Pursuant tnthe federal CdpyiifgttAet fUtfa 17 df the Unhed States Code), it is uRlaii#ll to

repidduce, or piihihdy dls|M»r 807 WCMIE |i»r my sy^siwiM pffirHsn

without idw permls^n of die copyr^ oMmer. The OMute> houaniei^)reie#rtE» » fair use diimetliM
has the effect of excusing an act of cofiyright inhfegement. it tsthe intention of the Ubeary to^wUMn
the paremetera of die fair use defense hi fliowintHmifed posttngof eo^Hghted tnatedefe in Secdenie
Course Management s^ms such as diis one. It is fee hHendonoffee Lferery, flioreower, dfet niife

purposes, fiirfeer r^wluc^nofiMs eopintl^

-;^#v

^ --r

I*,
16664- NOVEMBER 1981 ST11

FUNDAMENTAL ANALYSIS OF AGGREGATE

INTERLOCK®
By Joost C. Walraven'

INTRODUCTION

capacity of cracks in concrete to transmit shear forces is of vital importance

many structures. Cracks, which occurred as a consequence of bending moments,
can be subjected to shear forces as a result of other loading configurations.
Such a situation occurs, for instance, in long bridges, supported by hinged
columns in the longitudinal axis, where the torsional moments have to be resisted
Uy the abutments. In order to guarantee the full bearing capacity of the structure,
cpen cracks have to transmit shear stresses. In fully prestressed concrete this
Applies to the ultimate loading state, but in partially prestressed concrete this
mechanism will already occur under serviceabiUty conditions.
Shear forces in cracks can also occur due to other reasons. The anisotropical
Properties of cracked reinforced concrete itself can directly give rise to shear
forces in cracks; this is the case, for instance, in beams without shear rein­
forcement, the shear resistance of which relies, by virtue of redistribution of
forces, for a considerable part on the resistance of the crack faces to shear
displacements (8,22,23).
In the design of relatively simple structural elements subjected to shear, the
shear capacity of the cracks can be used as an active component by applying
the shear friction analogy (3,7,15,16,17). The shear friction concept, however,
applies only to the ultimate loading state and does not relate the stresses on
and the displacements of the crack faces. The interest in the constitutive relations
of cracks in concrete was considerably enhanced during the development of
modem numerical calculation programs, which enable the analysis of the behavior
of stmctures during the whole loading history, provided that adequate formulations
of the basic material properties are available. Because of the lack of a satisfactory
model for the behavior of cracks in concrete subjected to shear forces, widely
varying formulations are encountered. Most authors characterize the shear
resistance of cracks in their finite element analyses by simply reducing the
Presented at the April 18, 1980, ASCE Convention and Exposition and Education
Program, held at Portland, Oreg. (Preprint No. 80-011).
_ Sr. Research Engr., Univ. of Technology, Delft, The Netherlands.
Note. Discussion open until April 1, 1982. To extend the closing date one month,
^ a written request must be filed with the Manager of Technical and Professional Publications,
ASCE. Manuscript was submitted for review for possible publication on December 30,
^ 1980. This paper is part of the Journal of the Structural Division, Proceedings of the
American Society of Civil Engineers, ©ASCE, Vol. 107, No. STll, November, 1981.
^ ISSN0044-8001/81/00n-2245/$01.00. .
2245
2246 NOVEMBER 1981 ST1I
shear modulus G, after cracking, by a factor (0 < < 1). Sometimes
a constant value is chosen for (5,11,14), while in other cases is defined
to be a function of one or more parameters, such as, for instance, the strain
normal to the direction of the cracks, the level of the shear stress, the concrete
quality, and the distribution of crack widths and distances (6,9,20,21).
However, hardly any attention has been paid to the fundamental property
that shear displacement of crack faces at constant crack width inevitably provokes
stresses normal to the crack due to wedging action of the irregular crack faces.
Only recently Bazant and Tsubaki (2) pointed out that if this effect is disregarded,
an unsafe design of structures may result.
Therefore, it should be emphasized that the mechanism may not be represented
by the traditional relation, T = A, in which T = the shear stress; A =
the shear displacement; and B„ = a constant, but that both the shear stress
and the normal stress always have to be taken into account as essential
components. Assuming that the irregular faces of a crack can be deformed,
both the shear stress, T, and the normal stress, CT, are functions of the crack
width, w, and the shear displacement. A:
T=/,(W,A) and a=/„(w,A) . . (1)
The constitutive equations for a single crack can, for instance, be formulated
in an incremental way as proposed by Bazant and Gambarova (1):

....... (2a)
[daj _B,„ B„_ UAJ
in which B„„; B„,; = the crack stiffness coefficients. Differentiating Eq.
1 and comparing the result with Eq. 2, it is concluded that
df„ df, df.
= —= = . .- (2b)
dw dA dw dA
The basic problem is the formulation of Eq. 1. In this paper a fundamental
model is developed, based on a statistical analysis of the crack structure and
the associated contact areas between the crack faces as a function of the
displacements, w and A, and the composition of the concrete mix. Dowel action
can be shown to be of minor importance in practical circumstances (26) and
is therefore neglected; so this study concentrates on the mechanism which is
generally denoted by the term aggregate interlock.

FUNDAMENTALS

Concrete can be represented as a two-phase System: In a matrix (hardened

cement paste) a collection of aggregate particles are embedded. Generally the
strength and stiffness of the aggregate particles are greater than those of the
matrix. However, the contact area between both materials, the bond zone, is
the weakest link of the system. Therefore, cracking occurs commonly through
the matrix, but along the periphery of the aggregate particles. These particles
are simplified as spheres, which can be intersected by the crack plane at all
depths with the same probability. This results in a crack structure, as shown
in Fig. 1.
ST11 AGGREGATE INTERLOCK 2247
Considering this figure and taking into account that the size of most particles
is considerably greater than the crack width, it can be concluded that the
microroughness of the crack, caused by the aggregate particles projecting from
the crack plane, dominates the macroroughness, due to overall undulations of
the crack faces. Therefore, the overall crack plane is assumed to be a flat
plane.
Hardened cement paste is a viscoelastic material; the deformations provoked
by stresses are only partially elastic, for the other part plastic (13). Under
multiaxial stresses, as in the area between the aggregate particles in concrete,
large plastic deformations can occur as a result of pore-volume reduction. Since
the plastic deformations are expected to dominate the elastic deformations,
the stress-strain relation of the matrix material, consisting of hardened cement

FIG. 1.—Generally Observed Structure of Crack Plane

a
pu

FIG. 2.—Rigid-Plastic Stress-Strain Relation for Matrix Material

paste with aggregate particles smaller than 0.25 mm (an arbitrary limit between
the two phases), is assumed to be rigid-plastic, as shown in Fig. 2.
The stress at which plastic deformation occurs is denoted as Thus, it
can be expected that, during shear displacement of the crack faces, contact
areas develop on the surface of the particles, due to plastic deformation of
the matrix. Fig. 3 shows the formation of this type of area as a result of
a shear displacement in the direction of the x-axis. The stresses at these contact
areas produce reactions in the directions of all principal axes. As a result of
the assumption of spherical particles, the resulting component in the Z direction
is zero, just as for a real crack face, if the crack area is not too small. As
a result it is possible to consider a cracked concrete body, as shown in Fig.
4(a), as an assemblage of a large number of slices each of finite width [Fig.
4(c)], and it is possible to derive the overall behavior of the crack by first
2248 NOVEMBER 1981 ST11
studying the properties of this thin slice.
Fig, 5 shows a cross section through a particle lying in a Z-plane in which
there is a line of contact between the opposite crack faces. The projections
of this line of contact on the x and y directions are and Oy. The shaded
area represents that part of the matrix where plastic deformation has occurred.

FIG. 3.—Contact Areas during Sliding

Z-section

iift
a b c

FIG. 4.—(a) Cracked Concrete Body; {b) Z-Plane of Intersection; (c) Representative
Slice

(b)
(a)

FIG. 5.—(a) Contact Area between Matrix and Aggregate; (b) Stress Conditions

If the shear load on the plane of cracking is increased and the crack opening
is counteracted by restraining forces, a mechanism will develop which can be
described as follows. The contact areas tend initially to slide; as a result of
this sliding, the contact area is reduced, which results in too high contact stresses.
Thus, further plastic deformation occurs, until equilibrium of forces is obtained
in the x and y directions.
The stresses at the contact area are resolved into a stress, cr_„, normal to
ST11 AGGREGATE INTERLOCK 2249
the contact area and a stress, Tp„, tangential to this area. The stresses, cTp„
and are interrelated by the condition that the contact areas are about to
slide. Therefore the equilibrium conditions are formulated, based on a uniform
critical stress combination, with
(3),

Next, the components of the contact forces in the x and y directions can
be derived, based on the previous assumptions. Fig. 5(h) shows the equilibrium
conditions at a particle surface. The reactions in the x and y direction can
be formulated as
F y — <j pu'd — T
* pu ^ '

F* = fj pu 'fl+T
' ' pu' ***

If not only the contact line of a single particle is considered, but of all particles
over a unit length of the crack [Fig. 4(h)], so that intersection circles of different
diameters are encountered, the formulation is generalized to
= ^pu^a^ -rp„'2a^ . (5«)
25, (5h)
in which 25,, 25,, = the most probable, average projected contact lengths
over the unit crack length [Fig. 4(h)] considered. If the lengths, 25, and 25^,
are known, it is possible to proceed to surface areas: a unit surface can be
considered to be composed from an infinite number of lines of unit length,
which have all the same expected average values of 25, and 25^. So the most
probable contact areas A ^ and ^4 ^ for a unit crack area are obtained by multiplying
S5, and 25_y by a unit width I, so that = '2a1 and Ay = 25^T. Similar
considerations apply to the transition from the forces 2Fy and 2F, to stresses,
CT and T. Inserting these values and Tp„ from Eq. 3 into Eq. 5, we can write
- 1^^,) ('«)
T= + M.) (6l>)

CALCULATION OF PROJECTED CONTACT AREAS A^ AND Ay

At first the distribution of the circle diameters in a Z-plane, crossed by the

crack, is established. To be able to calculate the distribution of the diameters
of the intersected circles in the Z-plane, a Fuller curve is adopted for the
distribution of the aggregate particles. This cumulative distribution function is
represented by

(7)

in which p = the fraction, passing a sieve with an opening diameter D; and

^max = diameter of the greatest aggregate particle. The probability that
an arbitrary point in the concrete is located in an aggregate particle is denoted
withp — Pk- Properly, p^ is the ratio between the total volume of the aggregate
and the concrete volume. The probability that a point, if located in a particle.
2250 NOVEMBER 1981 ST11
lies also in a particle with a diameter smaller than an arbitrary choosen value,
can be expressed, using Eq. 7, as

p(^D<D,)=pA-p-A (8)

By differentiation of this function the probability density function p'{D^) is

obtained:

P'(DA= dp{D<DA 1 / D, V 1
^ = -1/2 (9)
dD, 2
1
with c =

Subsequently an analysis is made of the probability that an arbitrary point in

a Z-plane lies in an intersection circule with a certain diameter, DQ .The probability [
that a point, if located in a particle with a diameter D^, lies also in an intersection
circle with a diameter > DO{DQ < D^), is equal to the ratio between the volume
of the sphere section A (Fig. 6) and the volume of the hemisphere. The volume
of the sphere section B is equal to

VB = ^'IT h{3a''+ h^) with A = — Z), - "y — Z)'- a' ....... (lOn) |
6 2
while the volume of the hemisphere is
1
irZ)^ ...... (10Z»)
12
The probability that a point in this sphere lies in an intersection circle with
D > DQ is equal to

= ••••••

Substitution of Eqs. 10a and lOh into Eq. 11 and elaboration results in

.....,....(.2)

So the probability that an arbitrary point in a Z-plane [Fig. 4(h)] lies in an

intersection circle with a diameter, D> Do, is obtained by integrating the product |
of Eqs. 9 and 12 over the interval : ,

p,(D>Do)= \
i
^max

Do
p'iDAPDAD>Do)dD,
P'

Inversely, the probability that an arbitrary point in the Z-plane is located in

(13)

an intersection circle with a diameter, D < Do, is obtained from Eq. 13 as

pAD<^o)=^-PciD>Do) (14:
Substitution of Eqs. 9 and 12 into Eq. 13, integration of this expression and
substitution of the result into Eq. 14 yields (26)
ST11 AGGREGATE INTERLOCK 2251

PXD<D^)=P,{\M5 0.053

- 0.012 2)^OZ);;L- 6.0045 Z)o'Z);L-0.0025 (15)

This function is shown in Fig. 7. The density function for the probability that
an arbitrary point in the concrete body, laying in a Z-plane of intersection,
is situated in an intersection circle with a diameter D = DQ is obtained by
differentiation of 15 to DQ:
P'AD,) =p,{0.5^1 D-'' - 0.212 Z)^Z)J, - 0.075
- 0.036 - 0.025 DlD-JZ) (16)
Since the distribution of circles of intersection in the Z-plane is now known,
it is possible to establish the frequency of circles which both lie in the Z-plane

FIG. 6.—Intersection Circle with Diameter, D^, on Particle with Diameter,

p(D<DJ

Q2 0.A 0.6 0.8 1.0 "Q

max

FIG. 7.—Cumulative Distribution Function for Diameter of Intersection Circles

and are crossed by the crack [Figs. Mb) and 8]. The average length s of the
intersection line, AB, for a circle with a diameter, DQ, crossed by the crack
is
IT Z>n
(17)

Considering a crack section [intersection of the crack plane and the Z-plane,
Fig. 4(c)] with a unit length, the probability density function for the expected
part of that length, containing merely points which are located in an intersection
circle with a diameter, DQ, can be obtained by multiplying this unit length with
P'{DQ) (Eq. 16), so that
HDQ)=P'(DQ)-I^P'(DQ) (18)
2252 NOVEMBER 1981 ST11
Consequently, the probability density function for the expected number of
intersection circles with a diameter, DQ in the Z-plane, which intersect also
the unit crack length, can be obtained from Eqs. 17 and 18 as

n(Do) = (19)
0.25IT DQ
In what follows the contribution of the individual intersection circles to the

FIG. 8.—Circles of Intersection Crossed by Crack

Ap .y

//////i Vr^ /////^

777777, 777777

FIG. 9.—Position of Intersection Circle Characterized by u, Randomly Varying between

0 and R

contact area between the crack faces is analyzed. The results of this analysis
will be combined with the results of the previous part (Eq. 19), which will
yield the total contact area for a unit crack area, as a function of the displacements
between the crack faces.
At first the role of a single intersection circle with an arbitrary diameter,
D, is studied. The distance of the center of the circle to the central crack
ST11 AGGREGATE INTERLOCK 2253
line is denoted as u (Fig. 9). This value is assumed to be a random variable
in the interval 0 < « < JR. For —R < M < 0 the circle is in the opposite
crack face, for which the same considerations are valid. If w + i/ > /?, then
for no value of the shear displacement, A, any contact area is formed. If w
+ u < R, then two characteristic values for A can be found. For A < A,
no contact area is found [see also Fig. 10(a)]. If AQ < A < A^, a contact
area is obtained, which increases for increasing value of A. This phase is denoted
as the growing contact phase [Fig. 10(A)]. If A > A^, no increase of contact
area by further shear displacement can be obtained. This phase is denoted
as the maximum contact phase [Fig. 10(c)]: (1) Phase no contact (0 < A <
AQ); (2) Phase growing contact (AQ < A < A^); and (3) Phase maximum contact
(A^ < A). The values AQ, A^, are determined with the help of Fig. 9:
Ao = OA - OB = - (« + wf (20a)
A^ = IRw — w^ (20A)
The associated projected contact areas, a^ and a^, are for 0 < A < AQ
= «3' = 0 (20c)

for An < A < A,.

— (w + A ) > Y W -u (20fi?)
4 Vw'+A^ 2

_ (w^+A^) . ^ ^ +Vi;^-(a + w)' (20e)

2 V 4 Vw^+A^
for A > Aj
a^ = R-{u+w) (20/)
a^ = y/ R^ — {u + w)^ (20g)
Eqs. 20 contain the embedment depth, u, as a variable. However, in order
to get results which can be directly combined with Eq. 19, average values,
Oy and a^, are needed, taking into account the full variation interval of u.
An intersection circle must be taken into account if it offers a contact area
in its most favorable position. It is evident that the most favorable position
is obtained if the embedment depth u is zero. If a circle even in this extreme
position is not in contact with the opposite crack face, it may be excluded
from the calculation. The first demand, if contact is requested, is that R >
w; if i? < w, there is no contact for any value of A, even in the extreme
case that « = 0.
If contact is possible in advance, so that J? > w, an analysis has to be made
of how the circles contribute to the total contact area, taking into account
the different modes of contact shown in Fig. 10. The value i?,, up to which
the radius of the intersection circle has to grow (Fig. 11) to provide at least
one point of contact, can be calculated from Eq. 20a.
For M = 0 one can derive that only one point of contact is obtained:
w^ + A^
=— (21)
2A
 2254 NOVEMBER 1981 ST11
The value i?2, providing an upper limit for the range in which maximum contact
(Fig. 10) is obtained, is found from Eq. 20h:
+ A'
R. = (22)
2^

Comparing Eqs. 21 and 22 it is seen that for A < w, denoted as case A, the
value /?2 is smaller than the value /?,, so that the maximum contact phase
has no significance. So for A < w all circles with a radius /?, < R < R^^^
can contribute (if this really happens depends on the value of the embedment
depth, u). If, however, A > w, case B, then i?2 is greater than /?,, so that
both the growing contact phase and the maximum contact phase are of practical

V77777777

FIG. 10.—Different Phases for w + v < R: (a) Phase No Contact (0 < A < AQ);
(b) Phase Growing Contact (AQ < A < A^); (c) Phase Maximum Contact (A^ <
A)

/y/yxl A
mm

u=0
R\//////

FIG. 11.—intersection Circle in Its Most Favorable Position, u — 0

importance. Further on it must be noted that now the hmit value i?, does
not apply, since contact is always guaranteed for /? > w (see also Fig. 11
for A > w). So for A > w all circles with a radius w < i? < /?2 can contribute
(again dependent on «), and if they do maximum contact is obtained; further
on all circles with a radius i?2 — ^ — ^ma* can contribute, and if they do
growing contact is obtained. The two cases A and B are shown in Fig. 12.
 ST11 AGGREGATE INTERLOCK 2255

The fundamental cases A and B have to be distinguished, establishing the average

expected contributions of the intersection circles to the contact area. This is
done in the following derivation.
"growing^ contact "
A<w
(R = R = »
2 2W 1 2A

"maximum contact" "growing contact"

A>W u
R=W p _ -t-A^
2^

FIG. 12.—Schematlcal Representation of Fundamental Contact Modes for Varying

Values of Radius, R, of Intersection Circles

////////Jh

////////A

FIG. 13.—Calculation of Maximum Embedment Depth, Which Still Contact

Exists

u =0 U = Umax =R

FIG. 14.—Calculation of Average Contribution of Circle with Radius, R, to Contact

Length,

Case : A < w.—If there is any contact area, this is at least the case for
the minimiim embedment depth, « = 0. Solutions are found up to an upper
bound, derived using Fig. 13. For constant values of A, w and R, the
variable uls increased so far that only a single point of contact remains; in
 2256 NOVEMBER 1981 ST11
that stage is reached. With the help of Fig. 13 one can derive this value
"max to be equal to

+ A^) + -\/w^{w^ +A^)^ - + A^lKw^ + A^)^ - 4A ^R^\

(23)
w +A'
So values for and are found for the range
0<M < (24)
The probabihty density function for the occurrence of a value, u, is assumed
to be equal to
1
P(«) = J (25)

The expected value for the average contribution of a circle with radius, R,
to the contact lengths, and can be formulated, using Eq. 25, as shown
in Fig. 14
1 r " = "max
= —\ ^yR du (26)
^ J «=o
1 r " = "max
and =— \ a,^du (27)
^ J «=0
in which Oy^ and a^jf = the contact lengths, Uy and a^, for a circle with radius,
R, according to Eqs. 20d and 20e.
Substitution of Eq. 20d in Eq. 26 and Eq. 20e in Eq. 27 and integration
results in

.l/,!,, A Iw 1 «^ax
=7 7 T7=f==Tr-«max - T7 ««ax " T ^^S)
R y 4 Vw+A 2 R 2 R

1 / 2 1 2 2 ^
— A-\ R^ (w" + A^)-^ ,
2 V 4 Vw'+A' R

VR^ - (W + Mmax)^
2R

w J— R w+ R w
vR — w 4 arcsin arcsin — (29)
2R 2 R 2 R
Case B: A > w.—In Fig. 12 it is seen that for /? > (w^ + A^)/2w the calculation
can be carried out in the same way as in case A. For the range
w' +A-^
w < R< (30)
2w
the maximum contact phase is valid. Similarly, as in the growing contact phase,
a circle is in contact with the opposite crack face if the embedment depth.
 ST11 AGGREGATE INTERLOCK 2257
u, is greater than zero and smaller than a certain upper bound. In Fig. 10(c)
it can be seen that this upper bound is obtained for u = = R — w. For
values of i? in the range indicated by Eq. 30, and are obtained by
substitution of Eq. 20/in Eq. 26 and Eq. 20g in Eq. 27 and integration: This
results in
(R-w) J
{R — u — w)du = (i? — (31)
0

_ IT w R w
——R V R —w arcsin — (32)
4 2/? 1 R
For the range R > (w^ + A^)/2w Eqs 28 and 29 are valid.
For a unit length of the crack line, line of intersection of the crack plane
and a Z-plane [Fig. 4(c)], it was shown that the probability density function
for the expected number of circles with a diameter, /)„, intersected on this
length, can be expressed by Eq. 19. The total contact areas in the X and y
directions, provided by all circles intersecting the unit crack length, can be

•s:=s:
expressed as

( ^max
niR) a^j^ dR (33)

\ n{R) a^^ dR (34)

^ ^min
Substitution of the concerning values for /?„;„ and R^^^ (Fig. 12), the associated
expressions for dy^ and (Eqs. 26-29) and n{R) (Eq. 19) in Eqs. 33 and
34 and some rearranging of terms results in the final expressions for the projected
contact areas Ay (= l^dy) and (= '^d^), as a function of crack width, w,
and shear displacement A for a crack area with a unit length and a unit width:

Case A: A < w

= S ^max

.2,^2
A
^
4 f D \
]G,(A,w,D)dD (35)
j

A^= " " Pk ^("T ] G2(A,w,DydD (36)

\2?„,ax /

Case B: A > w
f 4 / D \
Ay,= \ p,— F(-^]'G,(A,w,D)dD
J 2H. \^max /

+
S^m*x

+
Pk-^P\
4

F
( D

\-^max /
\
yGi{A,w,D)dD (37)
 2258 NOVEMBER 1981 ST11
_ f 4 / 2) \
\ Pk G4 (A ,w,D)- dD
J 2w '''' \ ^max /

r ^max 4 / D \
\ P, — F[- .G,(A,w,i)) dD (38)
J W2 + A2/W \-^max /

with G i(A,w,D) = D \/Z)'-(w^+A')

V +A^
- max -u
••max

^ ' • iv
GJCA, W,D) = D' - (Vf^ + A^) / 2 , X2 '"max + ("max + W)
V w +A

'sj—D^-iw + u^^^f- w •\j^D^- w 2 H ^2

D arcsin
w + u,

2M;-I
— arcsin — dD . (39)
4 Z)

Gj(A w.B) = Z) Z) - iv)

G,(\W,D) = D-'(^D'-WJ-D'-W'-— arcsin —)

-0.036[-^] -O.OZsf-^Y"
V;>„„ / vz>„„ /

= volume aggregate/volume concrete

"max is expressed in Eq. 23
Integration of Eqs. 35-38 has been carried out in a numerical way with an
Algol program (25).
Fig, 15 shows, as an example, the calculated relations for and A^, or
SOj, and Id^ for a unit crack area of 1 mm^ (0.00155 in.^), as a function of
crack width, M;, and shear displacement, A, for a concrete mixture with a maximum
aggregate particle diameter of 32 mm (1.26 in.) and a value of 0.75. The
shear and normal stress across the crack faces are related to the contact areas
 2259
AGGREGATE INTERLOCK
ST11

IAy
(miT?) if
^ CN
0.016-
(10.3)
.5. .c-v .<v\
0.012.
(7.7A)

0004
(5.2)

Cm.)
0.6 0.8 1.0 1.2 1.A 16 (mm)
shear displacement A —»

Aj, in.^(mm^)

FIG. 15-—^Total Projected Contact Areas, and ^;,/for 1-mm Crack

of Crack Width, w, and Shear Displacement, A, Calculated with Eqs.

•—CTJTSW E
& ft .E
o
1 ^
^ II o
1 II crack o
L.
3
i' .
•U O restraint bors
m

HG. 16.-Pushoff Tests on Externally Restrained Specimens

r end A with Eq. 6. In order to evaluate the theoiy and to establish the
' V and p. a number of experiments have been earrted out.
na
 2260 NOVEMBER 1981 ST11
were the concrete strength and composition, the external restraint stiffness and
the initial crack width. Five different concrete mixtures were tested. Three
of these had the same maximum particle size of 16 mm (0.63 in.) but different
cube crushing strengths of/;^ = 13 N/mm^ 37 N/mm^ and 59 N/mm^ (1,915
psi, 5,286 psi, and 8,428 psi), one had a maximum aggregate particle diameter
of 32 mm (1.26 in.) and a strength of 33 N/mm^ (4,714 psi), and one was
a sanded lightweight concrete. The aggregate in the first four mixes was distributed
according to a Fuller curve, to enable a direct comparison between theoretical
and experimental values. Detailed data about these experiments are found in
Ref. (25).

COMPARISON BETWEEN THEORY AND EXPERIMENTS

The relations between the stresses in a crack on the one hand and the
displacement components on the other hand have earlier been formulated as
(see also Eq. 6)
^ 1
T = ^pui^y + \X-A^)i •
in which and Ay depend on the crack width, w, the shear displacement,
A, the maximum particle diameter, f),nax the total aggregate volume per
unit volume of the concrete, as expressed in the functions, Eqs. 35-38.
The parameters, the matrix yielding strength, and p., the coefficient of
friction, are established by fitting Eqs. 40 to the experimental results. It appears
] that the best results are obtained for a friction coefficient of p, = 0.4 for all
mixes. This value is of the same order as was experimentally established by
Weiss (27) in friction tests on concrete, mortar, and particle surfaces. The matrix
yielding stress, CT^„, which has to be inserted to get optimal fitting, is a function
of the uniaxial concrete strength. The best results are obtained for
or^„ = 6.39/;°''^N/mm^ (56.7/;° "psi) (41)
The matrix yielding strength turns out to be somewhat higher than the strength
of the concrete itself. This must be considered as regular: The weakest link
of a hardened concrete is the interface between aggregate particles and matrix,
where microcracks initiate the deterioration of the concrete; as a result the
concrete strength is lower than the strength of its constituting components.
Also the fact that the ratio between matrix strength and concrete strength decreases
with increasing concrete strength, as results from Eq. 41, is a generally observed
phenomenon (4,19). All experimental results could be well described with Eqs.
35-39 and the material constants, p, = 0.4 and cr^^, from Eq. 41. Examples
are given for two mixtures (Figs. 17 and 18). It must be emphasized that, to
fit Eqs. 40 to the experimental results, only two degrees of freedom exist,
p., so that only two lines per diagram (e.g., for w = 1.0 mm) can actually
be fitted. The fact that automatically aU other lines are found to fit the experimental
results supports very well the validity of the theory.
FURTHER ANALYSIS OF AGGREGATE INTERLOCK

The model which has been developed is in adequate agreement with the
 ST11 AGGREGATE INTERLOCK 2261

f^=8A43 psi (59.1 N/mm^); D„„ = 0.63 In. (16 mm)

(psi) (N/mnri^)

W=a035 in
Q9mm(Al
w=0.039in.
in 600 1.0inm(®)

K 600

S 400
in 200
ops (in.)

displacemen t

w=Q039 in
1.0 mmC«)
=0035 In.
0.9 mm M

(NImrf)

FIG. 17.—Comparison between Experimental Values for Concrete with = 59

N/mm^ (8,428 psi), = 16 mm (0.63 in.), = 0.75 and Theoretical ^Values,
with o^„ = 65 N/mm^ (9,285 psi) and p. = 0.40

(psi) (N/mrrft ^ = 4771 psi (33.4 N/mm^) ; D = 1.26 in. (32 mm)
1800
.~.vj

* .0.035in./0.9mm(A)
's0.039in.'l.0mm(")

0Q3 0P4 005 006 007 008 (inl

20 (mm)
shear displacement A

<5 600
•K 800

nllOOO
E
O1200

4>s0 (N/mm^)

FIG- IS-—Comparison between Experimental Values for Concrete with/' = 33.4

fSI/rnm^ (4,714 psi), = 32 mm (1.26 in.),= 0.75 and Theoreticaf Values,
with cr^u = 44 N/mm^ (6,286 psi) and p, = 0.40
2262 NOVEMBER 1981
experimental results. As such it can be used for parameter studies, yielding
a better insight into the fundamental mechanics of aggregate interlock.

=a63 in. (16 mm); qpu=7143psi (50 N/mm^); f^=0-75

(N/mm^)
•w=01008 in. ((X2 mm)

in. ©.6 mm)

'|i=O.A
W = 0.039 in.
(10 mm)
|i=O.A

0.03 Op A 0.05 0.06 Op7

0.Q8 (in)
2D (mm)
displacement A

tt039ln.(l0mrri
(0.2 mm) |i=O.A
U=0
(N/mm^)

FIG. 19.—Role of Friction between Aggregate Particles and Matrix in Transfer of

Stresses across Crack, for Concrete with = 40 N/mm^ (5,714 psi), = 16
mm (0.63 in.), and = 0.75, Calculated with Theoreticai Model

shear stress X
(N/mn£)
w.OJO* In. (0.1 mm)!

0*® Onux
^ DmIX
•® Om»x lw=0.024in. (0.6 mm)l
5°ini«
4 Dm.. I w.0.039 in. (10 mm)|

0.3 0„
..
_ _0.6 Dmix ' •0.8 D„

=0.4_p„,, --a6D„

•0X)»020„„ (in) ,
oirt '—t A.
0.1 D
0.4 "max
0.6
<XC-0.6 i'mw
D,
-CD-eo
1.0 D
- 10 D„.,
(NAwri')
normal stress 0

FIG. 20.—Contribution of Various Aggregate Fractions to Transmission of Stresses

across Cracks for Crack Widths w = 0.1 mm (0.0039 in.); w = 0.6 mm (0 0236 inl­
and 1.0 mm (0.039 in.) for a Concrete with/^, = 30 N/mm^ (4.286 psi) D =
32 mm (1.26 in.) and = 0.75 Calculated with Theoreticai Model'(Inserted ValLes
= 44 N/mm (6,286 psi) and ji = 0.40)

Role of Friction between Aggregate and Matrix.—It was shown that equi­
librium in the contact area was obtained by combinations of normal (yielding)
stresses and shear (friction) stresses. It was shown that a friction coefficient
equal to 0.4 resulted in the best fitting of the curves to the experimental results.
I ST11 AGGREGATE INTERLOCK
By conducting a calculation with a friction coefficient. ,i - 0. the innucnc^
of faction can be visuaUzed. A calculation is carried out for a mixture wnh
2263

I maximum a^regate size, /)_ = 16 mm. = 50 N/mm\ p, = 0.75 and

I R 0.0 and 0.4. The results of the calculation are shown in fig. 19 for some

?039 ^
, ^ considerable part of the shear resistance is provided by
friction. If this friction would be eliminated e.g., by oil between the crack
faces, and the external restraining stiffness would remain the same (compare
equal <r-w combinations in Fig. 19), the shear stress which can be transferred
is considerably smaller (even more than 100% for small crack widths), and
equilibrium is obtained for smaller values of the shear displacement. A. which
shows also that more overriding of particles and less deformation of the matrix
will occur.
Contribution of Various Aggregate Fractions to Transmission of Stresses In
Crack. By a slight modification in the derivation of the equations, representing
the relations between stresses and displacements in the crack, it is possible
to find the contribution of only a part of the aggregate particles, l-q. 15 was
found by integrating the product of Eqs. 9 and 12 over the interval. l)„
-C'max- By integrating only over the range, Z)„-» n n < I, the influence
of aggregate fractions is visualized. An example is given in Fig. 20: With the
modified expressions 35-38 the contribution of fractions has been established
and represented for a small [0.1 mm (0.0039 in.)], an average (0.6 mm (0,02.36
in.)], and a^great [1 mm (0.039 in.)] crack width, for a mixture with a cube
crushing strength of= 30 N/mm^ (4286 psi) and a maximum particle diameter
of = 32 mm (1.26 in.) (inserted values {T^„ = 44 N/mm' (6286 psi). u
= 0.40, = 0.75).
The lines, represented in these figures, describe the relations between o.
T, w, and A if only the particles with a diameter between 0 and a s-arying
fraction of are considered. It is seen that the small aggregate fractions
loose importance if the crack width becomes greater.
Influence of Size of Aggregate.—To obtain an impression about the influence
of the size of the aggregate, two mixtures have been compared. Both ha>c
the same properties, except for the maximum particle diameter, which is once
16 mm and once 32 mm. The results of this comparison arc shown in I ig
21. It is seen that the normal stress, a, is not very sensitive for this variation,
but that the shear stress, T, is influenced, the more if the crack width is greater.
This tendency was confirmed by the results of the experimental part o t is
investigation.
Influence of Grading Curve.—In the previous analyses and in the experiments
always a Fuller curve was adopted. However, in practice most ct cs a
a permissible grading curve region. The ideal Fuller curve is near to t e
boundary of this area. To study the influence of the gra
is chosen which approximates to the upper limit given tn t c i
of Practice, the VB' JA, for = 16 mm (0.63 in.). c^^nfirminr
The relations between stresses and displacements for c Other values
to the grading carve B were calculated, as show., tn
adopted were p. = 0.75, u = O '"'. «"<• "7- " . p ,hc comparaMe
concrete has the same maximum particle dia . ma.
 2264 NOVEMBER 1981 ST11
Fuller mix, it contains a much higher proportion of sand particles.
The results previously obtained under the aforementioned section on the
transmission of stresses, where the contribution of the individual aggregate I
fractions to the transfer of stresses in a crack was established, were used for
shear stress X
(psi) (N/mm^ s 1.26 in. (32 mm)
02 02 =0.63 In. (16 mm)
1200

600

crack width w (mm)

(1mm =0.039 In.)

003 0.0A 005 0.06 0.07 0.06 (In)

(mm)
displacement ii

05 0-6 0.7
,1 0.2
(psi) (N/mm')
normal stress 0

FIG. 21.—Influence of Size of Aggregate on Transmission of Stresses across Crack

for Concretes with/;, = 30 N/mm^ (4,286 psi), = 0.75, = 16 and 32 mm \
(0.63 in. and 1.26 in.). Calculated with Theoretical Model (Inserted Values o- „ = /
44 N/mm ^ (6,286 psi) and |x = 0.40) ^ '

shear stress X
sieve aoerture -•
(N/mm*)
0.1 0.1 QL2 l.llQ2

02 OJ* 06 OS 1.0
ro D/D

crack width (mm)

(1 mm s0039 In)

0.2 0.3
grading curve A
NArim^) grading curve B
normal stress 0

FIG. 22.—Influence of Grading Curve on Transfer of Stresses across Crack for Two
Comparable Mixes Confirming to Different Grading Curves, Calculated with Theoreti­
cal Model (Inserted Values = 32 mm (1.26 \n.),p^ = 0.75, and |x = 0.40)

the calculations. The results for the concrete mix with the grading curve B,
are shown in Fig. 22 by the dashed lines, and are compared with the results
for the Fuller-type concrete, calculated earher (Fig. 21).
 ST11 AGGREGATE INTERLOCK 2265
It is seen that in both cases the influence of the sieve line on the normal
stresses, CT, is not great, but is significant for the shear stresses. The most
pronounced differences are obtained for greater crack widths. This can be
expected, since the sand mixtures according to grading curve B provide a smaller
potential contact area at greater crack width.
Cyclic Loading.—From tests (12) it is known that in the case of cyclic loading
a considerable difference exists between the behavior of the crack plane during
the first loading cycle and the subsequent cycles. The shear stress/shear
displacement relation of the initial cycle is nearly linear, and after unloading
a considerable amount of hysteresis is observed. The shear stress/shear displace­
ment relationship for the later loading cycles is highly nonlinear and a hardening
type of behavior is observed. This overall behavior can be explained with the
theory.
This is done on basis of a fictitious specimen (Fig. 24) with a concrete (quality
oif[^ = 33.4 N/mm^ (4,770 psi) and a maximum aggregate particle diameter

FIG. 23.—Three Characteristic Stages during First Loading Cycle: (a) Before Loading;
{b) at Peak Shear Stress; (c) after Unloading

shear stress X
(psi) (N/mrr?)

shear.
[./ displ. A
0.6 (mm)
loading (0JQ16) 0024) (in.)
crack plane

(psi) (N/mm')

FIG. 24.—(a) Fictitious Specimen; (h) Calculated Response to Cyclic Loading

of 32 mm (1.26 in.), so that . Fig. 18 can be used, a preset crack width of

Wj, = 0.5 mm (0.020 in.), and external restraint bars, providing an equivalent
restraint stress of 0.5 N/mm^ (71 psi) for a crack width increment of dw =
0.1 rnm (0.004 in.). The specimen is subjected to a maximum (cyclic) shear
stress of T = 3 N/mm^ (430 psi). Fig. 23(a) shows the positions of the crack
faces before loading. When the shear load is increased, the crack faces engage
in contact. Fig. 23(b). The relation between T, shear stress, and A, shear
displacement, can be calculated using Fig. 18. If the maximum shear stress
is reached, point A in Fig. 24(b), the friction between particles and matrix
is still a maximum, p- = 0.40. If the shear force is decreased, a relative movement
 2266 NOVEMBER 1981 ST11
between the crack faces can only occur if the maximum amount of friction
in the opposite direction is exceeded. With the theoretical model it was calculated
that this point is reached when the shear stress is reduced to 0.50 N/mm^
(71.4 psi), point B in Fig. 2^{b). Fig. 23(c) shows that, after unloading, the
no contact phase is reached before the shear displacement is back in the neutral
position. This no contact phase is considered to be reached if in the most
favorable case, R = \/2 « = 0, no contact exists any more. For the
case considered it was calculated that the remaining shear displacement, A,,
is equal to 0.41 mm (0.016 in.), point C in Fig. 2A{b).
To bring the two halves of the specimen back in their neutral position a
small shear force may be necessary, since the rubble between the crack faces
due to the deterioration of matrix material during loading, may cause some
frictional resistance, point D in Fig. 2A(b). If the shear force is applied in the
opposite direction, the same type of behavior can be expected since those parts
of the crack surfaces which get in touch in this reversed cycle are not yet
damaged (Fig. 23). Thus, a similar loading and unloading curve can be expected
(Fig. 24(6), points A', B', C, D').
In the subsequent loading cycles the presence of the cavitations, worn out
in the first cycle of loading, influence the behavior of the specimen considerably.
At first a shear displacement will occur under a low shear force, until contact
between the opposing areas is obtained (A > 0.41 mm, point E). Then in a
short interval of A full contact between the cavitations will be obtained. In
this short interval a proceeded wearing off will take place at places of high
contact stresses, point x in Fig. 23(c). Hence a steeply ascending branch (EFG)
may be expected, slightly shifted from the previously obtained loading line.
At unloading a similar behavior as in the first cycle may be expected, GHI,
Fig. 24(6). A comparison of this constructed T-A relation with experiments,
carried out by Laible, White, and Gergely (12), shows a rather good agreement.

CONCLUSIONS

1. The mechanism of aggregate interlock can only be adequately described

if normal stress, shear stress, crack width, and shear displacement are all involved.
2. Aggregate interlock is characterized by two fundamental modes of behavior:
sliding at the contact area between particles and matrix at opposite sides of
the crack {overriding) and irreversible deformation of the matrix by high contact,
stress.
3. The experimental results can be adequately described by a physical model,
simplifying the aggregate particles to rigid spheres of different sizes, protruding
from a flat crack plane, and adopting a rigid-plastic stress-strain relation for
the matrix between the particles.
4. All particle fractions contribute quahtatively in the same way to the transfer
of stresses across the crack; only fractions containing merely particles with
a diameter smaller than two times the crack width may be considered to be
inactive.
5. For the friction between particles and the matrix, a friction coefficient,
)x = 0.4, independent of the concrete quality and the level of the stresses,
gives the best results. The yielding strength of the matrix, which has to be
used in order to get the best results, is a function of the concrete strength.
 ST11 AGGREGATE INTERLOCK 2267
The values for both the friction coefficient and the matrix yielding strength
are in good agreement with expectations, based on experiments described in
( the literature. . r
i 6. The resistance of the crack faces against shear displacements is most or
all governed by the strength of the concrete. The size of the aggregate particles
[ in the range tested, = 16-32 mm (0.63-1.26 in.), plays a subordinate role.
7. The grading curve of the aggregate has some influence: A theoretical
f comparison between two mixes, one composed according to a Fuller curve
I (large particles) and the other containing a large sand fraction, displays an influence
on the shear stiffness at larger crack widths. The development of the normal
I stresses is hardly influenced.
! 8. If the friction between the crack faces would be reduced e.g.,_ by oil,
I this would result in a considerable reduction of the resistance against shear
j stresses and an increase of the overriding at the expense of the deformational
, mode of behavior.
9. The behavior of cracks, subject to cyclic loading, as known from the
literature, can be well explained and described by the theoretical model, at
least in a qualitative way.

SUMMARY

A mathematical model has been developed describing the mechanism of

aggregate interlock in cracks. The model is based on the behavior on particle
level, taking into account the deformation of the hardened cement matrix and
frictional forces between aggregate particles and matrix during sliding. The
relations between the displacements of the crack faces and the stresses across
the crack have been related to the structure of the crack faces, which has
been established with a statistical analysis. The results of the theoretical model
are in adequate agreement with the results of experiments, carried out separately.
Thus, the model enables a further analysis of the mechanism of aggregate interlock.
In this way insight could be obtained into the role of the aggregate distribution,
the size of the aggregate, the contribution of aggregate fractions, and the role
of deformation and friction on particle level. The model makes possible an
explanation of the experimental tendencies, found in tests of other investigators
for instance, for cyclic loading.

ACKNOWLEDGMENT

The financial support and encouragement of the CUR, Research Committee

of the Dutch Concrete Association, is gratefully acknowledged.

APPENDIX I.—REFERENCES

I Razant Z P and Gambarova, P., "Rough Cracks in Reinforced Concrete," Journal

'• ^/Zk'ru:,urTDMsion. AS6E, Vol. 10«, No. ST4, Proc. Paper 15330, Apr., 1980,
, P and Tsubaki, T., "Optimum SUp-Free Limit Design of Concrete
IScSg of ,He sJucural D„islon. ASCE, Vol. 105, No. ST2,
Proc. Paper 14344, Feb., 1979, pp. 327-346.
2268 NOVEMBER 1981
3. Birkeland, P. W., and Birkeland, H. W., "Connections in Precast Concrete Construc­
tion," Journal of the American Concrete Institute, Vol. 63, No. 3, Mar., 196 , pp.
345-368. „ ,
4. Bremer, F., and Steindorfer, P., "Bruchfestigkeit und Bruchverformung von ^e^on
unter mehraxialer Belastung bei Raumtemperatur," Deutscher Ausschussfitr Stahlbe-
ton. Heft 263, Ernst und Sohn, Berlin, West Germany, 1976. ,^ »
5. Cedolin, L., and Dei Poli, S., "Non-Linear Plane Stress Analysis of Reinforced Concrete
by the Finite Element Method," Construzioni in Cemento Armato, Studi e Rendiconti,
Politecnico di Milano, Vol. VIII, Milan, Italy, 1977, pp. 3-33.
6. Cedolin, L., and Dei Poli, S., "Finite Element Studies of Shear Critical R/C Beams,
Journal of the Engineering Mechanics Division, ASCE, Vol. 103, No. EM3, Proc.
Paper 12968, June, 1977, pp. 395-410.
7. Fattah-Shaikh, A., "Proposed Revisions to Shear-Friction Provisions, Journal oj
the Prestressed Concrete Institute, Vol. 23, No. 2, Mar./Apr., 1978, pp. 12—21.
8. Fenwick, R. C., and Paulay, T., "Mechanisms of Shear Resistance of Concrete Beams,
Journal of the Structural Division, ASCE, Vol. 94, No. STIO, Proc. Paper 2325,
Oct., 1968, pp. 2325-2350. • t, • u
9. Geistefeld, H., "Stahlbetonscheiben in gerissenem Zustand-Berechnung mit Beruck-
sichtigung der rissabhangigen Schubsteifigkeit im Materialgesetz," thesis presented
to Technische Universitat Braunschweig, at Braunschweig, Germany, in 1976, in partial
fulfillment of the requirements for the degree of Doctor of Philosophy.
10. Goto, Y., "Cracks Formed in Concrete around Deformed Tension Bars," Journal
of the American Concrete Institute, Proceedings, Vol. 68, No. 4, Apr., 1971, pp.
244-251.
11. Hand, F. R., Pecknold, D. A., and Schnobrich, W. C., "Non-Linear Layered Analysis
of RC Plates and Shells," Journal of the Structural Division, ASCE, Vol. 99, No.
ST7, Proc. Paper 9860, July, 1973, pp. 1491-1505. .
12. Laible, F., White, R. N., and Gergely, P., "An Experimental Investigation of Seismic
Shear Transfer across Cracks in Concrete Nuclear Containment Vessels," A CI Special
Publication SP-53, Detroit, Mich., 1977.
13 Locher, F. W., and Wischers, G., "Aufbau und Eigenschaften des Zementstems,
Zement Taschenbuch 1976/1977, NCH Nederlandse Cement Handelsmaatschappij
B.V., The Hague, The Netherlands, 1977, pp. 43-59.
14. Lin, C. S., and Scordelis, A. C., "Non-Linear Analysis of RC Shells of General
Form," Journal of the Structural Division, ASCE, Vol. 101, No. ST3, Proc. Paper
11164, Mar., 1975, pp. 523-537.
15 MacGregor, J. G., and Hawkins, N. M., "Suggested Revisions to ACI Building Code
Clauses Dealing, with Shear Friction and Shear in Deep Beams and Corbels," Journal
of the American Concrete Institute, Proceedings, Vol. 74, No. 11, Nov., 1977, pp.
537-545. ^ . ,, , I
16 Mast, R. F., "Auxiliary Reinforcement in Concrete Connections, Journal oj me
' Structural Division, ASCE, Vol. 94, No. ST6, Proc. Paper 6002, June, 1968, pp.
1485-1504. ^ ,
17 Mattock, A. H., "Shear Transfer in Concrete Having Reinforcement at an Angle
to the Shear Plane," ACI Special Publication SP-42, Shear in Reinforced Concrete,

18 Rei^ardt!^H!*^W., and Walraven, J. C., "Shear Transfer in Reinforced Concrete

with SmaU Crack Widths," presented at the April 14-18, 1980, ASCE Convention,
Portland, Oreg. (Preprint 80-012). „ „ n. v,^l 97
19. Reinhardt, H. W., "Anspriiche des Konstrukteurs an den Beton, Beton, vol. z.i,
' No. 5, May, 1977, pp. 195-199. „ ^ the
90 Schafer H "Zur Berechnung von Stahlbetonplatten, thesis presented to tnc
' Techniske Universitat Darmstadt, at Darmstadt, W. Germany, in 1976, m partial
fulfillment of the requirements for the degree of Doctor of Philosophy. , y,. "
21 Schimmelpfennig, K., "Bruchsicherheitsberechnung von Stahlbetondruckb^altern,
D^Z7AZc%ssfur Stahlbeton, Heft 257, Ernst and Sohn, Berlin, West Germany,

99 iwfmv R N and Andriopoulos, A. D., "Contribution of Aggregate Interlock and

Dowel Forces" to the Shear Resistance of Reinforced Beams without Web Rem-
 ST11 AGGREGATE INTERLOCK 2269
forcement," ACI Special Publication SP-42, Shear in Reinforced Concrete, Vol. I,
1974, pp. 129-166.
( 23. Taylor, H. P. J., "The Fundamental Behavior of Reinforced Concrete Beams in
Bending and Shear," AC/ Special Publication SP-42, Shear in Reinforced Concrete,
Vol. I, 1974, pp. 43-78.
24. Walraven, J. C., Vos, E., and Reinhardt, H. W., "Experiments on Shear Transfer
in Cracks in Concrete, Part 1; Description of Results,'' Report 5-79-3, Stevin Laboratory,
Delft University of Technology, The Netherlands, Dec., 1978.
25. Walraven, J. C., "Experiments on Shear Transfer in Cracks in Concrete, Part II:
Analysis of Results," Report 5-79-10, Stevin Laboratory, Delft University of Technolo­
gy, The Netherlands, Nov., 1979.
26. Walraven, J. C., "Aggregate Interlock: A Theoretical and Experimental Investigation,"
thesis presented to the Delft University of Technology, at Delft, The Netherlands,
in Oct., 1980, in partial fulfillment of the requirements for the degree of Doctor
of Philosophy.
27. Weiss, R., "Ein Haufwerkstheoretisches Model der Restfestigkeit geschadigter Be-
tone," thesis presented to the Technische Universitiit Braunschweig, at Braunschweig,
West Germany, in 1978, in partial fulfillment of the requirement for the degree of
Doctor of Philosophy.

APPENDIX ii.—NOTATION

The following symbols are used in this paper:

A, = expected average value of projection on y-plane of total contact

area, obtained over unit crack area, for certain combination of
w. A;
A.. = expected average value of projection on the x-plane of total contact
area, obtained over unit crack area, for certain combination of
w. A;
= projected contact length in z-plane on x-axis;
projected contact length in z-plane on x-axis for circle with radius
R;
= expected average value of a^^;
Oy = projected contact length in z-plane on y-axis;
= projected contact length in z-plane on x-axis for circle with radius
R\
dyR = expected average value of ay^,
... = crack stiffness coefficients;
...
D = diameter of intersection circle;
D = diameter of maximum intersection circle (maximum particle);
Dy = arbitrary particle diameter;
Do = arbitrary particle diameter;
F = force in X direction per unit width in z direction;
p — force in Y direction per unit width in z direction;
G — shear stiffness modulus of uncracked concrete,
= relative aggregate volume fraction;
R = radius of intersection circle or particle;
o = maximum value of R',
R^" = minimum radius of an intersection circle for which contact can
' occur at a certain w, A combination;
R = maximum radius of an intersection circle for which maximum
2270 NOVEMBER 1981 ST11
contact mode can occur at a certain w, A combination;
u = embedment depth, defined as distance from center point of
iijtersection circle to central crack face (see, e.g.. Fig. 10);
"max = maximum value of u for which, at certain w, A, R combination
contact is possible;
w = crack width;
= shear stiffness reduction factor;
A = shear displacement;
AQ = shear displacement below which no contact is possible at certain
combination of u, w and R ;
A^ = shear displacement above which the maximum contact mode
occurs at certain combination of w and R;
(jL = coefficient of friction between particles and matrix;
a = normal stress;
CT = normal stress at plastic deformation of matrix;
T = shear stress; and
T_„ = shear stress at plastic deformation of matrix.