21
KONINICL.  NEDERL.  AKADEMIE  V A N  WETENSCHAPPEN   AMSTERDAM Reprinted  from  Proceedings, Series B, 70, N o. 1, 1967 MECHANICS A  STRIP  THEORY  F O R  ROLLING  WITH  SLIP  A N D  SPIN.  I B Y J. J.  KALKERi) (Communicated  by Prof. W. T.  KOITBB  at the meeting of June 25, 1966) SUMMARY I n  recent years, several authors  [3], [4]  have proposed a strip theory t o  deal  with  the transmission of force  during  he  rolling  of two elastic bodies. These theories have been hampered by the fact that  only slip i n  th e  direction  of  rolling  could  be considered. This was due to the  circum stance  that the basic  CARTER-PORITSKY  theory  [ 1 ]  does  not take lateral slip,  or spin  into  account. The present paper supplies an approximative theory  to remove this  restriction. This  theory is  valid  f or all values of the creepage, bu t  only  for a  limited range of the spin parameter. I t is tested b y the  case  of pur e creepage, an d  by the  case  of pure spin. I n  t he  case  of pur e creepage, it is  found  that the  total  tangential force is  cl ose to the tangential force obtaine d  from  other  sources  [6],  when the contact ellipse is narrow in the  direction  of  rolling.  I n the  case  of a circular  contact  area,  the agreement is much worse: rela tive errors of up to 2 5 % occur,  while  i t also  appears  that the influence of Poisson's ratio  is  incorrectly  accounted for by the strip theory. The adhesion  area, however, is i n excellent agreement  with  the experiments  [3]. I n  th e  case  of pure spin the  total  force has a relative error  with  respect t o  the results of  KALKER'S  numerical theory  [ 6 ]  of at most 2 0 % when th e  axial  ratio of the contact ellipse is 0.2  (minor  axis in  rolling  direction), while  in the  case  of a circular contact  area  the error is up to 4 0 %. In both cases  considered, Poisson's ratio is  0.28.  I t also  appears  from  a comparison with  theories of  infinitesimal  spin tha t even for slender ellipses the dependence on Poisson's ra ti o is more complex tha n tha t envisaged by the strip theory. Judging  from  the experi men tal evidence of  POON [10] ( 1  case) , i t  would  appear  that the adhesion  area  is exce llen tly predicted b y  the strip theory in the  case  of a circular contact  area. 1 )  Laboratorium voor  Technische Mechanica, T . H .  Delft,  Mekelweg  2. Report nr .  327.

1967-13 a Strip Theory for Rolling With Slip and Spin

Embed Size (px)

Citation preview

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 1/21

K O N I N I C L .  N E D E R L .  A K A D E M I E  V A N  W E T E N S C H A P P E N  —  A M S T E R D A M

R e p r i n t e d  f r o m  Pr oceed ings, Ser ies B, 70, N o. 1, 1967

M E C H A N I C S

A  S T R I P  T H E O R Y  FO R  R O L L I N G  W I T H  S L I P  A N D  SPIN.  I '

B Y

J. J.  K A L K E R i )

(Communicated  b y Prof. W . T .  K O I T B B  at t he me et in g of June 2 5, 1966)

SUMMARY

I n  recent years, several authors  [ 3 ] , [ 4 ]  have proposed a strip theory

t o  deal  w i t h  t he tra nsmission of force  during  the  r o l l i n g  of two elastic

bodies. These theories have been hampered by the f act tha t  only slip

i n  the  direction  of r o l l i n g  could  be considered. This was due to the  circum

stance  that the basic  C A R T E R - P O R I T S K Y  theory  [ 1 ] does  not take lateral

slip,  or spin  into  account. The present paper supplies an app roxi mat ive

theory  to remove this  restriction.

This  theory is  v a l i d f or all values of the creepage, bu t  only f or a  l i m i t e d

range of the spi n parame ter. I t is tested b y the  case  of pure creepage,

an d  by the  case  of pure spin.

I n  the  case  of pur e creepage, i t is  found  that the  total  tangential force

is   close to the ta nge nti al force obta ined  f r o m  other  sources  [ 6 ] ,  when

the contact ellipse is narrow in the  direction  of  r o l l i n g .  I n the  case  of a

circular  contact  area,  the agreement is much worse: rela tive errors of

up to 25 % occur,  while  i t also  appears  that the influence of Poisson's

ratio  is  incorrectly  accounted for by the strip theory. The adhesion  area,

however, is i n excellent agreement  w i t h  the experiments  [ 3 ] .

I n  the  case of pure spin the  total  force has a rela tive error  w i t h  respect

t o  the results of  K A L K E R ' S  numerical theory  [ 6 ]  of at most 2 0 % when

the  axial  ratio of the contact ellipse is 0.2 (minor  axis in r o l l i n g  direction),

while in the  case of a circula r contact  area  the error is up to 4 0 %. I n both

cases  considered, Poisson's ratio is  0 . 2 8 .  I t also  appears  f r o m  a comparison

w i t h  theories of  infinitesimal  spin tha t even for slender ellipses the

dependence on Poisson's ra ti o is more complex tha n tha t envisaged by

the strip theory. Judging  f r o m  the exp eri men tal evidence of  P O O N [ 1 0 ]

( 1   case), i t  would  appear  that the adhesion  area  is exce llen tly pred ictedb y  the strip theo ry i n the  case  of a circular contact  area.

1 )  L ab o r a to r ium v o o r  Technische Mechanica, T . H .  D e l f t ,  Mekelweg  2. Report

n r .  327.

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 2/21

1 1

T A B L E   OF   CONTENTS

Page

Summary  10

List of Symbols  11

1 .  Introduction  13

2 .  The Theory of Hertz  14

3.  Boundary Conditions of Rolling  16

4.  Solution of the Elasticity Problem under Three Simplifying Assumptions  17

4 . 1 .  T h e  Connect ion  between Surface  T r a c t i o n a n d  D i sp la c e me nt  in a

Ha lf spa c e .  T h e  Cross-Dependence  o f N o r m a l a n d  Tangen t ia l Tract i ons  18

4.2. The  S t r i p t he ory  A s s u m p t i o n 2 1

4.3. The  Carter  A s s u m p t i o n 2 3

5.  The Solution of the Problem in Strips with an Adhesion Area  25

5 . 1 .  T h e  Basic  S o l u t i o n i n  S t r ips  w i t h  a n  Adhesion  Ar ea 25

5 .2. The Fr ic t i on La w in the Are a o f S l ip 29

5 .3. The Fr i c t io n L a w in the Are a o f  Adhesion  32

5 .4. The So lu t ion  u n d e r  a Res t r ic t ion 34

6.  The Solution in Strips without Adhesion Area  37

7.  The Total Force Exerted on the Lower Body  41

7 . 1 .  T h e P r i m i t i v e s o f  Fx  a n d  F

y  41

7.2, The  In t e r se c t i on  o f the   S e pa ra t r i x  w i t h  t he  L e a d i ng  Ed ge  of t he

Cont a c t  Ar ea 43

8.  Two Special Oases  44

8 . 1 .  T h e  Case  o f  Creepage  w i t h o u t S p i n  (y = 0)  44

8 .1 1 . T h e D i v i s i o n o f t h e  Cont a c t  Ar ea 45

8 .1 2. T h e T r a c t i o n D i s t r i b u t i o n 4 6

8.13. Th e Sli p 47

8.14. The Angle  between  S l ip and Tr ac t io n 47

8 .15 . The T ot a l  Force  T r a n s m i t t e d  b y t h e  U p p e r  Bo dy to the Lo we r 49

8.2. The  Case  o f  P u r e  Spi n 52

8 .21 . The Di v i s i on of the  Cont a c t  Ar ea 56

8 .2 2. T h e T r a c t i o n D i s t r i b u t i o n 5 7

8.23. Th e Sli p 57

8.24. The Angle  between  S l ip and Tra c t i on 58

8 .25 . The T ot a l  Force  T r a n s m i t t e d  b y t h e  U p p e r  Bo dy to the Lo we r 59

8.26. The  Case  | y | > 1   6 0

9.  Conclusion  61

References  62

L I S T   OF S YMBOLS

B o l d  face  symbols  designate  vectors.

A  supe rsc r i p t  +  indicates tha t  t h e q u a n t i t y  belongs  t o the low er  body.

A  superscr ipt  —  indicates that  t h e q u a n t i t y   belongs  t o the  up p er  body.

Reference

S y ^ 0 1 M e a m n g

  equat ion

A  dimensionless  constant  connected  w i t h

n o r m a l  pressure.

(2.7)

Semiaxis  o f  contact ellipse  i n r o l l i n g  (x)

di rec t ion . (2 .1 )

Semiaxis  o f  contact ellipse  i n  la t e ra l  (y)

direct ion (2.1)

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 3/21

12

Symbol MeaningReference

equat ion

c

E

E'E

Fa

Fx,Fy

67, G+, 67¬

9

h

h'

k

I

M

N

P{x,  y)

Rx>  B

v<B

x 'R

v

s

u+,  v+, u>+,

u~,  v~, u>-

u(x,  y), v(x, y), w(x, y)

V

V+, v-

w

(WX,  Wy)

X , Y

X ' ,  Y'

X ,  Y

x,  y, z

H a l f w i d t h  of contact  in terva l .

Contact  area,  —  in terva l .

Contact  area  in st r ip approxi mat i on.

Complete  ell iptic  int egr al of the second

k i n d  (no  vector).

Slip  area,  —  in terva l .

L o ck ed  area,  —  in terva l .

E x c e n t r i c i t y  of cont act elli pse.

U n i t  vectors.

T o t a l  tan gent ial force per  u n i t  l ength in

indirect ion.

M o dul us  of r i g i d i t y :  combined, lower body,

upper body.

R a t i o  o f the  axes  of the co nta ct ellipse,

H a l f w i d t h  of adhesion  in terva l .

Dist ance of center of  Eh  to leading  edge

o f  E.

Distance of t r a i l i n g  edge  of E  to separatrix.

M a j o r  semiaxis of cont act e llipse.

A  large length.

N o r m a l  force.

Dis s ipa t io n  pe r  u n i t  t ime per  u n i t  area.

Coordinate  yjb  of bou nda ry between stri ps

w i t h  Eh  and st r ips  w i t h o u t  Eh.

R a d i i  of curva tur e of lower an d upper bod y.

M i n o r  semiaxi s of cont act ell ipse.

R e l a t i v e  sl ip (of upper bod y over  lower).

T o t a l  tangential force.

(x,  y, z)  compone nts of elastic disp lace ment.

Displacement differences.

Rolling velocity.

V el o c i t y  o f a par tic le of the bodies.

R i g i d  v e l o c i t y  of the bodies.

A   f o r m  i n v o l v i n g  creepage  an d spin p ara

meters  a n d  y.

U n i t  vector i n the di rect ion of the  slip.

(x,  y)  components of tangen t ia l t ract ion

ac t ing  on lower body.

The par t of tangent ia l t ract ion act ing on

whole  E.

The par t of tangent ia l t ract ion act ing on  Eh.

Cartesian coordinate system.

(4.8)

(2.1)

(4.9)

(2.7),

Table 1

Sec. 3.

Al so :  (5.42)

Sec. 3.

Al so :  (5.42)

(2.2)

(8.16b)

(8.21b)

(5.7)

(2.6)

(2-2),

Table 1

Sec. 5 1,

E i g .  1, (5.40)

Sec. 5.1

F i g .  1, (5.40)

(5.41)

Sec. 2

(4.9)

(2.7)

(6.10),

(6.11) ,  (6.16)

(7.10),

(7.12) ,  (7.14)

Sec. 2

Sec. 2

(3.3), (3.4),

(5.2a)

Sec. 3

(4.2)

(3.1)

(3.2),  (3.3)

(3.1),  (3.2)

(5.19a)

(3.5)

(5.4)

(5.5)

Sec. 2

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 4/21

1 3

Symbol M eaningReference

equation

x' ^-coordinate  measured  f r o m  center  o f  En- (5.5)

Z N o r m a l  pressure  acting on the lower body. (2.7)

Dimensionless  constants  connected  w i t h

X',  Y'. (5.4)

<*', p Dimensionless  constants  connected  w i t h

X ,  Y . (5.5)

Ö Angl e between slip an d tan gent ial tr act ion. (5.24), (5.26)

e Angle  between slip an d  a;-axis;  also, in

8.26, a small posit ive q ua nt it y. (8.14)

•Q Angle  between tangenti al t ract i on and

a-axis. (8.14)

V L a t e r a l  creepage  p arameter . (5.13)

n =  (l-a)n

e Angle of ap ex  of adhesion  area  i n p ure  spin. (8.33)

A   combined elas t ici ty  constant. (4.3), Ta bl e 2

h(x') A   posi t ive fu nct ion . (5.19b)

h{x) A  posi t ive fun cti on. (5.20)

ft Coefficient of  f r i c t i o n  ( take n consta nt). (3.5)

1 L o n g i t u d i n a l  creepage  p arameter . (5.13)

1 E ffec t iv e l o ng i tud ina l  creepage  p arameter1i n   a str ip. (5.13)

Q Characteristi c le ngt h of the bodies. (2.3)

•y,  cr+, o~ Poisson's  rati o. Combin ed, lower body,

up p er  body. (2.6)

X Angle  of Her tz . (2.5), Ta bl e 1

I n   8 . 1 :  p a r a m e t e r  t o  describe  total force. (8.2)

Vx L o n g i t u d i n a l  creepage. (3.2)

Vx E ffec t iv e  l o n g i t u d i n a l  creepage  in a str ip . (5.1)

Vy L a t e r a l  creepage. (3.2)

<P Spin. (3-2)

X

A  p a r a m e t e r  connected  w i t h  p ure  spin. (8.43)

V S pin  p arameter . (5.13)

1 .  I N T R O D U C T I O N

Consider  two  bodies  of revolution which are  pressed  together  w i t h  a

force N , and whi ch  r o l l  over  each  other.  Owing to the normal force  N,

a  contact  area  E  is form ed in which the  bodies  touch. I f the conformi ty

o f  the  bodies  is not too  strong,  and  changes  of  curvature  are small, the

contact  area  and the normal  pressure  t ransmit ted  across  i t are given by

the Hertz  theory, according  to which  E  is an ellipse, and the norm al

pressure  distribution is half ellipsoidal. The  results  of the H ert z  theory

are given in  Section  2. I n additi on to the norma l  pressure,  a  tangential

force can be  t ransmit ted  from one  body  to the  other,  owing to friction.

Wh en the velocities of the  bodies  at the  contact  area  are  almost equal,

slip  w i l l  occur  in part  of the  contact  area,  called the  area  of slip  Eg,  while

i n   the  other part,  the locked  area  Eh,  there  is no velocity of one  body

w i t h  respect  to the  other.  This is a  consequence  of the fact  that  the  elastic

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 5/21

1 4

deformation  modifies the  local velocity  pattern  near  the contact  area.

I n  the  area  of  slip, work  is done by the  f r i c t i o n  forces; macroscopically

this  results i n a difference of the  overall  circumfer ent ial velocities of

the bodies. Thi s differ ence is described by the concepts  creepage  and  spin.

A  rigorous  definition  of  these  concepts  w i l l  be  found  i n Section 3 .

The  f i r s t  study of  r o l l i n g  w i t h  creepage  was made by  CA RT ES , [ 1 ] , i n

1 9 2 6 .  He considered the  case  th at the contact  area  is a str ip (as it is

i n  two cylinders  w i t h  parallel axes), and that the bodies  r o l l  over each

other  w i t h  creepage  in the direction of  r o l l i n g .  I n recent years, Carter's

results have been applied by  H A I N E S  an d  OL LE R T ON [ 3 ] , an d also by

J O HN SO N [ 4 ] , to calculate the more general  case  of an  elliptical  area  of

contact. The process,  which  is due to Haines, is as  follows.  The contact

area  is  split  up in very narrow strips  w i t h  the  long  side par all el to the

direction  of  r o l l i n g .  For each  strip,  the  stress  distribution  is calculated

according  to Carter;  these stress  distr ibut ions, pu t beside each other,

give  a n app roxima tion of the t rue  stress  distribution.  T his appr oximat ion

is  good for slender contact ellipses  w i t h  the narrow side in the direction

of  r o l l i n g ,  while  for non-slender ellipses i t gives a good qual it at ive insighti n  the real  stress  distribution.

This  process  was applied by  H A I N E S  and  O L LE R TO N [ 3 ] to  creepage  i n

the direction of  r o l l i n g ,  in an  elliptic  contact  area.  Then, Carter's theory

can be used without modification. Carter's tre atm ent of the elastic problem,

however, is restricted to displacements in the  r o l l i n g  direction  only,

an d  hence  cannot be used when there is lateral creepage  or spin. I t is the

object  of the present  paper,  to give a trea tme nt of the elastic proble m

without  this defect. As a  consequence,  we are able to give strip solutions

applicable  to  longitudinal  creepage,  lateral  creepage,  and spin. I n our

considerations of ela sticity, the bodies are app roximated by halfspaces.

2 .  T H E  T H E O R Y  OE  HE R T Z

When  the bodies are  pressed  together, a contact  area  E  is form ed

between them. This contact  area,  and the normal pressure  Z  carried by

i t  are the object of study i n the t heory of He rt z (see, for instance,  L O V E  [ 8 ]

p .  1 9 2  sqq).

I n  thi s theory, the bovmdary conditions are set up for Moti onless,

counterformal  f i n i t e  bodies. Then, in order to  s i m p l i f y  the solution as

much as possible, the geome try of the bodies is simplified by approxi matin g

the bodies by elastic halfspaces,  which  in a cartesian system  (x, y, z)  are

given  by z > 0 and z < 0 . The boundary conditions are retained  w i t h  only

slight modification,  the most important of  which  is, th at t he elastic

displacement  u  and the  stresses  become unnoticeable far  f r o m  the contact

area.

I t  is the n  found  that the contact  area  is  elliptic  i n  f o r m .  We adopt

a coordinate system  (x, y, z)  i n  which  the  o r i g i n  lies i n the centre of the

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 6/21

15

contact  area,  and in which  the  axes  of the ellipse are the  axes  of x  an d  y.

The axis of z  coincides  w i t h  the inner nor ma l of the lower body (halfspace

2> 0) at the centre of the contact ellipse. So, the contact  area E  is given by

(2.1)  E={x, y, z:  2=0, (a;/a)2 +  ( y / 6 ) 2 < l } .

The ratio of the  axes  of the ellipse is complete ly deter mined by thegeometry of the mid efor med bodies,  a  an d  b  are the semiaxes i n  x  and  y

direction;  apart  f r o m  that we  write  I fo r the ma jo r semiaxis of the ellipse,

an d  s  for the min or semiaxis; fur ther , we set

(2.2)  g-sjl, e =  ]/l-g*.

We  assume  that the planes of principal  curvature almost coincide. This is

the  case  when two bodies of  revolution  r o l l  steadily over each other:

indeed, the parallel circles of the bodies lie then almost in parallel planes.

Under  this assumption, the  principal  directions of the ellipse lie almost

i n   the planes of principal curvat ure. I t should be note d that the assumption

is n ot necessary fo r the  v a l i d i t y of the He rt z theory (see  L O V E  [ 8 ] ,  p. 194),bu t  the analysis is somewhat simpler th an i n the general  case.  We  write

R£,  Rf  for the  principal  radii  of cmvat ur e of the lower body  which

correspond to the plane  y = 0, x = 0  respectively. We take  R+, R+  positive

when  the corresponding centre of curvat ure lies i n the lower halfspace

z>0.  I n the  same  manner,  R~  an d  R~  are denned as the  principal  radii

o f  cu rvatu re of the upper body; they are take n positive, when the corres

ponding  centre o f curvatu re lies i n the halfspace z <0. We define the

characteristic length  q  of the bodies as  follows:

(2.3)  l /e  = | ( W + W + W +  l / i 2 v - ) .

I n   the  case  of a  sphere  on a plane,  q  in the diameter of the  sphere. Now,we  have:

(  I f  (1/R+ + 1/R-)>(1IR+  + 1/R-),  then  l =  b>a=s,

(2.4)  I( I f  {1IR+ + IIR-)<{1IR+  + 1IR-),  then s =  & < a = Z.

The ra tio of the  axes g is tabulate d by He rtz  w i t h the aid of Hertz's angle  T ,

which  is defined as

(2.5)  cos r =  fc[(l/i?+  +  l / ^ - ) - ( l / i ? +  +  l / i 2 - ) | .

He rtz 's table is reproduced i n table 1.

The  pressure  distribution,  and the length of the semiaxes depend onthe elastic constants of both bodies, combin ed i n special manner . L et G+,

G-  be the modulus of  r i g i d i t y  of lower, upper body, and let  a+, a be

Poisson's  ra ti o of lower, upper body. Then the resul ts can be described

i n   terms of a combined modulus of  r i g i d i t y  G  and a combined Poisson

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 7/21

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 8/21

17

unstressed  state  so th at t he elastic deformat ion dies out fa r  f r o m  the

contact  area.  We denote by  V+,  V ~  the real velocity  of lower, upper body.

The   relative slip  is  defined  as the  velocity  of the  upper body with respect

to   the  loiuer, divided  by the  rolling velocity:

( 3 . 3 )  s(sx,sy)  =  (l/V)(V--V+).

According  to  K A L K E R  [ 6 ] ,  Sec. 2, the re lat ive slip is, under condi tions

o f  steady  r o l l i n g ,

..  b(u+ —  u~)  b(v+—v~)( 3 . 4 )   S x  = vx-cpy  +  -

5

—^ - %=-%+p g; +  -  v

" —

where  (u+, v+)  are the  (x,  ^-compone nts of the elastic disjDlacement of the

surface of the lower body,  while  [vr, v~)  are the corresponding quantities

of  the upper body. D enoti ng the slip  area  by  Eg,  the locked  area  by  En,

the tange ntia l tr act ion exerted on the lower body by  ( X ,  Y),  the boundary

conditions  are, according to  K A L K E R  [ 6 ] ,  Sec. 2 :

(X ,  Y) = (ivx, wy) ixZ,

/u :  coefficient of  f r i c t i o n ,  ta ken constant, \ i n  Eg,

( 3 . 5 )x  =  SxjS Wy =  SyJS S = I  SX, Sy)  | = + S^

s x  = sy  = 0, \{X,  Y)\<pZ  in  Eh

X=Y=0  on  2 = 0, outside  E,

stresses  and displacements vanish at  i n f i n i t y .

4 .  S O L U T I O N  OE TH E  E L A S T I C I T Y   P R O B L E M U NDE R  T HR E E

S I M P L I F Y I N G -  ASSUMPT IONS

I n  the  present  Section we  w i l l  solve the elast icit y problem under three

simplifying  assumptions.  Firstly,  i n Sec.  4 . 1 ,  we  w i l l  give the connection

between the surface tractions and the elastic displacements. A  plausibility

argument  w i l l  show us to do away  w i t h  the observed  cross dependence

of  t he problem of nor ma l tr act ion and the tangent ial tractions  present

i n  the contact  area. This is done  w i t h  greater confidence, since the  f o l l o w i n g

assumptions are of a mu ch more sweeping nat ure.

Secondly, i n Sec.  4 . 2 ,  we  assume  that the traction distribution changes

very  slowly  w i t h  the lateral coordinate  y,  so that the slip can be calculated

as i f thi s  change  does  not exist.

T h i r d l y ,  i n sec.  4 . 3 ,  we  assume,  led by the example of Carter's solution,

that the traction distribution  for f i x e d lat eral coordinate  y  is the difference

of  two halfelliptical tr ac ti on dist ribu tion s, one acting on the whole contact

w i d t h  and the other on the adhesion  interval,  and the displacement

functions  needed  are calculated  e x p l i c i t l y .

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 9/21

1 8

4 . 1 .  The  Connection Betwee n Surface T raction  and  Displacement  in a

Halfspace.  The  G ross D ependence   of  Normal  and  Tangential Tractions

According  to  L O V E  [ 8 ] , p. 19 1 and 24 3, the connection between the

elastic displacement  which  vanishes  at  i n f i n i t y ,  and the surface tractions

acting in the contact  area  E,  are  given  by

v?{x,  y, T \z\) =  J L ƒ ƒ   ̂ x{x', y')  •

\r  \z\ + r  +  t* r{\z\ \-r}* }  +

( 4 . 1 )

(x -x ){y-y ) ( l - 2 c f f ) ( x - x ) ( y - y )

r3 r{\z\+r)z +

( s - s ' ) | z |  ( l - 2 g T ) ( 3 - a . ' )

KN+r+

• =  ] / (» —  . r ' )2 +  (?/ — ?/')2 +  |z [

2

.  Upper sign for upper halfspace,

lower  sign for  lower  halfspace.

F o r  the  solution  of the boundary value problem, we need  only  the

displa cement at the boundary of the bodies, th at is, at 2 =  0 .  Furthermore,

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 10/21

1 9

w e a r e o n l y   interested  i n

u(x , y) = u+(x , y ,  0 ) -  u~(x , y ,  0 ) , f o r t h e  problem  o f  tangential

traction,

v(x, y) = v+{x, y,  0 ) -  v (x, y,  0 ) , f o r t h e  problem  o f  tangential

( 4 . 2 ) traction,

w(x, y) = iv+{x, y, Q)-w-{x, y,  0 ) , f o r t h e  problem  o f  normal

traction.

Here  w e  have included  t h e  problem  o f normal  traction,  i n order  t o  motivate

our use of  Hertz's  solution  o f  that  problem.  W r i t i n g

1 / l 1 \  a , ( a+ a ~\ n ( l ~ 2 a + 1~ 2a~

(4 .3 ) ê = i ( ^ + ^ J > ë = H ^   +  ^ r

w e  calculate  u(x, y) , v(x, y)  a n d  w{x, y) ,  an d f inal ly  find

u(x , y) = u+{x , y, 0)-u~{x, y,  0 ) =

(4 . 4 )

+ Y(x ,^ g ( * - * ' ) ( y - y ' ) _   w ^ « ( « - « O

i 2 3 i J 2

©(a;, ?/) s «+ (» ,  i / , 0 ) - « - ( » ,  y,   0 ) =

Ü. (a;  , y ) rR*

/ \   1  f f  [r<r> i / \H ^ X , )  I F ( x '  V

' ) X ^ ~ ^  I>(x,y)=^Q  li  — — + J , i * > y j ^ 2  -t -

+ Z(x , y )  i ^ ] cfo ' i ? = l /O ^ ' - ^ T +  ^ - y 7 ) 5 -

We see  that  u , v  a n d  w  near  t h e  contact  area  a re  composed  o f three  terms,

of   t h e  order  o f  magnitude  o f

(4 .5)

u   =  0( TxjnGl) + 0{oT y\nGl)  +  0 ( K J V > 6 7 7 ) ,

v  =  0(aT x\nGl) + 0(T y\nGl) + OixN jnGl),

\w   =  0[nT x\nGl) + O^Ty/nOl)  +  0({l-a)NjnGl),

T y , T y , N:  t o t a l  force  i n t h e ( * ,  y,   z)-directions.

W e s e e f r o m ( 4 . 3 )  that  % vanishes  when  t h e  elastic  constants  o f t h e

bodies  a re  equal:  a+  = a~ = a,   G+  = G ~ = G ,  o r  when  both  bodies  a re i n

compressible:  a+  = o~ =  i -  I n  these  cases,  u   a n d  v  a r e n o t  influenced  b y   Z

n o r i s   w  influenced  b y   X   a n d  Y.   This  implies  that  t h e  Hertz  theory  c a n

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 11/21

2 0

be used for the det erm ina tion of the contact  area  E  and the normal

pressure Z,  while Z  does not  directly influence the relative slip  (see eq.  ( 3 . 4 ) ) .

I n  table 2, we give the value of n for a few combinations of materials for

which  i t  does  not vanish. We see  f r o m  this table that  n  is always quite

small.  Thi s jus tif ies us to neglect it s influence altogether, an d we are the

more  justified  i n doing thi s, since we  w i l l  make assumptions of a much

more sweeping na tur e i n Sections 4. 2 an d  4 . 3 .  I n a more precise theor y,

one  would,  i t  seems,  always be  justified  i n neglecting the influence of

X  an d  Y  on  w.

T A B L E  2

T h e  value of  x  f o r  different  combinations of materials . T he fi rs t- menti oned is t he

ma t e r i a l  of the upp er bod y  ( — ) ,  the second-menti oned is the ma te ri al of the  lower

b o d y  ( + ) .

Mater ial Steel/Brass S tee l / Al um inum Steel/Glass

+  0.03 +  0.06 +  0.13

Mater ial B r as s / Al um in um Brass/Glass Al um inum / G l as s

X +  0.02 +  0.10 +  0.03

R e m a r k . = 0 whe n  (a+,  67+)=  (o",  G~),  or when  <x+=o—=Jj

; 0  wh e n o- *= J, 67* > 67±.

Indeed,  T x  an d  T y  are of th e order  /uN,  where  p  is the coefficient o

i f r i c t i o n   which  is  ordinarily  much smaller than  unity.  So the influence o

F x  an d  Fy  on  tu is of order  O^inNjnGl),  which would  seem to be negligible

w i t h  respect t o the influen ce of Z  which  is 0((l-a)NjnGl).  Bu t the direct

influence  of the normal pressure  distribution  on the tangential displace

ment differences u  and v  should be retai ned i n a more precise t heo ry; i t is

of  the order  0(xNjnGl),  while  the influence of the tangential traction

[X ,  Y)  is of the order  0{/.iNjnGl).  So the relati ve influence of the te rm

w i t h  H is larger by a fa ctor  0 ( [ 1  - o - j / y t 2 ) ,  as compared  w i t h  the problem

of  the nor mal force.

A s  we said before , we  w i l l  here  neglect the terms  w i t h  %. This implies

tha t the H er tz theo ry can be used for the de ter min ati on of the contact

area  and the normal pressure,  while  the normal pressure  does not  directly

influence  the relative slip.  The hori zont al displacement differences u  an d  v

are  hence  given by

1  »» P „ , . .. f  l - o -  a{x-x'fu(x,  y) =

7lG E  LX(x ,  y')

R

+  y')o{x-x')(y-y')'

i?3

i?3

dx dy',

+

( 4 . 6 )

v(x,  y) X ( x ,  y>)  «*-<o<*-jn  +

+  Y ( x W ) [  ̂ +

  a - ^ } ] d x ' d y ' ,

R  =  y-fiZ-x')z  +  (y-y')z.

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 12/21

21

4.2.  The  Strip Theory Assumption

We  w i l l  now  assume  that the traction distribution  changes  very  slowly

i n   the lateral  (y)  direction.  One  would  judge that the results  based  on

this hypothesis are  best  v a l i d for ellipses which are narrow in the direction

o f  r o l l i n g .  This is indeed  partially  borne out by the results.

On t he basis of thi s assumpt ion we calculate the displacement differences

u  and  v  in the point  {x, y),  as i f the tan genti al tra ctio n  distribution

X(x',  y'), Y(x', y')  were given by

(4.7)  X(x',  y')=X(x',  y), Y(x', y') = Y(x', y),

where we let  y'  run in principle  f o r m  - c o to +oo . Le t the  w i d t h  of the

contact  area  at the  line  yr—y  be given by  c,

(4.8)  c =  a l / l - ( # ) 2 ,

then  the new contact  area  E'  is given by

(4.9)  E' = {x',y': -c<x'<c, -M<y'-y<M,  M  -> oo}.

We  do not set  M = oo,  since this  would  lead to  i n f i n i t e  u  and  v.  The large

term  vanishes, however, when we differentiate u  and  v  w i t h  respect to  x,

so that after the  differentiation  is performed, we can let  M  -> oo.

We  f i n d  the  f o l l o w i n g  expressions for  u  and  v:

(4

'I0

>  * v) - ± Uw, v) & V+

1 f =  }/(x-x'y+y'z.

The integrals  w i t h  — as in tegra nd vanish, since the integrands

are odd i n  y'.  Furthermore, we can  write  y'*=fz

—{x-x')z

.  So we  find

{ f =  ]/{x-x')z

+y'z

.

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 13/21

5 4

spin

direction of traction

y=-qb

direction of sli p, y<0

x, rolling

direction of slip y>0

y = qb

direction of traction

(a)

spin

direction of traction

direction of sl ip, y=0

x, rolling (b)

direction of slip , y >0

direction of traction

F i g .  7 . Tr act ion an d s l ip for  p u r e  s p in ,  a/b = 1, a —  0.50.  (a): ip  = 0.305.

( 6 ) :  yi =  0.800. Open arrows:  slip.  Closed arrows: traction.

Negative diss ipat ion  areas shown  shaded. T rac t i on and s l ip are  m i r r o r  an t i - s y m m et r ic

a bout  the  a-axis.  T he  scale  of t he ar row s i n (&) is  h a l f  t he  scale  of the a rrows in  (a).

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 14/21

5 5

F i g .  8. Tr ac ti on an d slip for pur e spin ,  a/b  = 0.2,  a =  0.50.

(a)  y> = 0.305,  (b) if  = 0.800. Op en ar row s:  slip.

Closed ar rows: t rac t io n. Nega tive diss ipa t ion  area  shown  shaded.  T r ac t i o n and

slip  ar e  m i r r o r  ant i- symmet r ic about the  s-axis.  T he  scale  of the arrows i n the  same

i n   (a)  as in (b).

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 15/21

56

itiN

05

03

0-2

0-1

numerica

o numerica

values £ =b

values £ =

D

o-=0-28

>2 a=0-28

©

/

/o

of/

S~—5— •g=0-2/o

of/°  o

o o ^ So

7

0 0-2 0« 0-6 0-8 10

Fig.  9 . The tot al force for pure  spin;  a  =  0 . 2 8 .  Also shown are nume rical  values

o f  the force  from  K A I K B B  [ 6 ] .

I t  is seen  f r o m  f i g .  9 that in the  case a/5 =  0 . 2 ,  cr =  0.28  the strip theory

gives  f a i r l y  good results for values o f  \ip\ up to 0.9 (err or: up to 20  ) .

I n   the  case  of a circular contact  area  (ajb=  I,  c r = 0 . 2 8 )  of fig. 9 the error

is  considerably larger (up to 40 % for  \y>\<0.8).  The fac t th at the soluti on

gets  worse when  \y>\  increases  is borne out i n th at the character of the

curve of the strip solution for large spin is not the  same  as that of the

numerical  theory.

From a comparison of fig. 9 w i t h  fig. 6 (the to ta lf or ce i n case of vani shing

spin),  it is  seen  t ha t especially for ellipses nar row in the dire ction of

r o l l i n g the str ip theory is much better for pure  creepage than for pure spin.

8 . 2 1 .  The  Division  of the  Contact Area

As  always i n the  case  when  \y\< I,  the  area  of adhesion  J$n  borders

on  the leading  edge  of the contact  area.  The separatrix of Eg  an d  E% is

most easily constructed by  means  of the distance  h  f r o m  the  trailing

edge  of  E  to the separatrix:

(8.26) (8.32)  h =  2 \ f y \ j { l - a ) ] / Y ^ f .

The division of the contact  area into  an  area of slip  and an  area of adhesion

is  shown i n fig. 7a for the circle  case  a/b=l,  Poisson's  ratio  cr =  0 . 5 0 ,

and the spin  y> =  0 . 3 0 5 . This is the  case treated by  P O O N  [ 1 0 ] .  The  division

o f  the contact  area  is shown i n fi g. 7b for a/6 =  l ,  o - = 0 . 5 0 ,  f = 0.8,  i n

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 16/21

5 7

fig. 8a for a/6 = 0.2,  rr = 0.50, y =  0.305,  and in  f i g . 8b for a/b = 0.2,  o- = 0.50,

y> =  0.8.

The  angle  9 of the  apex  of the  area  of  adhesion  is

O I „ . l

(8.33) 9 = 2 cot"1

( l - o - )  Vi-y>2

i n  the  case of fig. 7a, 9 =  7 6 ° . Thi s is precisely the value fou nd by  P O O N  [10].

The  angle  9 diminishes monoton ically as  tp increases,  an d  finally  vanishes

f o r  tp — l.  I t should be noted tha t neither  h  nor 9  depends  on the axi al

ratio  g  of the  contact  ellipse.

8.22.  The  Traction Distr ibution.

The traction distribution is given in (5.44)  w i t h  the  parameters  a, ft,

as in (8.28). We  find:

X  = (1/a)  f.iAG   ] / l - y2

  sign ( -  y>y)  {]/c 2-x 2  +

- Yhz-x *},  x' = x-c + h,  in  Eh

;

=  / . i Z ] / l - f 2

 sign  {-tpy)  | # I < # , i n  Eg',

(8.34) < =  fiZ  sign ( -  y>y)  \y\ > qb,

T ^ f i Z f \y\<qb,

=  o  \y\>qb,

Z  =  AG]/l-{xla)2-(ylbf  = ^t ? (c/a)  H ' l - ( * / « ) » .

We see tha t i n the locked  area,  Y  is a  constant  m u l t i p l i e d  by  fxZ,  the

constant  being the spin  parameter  y>  (see fig. 2c). X as always has the  same

character  as i n the  case  of vanishing spin. We see  f r o m ( 8 . 3 4 )  tha t the

traction is mirror antisymmetric  about  the  £c-axis  (i.e., the  .^-componentis odd i n  y,  the  ^/-component  is  even  i n  y).  I n  figs.  7 and 8, the t ra cti on

has  been sketched  for the  values  of the  parameters  which prevail  there,

f o r  \y\<bq.  I n strips wit hout locked  area,  \y\>bq,  the  ^-component  of

the traction  vanishes  entirely.

8 . 2 3 .  The  Slip

The slip is given by  ( 8 . 2 9 ) :

(S

X ,  Sy) =

= [2/ tA(l  -  ff)l/T^2  ( l /o) Vz *-Jfi  sign  ( -  rpy), 0) in Eg,  \y\<qb,

=  ( f ^ é [ - w - | - (1  a)x  sign  ( f y ) ] ,  ipx) ,  \y\>qb.\  a a /

We see tha t i n strips  w i t h  locked  area  (\y\<qb),  the slip is paralle l to

the .r-axis, while in strips without locked  area  (\y\>qb),  i t  does no t  have

a  f i x e d  direction. The slip is mirror antisymmetric  about  the  a;-axis,

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 17/21

58

as is the traction. The  slip  has been sketched i n  figs.  7 and 8 for the stri ps

without  locked  area  \y\>qb.  I t is  seen  f r o m  fig. 7b, that large slips occur

close to the adhesion  area  whe n |y| = 0.8. This underlines once more the

ba d  quality  of the solu tion fo r large values of  \%p\.

8.24.  The  Angle between  $%> and  Traction

We  first  consider the angle between  slip  and traction in the  area  of

slip  of the strips  w i t h  adhesion  area  (\y\<qb).  This is act ual ly quite  easy

to  do, since the angle between  slip  and  r o l l i n g  direction (the positive

.T-axis)  vanishes when  fy<0,  and is 180° when  ipy>0.  I t can be  found

f r o m   (8.34) that when  \y\<qb  the tra cti on in the  slip  area  makes an

angle f  w i t h  the positive rc-axis,  which  is given by

(8.36)t  = sin"

1

  sign  (-ipy)]+  [1+sign  (tpy)]  •  90°

— 90°  < sin-1  x  < 90°, sin"

1

  i n  degrees,

while  the angle of the  slip  w i t h  the positive  a;-axis  is given by

(8.37)  e =  [1+sign  {yy)] •  90°.

The angle <5 bet ween  slip  and traction  becomes  then

(8.38)  8 = e-£ =  sin-i  sign  (y)].

We  see  f r o m  this that  \d\<  9 0 ° ,  f r o m  which  i t  follows  that the dissipation

is  always pos it ive . |<5 increases  f r o m  0° when  yj = 0  to 90° when  | f | = l .

Int erme diat e values are given i n table 4. I n the  case  of  figs.  7a and 8a,

|<5 =  17.8°  (y = 0.305) i n  \y\<qb.  I t should be note d th at the angles e, <5

and f are independent of the  axial  ratio  ajb  of the contact ellipse.

T A B L E  4

The angle  \d\  be tween  slip  a n d  t r a c t i o n  i n  Eg,  \y\<qb, £=ri  =  Q.

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

|«5|  = 5.7° 1 1 . 6 ° 1 7 . 4 ° 2 3 . 5 ° 3 0 ° 3 7 ° 4 4 ° 5 3 ° 6 4 ° 9 0 °

I n  the strips  \y\>qb,  the angle between tr act ion and  slip  varies. The

traction  is always paral le l to the a-axis. We see  f r o m  figs.  7 and 8 that

the angle  becomes  smaller as  \y\  increases.  Analytically  this can be  seen

f r o m   the fact that on lines  w i t h  x = constant , the  «/-component  of the

slip  is constant,  while  the term  —ipy  of the ^-component  increases  i n

absolute value.  Also,  as we approach the ti ps of the ellipse, the ma xi mu m

absolute value c of  \x\  decreases,  which  also has the effect of  improving

the angle.

I t  can be shown that in all  cases  0<y><  1  there is an  area  of negative

dissipation  present i n the region  \y\>qb  which  is shown shaded i n  figs.  7

and 8. I t is dete rmin ed by the  condition  (see  (8.31)):

( 8 . 3 9 )  P(x,y){ I vyl  — ( i — o r ) <  o

\w\

1-0<  X < c.

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 18/21

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 19/21

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 20/21

6 1

that  w i t h  increasing spin the  area  of adhesion  Eu  becomes  smaller and

smaller, and that i t vanishes  in the  l i m i t i n g  case  \y\ =  1 . I t can be a rgued

that it would  seem that i n the  case \y\ >  1 no  area  of adhesion form s at a ll :

according to  ( 8 . 2 6 ) ,  k  is then purely imaginary. This  would  lead to a

completely vanishing total force. However, we see  from fig. 9 th at according

t o  the num erica l theory, a nd also according to  J O H N S O N ' S  experiments  [ 5 ] ,the tangential force does not vani sh: it vanishes only i n the  l i m i t  \ip\  -^oo.

So there is  l i t t l e  poin t in applying the st rip theory to the  case  of pure

spin,  when  | t y > | > l .

W e  saw here  tha t for pure spin the  area of adhesion  vanishes  as  \ip\  — >• 1.

This is no t always so. Ta ke, fo r instance, the  case  ip—l  e, e | 0, f = 0,

?]<0.  Th en we have, according to  (5.41),

2a{Yrj* +  { l - y j2 ) i 2

 + fjy>} -> a\rj\

( 1  — a )2a\7]\  Y

W e  see tha t when ?/ =  0  and  | ? ? | <2 ,  k<2a = 2c\y= 0,  so  t ha t ther e is an

adhesion  area.  Such an adhesion  area  is sketched i n fig. 1 0, fo r  a/b = l,

o- =  0 . 2 9 ,  ip=l, f)=-rQ. ,  1 = 0 .

x, rolling

F i g .  10. T he  d i v i s i o n  of the con ta ct area fo r

a/ b  = l , o - = 0. 29,  y>  = 1, t j =  —  0.5, f  •=  0 .

9 .  CONCLUSION

I t  is possible to exten d the st rip theor y of Haines and  Ollerton  i n

such a wa y th at i t is possible to describe la te ra l traction s.  W i t h  this

theory, a solution can be fou nd for the p roble m of late ral  creepage  i n

r o l l i n g ,  and also for the p ro ble m of spin , as  long  as the spin is smaller

than a certain value  (\f\<l).  The solution has the  same  character as

Carter's  well-known  t ract ion  distribution  for  longitudinal slip  i n  r o l l i n g .The soluti on fou nd has the disadvantage tha t  slip  and tr acti on are not

continuous when one  passes  f rom  strips  w i t h  adhesion  area  to strips

without.  T here is indeed a the ory for the str ips  without  adhesion  area

which  does  no t have thi s defect, but the perfor mance of tha t solution

8/12/2019 1967-13 a Strip Theory for Rolling With Slip and Spin

http://slidepdf.com/reader/full/1967-13-a-strip-theory-for-rolling-with-slip-and-spin 21/21

62

i n  the to tal force was worse tha n the perfor mance of the discontinuous

theory.

The theory was  discussed  in detail for the  case  of pure  creepage  and

for  the  case  of pure spin. I n the  case  of pure  creepage  it was found that

the theory worked  w e l l ,  especially in the  case  of contact ellipses  which

are narr ow i n the d ire ction of r o l l i n g .  As to the total force, a good quanti

tative  agreement  w i t h  K A L K E R ' S  numeri cal theory [6 ] was fou nd ; thi s

agreement  was not so good i n the  case  of a circular contact  area,  where

apart  f r o m a nume rica l disagreement i t was  also foun d that the  dependence

on   Poisson's  ratio was given incorrectly.

I n  the  case of pure spin  there  is much less agreement  w i t h  the numerical

theory. Por contact ellipses  which  are na rrow i n the di recti on of  r o l l i n g ,

a rela tive error of not more t ha n 20 % is  present  i n the results when

\y>\ <  0.9 .  For the  case  of a circular contact  area,  the relative error can

grow  to as mu ch as 40 % when  j y i | < 0 . 8 .  From a comparison  w i t h  the

theory of infinitesimal  pure spin it is seen  that in the  case a = 0 there  is an

asymptotic  agreement  for  ajb^O.  I n the  case  o^O this is not true;

the error is at tr ibu te d to the stri p the ory assumption (see sec. 4.2) .

I n  the  case  of  creepage  without spin, see  H A I N E S  [2] , and  H A I N E S -

O L L E R T O N [ 3 ] ,  the  area  of adhesion is excel lently predicted by the s tr ip

theory. The same seems to  hold  in the  case  of pure spin, i f one may judge

f r o m  the exper iment of Poon on the  case  a/b—1,  a=0.5,  y > = 0 . 3 0 5 .

R E F E R E N C E S

1 . C A B T E R ,  F. W. , On the  A c t i o n  of a Loc omot ive  D r i v i n g  Whee l . Proc. R oy .

So c, A 112, 151-1 56 (1926).

2 . H A I N E S ,  D . J . , Contact  Stresses  i n F l a t  E l l i p t i c a l  Contact Surfaces  which

Support Ra dia l and Shearing Forces duri ng  R o l l i n g .  Inst . Mech.

En grs. , 179, Pa rt 3.

3. • , an d E .  O L L E R T O N ,  Contact  Stress  D i s t r i bu t i ons on  E l l i p t i c a l  Contact

Surfaces Subject ed to Ra di al an d Tan gen tia l Forces. Proc. I ns t .

Mech.  Engrs., 177, 95-114 (1963).

4 . J O H N S O N ,  K . L . , T a nge n t i a l T ra c t i ons a nd  M i c r o s l i p  i n  R o l l i n g  Contact. Proc.

S ymp.  R o l l i n g  Contact Phenomena. E d.  B i d w e l l ,  Elsevier 1962,

p .  6 sqq .

5. , The Inf lue nce of Elast ic De for mat ion up on the  M o t i o n  o f a  B a l l

R o l l i n g  bet ween T wo Surfaces. Proc. Ins t . Mech. Engrs ., 173,

795-810 (1959).

6.  K A L K E R ,  J . J . ,  R o l l i n g  w i t h  Slip  an d Spi n i n the  Presence  o f D r y  F r i c t i o n .

Wear 9, 20-38 (1966).

7. , O n th e  R o l l i n g  Cont act o f T wo Elasti c Bodies in the  Presence  o f

D r y  F r i c t i o n .  D octo ral Thesis, to  appear.

8.  L O V E ,  A . E .  H . ,  A Treatise on the The ory of E l as t ic i ty .  4t h E d . Cambri dge, 1927.

9.  M t r s K H E M S H V i M ,  N .  I . ,  S i ngula r In t e gra l E q ua t i ons .  Noordhoff  1953.

1 0 . P O O N ,  S . Y . , A n E xpe r i me n t a l S t ud y of t he  Shear  T r a c t i o n  D is t r ib u t io n  i n

R o l l i n g  w i t h  Spin. Wear, to  appear.

1 1 . V E R M E U L E N ,  P . J . a n d K . L .  J O H N S O N ,  Conta ct of Non spher ical Bodies Trans

m i t t i n g  T angen t ia l Forces. J .  A p p l .  M ech ., p. 338-340 (1964).