Int. J. mech.ScL PergamonPress. 1973. VoL 15, pp. 993-1010.

Printed in

Great

BrRain

PRESSURE DISTRIBUTION AND OF MACHINED COMPONENTS

DEFORMATIONS IN CONTACT

N. BACK, M. BURDEKIN and A. COWLEY

Mechanical Engineering Department, Centro Tecnologico, Universidade Federal de Santa Catarina, and Mechanical Engineering Department, University of Manchester Institute of Science and Technology, Manchester, England

(Received 1 February 1973, and in revised form 9 July 1973)

S u m m a r y - - T h e paper presents a general method for calculating the pressure distribution at contacting machined surfaces and the resulting deformations of components subjected to external loads. I t is well known t h a t when machined surfaces are compressed one against the other then due to the deformation of the asperities the approach of the surfaces is related to the interface pressure by a non-linear function. Another effect t h a t must be taken into account is t h a t the actual area of contact is not known previously and is, of course, different from the apparent area. F o r the method presented, the contact or connexion between the components is simulated with finite elements, springs or plates defined as a function of the surface roughness and surface deformations. The system so established is solved in an iterative way using the finite element method and this enables the pressure distribution at the contact and the resulting deformations of the whole component to be determined. Finally, examples of some problems solved b y different methods are given and the results compared with the experimental data. NOTATION area of influence of a pair of nodes cross-section of the springs C constant relating the surface deformation and interface pressure d(t:,D displacements E , E . , E ~ modulus of elasticity F , F . forces Fr friction force k modulus of the elastic foundation L length constant of the power law relationship between contact deformation and m interface pressure p a apparent interface pressure P.,P~ pressure at the surface in contact t thickness of the finite plates displacements x, y co-ordinates Iz coefficient of friction contact deformation O. rotation
a. I. INTRODUCTION

A~

MACHINE tools are n o t generally m a n u f a c t u r e d as one continuous casting or fabrication and the reasons for this are the difficulty in manufacturing and transportation, and also for functional purposes such as guideways. Such constructions, therefore, necessitate counexions between t h e basic elements o f
6e 993

994

N. BACK,M. BURDEKINand A. COWLEY

the machine and these can be classified as bolted or fixed joints and sliding joints in the case of guideways. In both fixed and sliding connexions, forces are transmitted across the interfaces and the overall static and dynamic characteristics of the machine tool can be considerably influenced by compliance at the individual connexions. Much of the work which has been carried out on establishing the characteristics of machine tool joints has concentrated on the normal compliance of contacting surfaces and its dependence on surface topography, hardness, interface pressure and materials. Relevant work is presented in refs. (1)-(7) and (13). Connolly7 and Levina T M have used simple structural models in an attempt to show the influence of the interface compliance on the overall stiffness of a structure. However, in their analyses the components in contact are assumed to be rigid which is invalid for most practical cases encountered. When the components are not rigid, the resulting model becomes more complicated since the pressure distribution between the surfaces will be quite different from the assumed and for certain areas of the surface contact can be completely lost. Another aspect t h a t will increase the complexity of the solution, as it is shown in the references mentioned, is that the surface compliance is nonlinear and given by the following equation = c(100P) m (1)

In equation (1), 2 is the approach of the surfaces in tma, c and m surface characteristics dependent upon the pair of materials and surface finish. ]t has been shown ~ t h a t for cast iron c = 0-3-2.0 and m = 0.3-0.7 for all conventional machined surfaces. P is the interface pressure in kgf/mm e. A more detailed analysis of the relationship (1) and the parameters c and m can be found in refs. (1) and (2). In order to include the nonlinear compliance characteristics of the surfaces, elasticity of the components and the fact t h a t the surfaces are stiff only in compression, the simplest way is an iterative technique using the finite element method. In this case the contact is simulated by some connecting elements t h a t represent the surface behaviour. The finite element method will not be discussed here since it is adequately covered in refs. (8-10) and (12). A discussion regarding the precision, preparation of the data and the computer programs of the finite element method used in this work are given in refs. (]), (8) and (9). Three methods used for the purpose of calculating the pressure distribution and deformation in joints, are described herein.
2. HYDROSTATIC METHOD

This terminology has been chosen since the representation of the sm'fkce interface is analogous to a hydrostatic bearing system. Consider the resulting pressure distribution at the apparent area of contact and the deformations of the components shown in Fig. l (a) due to the applied force F. For simplicity it is assumed that the components have the same unit width so that a plane state of stress can be considered but this is not a limitation.

Pressure distribution and deformations of machined components in contact

995

The first step in the solution is the division of both components into finite elements. As is shown in Fig. l(b), both components can be divided independently b u t along the apparent length of contact it is recommended that the nodes coincide forming pairs. The distance between the consecutive pairs of nodes along the contact can be different, b u t as is shown later the best way is to choose equal distances.
F

IA

.....

..........

""

lY'V x, u
o)
A

d) ~ ~ /Pl(i+n.J +n)

F I I Ill : I J.L_L 7--r --',---l-,--I-7-7-- i - - II- ]i +I- t ------i------i------i-- ~ - i - - ~ ; L--'-_I__ +~_4__,,-.-' -,-4--L

e)

.....~:g~ -i- i-t

I I 1 I I, ,_.,..L J_ ~

.--,---,,---,-,-,--,,-,.
I____~__~, . . . . t J,- -

b)

k L\

/5 //~

th iterotion ,h

"~
.I I

g)

c)

FIe. 1. (a) Joint model; (b) finite element division of the joint components; (c) displacements at the surface in contact of the joint components; (d) deflexions at the contacting surface used to generate a new pressure distribution; {e) calculated pressure distribution at the surface in contact; (f) concentrated loads at the nodes; (g) sequence of the pressure distribution in the successive iterations. I f the system is symmetric then the displacements u along A B are zero and the components are analysed independently with the points A and B completely fixed as shown in Fig. l(b). For the first iteration a pressure distribution must be assumed, for example, a uniform pressure, and from this pressure distribution the equivalent concentrated node forces F(i+,~ are calculated and applied as it is shown in Fig. l(b). At this point two independent problems are solved using the finite element method. Firstly, the deflexions v of the nodes along the surfaces in contact are calculated as shown in Fig. l(c). Duo to symmetry only the deformations at one surface, at the right-hand side of the component, will be represented. The next step is to calculate the sums of the displacements of each pair of nodes or d(~+n,~+~ = v(~+n)+ vu+~. (2) When the m i n i m u m sum is obtained, say for example d(~s~,then the differences given in the following equation are calculated
JD(i+,,t+n) = d[i+n,j+n)-d(i.~)

(3)

and when plotted result in the curve in Fig. l(d).

996

N. BACK, M. BURDEKIN a n d A. COWLEY

To generate t h e pressure distribution for the second i t e r a t i o n e q u a t i o n (1) a n d Fig. l(d) m u s t be analysed t o g e t h e r in t h e following way. F r o m Fig. l(d) it can be seen, w i t h o u t difficulty, w h e t h e r or n o t t h e components are flexible or rigid a n d m o r e or less if the c o n t a c t will r e m a i n o v e r all the n o m i n a l area of contact. H e r e it is supposed t h a t t h e c o n t a c t will r e m a i n o v e r t h e nominal area. T h e n a linear e x t r a p o l a t i o n is m a d e as shown in Fig. l(d). F o r practical purposes this is always a good a p p r o x i m a t i o n because t h e pressure at this p a r t of the surface will be small and consequently the surface will r e m a i n m o r e or less straight. A t the p o i n t where the m i n i m u m s u m of displacements occurs it is assumed t h a t t h e surface compression is A(i,j~, an a r b i t r a r y value t h a t in general m u s t be m a n y times larger t h a n t h e final value expected. F r o m this value, a straight line is d r a w n such as is shown in Fig. l(d) a n d at t h e intersection w i t h t h e curve t h e pressure falls to zero. The distances b e t w e e n these lines are t h e c o n t a c t deformations a n d can be calculated b y t h e following e q u a t i o n A(i+,,j+,) = A<i,j)- D~i+n,j+,~l. (4) A c c o r d i n g to e q u a t i o n (1) t h e pressure at t h e pairs of nodes can be calculated as follows
1 (.~(i+n,j+n,] lhn P(i+,,J+,) = ]-65 \ ~ ! "

(5)

P l o t t i n g t h e values f r o m e q u a t i o n (5) this gives the d i a g r a m shown in Fig. l(e). tVhen this pressure distribution is i n t e g r a t e d o v e r t h e area in c o n t a c t t h e force should be larger t h a n t h e e x t e r n a l applied force, otherwise the initial a s s u m p t i o n was wrong. To e v a l u a t e t h e forces t h a t should be applied at t h e nodes for t h e second iteration, t h e integral of the pressure distribution m u s t be corrected w i t h o u t altering t h e a c t u a l distribution so t h a t t h e result is in e q u i l i b r i u m w i t h t h e e x t e r n a l force. This can be done w i t h t h e following equation
n~OP

"-q = 7i- - A .+.,j+.)p.+.,,+.) = F , P( m '

(i,J)

(6)

where q is t h e n u m b e r of node pairs in c o n t a c t and A(~+.,j+n ) is t h e a r e a of influence of each pair. The u n k n o w n in e q u a t i o n (6) is p.,j~ and w h e n calculated it will define the pressure d i s t r i b u t i o n o v e r t h e surface b y t h e following e q u a t i o n P"+~'~+'~ = P"'J) P~'+"'~+"~ P~m (7)

t h a t is also p l o t t e d in Fig. l(e). The sequence of p r e p a r a t i o n of t h e d a t a for a n e w i t e r a t i o n can be simplified w h e n equations (2)-(5) are s u b s t i t u t e d in equations (6) and (7) a n d w h e n t h e l e n g t h of the finite elements along t h e surface in c o n t a c t are t h e same. F r o m t h e pressure distribution of Fig. l(e), t h e e q u i v a l e n t c o n c e n t r a t e d loads in Fig. l(f) are d e t e r m i n e d a n d again t h e finite element m e t h o d is applied and t h e deformations of t h e second i t e r a t i o n are obtained. To proceed w i t h t h e iterations one w a y is to reduce t h e v a l u e 2(i,~); it can be faster at t h e beginning and slower w h e n t h e convergence is reached. Similarly, t h e r e d u c t i o n in ~,,j) can be considered as reducing t h e l e n g t h of contact. The last f o r m is r e c o m m e n d e d . A t t h e p a i r of nodes, where t h e s u m of t h e displacements is a m i n i m u m , t h e n f r o m one i t e r a t i o n t o a n o t h e r , p~,.,~ decreases a n d Pc~,J) increases, and, therefore, for c o m p l e t e convergence these t w o values m u s t be equal. The following i d e n t i t y is t h e n obtained. P(~m = ]-0"6 - " (8)

The sequence of t h e pressure distribution o b t a i n e d f r o m t h e successive iterations can be represented as shown in Fig. l(g). F o r t h e last i t e r a t i o n t h e deformations of each c o m p o n e n t are d e t e r m i n e d and t h e surface compression is calculated. F r o m these results it is possible to assemble t h e whole d e f o r m e d s y s t e m w i t h t h e respective pressure distribution a t t h e c o n t a c t surface.

Pressure distribution and deformations of machined components in contact

997

If, for example, the relative displacement between points A and B on Fig. l(a) is required it is obtained from Deformation = d(i,~}Av,~). (9)

When non-syrametric systems are to be solved the so-called hydrostatic method described above is not recommended and to overcome these difficulties the following methods will be used. 3. P L A T E METHOD

To describe this method let us consider the deformations and the pressure distribution between the components shown in Fig. 2(a). Here again it is assumed that the two components are of the same unit width and that it is a plane state of stress.
Ft F3

a)

2.
i~ ~ ,

F2 F~ I

F3

[-~-+$-i-~-I--i-4 I-i:-i-i-/:~i-T?i
I'| , ,

I ',,Illl~
i i ~ ~ i ,

....
i .~_ ,

l---r--~i~ ,*~*~-i---, ---,r--I

I---*---/-~ * ~ - ~ '--~-~*:~--}...... *--4
I
/ I

,
I

~
~

I
i

I
,

~
,

b)

~:I' E? ,
e)

E.

. . . . . . . . . . .

. . . . . . . .

. . . . . . .

d)

? 8 9 I0 No of iterotior~

Fie. 2. ( a ) J o i n t model; (b)finite element division of the joint and representation of the d u m m y plate at the surface in contact; (c) uniform contact deformations for the first iteration; (d) displacements of the contacting nodes normal to the surface from the first iteration; (e) displacements of the contacting nodes normal to the surface for the trial iteration; (f) calculated pressure distribution; (g) sequence of contact deformations at a general pair of nodes without improving the convergence. The finite element division must follow the same rules described in the first method. However, in this case both components are connected b y a d u m m y plate with the behaviour equivalent to the surface in contact. When this plate is inserted the dimensions of the system are not modified. The thickness of the plate is only considered in the calculations in order to define the rigidity and no bending action is implied b y the term plate. The thickness and the elastic modulus of this plate must be chosen so that the

998

N. BACK, IV[. BURDEKIN

and A. COWLEY

bohaviour of the connexion remains the same as the surface in contact. The properties of the plato are such that it must have high rigidity and be nonlinear in compression, have zero rigidity in tension and no shear rigidity in the direction tangential to the surface so that the friction force can be taken into account. The method described below will not include all these complexities but the main properties can be taken into consideration. For the first iteration the thickness of the plate is chosen which can always remain the same. The modulus of elasticity is chosen as a function of the surface finish, pair of materials and pressure distribution. At the beginning some assumptions for the pressure distribution must be made. This can be constant or any variable form calculated so that it equilibrates the external load when the components are assumed to be rigid. For an isotropic material the following relationship is valid: hE P = t l O a"

(10)

W h e n equations (1) and (10) are c o m p a r e d tho following i d e n t i t y can be established. 100
or

t x 10 a IOtA(i/m-i) cl/m

E-~

I n t h e last relationship E is t h e modulus of elasticity in k g f / m m 2 and t the thickness of the d u m m y plato in ram. This e q u a t i o n can be w r i t t e n in a general f o r m as follows 10th~/m-l~
En cl/,,~

(11)

Considering t h e p r o b l e m shown in Fig. 2(b), and assuming t h a t the pressure is uniform, t h e n t h e surface compression Am at t h e nodes can be calculated f r o m e q u a t i o n (1). The m o d u l u s of elasticity En for t h e connecting fmito elements is calculated using e q u a t i o n (11). I n this e x a m p l e En for all t h e elements is t h e same a n d represented in Fig. 2(c). F o r t h e s y s t e m so established, w h e n t h e finite element m e t h o d is applied t h e displacements in t h e y direction of t h e nodes s e p a r a t e d b y t h e plate can be represented as shown in Fig. 2(d). F r o m this figure it can be seen t h a t all finite plates o v e r t h e surface are in compression. T h e a m o u n t of compression can be calculated b y the relationship
An = v(j+,~)-v(i+n). (12) W h e n these vMues are s u b s t i t u t e d in e q u a t i o n (I1) t h e n the m o d u l u s of elasticity at t h e pair of nodes is determined. W h e n the v a l u e of An becomes n e g a t i v e t h e n t h e value of E~ is m a d e e q u a l to zero a n d m e a n s t h a t t h e element is eliminated. F o r t h e n e x t iteration, the modulus of elasticity of t h e different finite plates will be cMculatod as a linear interpolation of t h e values d e t e r m i n e d at the pair of nodes t h a t f o r m t h e b o u n d a r y of each p l a t e or

E~-

E~+En+I 2

(13)

Mow all the d a t a are p r e p a r e d for t h e second iteration. This sequence is followed until t h e difference b e t w e e n t h e values of A, f r o m one i t e r a t i o n to a n o t h e r is less t h a n a p r e d e t e r m i n e d value. Suppose t h a t after a certain n u m b e r of iterations t h e deformations at She surface in c o n t a c t can be represented as shown in Fig. 2(e). I t is e v i d e n t t h a t the deformations of t h e c o m p o n e n t s are calculated at the same time. W h e n t h e values f r o m e q u a t i o n (l l) are calculated for t h e last i t e r a t i o n t h e n the pressure distribution at t h e surface in c o n t a c t will be d e t e r m i n e d b y the following e q u a t i o n ' p,~ = ]-0-6 This is given in t h e d i a g r a m shown in Fig. 2(f). (./

Pressure distribution and deformations of machined components in contact

999

The pressure distribution can also be evaluated at the pair of nodes with the equation P" =

(E'_I + E') ~.
2t x 10a

(15)

The difference between the values calculated with equations (14) and (15) is small, this will be seen later when some examples are solved. When the sequence described above is followed the convergence is always certain but very slow. If, for example, the surface compression at a certain pair of nodes is analysed in the consecutive iterations the change of the surface deformations can be represented as shown in Fig. 2(g) converging to a constant value. To increase the speed of convergence, the surface compression used to define the modulus of elasticity for the third iteration is obtained by averaging the values from the first and second iteration at each pair of nodes. F o r the fourth iteration )~, is defined as the average of the first average and the third iteration. Following this procedure the speed of convergence is increased m a n y times. 4. S P R I N G METHOD

This method is very similar to the plate method, instead of finite plates, springs are used at the surfaces in contact Thus, only normal stiffness is incorporated and shear effects are neglected. Here again the components are divided into finite elements in an independent form, and the only restriction is t h a t at the contact surface the nodes must coincide forming pairs. The computer programs1," used are general, where a system can be formed from rectangular and triangular plates, prismatic elements and beams. The beams can be presented with torsional, flexural, shear and normal stiffness. Here for the spring method, only a normal stiffness will be introduced and bending effects neglected. This normal stiffness is defined by the modulus of elasticity which can be the same as the components ; a length and a cross-section t h a t is calculated as a function of the surface finish and the pressure at the contact. The general procedure for preparation of data for the first and subsequent iterations and the convergence is the same as t h a t described in the plate method. However, in this case it is necessary to define the cross-section of the beam. When at a certain pair of nodes the pressure is known, AM is the influence area and if the mesh is fine then the force transmitted across this area is given by

F,, = p . A . .

(16)

I f between the two nodes a spring is introduced with the cross-section a n and length L then the following relationship is valid -Pn = E)tn a . 10aL
or

(17)

10aF.L
an = E~.

When $'n from equation (16) and p . from equation (14) are substituted in equation (17) the following equation is obtained a. =

IOLA.,~(1/~-I) c1/'~ E

(18)

Here again, when ;1~ is negative, then the spring is in tension, and a . is equated to zero or as small as possible t h a t permits the solution, depending upon the computer program. The pressure distribution, when the last iteration is prepared, is calculated also with equation (14) or b y the following equation a . EA. p. = 10aA.L. (19)

1000

N. BACK, M. BURDEKIN a n d A. COWLEY

The difference b e t w e e n these t w o equations for calculating t h e pressure distribution is also v e r y small and can be seen later in t h e paper. F o r t h e spring m e t h o d , instead of changing t h e cross-section of t h e b e a m f r o m one i t e r a t i o n to another, this cross-section can be m a i n t a i n e d c o n s t a n t a n d t h e modulus of elasticity redefined in t h e consecutive iterations. W h a t m u s t be observed in this m e t h o d is t h a t the springs h a v e no transversal r i g i d i t y and, therefore, t h e s y s t e m can become unstable. I f it is n o t possible to introduce restrictions for some nodes, t h e n it is necessary t o include some e x t r a rigidities or introduce t h e shear stiffness o f m a c h i n e d surfaces as described in ref. (1). 5. F R I C T I O N AT THE OF MACHINED CONTACT SURFACE COMPONENTS

I n general, a t t h e c o n t a c t of t h e c o m p o n e n t s r e l a t i v e displacements of t h e nodes t a n g e n t i a l to t h e surface are observed and, therefore, friction forces are developed w h e n t h e surfaces are assumed to be rigid in shear. These forces can m o d i f y t h e final pressure distribution and t h e whole deformations of t h e components.

Frfi ) . . /

;///////////A

a)

F//2///fi,//////////////////;
Fr(i) J
\
b)
Fro. 3. R e p r e s e n t a t i o n of t h e friction forces a t t h e c o n t a c t i n g surface. (a) Spring m e t h o d ; (b) h y d r o s t a t i c m e t h o d . I n t h e plate m e t h o d shear forces arise a t t h e c o n t a c t b u t t h e representation of friction forces is difficult. F o r the h y d r o s t a t i c a n d spring m e t h o d t h e friction forces can be i n t r o d u c e d v e r y easily. Consider for e x a m p l e a p a r t of a surface in c o n t a c t as shown in Fig. 3(a) w i t h springs connecting t h e pair of nodes. The r e l a t i v e displacements for a p a i r is g i v e n b y v~ t h e n as it is shown in Fig. 3(a) the friction force F r i is applied a t t h e nodes. T h e friction force Frt will be calculated w i t h t h e following e q u a t i o n

F(i)

v(i) I
!

\ \

Fri = I~pi At,

(20)

where/L is t h e friction coefficient. These forces are included in t h e general load v e c t o r a n d the procedure will be the same as it is w h e n o n l y t h e e x t e r n a l loads f o r m t h e load vector. I n t h e case of t h e hydrostatic m e t h o d t h e friction forces are a d d e d as shown in Fig. 3(b). A m o r e general and sophisticated m o d e l used to represent t h e shear stiffness a n d friction forces at t h e interfaces is described in ref. (1). 6. A P P L I C A T I O N OF TECHNIQUES OF PROBLEMS TO THE SOLUTION

Problem 1
T h e b e a m of m i l d steel shown in Fig. 4(a) was used in ref. (1). Measurements of t h e deflexions along t h e surface in c o n t a c t were m a d e w h e n t h e loads were applied at t h e centre a n d s u p p o r t e d b y a base t h a t was assumed to be rigid. The surface characteristics were g i v e n b y e = 0.29 a n d m = 0.5. F o r t h e theoretical solution of t h e problem, t h e base was also considered to be rigid and t h e c o n t a c t b e t w e e n t h e c o m p o n e n t s is elastic according to e q u a t i o n ( 1 ).

Pressure distribution and deformations of machined components in contact

F Width
[l I I I LI
r, ......... ,,

1001

25.4turn
I I iJl r .... "~- -1 II

I i l I I i I I , I I l I I I I ILI I a---",-a----rl,, ...... I -,-~--i, .............. I,I I "1

_ ~ ' 5 mm

~lllll/l#111lll#/11111ll#/11/lllllll#ll1411111111111/lll/h

aJ
120 100
,~.

I"

3oo,.,~

~1

60
40 2O
i I I I I I I I I i

~
~
J~.lr"~l i i i "..I,..LI i

132 kgf l crn2

6.6

2"65 1"32
i i i i I i i I I

b)

150

300mm

-81- \ / -6

'65

:t
c)
I0

.o_
_~ 5 th iterotion ~th

9 8
?

"~

3 2

I
. . . . . .

/ J S

. , , , , , ,

. . . . . . .

d)

150

300turn

FIG. 4. (a) Finite element division and dimensions of the beam on a rigid base; (b) pressure distribution for different apparent interface pressures using the hydrostatic method; (c)contact deformations for different apparent interface pressures using the hydrostatic method; (d) sequence of the calculated pressure distribution using the hydrostatic method a n d apparent interface pressure of 1.32 kgf/cm =. Applying the hydrostatic method, the pressure distribution and contact deformations for different interface pressures are shown in Fig. 4(b) a n d (c). Fig. 4(d) shows the sequence of the pressure distribution for t%he consecutive iterations in the case of 1.32 kgf]em 2 apparent interface pressure. The n u m b e r of iterations for the other loads was in the same order. I n Fig. 5(a), the deformed beam for the interface pressure of 6.6 kgf]cm 2 using the plate method is shown. Similarly, the deformed beam could be drawn for the other methods and loads.

1002

N. BACK, M. BURDEKIN and A. COV~'LEY

I n Figs. 5(b) and (c) the pressure distribution and the contact deformations are represented using the plate and spring method for the interface pressures of 1.32 and 6.6 kgf/cm =. Figs. 5(d) and (e) show the sequence of the contact deflexions in the successive iterations, here using the plate method but for the spring method it is similar.
Delormohon scole 0 10 20 30 40.urn

F= 510 kgf

40
30

Pa = 6 6

+ Sprln9 method

2o
I Pa =1150

bJ

300 rnm

-6 -4

Po

:6"6

Pa :1"32

-2

+~+

:Jt
e)
-5 -4 -3 -2 4 th iteration 3 rd

2 3 4 d)

'L
t

\i ,t~"~ '''5 th iteration 4 ~ 3 th 2nd

,,,,

,,,,

-2 -t-5 -I0
0

"5

~"Ist

1-0 15 e)

~.~.,.x~ f

FIG. 5.

(a) F i n i t e element division and deformations o f the beam

using the plate method; (b) pressure distribution using the spring and plate method for the apparent interface pressure of 6.6 and l'32kgf/cm2; (c)contact deformations; (d) sequence of the contact deformations using the plate method and 6.6 kgf/cm 2 apparent interface pressure; (e) sequence of the contact deformations using the plate method and 1-32 kgf/cm s apparent interface pressure.

Pressure distribution and deformations of machined components in contact

1003

Due to the small differences between the results using the different methods and the evaluation of the pressure with equations (14), (15) and (19) the comparison is shown in Table 1. I n this Table are also represented the results of the integrated pressure distribution. Only when the hydrostatic and spring methods are used in conjunction with equation (19) will the integrated pressure be equal in magnitude. The reason for this can be appreciated when the assumptions for each method are analysed. F o r the alternative ways of calculating the pressure distribution the differences depend upon the number of iterations performed. I t must be remembered t h a t the pressure is a function of the surface deformation to the power of approximately 2 and, therefore, any difference in the final deflexion gives a larger difference in the pressure.
TABLE 1. CO~PARmON

OF THE CONTACT DEFOR~IATIONS AND PRESSURE OVER THE HALF LEI~GTH IN CONTACT

DISTRIBUTION

Hydros method Force (kgf) 102 ~ 0-87 0.79 0"60 0.40 0-20 0.00 Tot. F. 2.20 1.93 1.35 0.72 0.10 0-53 Tot. F.
- -

Plate method

Spring method )l 0-88 0"79 0.61 0.41 0-21 0.01 Tot. F. 2-20 1-93 1.34 0.70 0.06 -- 0-59 Tot. F. p (eqn (14)) 9.35 7.69 4.54 2.01 0-52 0.00 98.73 59.3 45-6 21.9 5.92 0.47 0.00 526.1 p (eqn (19)) 9.65 7.91 4-67 2.09 0-59 0.04 102.0 59.1 44.8 20-7 5.33 0.41 0.00 511.9

p 9.50 7.90 4.79 2.14 0.53 0.00 102.0 59.2 45.3 20.4 5.09 0.00 0-00 510.0

~ 0.86 0.77 0-59 0-40 0.21 0.02 Tot. F. 2.18 1.90 1.30 0.69 0.10 -- 0.50 Tot. F.

p (eqn (14)) 8.94 7"31 4.28 1.90 0-51 0.00 93"83 57.8 43-9 20.7 5.75 1.20 0.00 509.9

p (eqn (15)) 8.76 7-29 4.39 2.00 0.59 0.18 95-65 53.4 41.2 19.8 5.59 2-30 0.00 485-2

510

p, kgf/cm ~, 2, ~ m ; Tot. F., total force. For this particular problem the results of the three methods gave very good correlation but this is not general; the hydrostatic and spring methods behave similarly for any problems where they can be applied but the plate method will in general differ in particular when the relative tangential displacements of the surface are expected. As was seen before, the number of iterations is in the order of four to five b u t when for a particular problem the deformations for several loads are required then the number of iterations can be decreased as follows. F o r the lowest load the procedure is the same as described before or beginning for example with a uniform pressure distribution. Fo r the second load amplitude, the pressure distribution from the first load can be used to define the characteristics of the finite element connexions and so on. When this procedure is followed then, depending upon the difference between the loads, the number of iterations can be reduced to one or two.
Problem 2

This problem, which is shown in Fig. 6(a), was examined experimentally by Levina. a The beam of cast iron was placed on a rigid cast iron base, and the surface finish at the contact was hand-scraped b u t this was not specified in detail. As is known for handscraped surfaces the compliance depends upon the quality of the hand-scraping. Fo r the

1004

N. BACK, M. BURDEKIN a n d A. COWT,EY

c a l c u l a t i o n s t h e surface c h a r a c t e r i s t i c s t h a t were u s e d r e p r e s e n t a p p r o x i m a t e l y t h e r a n g e for t h i s t y p e o f surface finish 3 a n d t w o v a l u e s o f c were chosen, 0.6 a n d 0.8, a n d m = 0.5. T h e m o d u l u s of e l a s t i c i t y for t h e c a s t i r o n was t a k e n as 9500 k g f / m m ~.

Dfr oo c l e mhnae o s F 2 0Q = 0k f
I

11-4--l--'J-'-'-q-d-d-4--l-+-F n._L__L--J_J_J_J_j_d_d_ 4_

--- ',-d ',--4t J _1__1_J L L R

4-t "
j

[,

20ram

L
o)
12 tO
#

....

~s
2

II Poe eo I tmt d h
Spring method

250

500 m m

b) -1 .[ 0 ~.
, , , I ~ , ,

+
,~,

~2 3 )
0

F LeL

L.~

5 "- 3

32
e)

1
i

FIG. 6. (a) F i n i t e e l e m e n t d i v i s i o n a n d d e f o r m a t i o n s of t h e b e a m u s i n g t h e p l a t e m e t h o d , p a = 4 k g f / c m 2 a n d c = 0.8; (b) c a l c u l a t e d p r e s s u r e d i s t r i b u t i o n u s i n g t h e s p r i n g a n d p l a t o m e t h o d for p a = 4 k g f / cm~; (c) c o n t a c t d e f o r m a t i o n s for p a = 4 kgf]cm*; (d) c o n t a c t d e f o r m a t i o n s for p a = 2 kgf/cm*; (e) p r e s s u r e d i s t r i b u t i o n for IDa = 2 k g f / e m *. F o r t h e t e s t s 3 a n d c a l c u l a t i o n s t h e a p p l i e d loads a t t h e c e n t r e were e q u i v a l e n t to 2 a n d 4 k g f / c m ~ n o m i n a l i n t e r f a c e pressure. Fig. 6(a) s h o w s t h e c a l c u l a t e d d e f o r m a t i o n s o f t h e b e a m for 4 k g f / c m 2 i n t e r f a c e p r e s s u r e a n d b y u s i n g t h e p l a t e m e t h o d w i t h c - 0-8. I n F i g s . 6(b) a n d (c) t h e p r e s s u r e d i s t r i b u t i o n a n d c o n t a c t d e f o r m a t i o n s a r e r e p r e s e n t e d for t w o v a l u e s of c a n d u s i n g t h e

Pressure distribution a n d deformations of machined components in contact

1005

plate a n d spring methods when the interface pressure is 4 kgf/cm 2. Similarly, results for 2 kgf/cm 2 are shown in Figs. 6(d) and (e). I n Fig. 7 the same problem was considered using the hydrostatic method.

P ,,-~ ~i ~..i- "--t ~-44=4-! i-iJ !14i1-~, q-i 4 4 ~-'~-~-]-'-'4-~L] J J J_J_~]..Jd-J-J-J-J-'--LJ-~--

20m,~

a)

12 ~0 8

+ . C= "8 C='6 po=4kgf/ crn2

2
i

b)

250

500rnm

0 2 3
)

" '

'

. . . .

5 thiterotlon

"~ t TM

d)

iLS... , . \ .
11

~ ~ 2 n d

tO 9 8

Tthiteration 3rd

~7
"~. 6

~5
2

I i list

e)

7. \
! , |

Fzo. 7. (a) Finite element division of the beam; (b) pressure distribution using the hydrostatic method; (e) contact deformatious f o r / ~ -- 2 a n d 4 kgf/em 2 b y using the hydrostatic method; (d) sequence of the pressure distribution for p a - - - 2 k g f / c m ~ using the hydrost~tie method; (e) sequence of the pressure distribution for p a = 4 kgf/em =. The comparison of the calculated contact deflexions using the three methods and the measured deflexion8 are shown in Fig. 8. Here the correlation between the calculated a n d measured contact deformations is quite good. Comparing the calculated results the hydrostatic a n d spring methods give approximately the same values which are larger t h a n the plate method as can be expected.

1006

N. B a c K , M. BURDEKIN a n d A. COWLEr

0 L ~,,,%.
2 3

250

. J

500mm

o)

b)

~--. + x

Hydrostatic method Plate method Spring method Measured method

FIG. 8. Comparison of the calculated and measured contact deformations c o n s i d e r i n g c = 0.8. (a) F o r p a = 4 kgf/cm2; (b) for p a = 2 k g f / c m 2.
Problem 3

T h e b e a m in Fig. 9(a) w a s also u s e d b y L e v i n a 3 in w o r k o n t h e m e a s u r e m e n t of c o n t a c t d e f o r m a t i o n s . T h e surface finish a n d m a t e r i a l was t h e s a m e as in p r o b l e m 2. I n Figs. 9(b) a n d (e) t h e p r e s s u r e d i s t r i b u t i o n a n d c o n t a c t d e f o r m a t i o n s a r e s h o w n for 2 a n d 4 k g f / c m 2 a p p a r e n t i n t e r f a c e pressure. T w o v a l u e s of t h e surface stiffness coefficient e were considered i n c o n j u n c t i o n w i t h t h e h y d r o s t a t i c m e t h o d . Figs. 9(d) a n d (e) s h o w a g a i n good c o r r e l a t i o n b e t w e e n t h e c a l c u l a t e d a n d m e a s u r e d a deflexions. T h e t h r e e p r o b l e m s so f a r p r e s e n t e d a r e n o t h i n g m o r e t h a n a b e a m o n a c o m p l e x f o u n d a t i o n or g e n e r a l l y s p e a k i n g , a n o n - l i n e a r f o u n d a t i o n . E v e n so t h e t r e n d of t h e r e s u l t s c a n b e c o m p a r e d w i t h t h e basic t h e o r y of b e a m s o n a l i n e a r elastic f o u n d a t i o n . 14 F o r a n i n f i n i t e l y l o n g b e a m , t h e deflexions d u e t o a c o n c e n t r a t e d l o a d F are defined b y t h e following e q u a t i o n = (F0/2]c) e -~ (cos 0 x + s i n 9x), where
0 = ~/(]c/4EI).

(21)

H e r e , ]c is a c o n s t a n t of t h e l i n e a r f o u n d a t i o n , E t h e m o d u l u s of e l a s t i c i t y a n d I t h e s e c o n d m o m e n t of area. F o r t h e l i n e a r case t h e l e n g t h o f t h e w a v e s of t h e c o n t a c t d e f o r m a t i o n s (or h e r e will b e t e r m e d t h e l e n g t h of c o n t a c t ) is d e p e n d e n t o n l y u p o n ~ or k a n d E x I . T h e s a m e h a p p e n s for t h e n o n - l i n e a r case t o a c e r t a i n e x t e n t , b u t n o w t h e f o u n d a t i o n m o d u l u s is d e p e n d e n t u p o n t h e deflexion a n d c o n s e q u e n t l y u p o n t h e load. I f 8 increases t h e l e n g t h of c o n t a c t decreases. Therefore, if/c is t h e l i n e a r model, a n d t h e r i g i d i t y of t h e surface a n d t h e l o a d i n t h e n o n l i n e a r m o d e l increase, t h e n t h e l e n g t h o f c o n t a c t will decrease. A n o t h e r w a y to increase ~ is t o d e c r e a s e t h e r i g i d i t y E I t h e n a g a i n t h e l e n g t h o f c o n t a c t decreases. T h e influence of t h e s e f a c t o r s c a n b e seen in t h e figures of t h e p r o b l e m d e s c r i b e d a b o v e .
Problem 4

T h e c o m p o n e n t s for t h i s p r o b l e m s i m u l a t e a sliding j o i n t a n d a r e s h o w n in Fig. 10(a). T h e f o r m u l a t i o n o f t h e p r o b l e m is as follows. T h e deflexions a n d t h e p r e s s u r e d i s t r i b u t i o n a t t h e c o n t a c t of t h e c o m p o n e n t s will b e c a l c u l a t e d w h e n a l o a d of 30 k g f is a p p l i e d in t h e x d i r e c t i o n as s h o w n i n Fig. 10(a). O t h e r c o n d i t i o n s a r e t h a t for t h e n o d e s a l o n g A B t h e v e r t i c a l d i s p l a c e m e n t s v a n d t h e r o t a t i o n s 0~ are zero. Also t h a t t h e n o d e s a l o n g E F a t t h e lower e l e m e n t are fixed.

Pressure distribution and deformations of m a c h i n e d c o m p o n e n t s in c o n t a c t

Deformation scale

1007

~=2oo,g,
I

b ;o 2'o Jo ~'o~,m

J-i-i-H-+Ti-"~-~'-T ~- '- -!-i--i-;--:-,.-_.-,-,--,,--,--.--,,--..-.--,-..-,--,-.--,--,-,-t _Fi_i_i_i_i_i_i_i_i_~.j_j_!_~j_+~__l__

._,_,__._,_,_,__,_,__._,__.,__._._.__,-,-,
l--r-r~--r ...... ~----i . . . . 7--i

I--i-~--.-"--= -i -,~-~---~-i--!-- .--~2 oo ..,.

-'~- - - P -

....

L
a)
7

OOmm
~ ~o

.I
+
x
C= '6

/ \ /
.+.
1 I I i i i ! ! i

Measured

\./po=4*gf/cm2

b)

250

500 mm

o~.;

. ,

'

. . . . .

c) .o| . . . . , . . . . .

"ot
d)
"0 [ j , , , I ~ ~

f
' ' I~

I~

"5~
7.5 kgf Icm 2

~ I'0 e)

FIG. 9. (a) F i n i t e element division a n d deformations of t h e b e a m using t h e h y d r o s t a t i c m e t h o d , Pa = 4kgf/em2 a n d c = 0.8; (b) pressure d i s t r i b u t i o n f o r p a = 2 a n d 4 k g f / c m 2 a n d c = 0.8; (c) pressure distribut i o n for c o n t a c t d e f o r m a t i o n s ; (d) comparison of t h e calculated a n d m e a s u r e d c o n t a c t deformations for p~ = 4 k g f / c m 2 ; (e) comparison of t h e calculated a n d m e a s u r e d c o n t a c t deformations for p~ = 2 k g f / e m 2. I t can be seen t h a t w h e n t h e c o m p o n e n t s arc u n d e r load t h e n c o n t a c t will only occur along t h e surface 01). B y a p p l y i n g t h e h y d r o s t a t i c m e t h o d w i t h e - - 0 . 6 , m = 0.5, E = 9500 k g f / c m * a n d u n i f o r m thickness of t h e joint of 10 m m t h e d e f o r m e d c o m p o n e n t s are represented in Figs. 10(b) and (e). W h e n t h e p l a t e m e t h o d is applied t h e n t h e connecting finite p l a t e is only i n t r o d u c e d along 01). F o r this case, a n d using t h e characteristics g i v e n above, t h e calculated deformations are represented in Fig. 10(d). F o r t h e ease of t h e h y d r o s t a t i c m e t h o d , k n o w i n g t h e pressure distribution t h e n t h e surface deformations can be calculated a t t h e c o n t a c t a n d w i t h c o m p o n e n t d e f o r m a t i o n

1008

N. BACK, M. BURDEKIN and A. COWLEY

then the whole deformed joint can be rebuilt. This could be done but will give approximately the same as Fig. 10(d). The total deflexion of 24.0 ~m shown in Fig. 10(d) was also obtained with the hydrostatic method. Fig. 10(e) shows the comparison of the pressure distribution calculated with these two methods. A

a)

i .

"1

L .........

,,
Deformotion scale , .,,,,h

;o 2o ~o.m

Dimens~onol scale

30 25 u 20 "~ t5 t K

--

Hydrostotic method X Plate method

tO

15

20

25

$flmm

d)

Fz6~. 10. (a) Finite element division of the component; (b) and (c) deformations of the components using the hydrostatic method; (d) deformations of the part using the plate method; (e) comparison of the pressure distribution using the plate and hydrostatic method. 7. C O N C L U S I O N S T h e first a s p e c t to be discussed in t h e precision of t h e results o b t a i n e d f r o m t h e m e t h o d s . O b s e r v i n g t h e theoretical a n d m e a s u r e d deflexions for p r o b l e m s 2 a n d 3 it can be seen t h a t t h e correlation is good b u t as w a s m e n t i o n e d t h e e x p e r i m e n t a l d a t a were n o t specified such t h a t t h e a c t u a l surface characteristics could be used for t h e calculations. I f t h e values o f c = 0.8 a n d m = 0-5 r e p r e s e n t t h e h a n d - s c r a p e d surface finish m o s t f r e q u e n t l y used in p r a c t i c e t h e n p r o b a b l y t h e differences c a n n o t be a n a l y s e d f r o m this p o i n t o f view. I f t h e s h a p e o f t h e differences b e t w e e n t h e calculated a n d m e a s u r e d deflexions is considered it can be concluded t h a t t h e surface c o m p l i a n c e used in t h e calculations are v e r y close t o t h e one in t h e test. 8 W h e n a t e s t is used t o c o m p a r e w i t h t h e theoretical results, as is t h e case here, one m u s t be certain t h a t t h e precision of t h e m a c h i n e d surfaces is good. F o r small d e f o r m a t i o n s a t t h e surface in c o n t a c t

Optimum fibre-reinforced sheets for two alternative loadings

1017

(al, a2) are of opposite sign then (%, a~) are also of opposite sign. I f the fibres are fully strained, then ac, = - - a ~ and Io~,I = Eel. v~ and v~ are then derived from (11), giving v ---[al[ + [a~ I I t a n a + t a n E~! tan ~ (16)

from (2). I n this case it m a y be noted that v is dependent on (a, fl). Also, since tan ~ tan
fl = -

a2/ax > 0

(cf. equation (10)) both ~ and fl lie in the same quadrant. I t m a y be seen, from {10) and (16), that v becomes infinite whenever
tan ~ = tan~ = 4(-

~,/~)-

Minimization of the functional form of v given in equation (16) gives least weight when (a,/g) ----(0 , 90 ) with [o'1 [+[a~ [ Vmin ~ E~I There is thus a contrast with the case, in Section 3.1.1, when the principal stresses are of the same sign. I n the present case an infinity of fibre arrangements is again possible, their weight being given by equation (16); a unique m i n i m u m occurs, however, when the fibre directions coincide with the principal stress directions. 3.2. Three fibre directions Since a three-fibre system is statically determinate it is possible to calculate b y static means alone the strains arising from a given applied stress; interest m a y be restricted therefore to fully strained systems since overstraining in a n y direction m a y be countered b y increasing the amount of fibre in that direction, with a corresponding increase in weight. 3.2.1. Principal stresses of same sign. First, consider a fully strained laminate in which the strain is isotropic in either tension or compression, with and Then
v = = [a,l+la,

Ee I

(17)

from equation (1). The other possible fully strained system is one in which one fibre direction (~, say) is strained in the opposite sense to the other two. I n this case
v ~ - v # + vy =

[al+a2 .Eez

from (1), and the weight is thus greater b y 2v~ t h a n that given b y (17). The optimum is thus given b y (17) and corresponds to a state of isotropic tensile or compressive strain, whichever is appropriate to the applied loading. There is an infinity of choices for (~, fl, 7) and hence an infinity of optima ; the choice of optimum (c~,fl, ~) is, however, not completely unconstrained since the only admissible angles are those which give rise to simultaneous tension or compression in all three fibre directions. Such admissible angles are derived from consideration of equation (8). The angles ~ and ~ m a y be regarded as given, b u t it will always be true that these angles and their conjugates (~', fl') lie in the sequence

,~,,8, ~',~', ~;
67

1018

P. BARTHOLOMEW a n d G. Z. HARRIS

f r o m e q u a t i o n (8) it m a y b e s h o w n t h a t p r o v i d e d 7 lies b e t w e e n a ' a n d / 3 ' , fibres in all t h r e e d i r e c t i o n s will b e s t r a i n e d i n t h e s a m e sense. A t t h e e x t r e m e s of t h e r a n g e (a', fl'), Y will f o r m a c o n j u g a t e p a i r w i t h e i t h e r a or ]3, fibres in t h e o t h e r d i r e c t i o n c h a n g i n g t h r o u g h zero t o t h e opposite s t r a i n s t a t e . As a n i l l u s t r a t i o n , Fig. 2(a) shows t h e v a r i a t i o n of t h e t e t a l w e i g h t w h e n ~2 = 2 ~ a n d t h e angles are c h o s e n so t h a t a = 50 ,/3 -- 160 (giving a ' = 120.2 , ]~' = 79.7 ) ; t h e r e g i o n of c o n s t a n t o p t i m u m weight, f i ' < 7 < a ' is e v i d e n t . Fig. 2(b) shows t h e v a r i a t i o n of s t r e s s r e s u l t a n t s w h i c h d e t e r m i n e t h e c o n t r i b u t i o n s v~, v~, v v t o t h e t o t a l fibre weight.

I
5LFe~S resultant

l / l

t
I I ~

'
i :

I
I

1
I I I I I I I I
I I
I
0

<..,.=,.b(
/I ~

',
I \]

}tli

"l

:
:

t
18o if

,/
/ I i t l<<-,,,o:

\
7/i~j I_,'_A

'
'

<:<:
:

I t ?

l{ /i

FIG. 2. (a) V a r i a t i o n of t o t a l w e i g h t w i t h 7 for % = 2a 1 w h e n a a n d ]3 are specified. (b) D i r e c t stress r e s u l t a n t in t h e fibre d i r e c t i o n for e a c h s y s t e m o f plies.

3.2.2. P r i n c i p a l stresses of opposite sign. W h e n 0.1 a n d 0.2 are of different sign, it is e v i d e n t f r o m e q u a t i o n (1) t h a t 0.~, 0.~, 0.v c a n n o t all b e of t h e s a m e sign. I f t h e fibre d i r e c t i o n s are so n a m e d t h a t t h e f~ d i r e c t i o n is t h e one in w h i c h t h e fibres are s t r a i n e d in t h e o p p o s i t e sense t o t h e o t h e r t w o directions, t h e n
v~-v~+v v : + +

0.1 +

0.2

f r o m (7) ; t h e u p p e r sign c o r r e s p o n d s to c o m p r e s s i v e s t r a i n in t h e t7 direction. T h e v o l u m e f r a c t i o n of t h e whole s h e e t is t h u s g i v e n b y

v: zvt~T
-

(Yl + 0"2

--

E~ s

tlS)

f r o m (2). T h u s , t o m i n i m i z e t h e w e i g h t of t h e s h e e t it is sufficient to m i n i m i z e v~, s u b j e c t t o t h e c o n s t r a i n t t h a t t h e / 3 fibres c o n t i n u e t o c a r r y a stress of o p p o s i t e sign t o t h a t c a r r i e d b y t h e (a, 7) fibres. T h e o p t i m u m i n t h i s case is f o u n d to c o r r e s p o n d t o two-fibre d i r e c t i o n s only, rt~ther t h a n t h r e e ; t h e s e t w o d i r e c t i o n s coincide w i t h t h e p r i n c i p a l stress directions. A n o u t l i n e p r o o f is n e x t g i v e n of t h i s s t a t e m e n t . I n i t i a l l y i t is c o n v e n i e n t t o r e g a r d a a n d Y as g i v e n while ]5 varies. T h e c o n s t r a i n t o n fi is t h a t t h e fl d i r e c t i o n s h o u l d b e s t r e s s e d in t h e o p p o s i t e sense t o t h e (a, 7) directions. I f ]5 is c o n j u g a t e t o a, t h e n ey = 0 (eL Section 2.2). T h u s a c o n s t r a i n t e x i s t s a t ]3 = c( since s v c h a n g e s sign i f f~ passes t h r o u g h a ' ; e v i d e n t l y a s i m i l a r c o n s t r a i n t also exists a t / 7 = y'. A m i n i m u m c a n b e f o u n d w h e n e i t h e r of t h e s e c o n s t r a i n t s b e c o m e s active, b u t t h i s results i n t h e t h i r d fibre d i r e c t i o n b e i n g u n s t r a i n e d . T h i s m a y b e d i s r e g a r d e d since t h e a r r a n g e m e n t so d e r i v e d is a two-fibre s y s t e m in w h i c h t h e d i r e c t i o n s are c o n j u g a t e , w h i c h is precisely t h e case c o n s i d e r e d a l r e a d y i n Section 3.1.2.

O p t i m u m fibre-reinforced sheets for two alternative loadings A n u n c o n s t r a i n e d o p t i m u m is t h u s sought. Since v~ a~ sin ( ~ - 8 ) sin ( ~ - 8 ) = al sin y sin a + a~ cos y cos a

1019

from (8b), the angle 8 which minimizes v~ is t h a t which maximizes [sin ( y - 8 ) sin ( a - 8 ) [. This gives a free m i n i m u m w h e n = 8 = (a+Y)+ 90.

or

8 (a-l-y)

(19)
(20)

Considering the first of these equalities

v~Eef = ~ (al +as)

a 1 sin s (a + y) + a s cos s (a + y) sin 2 ( y - a )

from (Sb). A l t h o u g h the lowest value of v~ is required, this does n o t necessarily correspond to the free m i n i m u m of (20). The two constraints t h a t 8 = (a+y) should be between a" a n d y" each give ~<[al sin s (a+y) + a s cos s (a + y)] s B y t a k i n g sin (y--a) as large as possible, the m a g n i t u d e of the second t e r m on t h e right of (20) m a y be m i n i m i z e d with respect to ( y - a ) . Thus a constraint becomes active, a n d a two-fibre s y s t e m is again derived. E q u a t i o n (18) t h e n becomes 4a~ a~ r e e l = _+( a l + a s ) ~+ (al + a~) + (as - al) cos (a + y)" This form of v depends o n l y on the angle (a + y ) a n d it m a y be shown t h a t the m i n i m u m weight is given b y

v
with

lal]+lasl Es~, o, 8 = 0

a=y=9o
or

a=y=

0 , 8 =

90"

A similar proof m a y be given if 8 = ( a + y ) + 9 0 (ef. e q u a t i o n (19)). Thus the two-fibre system using t h e principal stress directions is lighter t h a n a n y three-fibre system. 3.3. Four fibre directions The first case considered is t h a t in which the principal stresses have the same sign. I n this case, systems with four (or more) fibre directions are o p t i m u m only if the s t r a i n state caused b y t h e load is one of isotropic t e n s i o n or compression. I f in e q u a t i o n (1) t h e direction ~ a n d v o l u m e fraction v8 are regarded as given, t h e n the stress carried b y the 8 fibres m a y be s u b t r a c t e d from the applied stress (al, a s, 0) leaving t h e p r o b l e m of d e t e r m i n i n g t h e fibre directions (a, fl, y) a n d volume fractious (v,, v~, v0 to carry the modified load (al + va el cos ~8, as -+v~ el sin g 8, + v8 el cos 8 sin 8). (21)

I f this modified load has principal stresses of the same sign as (al, as), i.e. if 8 a n d v~ are such t h a t v~< eI(al s i n 2 - ~ a s cos28) I t h e n the p r o b l e m reduces to t h a t of d e t e r m i n i n g a n o p t i m u m three-fibre system, which has already been discussed in Section 3.2.1. Thus there is a n infinity of possible choices

1020

P. BARTHOLOMEW a n d G. Z. HARRIS

for ~ a n d v a a n d each such choice combines with a n infinity of choices of (a, 8, ~'), (v~, v~, vv) a n d gives rise to the m i n i m u m weight already given in e q u a t i o n (17). As in Section 3.2.2, if the principal stresses are of opposite sign, t h e n the o p t i m u m is the two-fibre system in the principal stress directions. 4. O P T I M U M ARRANGEMENTS FOR LOADINGS TWO ALTERNATIVE

4.1. General The lay-up considered is t h a t required to carry either of the loadings a ' = ( ~ , a~, ~ v ) or a" = (a~, a~, T~,). I f a three-fibre lay-up (~, 8, ~) is taken, with the fibre stresses due to each loading being respectively (a'~,a~, a~) a n d (a:, a~, a'~), t h e n (for example) the m i n i m u m for v~ is d e t e r m i n e d b y v~ = But m a x ( l a ~ I , ] a : t ) = 15~[+[a~[, where 5~, a~ are the a-fibre stresses due to applied loadings. a = (~'--a"). ) (23) m a x ( !a~ I, l a~ t ) E~ I (22)

= (-'+~"), I

(24)

5 = (5~, 5v, e~v) a n d a = (a~, a~, T~) represent " m e a n " a n d "difference" loadings. Thus

v~ = ~ - ( 1 ~ 1 + I ~ 1 )
from (22) a n d (23). Similar equations follow also for v~ a n d vr; on s u m m a t i o n
v = v~+v~+v~

(25)

= ~-e ( i ~ l + ] ~ i + l ~ . L ) +

E---~(I La, l+ia~l+lavl).

(26)

The terms on the r i g h t - h a n d side of (26) are obviously the i n d i v i d u a l weights of the two (a, 8, ~) systems which separately equilibrate 5 a n d a. Thus to o b t a i n m i n i m u m weight in (26) it will be sufficient to m i n i m i z e weight for each of 5 a n d a separately, provided t h a t lay-ups (a, 8, )') can be found which simultaneously do this. I n the twoa n d three-fibre systems considered in Sections 3.1 a n d 3.2 the m i n i m u m weight for a single applied load a was, in all cases, found in terms of the principal stresses (al, a2) as EE! Thus, the lowest possible weight to be expected is v = ~--et ( [al I+ [a2 ]+151 [+ [~2[) from (26), always provided t h a t appropriate lay-ups can be found. A n a r g u m e n t similar to the foregoing also holds for two fibre directions (a, 8). I n view of t h e form of e q u a t i o n (24), the pairs of loadings
1

(27)

(a',~"),

(-a',a"),

(-~',-a"),

(~',-a")

all determine the same pair (5, a) of " m e a n " a n d "difference" loadings. Thus a n o p t i m u m derived for a n y one of these (the pair (a', a") in the present case) will also provide a n o p t i m u m for the other three pairs. I n the r e m a i n d e r of this section, a n d in Section 4, the o p t i m u m lay-ups will be classified according to the signs of the principal stresses (5~, 5~) a n d (ap a2). The three separate cases considered will be (i) 5152>0, (~l a2>O, (ii) 51 5~ > 0, a~ a~ < 0 or 5x 52 < 0, o* a 2 > 0, 1 (iii) al 5~ < 0, al az < 0.

Optimum fibre-reinforced sheets for two alternative loadings

1021

4.2. Two fibre directions 4.2.1. 51 5s> 0, al a s > 0. For a two-fibre lay-up (a, fl) to supporg each of the loadings 5 and a, the angles (a, fl) must be chosen to form a conjugate pair with respect to each load separately (cf. Section 3.1.1.). That is 5~tan~tan8+5~-fx~(tana+tanS)
ax t a n a t a n 8 + a~ -- ~ ( t a n

-- 0, /

(28)

a + t a n 8) = 0

from (9) ; when 51 5s > 0, al as > 0 it may be shown that these equations always determine a unique pair of directions (a, 8)" From Section 3.1.1, this pair (a, 8) necessarily corresponds in each case to the optimum weight ( 151 ]+ 15s ] )~Eel and ( [ a 1 ]+ ] a s I )~Eel. I n view of the remarks of Section 4.1 a true optimum is thus derived, v being given by equation (27). I t m a y be noted that this arrangement allows a third fibre direction to be added without changing the statically determinate nature of the lay-up. This additional direction could be used to accommodate a further load, and it would then be possible to consider fibre arrangements for three alternative loadings. This topic is not pursued a n y further in the present paper 4.2.2. 515s<0 or a l a ~ < 0 . I f alas<0
or

withalas>0 w i t h a 10 s < 0

515s>0

then it m a y be shown that the equations (28) again define a unique pair of directions (a, 8) and a two-fibre system exists which equilibrates both 5 and a. However, one of 515s a n d a l a s is negative and the corresponding optimum solution (cf. Section 3.1.2) must then have orthogonal fibre directions parallel to the relevant principal axes; in view of the remarks of Section 4.1, only if the solution of (28) coincides with these orthogonal directions will an optimum be derived. If 515s<0, a l a s < 0 , then equations (28) do not always have a solution; furthermore, even when a solution exists it is optimum only when the principal stress directions of (51, 5~) and (al, as) coincide. Thus, when 51 52 <: 0 or al as < 0 the optimum does not in general correspond to two fibre directions only. 4.3. Three fibre directions 4.3.1. 515s>0, a l a s > 0 . A n y optimum solution must have the property t h a t the angles (a,8, y ) give an optimum for each of 5, a. As shown in Section 3.2.1 there is an infinite n u m b e r of choices in such cases, and several ranges of angles in fact give rise to an optimum. (i) Combination of a pair of two-fibre systems. The simplest system to construct is obtained b y selecting a n angle a together with its conjugate angles 8, Y with respect to the loads 5 and a. From Section 3.1.1, the pair (a, 8) will necessarily correspond to the optimum weight ( 151 [+ [52 I )lEe1 while the pair (a, y) will correspond to the optimum ( [al [+ [ as [ )/Eep Since a is arbitrary, a n infinity of optima is thus derived, the weight in each case being given by equation (27). (ii) Combination of two general three-fibre systems. I f the two angles (a, 8) are taken as given, suppose (a', 8') and (~", 8") to be the conjugates of (~, 8) with respect to 5 and a, respectively. I n view of the remarks of Section 3.2.1, if the ranges (a',8') and (a",8") overlap then any angle ~ in the intersecting range will give an optimum solution. I f only a is regarded as given, then the condition that the two restrictions on y should overlap m a y be used to define a constraint on 8- I n this manner, a n infinite number of optima can be derived for a n y given a. 4.3.2. 515s>O, alas<Oor515s<O, alas>O. Suppose 515s > 0, a l a ~ < 0 . I n this case we seek to use the optimum two-fibre solution for a which (cf. Section 3.1.2) is found when the fibre directions (a,8) and the principal stress directions (al as) coincide. The

1022

P. BARTHOLOMEW and G. Z. HARRIS

angles (a, ]3) so derived m a y be used in a three-fibre system to give an o p t i m u m for 5 p r o v i d e d the t h i r d angle ~ lies between a' and f f (cf. Section 3.2.1). Since the lay-up (~, fl) is o p t i m u m for a and t h e system (a, f~, y) is o p t i m u m for 5, t h e derived system is o p t i m u m w i t h weight given b y e q u a t i o n (27). I f 51 5~ < 0, ~1 a~ > 0 a similar a r g u m e n t holds, except t h a t in this case the two-fibre o p t i m u m is chosen to correspond to 5. 4.3.3. 5 1 5 2 < 0 , a1(72<0. The o p t i m u m for each of 5 and (7 is t h e two-fibre s y s t e m having, in each case, t h e a p p r o p r i a t e principal axes as fibre directions (cf. Section 3.1.2). Since these directions will n o t in general coincide, a n y three-fibre system will be nono p t i m u m , its weight exceeding t h a t g i v e n b y e q u a t i o n (27). 4.4. F o u r fibre directions F o r systems which are n o t statically d e t e r m i n a t e t h e analysis of Section 4.1 does not hold, a n d it can no longer be assumed t h a t o p t i m a can be o b t a i n e d b y superposition, Use m a y be made, however, of t h e slack fibre directions established in Section 2.4 when orthogonal directions are fully strained in opposite senses. 4.4.1. 51 52 > 0 or (71 (72 0. W h e n 51 52 > 0 and (71 (72> 0, an infinite n u m b e r of o p t i m m n four-fibre systems exists. Since o p t i m u m two- and three-fibre systems h a v e a l r e a d y been d e t e r m i n e d for this case in Sections 4.2.1 and 4.3.1 there is little to be gained b y t h e additional d e t e r m i n a t i o n of o p t i m u m four-fibre systems. F o r such a system, t h e choice of angles m u s t be such t h a t w h e n the loads 5 and (7 act separately ~,, e~, %, Sa are strains h a v i n g t h e s a m e sense. Also, in one of the loadings (5 + (7), (O - (7) the strain state should be one of isotropic limiting strain, w i t h

W h e n 51 52 > 0 and 01 u2 < 0, systems h a v i n g four fibre directions are generally less efficient t h a n a n y of the o p t i m u m three-fibre systems. F r o m Section 3.1.2, t h e o p t i m u m for the load (7 requires two fibre directions (f~, fl + 90), stressed in opposite senses, coincident w i t h t h e principal stress directions ((71, (72). I f a t h i r d fibre direction a is added, equilibrium considerations show it to be slack. A n o t h e r slack direction y exists, a n d m a y be t a k e n as t h e f o u r t h fibre direction; strain c o m p a t i b i l i t y considerations (cf. Section 2.4) show t h a t either or

and

] t

and

(29)

N o t e t h a t a and y are n o t necessarily orthogonal. The load 5 will d e t e r m i n e a, y, v~ and %. N o w t h e loaA-carrying c a p a c i t y of the a and y fibres is only fully used if, w h e n 5 acts alone, Also each of ~ , eB+90 h a v i n g t h e same sign as a~ ; the l a t t e r expression follows since one of 5+(7, 5 - ( 7 m u s t fully strain the (fi, rid-90 ) fibres, whence 5 alone c a n n o t fully strain t h e m . F o r t h e angular v a r i a t i o n of strain represented b y t h e foregoing expressions to satisfy strain c o m p a t i b i l i t y considerations t h e n either fl<a<f~d-90 and fl<y<fld-900 ]

or

and f -90<Y

(30

The inequalities of (29) and (30) clearly conflict, whence the o p t i m u m c a n n o t be derived on t h e a s s u m p t i o n of four fibre directions. A similar a r g u m e n t to t h e foregoing holds w h e n 51 52 < 0 and (71 a2 > 0. 4.4.2. 51 52< 0, (Tx(72 < 0. F r o m Section 3.1.2, t h e o p t i m u m for t h e load (7 gives rise to fully strained fibre directions (fl, fl d- 90 ) coincident w i t h t h e principal stress directions ((71, (72), t h e weight of the system being t h e o p t i m u m v a l u e ( [(71 [4-1% [ )/Eet; similarly, the o p t i m u m for 5 gives fully strained directions (a, a 4-90 ) coinciding w i t h t h e principal

Optimum fibre-reinforced sheets for two alternative loadings

1023

stress directions (51, 5~), the optimum weight being ( ]51 I+ ]52 ] )/E@. Since each fibre pair ( a , ~ + 9 0 ) a n d (f~,f~+90) equilibrates its corresponding load, the existence of pairs of slack directions, derived in Section 2.4, m a y in this case be invoked to superpose the two solutions..4, unique optimum is derived having the fibre directions (a, f~, a + 90 , fl+ 90), the optimum weight being given b y equation (27). 5. Y I E L D ENVELOPES

I n this paper, it is assumed to be the fibre strength which limits the load-carrying capability of the sheet. This assumption leads to a yield envelope expressed in terms of Cartesian strains, with each fibre direction introducing a constraint of the form - @~<e~ cos ~a + e~ sin 2 a + ~x~ sin a cos a ~<@ (eL equation (3)). Thus, taking (~, e~, ~ ) as co-ordinates, failure of a fibre system in tension is represented b y a plane passing through the point (@, @, 0), there being a parallel plane on the opposite side of the origin which passes through ( - @, - @, 0) a n d represents failure in compression. Each of these planes is distant @ from the origin. Thus, in the general case o f n fibre directions, the yield envelope will be a solid bounded b y n pairs of parallel planes ; a set of n planes, each passing through (@, @, 0) will delineate tensile failure while a corresponding set passing through ( - @ , - e l , 0) will correspond to compressive failure. Uniformly distributed fibres in all directions (n-~oo) would correspond to the double cone shown in Fig. 3; with a finite number of fibre directions the bounding planes of the yield envelope are all tangential to this double cone.

q~zy

......

\\\

/
I

/._--"

k---

......

,'f.. \
' L--

"'.. ;

"'-.X.

(o, o,-a E~)

FtG. 3. Double cone yield envelope for fibres distributed through all angles. I n the case of two fibre directions, the resultant strain envelope is simply a tube bounded b y two pairs of parallel planes. Fihre-strain assumptions b y themselves supply no constraint for the "scissoring" strain, although clearly a yield envelope having the form of a n infinite tube is not a practical proposition; in practice failure would occur in the matrix or in the matrix-fibre bond. The addition of a third fibre direction introduces a n extra pair of parallel planes, giving a box-qke yield envelope whose six corner points represent fully strained states. The addition of a fourth fibre direction introduces a n additional pair of parallel planes through (st, el, 0) and ( - ~ I , - - @, 0), and so on. By expressing the failure surfaces in terms of stresses rather t h a n strains, load envelopes m a y be derived and compared with the specified restrictions on load. The condition that the lay-up should withstand the alternative loads ~" and ~ merely requires t h a t the envelope should enclose a parallelogram (which is a two-dimensional load envelope) with corners defined b y _+~', _+~". When three fibre directions are used, for

1024

P. BARTHOLOMEW a n d G. Z. HARRIS

e x a m p l e , t h e load e n v e l o p e is of t h e general f o r m i l l u s t r a t e d in Fig. 4; i t is seen t h a t s o m e loads falling o u t s i d e t h e p a r a l l e l o g r a m ( + a', _+a") m a y b e w i t h s t o o d . A d d i t i o n a l l y , t h e e n v e l o p e is a t h r e e - d i m e n s i o n a l one a n d loads c a n b e c a r r i e d w h i c h lie o u t s i d e t h i s p l a n e b u t w i t h i n t h e envelope. I n s o m e o f t h e cases c o n s i d e r e d (in Section 4.3.2 for e x a m p l e ) a n i n f i n i t y of o p t i m a is possible ; clearly in s u c h a case differing yield e n v e l o p e s will follow for t h e differing o p t i m a , a n d in a n y p a r t i c u l a r case t h e choice of fibre l a y - u p could well b e decided b y t h e a b i l i t y t o resist specific loads o u t s i d e t h e design envelope. T~

~m

./

O"

O,m

F I o . 4. L o a d e n v e l o p e of a t h r e e - f i b r e s y s t e m d e s i g n e d to c a r r y t h e p l a n e load
envelope (r t, a't~ -- o J , -- r p .

6. C O N C L U D I N G

REMARKS

This paper shows how optimum fibre arrangements may be derived for reinforced sheets which are subjected either to a single load or to two alternative loadings. Alternative loadings are categorized in terms of a "mean" load ~ and a "difference" load a, the nature of the optimum differing according to the signs
TABLE 1.

Single load
NO.

A l t e r n a t i v e loads
51 ~ 2 > 0 , 0"10"2<0 or O" 0"2 < O, 0"1 0"2 :> 0 1

of fibre directions n = 2 n = 3

a 1 a S> 0 I n f i n i t y of solutions I n f i n i t y of solutions I n f i n i t y of solutions

a 1a s < 0 One s o l u t i o n Solution degenerates ton = 2 Solution degenerates ton= 2

a~ (r2 > 0 O" 0"2 > 0 1

515 2 < 0 O 1 0"2 > 0

One solution I n f i n i t y of solutions I n f i n i t y of solutions

No s o l u t i o n I n f i n i t y of solutions Solution degenerates ton= 3

No solution No solution One solution

n -- 4

Optimum fibre-reinforced sheets for two alternative loadings

1025

of the principal stresses (51, 59) and (al, as). In all cases, optima for alternative loads are derived by superposing the separate optima derived for the individual loads 5 and a, the optimum weight in all cases being the sum of the optimum weights for these loadings taken separately. The general solutions derived are summarized in the Table 1. REFERENCES G. Z. HARRIS,Aeronaut Q. 18, 273 (1967). H. L. Cox, Br. J. appZ. Phys. 3, 72 (1952). H. L. Cox, The Design of Structures of Least Weight. Pergamon, Oxford (1965). H. S. Y. CH~, Q. appZ. Math. 25, 470 (1968). D. C. DRUCKERand R. T. SHIELD,Q. appl. Math. 15, 269 (1957). C. T. MORLEY,Int. J. mech. Sci. 8, 305 (1966).

1. 2. 3.
4.

5. 6.