer ring rotation
- for inner ring rotation
in which N is the bearing operating speed (rpm)
D is the diameter of the rolling elements (mm)
dm is the bearing pitch diameter {mm)
B is the bearing contact angle (degrees)
B.2.3 FLUID FLOW
The rate of heat flow that is transferred from fluid
: node i to fluid node j by fluid flow is
qfi,j" P_ cp (ti - tj)
_ is the volume rate of flow from i to j. It must be observed
t the continuity of mass requires the following equation to
be satisfied
o
z vii . 0
provided the fluid density is constant. The summation should
be extended over all nodes i within the fluid which have heat
ex:hange with node j by fluid flow.
B.2.4 HEAT RADIATION
The rate of heat flow that is radiated to node j from
node i is expressed by
= 6i,j { (ti @*T 4 - (tj*ZT3) }
_1 qRi,j
where
!
Ti " temperature of node i in C.
m
T = temperazure of node j in C
]
B:6
i
1982024093-105
f
C
i ORIGINALPAGE iS AT81D040
OF POOR QUALITY
and the value of the coefficient _i ' depends on the geometry
and the emissivity or the absorptiv_ of the bodies involved.
For radiation between large, parallel and adjacent surfaces :
of equal area, Ai, j and emissivity, el,j, 6i,j is obtained from
the equation ,

6i,j " i,j Ai,j '_

i
where _, the S_efan-Boltzmann constant, is
a . 5.76 i0"8 Wlm 2 /(degK) 4
For radiation between concentric spheres and coaxial cylinders
of equal emissivity, Ei,j, 6i, j is given by the equation
i,j aAi_i
_ij "- I (i-5,j)
1,3
where is as above, A_ _ is the area of the enclosed body
and A*-
1,j is the are'a o_Jthe surrounding body, i.e. Ai, j <A'i, j.
Expressions for 6i, j that are valid for more complicated
geometries or for different emissivities may be found in the
heat transfer literature.
B 2 5 CALCULATION OF AR_AS
.
In the case of heat conduction heat transfer in the axial directior
Ai, j is given by the equation LFig. B-3).
Ai, j = 2xrm .%r
Referring to the input instructions, Section 3.4.7, but recalling
L must be input in rnm not m_
L1 - rm = rI * r2
2
L2 _r r2 - rI
b
B:7
1982024093-106
t ATSlD040
ORIGINALPAGEIS
[ OF POORQUALITY
t
In the case of heat transfer in the radial direction, Ai, j
is obtained from the expression
Ai, j = 2_ rm H; L1 = rm; L2 = H
and similarly for the radiation term above
A* - 21T r* H
i,j m
L3 " r* m
_ L2 = H
in which H is the length of the cylindrical surface; where heat
is conducted between i and j, r m is given by the same equation
as above (Fig. B-4 (a)); where heat is convected betweeni and j,
rm is the radius of the cylindrical surface (Fig. B-4(b) ); where
heat is radiated between i and j, rm is the. radiu_ of The enclosed
cylindrical surface and rm* the radius of the surrounalng cylindri.,
cal surface (Fig. B-4(c) ) H
r
J
j
, I
" T2
. L qr _ _
Fig. B-5 Fig. B-4(a)
B:8 _.
[
i
1982024093-107
p_.-- _-, AT81D04 0
oRIGtHIkI..
oe pooe _n_
H L
oJ d //// ////
TT_,_ " _ '\
\\\\ \\\ _ \\\\\ 'N\\\ *, ,,I -
! _
Fig. P -4(b) Fig. B-4(c)
B.3.1 TRANSIENT ANALYSIS
For the transient analysis all of the data pertaining to
the node to node heat transfer coefficients must be provided by
the input. Additionally, the voltage and the specific heat at
each node is required.
o .
B:9
.
I
1982024093-108
AT81D040
. APPENDIX C
I
,[ SKF COMPUTERINPU T DATAPROGRAMFoRMsSHABERTH
I
I
I
t.
!i c._
1
1982024093-109
I ! ORIGINAL PAGE IS AT81D040
OF POORQUALITY
It.
, t 1
' i
|"
!
!
!
!
!
!
i
!
w
I !
|
I" i ,
,. C:2
1982024093-110
r,,.
_' ORIGINAL PAGE IS
OF POOR Q_I.IAL'TY
ATSID040
,O
_Crown Dro. Flsa
#.
=" '__" : AT81D040
q_ O _,O E L_
'" ; ,_- -_.;-,,- - : ORIGINALPAGE IS
; _ _- _ : : _._ QUALITY o -- OF POOR
p
_ e' iw'l
== C _ _ z.. I .._ O II O
II la. _ .I _,d i e t._
_ ._r" 0
e
..,- _ .- _ ,._ -
o
.o,,
,_ _ : ,- I
ii,
_= _ ' =,.
O k,
- _ _ _ _
- =
C:5 i.
] 982024093-] ] :3
I
J ORIGINALPAGEIS
: OF POOR QUALITY AT81D040
_. I -
I
b
L
.i .....
_i
",' ORIGINALPAGEIS
OF POORQUALITY
AT81D040
I
1982024093-115
ATSID040
Ii
ORIGINALPAGE'IS
i- OF POORQUALITY
1
t
i
ORIGINALPAGEIS
: OF POORQUALITY ATSlD040
1982024093-119
ORIGINALPAGEIS
i OF POORQUALITY
I
' AT81D040
|
",q !_
1982024093-120
AT81D040
ORIGINALPAGE M
MCa _ POORQUALrrY
I,
AT81D040
I
ORIGINALPAGI! '_ 18
OF POOR QUALITY
k
OflIGINAtPAGE IS AT81D040
OF pOOR QUALITY
",-' =" ORIGINAL ?AGE IS
_L. ...d
-=-_ OF POOR QUALITY AT81D040
M
2""
, :
8 _
S "" 1
_ L
I
I
!
I
_" I
!
,r
.t. " 1
? t I
o o _
I
!
i_r 0
I--
I
I
l
!
ol I
I
_._ _ ,
I
II I
.'2
_" C:20
_ - .;
, ........... _,,.,_,,,_,_,_o__,_o
I_O_U/_U_O //u
/
' _ AT81D040

i
O
_ ORIGINALPAM B I
i j o, OF POORQUALITY
'i" _tt] I
" i
l b I_I i
!_ ''
f _ L-,,
Ir'l IIl ,,,._'_ I
t!i! _!I -=,
|-. !
!1
I _,..ii I I
,.i,> it i!.xl.'.::I
"=" I I ti::::
= ' ] I I ..uuuu
, i- 141I Q
m 1.ira....
I .,. I .r i
II i
'i I
-,. : |
!
11
81
I :i ! !1 "
L
I '. . ' ,
C:25
'1982024093-'133
i
L
AT81D040
O_NAL PAGEIS
OF POOR OUALffY
r
!
z
ORIGINALPAGE_,I &TOLD040
f OF_ QUALm'
(
!_ omelet- pAeII _,TeZD040
_ O#i,OOI1 qU_IW
_ ---*---t
b
I ORIGINALPAGE IS AT81D040
[ OF POORQUALITY
i .
i
1
APPENDIX D
I
T .
._KF COMPUTER PROGRAM SHABERTH
SAMPLE OUTPUT
.
7
F
l
D:I
v
I
1982024093-141
I
AT81D040
APPENDIX D
I SKF COMPUTER PROGRAM SHABERTH
SAMPLE OUTPUT '_
1
The SHABERTH output samples which are displayed on Pages D:3
I
I to D:94 SKF and NASA versions of the code for three
represent
different shaft-bearing systems. The first set of examples
pre_ent the results obtained with the SKF and NASA versions for
I a system in which an input pinion is supported by a preloaded
pair of taperod roller bearings in a straddle configuration.
'I The _d set of examples display program output for a
system ,. _nich an i. r,tt pinion gear load is supported by a
r
I flanged cylindric?, , iler beari_ig in conjunction with two
an,_lar contact ba+[ bearings.
The third _et cf examples represent execution for a single
ball be3rine, sy:+em operating under a combined radial and
thrust lo_d.
The differences between the SKF and NASA versions of the code
which are described in detail in Appendix F, are reflected in
the following output parameters:
Value of EHD film thickness.
Ratio of film thickness to surface roughness (H/Sigma) printed
for most heavily loaded rolling element.
Lube-life factor.
LIO fatigue life.
Frictional heat generation rates.
Ratio of load carried by asperities to total contact load
(QAsP/QTOT) [for NASA version this ratio : O, see
Appendi F].
D:2
1982024093-142
AT8 ID04 0
SKF
m
2 TAPERED BEARINGS
LEVEL 1
D:3
L
I
1982024093-143
AT81D040
omGtNAL pAGEIS
OFpOORQUALrrY
'i
J
\
".1 :
o
.J
I-
kJ
_ z
o
I
f 0 ::[
c
o
o 41K Ig _
w
LJ _
911
.J I_ _ ! ! o.I
[ "
D
, Z 0
V"
[_ "
D.4
] 982024093-] 44
1982024093-145
!
ORIGINAL PAGE IS
r
t OF POOR QUALIFY AT81D040
ORIGINAl. PAQ[ m
: OF POOR QUALITY AT81D040
AT81D040
ORIGINALPAGEIS
OF POORQUALITY
oRIGIN_-LpAQI, IS
OF pOOR QUALITY AT81D040
AT81DO_O
ORIGINAL PAGEm
OFPOORQUALrrY
0
T W
_ * tO
E
_ O
O
_ x E ,-
- _
=.;= - _
D:IO
i.
] 98202409_-] 50
ollml_ pAOI IS
{_jI_.IT_' AT8ID040
J_
t ORIGINALPAQEIS ATS 1D040
OF POOR QUALrir'Y _. *
Ln
t
t
W r,_
f _ LJ
q _ _
-J _3t e_
T
I ,u
fO _ _o 4 o
E _ 0
I L_ ,u.e (%_ _ e,_
d _" o _
.J _. 0 .j
: w
" " D'I2
l
1982024093-152
1982024093-153
1
AT81D040
ORIGINALPAGE IS
OF POORQUALIT'!
4
1982024093-154
ORIGINAl- PAO( IS ATSID040
oFpool;QuAUTY
!
v _ oeeooeooeototeeo o_ _
: _ _
. _ _ ooooo_- _ _ _ _:_,.... ..... :. _ _o"
" ''''''7_'_??_?_'' .... t
_ eo oo
2 _ _ o*.eoeeeeeoeeeoe
_ _ _ oeeeoeeoeeeo_eee _
+] _ 1
i
D:15
, !_
] 982024093-] 55
i ORIGINALPAGEIS
OF POORQUALITY
i. AT81D040

c q , , o * , , _ "_
_o '\
*
0
z
" gl"
_o
., _;2;Zg2gZ;2;Z;2
: D:I6
[
1982024093-156
;_I oRIGINALpAQEIS
OF PoORQUALITY
i
re* ! e.,ee=G1,.qq,_ 1.q m.oq,=l e.,e.q .q
W .t
t=.-
2
i _ !" I=,,-4r e 41" _Po 4re 4"0 4Po 4_e _r 4J" e "_e
% e..
:. D:I7 [_
U _
] 982024093-] 57
i"
1
ORIGINAL PAGE IS AT81I)040 i__
OF POOR QUALITY i_
'_, D:I8
1982024093-158
ORIGINALpAQI[ iS
OF POORQUALITY AT81D040
t
D:I9
1982024093-159
i
AT81D040
]
_ .
J
,d
'7 2 TAPERED BEARINGS
LEVEL 1
D:20
_J
J
1982024093-160
i .
AT81D040
f
* ORIGINAL PAGE iS
OF POOR QUALITY
ORIGINALPAQ| 18 AT81DO40
OF POOR QUALITY
L
f
F. i
AT81D040
(! ORIGINALPAGE,
t.
OF POOR_NLJALrrY
AT81D040
_,
ORtGINALPAG| IS l!
OF pOOR QUALMS' it
W OO E Oa _ gO
to ee _ oe
W
w !
m
m m_ m_
AA
co
c _; _ _ _ _ .....
o _ o - x. ;
U
m
D:24
] 982024093-] 64
[+i
AT81D040
1+
1982024093-165
1982024093-166
,l
i ! ORIGINAL PAGE IS AT8!D040
: OF POOR QUAI rrY
]
i
t ,
_ _ :
0
i1 "
O
t
j _
t
I
"J D:27
] 98202409_-] 67
oRIGINAl..IPAOf.IS AT8ID040
O_ pOOR QUALITY
ORIGINAl PAGEIS
OF pOOR QUALITY AT81D040
i
i!.
Jl u,l
U
0
-J _ )-
Z
im ,_ _-) r,-
-r 0 i_l Cl ql ii
: =
IC I lied lie I.) l'%i ,i"
&._ $ _ I'-- I,.'- !
Lc
r._ + ""
- _ _ _-.:
_,. _ _ LL
..I _r 0
0 _ _-_ " " _ _ _
6a
, __.
J _a
_ g
_ - ; -
: _" 0:29 [
1982024093-169
ORIGINALPAl II
OF POORQUALITY llT81_)O_o !
,_ 0'_ l,t.' _,k.. (,,a
t
"_ ,
.._ (.a
_ l.J
l
;._,"
La
_.. _ _ _
I,,, .. "/
_ _ _ -
t.-,
1982024093-170
i
I
ORIGINALP&G! IS
i OF POOR QUALITY AT81D040 .
"3
.j
: OR_INALPAGE m
OF POORQUALITY AT81D040
W Illillllilllllll
0
'''llllllllllll, .... [_
" ti
i
_ oooooeoooooeeeee _o _
m
. ; : _==g:_::g_:ooo ==oo o o
Z
_ _ _ eeeeeeeeeeleeele _ _
U 0 _0 _0
d _
_ _ _ eeleeeeeeeteeeee
_ Ioooooeeeoeeleee _
0
_ 00000_00000000_0 _0 _0
_ _ leeeeoeooleeeeee _
: D:32 =
]982024093-]72
I .
ORIGINALPAGE IS
- OF POOR QUALITY ATS1D040
1982024093-173
Of pOOR q u_'_l AT81D040 i
i!
II
\
x _ _ 1
Z X I
. = -rrrr
D _
U _
_ W
_eeoeoieeoo
D
, _ _ ___
q II
_leioiOOee_
H
i IIIlIIIIII
(r
. : D:34
\;
1982024093-174
-i[
:! ORDINAL PAGEm
il OF POORQUALITY
I AT81D040
'I v Z * * * , , * * * o _o _0
' _ _ O0 O0
W
W
g
_ Z O0=O00000a @
_' _ _ _ oeoeoeoeoo _0 O0
_ Oa=_O00=O= NO _0
0 oo oO
U
k liOOillotl _0 _=
eo oo
= ooo:=oOOO = o
==00=0=0= 0 _0
_ _ 0 oeooooeooe _0 _=
d =
_ oeeooooooe _
_ eeoooooeee _ 0
' _ _ eoooooeooo _l
]
_ ,
]982024093-]75
1982024093-176
{-
AT81D040 ,_
J
L
i
! .
! 1 CYLINDRICAL
! 2 BALL BEARINGS
LEVEL 1
T
I
r
f .
I.
.
.
D:37
#
L
1982024093-177
I
I
ATSID040
I ORIGtHAI..pAGE IS
: Of pOOR QUALITY
" ORIGINALPAGE IS I '
OF POOR QUALITY AT81D040
L
AT81D040
ORIGIMALPAGELS OF POORQUALITY ;,
Ol_Gtt___ (_jALt_I AT81D040
1982024093-183
ORIGINALPAOE m AT81DOqO
OFPOOR_.IALJ'll'Y
T
0
E _ 0_
oO
O O0
a
_ oe
q_
oo
O
d _
0
_ _ _ 0 _ oeeoeeeee+ee
_ _ toe_ooeoeloo
i D:_
t
1982024093-184
: OININAL PAGEIS
OF POORQUALITY ATSIDo_;o
1982024093-185
D:46 o
1982024093-186
01:pOOR (_%JI&L_ AT81D040 i
_o
GN _ Z_
oeo Z oee Z _ !
0
o o o 0 o o
- o _ _ _ _
- _
D:47
i
1982024093-187
AT81D040
ORIGINAL PAGE IS
OFPOORQUALIFY
_'J 41"
,4qm
4"4"
q,,i i,,0

mng',
clO
41' dr
_m
OlD
J'_ tP,
QDeD
4r4r
_tfJ
_CD
gml,_
I I
OIID
D:48_
t
1982024093-188
ORtGINAI,-PAGE
OF pOOR (_jALffY AT81D040
:
/
W m
Z O
vJ _
0 _oeoooooooo ,
z > z
_ _oogoeeo_ee
0 _ _eeeeeeeeee
_ . _;;;;;gggg;
_ _oeooeoeooo
m
z D:_9 _
I'
I:
_ -
, 4
] 982024093-] 89
>
OR|G|NAL PAG B AT81DO_0 :
_ OFPOORQUAL_y
J
/ o
E eooooooo.. Q
N _ Qe ee \
o O
_ oeeoeeeooo _0 _ .
ee eo
m
I -
a,
v _O _O
_==: .... ;
- _ "''g& ......
- _ o =o.==o...= _=_ =:.
q
D:50
] 982024093-] 90
OF pOOR qUALiTY AT81D040 ! ,.
\ /
o
_ llllllllll|l
W
Z _ _10__ 'r _
__'_
[ m I|lllllllll
%
N 2
!
" D:51
!i
]982024093-]9]
AT81D040
l ORIGINAL PAQEm
OF POOR QUALITY
{
R
_ q
K
0
_ U
Z
_ e_ooeooeeoe
0 _ M
T _
M _ e * *
: e
:I
_
D:52
] 982024093-] 92
I "
i ,:
ORIQIHN-pAGEm I
OF _ (_d_,ITY AT81D040
ORIGINAl. PAGEIS
OF POOR QUALITY AT81D040
_ oeoeeo0oeee
W
0
_ eeeeeeoeeeo
_ _ 0 qeeeoeeoOee
d =
e_eeeeeeoe
0 _
_ _ ooe_e_ee_e
I .----- - '
_ = - 0.o._..0...
eeee
: _ D:54
L
] 982024093-] 94
I ATSID040
NASA
1 CYLINDRICAL
I 2 BALL BEARINGS
i LEVEL 1
I
i.
[ -
i
t
J|
4
r
!
D:55
1982024093-195
1982024093-199
1982024093-200
; '+ _+ ,
PAM 13 t+
OF POORQUALITY ATglD040 I ,
t
O
0 _o
% 0 _0
. o oo } ;\
: w
O t
_4
_ W
X r
_C
N q_ oe
" " i
U
d -
!
_ _ _ _W l
n
0 i _ illllltiltti
0
N
D:61
] 982024093-201
t
ORIGINAEPAQEIS AT81D040
OF POORQUALITY
I
1982024093-202
il 'r_
ORII)INM. PAOf_ AT81D040
oF pOOllQUALITY i+
l
oao I
Qm JQee
1 *" "'QO i
r _ e
_DOC
00 W $_Q
t_
+ P ,m+ + o +++ ore= .
u ;
ooe
_ _ .... _ _ .
" + " g d
" D:63
_' N
]982024093-203
!
ORIGINALPAQE IS AT81D0tl0 _:
OF POORQUALITY _
4'41"
QD
QID
4P41'
C,t
IflW#
mwl
DO
! !
Ot
! I
._,
e,,l_
IP'g
J
D:66 ..
1982024093-206
_j
i .I
AT81D040
i ORIGINALpAGEIS
oFpoo_ QUA.'n' I!
,5. r _ _
.',... ORIGINALPAGE
OF POOR QUALIFY AT81D040
O
e +
_ v Z e e , _ _
_ eo oo
Z
: o o
_ _ eooeeeeeoo _ _
0 eo oe
m
_ ooooeo_eee _0 _0
4 0 ":
_ 0 0 eooeoeoooo _ O0
d _
oeeoeoeeoe _, _
_ _ eo ee
Z _ _ e_eoeeeooo '_
o ....
_ eeoccoo _
D:68 "._
/
1982024093-208
i.
ORIGINALPAGE IS _
OF POORQUALITY AT81D040 i
I
q
_ llllllllllll
v,
, ___
_o_oeoooo_oe
_ _ooeeeeeooo
_ |J||JJ|J||J
_ _ooeeeoeeooe
_ Illllllllll
R
0
_ D:69
/ i 2
1982024093-209
f ORIQINALPAGEm _:
t OF POORQUALITY
AT81D040 ,._
1
W
%
N sooo_oeo$4_
0
oo_o0ooo_
: = _o 9_ .... _9
O
d "
_ J
D:70 I
J
] 982024093-2] 0
!
oFpoo_Quarry i
'i
._ []
% m _|lltOllllll
v
_ _eee'_eeeeeee
i t
H
_ooeoeoooeoe i
_eeeeeeeeeee
v fllleJl|il|
| t
_ = _ o _ _; ;Zg2gggg=
___
_oeoeesoeeee
_ voeeoooeeeee 2
. . ,_
- = i o
: D:71 1
[
J
] 98202409g-2 ] ]
i
t ORIGINALPAOEI$ .,_"
OF POOR QUALITY
i AT81D040 -L
_ ,i _
-- _ _o__
Z eeeeseeeeee
I
/
_ q
0
_ _ o_o0o0ooo_o
_ oeeeeeeooee _"
i d =
_ _ ee*e*e_e_ee
_ _ eeeeeeeeo
q
0
D:72
1982024093-212
AT81D040 '_
SKF "':
1 BALL BEARING
LEVEL 2
i.
D:73 !
i !
1982024093-213
"....... _,+i +, _\i:+_....
oRiGINAl- PAGEIS I :
" AT81D040
QUALITY
,,, OFPOOR
e, v ._
,.z W r.,.
':' _" at W
: U.U
0 _ J
cc ar _ _'_ 6,
_..J >-
..
,- -,,,,-, _ I-..
m L.J _ I,.I I.,J I.-
-J _ (/" -- .,_ I-- _ CD la.
l] X
1982024093-215
AT81D040
1 ORIGIHALPAGE IS
, OF POORQUALITY
T!
1982024093-216
p_
AT81D040 ' __:
ORIGINAL PAGE/S _.,
or P_oROU4LrTy

.J
E
m
._
t-
Cg
l " =
O W 0
0 ,I O 2I
T _ f W
' N W CD W 0
W _ 0 E
1 _ Z Z N
0 0 0
LJ 11-) "
I -
0
[ -
0
U _ Lb I
II 0 _" 0 4-- I_
ill W 0
I"
, I _ I
I _, 0 0
I
"Jr 0
_II 11( tlI
0
t
0 0 Z
" _ "-:"u _o *"
1982024093-218
i D:80
I
] 982024093-220
i
OR_|NAL PAGE 18 AT8 IDOq0 _ :::
[ OFPOORQUALITY i:..
.[
Q V .
[i " _""'"""
_o$ooooo
% t llllltlll
|
W
W 0 _
_111111tl
[ - _ --_._----
_ 0__
[ " l _ .
J I _ eo4oo
!
_ _'_ _l
i + I llli_l_l
t + lilmll
t _ _ III
| _ W _##II_
_ __I
_ _ _I__
_ __._
IIIIIII1
_ _ WIIIIIIIII
I I I _oeeeooo
| m I _I_I_
I
I
[ -
I D:82 ,
i,
1982024093-222
ORI_NAL PAQEm
OF POORQUALrrY
AT81D040
_ ooeoooeO
W
-
H
_ _ _ ooooe_oo
_ _ eeoeooeo
_ oeooeoeo
-- eeeee_ee
_ _ C
_ _ ,eeoeoeo
- - II _ |,J
_ r
_:sa [i
1982024093-223
I
AT81D040
I
NASA
1 BALL BEARING
I
LEVEL 2
r
" D:84
,1",
1982024093-224
__ .. ill
OF POOR _ ATBID040
0
' _ il
U. u
L_
_O-w
.J _i--U
] ;: i
|
!
,_ I--- _,1CIg
CD
*
t_ L_ _ i 4.J
i
'" " _'" "F." . . . D:86
I
] 982024093-226
AT8ID040 ,++-,+
ORIGINALPAGE IS
_" OF POOR QUALITY -:_,+
-._ _
, <
"5
.t i
ql
: w
w o
l,- 3
_ 4
a
r _ z
.rl _'
",,
I -
[ + "
r _, .,.I i...'
I _ w+
- -, ' n. 87
W
1982024093-227
1982024093-229
OR!GINALPAGE IS
OF POORQUALITY "' "
_" _ ATelD040 t"_
k-
tJ
W W
'J 0
' b.I 31- _I,
q
u
_ X _ _ W U
0 0
_J _ ql
_ iv t
I q_ _,m
*m 0 t.-
b_ t_
_" I 0 Ilk
tk tJ I._ 0 ,._
(/) _ t.,t ,I
Ii_ 0- O" I" -- E (D
::_ _ kJ _ _ _
-" t_ _ e.
L.
Q _
_.ji
1 D:91
F-
..... I
] 982024093-23]
1982024093-232
++ --'--t_-., +-_
-_ ATSID040
ORIGINALPAGE II
OF POORQUALITY
% + _IIIIIIII
W G
X 0
vi
W
_*eeo*ee*
= _ - ????T???
D
+ _ llllllll
%
_ _ _eeoeeeoe
*s D:93
] : -
] 982024093-233
W
I
: . [] + "IIIII
O
" I]
Z
I illlllll
" 1
_ _ _ oo_oooeO
d "
_ 4OllOIOI
! k _ tlllllll
.
%
" ' tit
_ _ eoeeeeee
_ Dlgg
]982024093-234
"_ AT8 ID040 ._
t
APPENDIX E
r Calculation of Cage Pocket and Cage Land Forces .
|
m
I ._
E:I
1982024093-235
!
t of _R QUAI.I'rY
AT81D040 /'
r NOMENCLATURE _"
ii: ' APPENDIX E
q_UN ' A = 16.9706 . R _ I?!_
I 3 + 2k -.
B = 16. 9706qoUxRx % _'_"_ I ,
C = (A + B ) _
, C = cage land radial clearance (in.)
J
% ,
{ Co = nOlUyl (Rx_) KI
Cp = cage pocket radial clearance (in.)
k = R
R
X
K1 = 1 1
3 _ 2k) 2 + 2"-2 " _ / k(3 + _ Uy
= roller total length (in.)
R = ball or roller radius (in.)
R = cage land radius (in.)
=C+' Y'
-- (in.)
,, Rx + R + Cp
, U = Ux + U
y = entrainment velocity (in./sec.)
t .
E:2
1982024093-236
i ORIGINAL PAGE IS '
t OF POOR QUALITY
ATSID040 *
i NOMENCLATURE continued ......
I APPENDIX E
U
1
= entrainment velocity of the ring-land interface
U (in./sec.) _
[i "
.!
= R (_i + _c ) for inner ring riding cage
tl
= R E 1_o + _c ) for outer ring riding caae '
!
4
Vx (in Isec ) _
U - " "
, x 2 l ,
V , i
U = ___y_ (in./sec.) _ II
y 2
z 2) sliding velocity (in. il _ V = IV\..+ Vy = /sac.)
X
_m
V = sliding velocity at the ring-land interface (in./sec.) _ ,i
= R (_i - _c ) for inner ring riding cage
= R (_o - _c ) for outer ring riding cage [I
V = R_ (in.lsec.) I
Vy = -R_ x (in./sec.) [i
cos.-ikc } = -i i : sin i
= angle between the rolling speed vector and the [i
pumping force vector
_o = absolute ambient viscosity (Ib.-sec./in_)
E:3
g
1982024093-237
I _ -_o_qUAL,_Y
_ AT8 ID0 40 i:
T
,
I NOMENCLATURE continued ...... , APPENDIX ?
.
r
_c = cage orbital speed (rad./sec. } !\ _.
t
_' wi = inner ring orbital speed (rad./sec.)
_o = outer ring orbital speed (rad./sec.)
r
i
_x = ball or roller rotational speed, x component (rad./sec.)
, _y = ball rotational speed, y component (rad./sec.) ,
L
>
E:4
[.
1982024093-238
!
:' ATSID040
l! : i
) -
Calculation of Roller-to-Cage Pocket Forces 11
H
A numerically stable means for calculating cage pocket forces
in cylindrical roller bearings was developed in [24]. The analysis
is generalized here to include cage pocket and cage land forces in [_
ball and tapered roller bearings.
I)
Web geometry is taken as a radially outwardcylindrical cavity _
: for ball bearing simulations and a rectangular cavity for roller _i
: bearings, Figure E.I. The force exerted by a ball or roller is
considered to act at the web midpoint, and all pocket contacts lie !
the pitch circle. Interactions between roller ends and cage
on
pocket sides are neglected. The load supported by the web is con- t
sidered to have two components: one normal to the plane of con-
tact (the surface of the web), and a frictional component in the
: plane of contact.
t
: The normal force component is calculated as a function of rol-
ling element to cage pocket offset Z = Z (_I, mg,''', _,), where
w_ is the orbital speed of the i-th _olli_g _lem_nt. PoCket loads
a_e computed as a function of the hydrodynamic lubricant film, :_
h=h(Z_) that fills the gap between the cage web and the rolling !
element surface.
Numerical stability is gained by taking a linear approximation
to the force-displacement equation describing the roller-to-pocket _
normal load component
Fnormal = KZ (E.I)
where K is chosen such that equation (E.I) will match the exact
solution at Fnormal = 67N (15 lb.) for roller bearings, and
F
normal = 1IN (2.51b.) for ball bearings.
K = ii Ii (ball bearings) (E.2) _
p .33R_o _ Cp
K = 67 67 (roller bearings)
p .33R_ o _ Cp
E:5
r
I
1982024093-239
le
E:6
1982024093-240
r, '
AT81D040
U
Contact Inlet Re_ion Hydrodynamic Friction Forces
Ball/Ca_e Contacts []
In a properly lubricated ball bearing, a constant viscosity
fluid film separates the metal surfaces of the ball and cage web. []
Surface motion causes pressure to build up at the entry to the U
contact. Ball surface motion entrains the lubricant into a
gently narrowing wedge, creating a distributed traction on the If
ball surface. Area integration of the surface traction yields a
three dimensional load vector. Components of the load vector are
given in terms of the dimensionless quantities FR and _ [25,. ,i
t_ Forces :
Rollin_ Components i!
FRy
= CoF R cosy (E.3)
l
FRx = CoFRlsiny)_ (E.41
Sliding Components
= _snoVx(RxRy_ ) i FSy (E .5) "
b
FSx = FS novy(RxRy)% (E,61 i
I:i
Moments :
MF,= -R" (FRy. _'Sy_ (_..V) {i
M_,y = R (F_ _s.) (_") {t
L!
D
' E:7 U
i
1982024093-241
,' 9
OR P/ tB ",
Values for F_ are obtained as follows. From [25], the force
I directed normal _o the plane of contact is
= CoFRRxcosy
Fnormal _--
[ , , or
[ FR = Fnrmal R (E.9) :
CoRxCOSy
is known (E.I), and other terms are functions of geometry.
Fnormal
Therefore, FR is known. Functions shown in Figures E.2 and E.5 were
curve fit* to give
[ = FR
FS .26 + 10.90 (E.10)
-I
Roller/Cage Contacts
The analytic description of lubricated, rolle;-to-cage pocket
contact, Figure E.1 (b), is based on the lubrication of a rigid
cylinder near a plane [26]. The radially directed friction force
I is expressed as a function of the normal force J
Ffric = PhFnorm (E.13)
l
where Fnorm is computed from (E.I). _h is a hydrodynamic friction
coefficient
FR is shown vs. dimensionless meniscus distance (p,) in Figure
I E.2: FS is shown vs. 01 in Fig. E.3. The functions are closely
approximated by:
'I FR = 34.74 (nPx) - 27.60 (E.II)
I FS = 8.82 (np,) + 3.89 (E.12)
for hydrodynamic contact. We get (E.IO) by solving E.11) for
_n_ a and inserting this expression into (E.12).
g:8
] 982024093-242
40 _
: [i
20 __
L_
o m i , mmmill , , , mi roll
maim
_ 1 10 102
_ Fieure E.2 Variation of F_ with the Dimensionless Meniscus l_i
Distance Pl (takenfnmnreferer_e[25])
D
E:9
1982024093-243
P
AT81D040
U
Uh = _ _ (E. 141
41'oor- t
H
' 2noUy
17
The moment generated about the roller's x-axis is given by
Cage Land Normal Forces and Friction Moment
|\
Forces which develop between a cage rail and its supporting [!
ring surface are obtained using the hydrodynamic solution for
self-acting, short journal bearings [27]. Forces are assumed to I _
act in the plane of rotation, i
The resultant of the pressure distribution on the cage can be
desccibed by orthogonal force components along and perpendicular to I;
a line passing through the cage center and point of closest approach
to the land.
I
Figure E.4 shows th_ relevant parameters for an inner ring land !
riding case, and Figure E.5 illustrates the outer ring land riding
' case.* The cage undergoes a radial displacement in the bearing
#
An xyz frame is at- 1 i XYZ frame, of magnitude e and direction 8c.
i tached to the cage, such that the y axis passes through the point i
of minimum film thickness. J
Assuming an isoviscous, Newtonian fluid, the lubricant forces J
which develop at the guide ring are given as a function of ec- I
centricity [27] i
= noUL 3 2 (E.16)
Y _ ' _')2 i C_ (1-c
!
FIz = _n"ULL3 _c (E.17) i
2 4 (l-C _)_
* These figures were repeated from section 2.4.1. for convenience.
' D
1982024093-245
j'
_
ORb|HA1- pAGE
I AT8 ID040
f A/
, FIGURE E.4 INkiER R=NG-CAGE,_ LAND v_,.uc_'-'eT,._. GEOMETRY
i i
t
E:12
I
-L ;
1982024093-246
FIGURE E.B OUTER RING-CAGE LAND CONTACT GEOMETRY
E:I3
P,
1982024093-247
__ _ ":' L iL I[
ORIGINAL PAGE iS
"{ OF POOR QUALIFY
&T81D040
, _,_
I The upper sign applies to an inner ring and the lower to an ouger !._{
ring land riding cage. Eccentricity is related to translation e, ,_
= e (E. 18) _':_
For level i solutions, the radial force F is set equal to the --
Y
[ cage weight. The eccentricity" e necessary to support that weight
is used to determine F and th_ cage-land friction torcpte . " :
Z ,_
c (E.191
C_ (1_e2_/2
The angle e is used to transform orce8 [tom the local xyz
C
frame to _.he cage reference frame:
McX 1 0 0 c
i't:
Pc 0 cose c -sine c Y (E.201
FEZ) 0 sine c cose c F'
f
t
E:14
{
1982024093-248
[ _
%
\
I 'i
( APPENDIX F
! SKF AND NASA VERSIONS Or SHABERTH -
{ ,
!
:l
[
._
- F:I
1982024093-249
F. 1 I NTRODUCT I ON _'_
This appendix describes the SKF and NASA versions of
- SHABERTH. The primary differences between the two versions en- _);.
_" compass the EHD film thickness and the concentrated contact trac-
"i
tion force calculations The relevant mathematical models are
_o
discussed in Section _'.2. The differences with respect to pro- " _
ii gram input data are explained in Section F.3. :i
F.2 _THERATICAL MODELS
F.2.1 EHD Film Thickness
In calculating the elastohydrodynamic film thickness,
-; SHABERTH/SKF uses the Archard-Cowking equation (3) for point contact
- and the Dowson-Higginson equation (4) for line contact Two film
thickness reduction factors are then multiplicatively applied:
I) a thermal factor due to heating in the _
_: contact inlet using the formulation of
Cheng (II) t_ :
L 2) a factor accounting for starvation at
the contact developed by Chiu (12).
SHABERTH,/NASAuses the film thickness equation developed by
: Loewenthal et al (6) This equation is applicable for both
point and line contacts.
F.2.2 Concentrated Contact Traction t
: The concentrated contact traction model used in SHABERTH/ [,_
F:2
]982024093-250
0
U
SKF accounts for lubricant shear and asperity interaction. _ _
A semi-empirical model developed by Chiu, discussed in (25),
is used to calculate an EHD lubricant shear coefficient. U -
Asperity effects are introduced by determining the portion _ _._i
U
of the contact load carried by the asperities, using the
analysis of Tallian (5), and then calculating the resulting _ i_":/
traction as the product of the normal load carried by the
asperities times the asperity friction coefficient. In _:_,
equation form the traction force is: i. I_ --'_
F = QEHD PEHD + QASP _ASP (1) r :--
i
i"
Q = QEHD + QASP (2)
where
F is the traction force
QEHD is the normal load carried by the EHD film _',
_EHD is the friction coefficient which develops
fr_, lubricant shear [
:, QASP is the normal load carried by the asperities
_j
_ _ASP is the asperity friction coefficient
i Q is the total load [
' SHABERTH/NASA calculates concentrated contact traction across
the EHD film only, using the model developed by Allen, et. al. U
(7). This model determinc_ the traction force by first calcu-
F:3
1982024093-251
ORtGtNAt-pAGE IS
I OF pOOR QUALITY
i'
I stress over the respective contact area. For a Newtonian fluid
the shear stress is given by the equation :
T = 1.1_
where
T is the shear stress
I n is the dynamic viscosity
i ,
v is the surface relative sliding velocity
' h is the film thickness
The lubricant viscosity is assumed to be an exponential func-
tion of pressure of the form:
1
1
where
-!
+ no is the dynamic viscosity at atmospheric pressure
,i ,:+. is the pressure viscosity coefficient ,
;!
-i s is the normal stress
The Allen formulation requires that the shear stress not exceed
a specified fraction of the normal stress such that
+i
+i _ T = n _v if n_ < T
h -
'+ V
'i r -- _'s i fr --- - +fs and n_ > _c
; h
]982024093-252
n
: u
where
r is the critical shear stress for which a value of
c
0.0069 N/mm2 (1000 psi) is normally used. --[[
_e is called the lubricant friction coefficient and
has been determined for specific lubricants. [j
Typical values of t" lie in the range 0.05 .< ]r .< 0.08. [-i
i
F.3 PROGRAbiUSE ,
The selection of the desired SIIABERTH version has been
made possible by the inclusion of two separate Map statements }
for the Univac 1100 computer:
i
1) The original Map statement ! :
@blAP,S AI.WAYS/MAP, ALWAYS/ABS
[ - enables execution of SHABERTH/SKF
: 2) A new Nap statement [
@btAP,S NASA/bIAP, NASA/ABS
enables execution of SllABERTtt/NASA _ti
The only difference between the two versions with respect
[/
to input data is on card type B16. For the NASA version, two [I
additional lubricant data items are specified for NCODE _<0: l]
1) AKN Empirical lubricant constant columns
71-75 [}
i I,
2) FRIC - l.ubricant Friction Coefficient columns I
I,i!
q
F:s i
i ! ,
i ! '
r
III "
1982024093-2S3
Default values ofAKN-S0.0and FRIC-0.07 will be used if these ',_
spaces are left blank. For NCODE values of 1 to 4, the following
I values of AKN and FRIC are assigned.
I NCODE AKN FRIC
1 18.2 0.07S
2 18.2 0.045
3 24.9 0.070
4 18.2 0.070
Y
F:6
i
1982024093-254