Professional Documents
Culture Documents
by
November, 1964
2
ABSTRACT
Mr. R. Glenn, who made much of the apparatus, and Mr. J. R. Turner,
who helped in the preparation of the specimens.
Miss M. Gregory for the care and ability with which she
proof-read the original manuscript; Miss P. Kerridse for her
painstaking care in typing the thesis; and Miss M. J. Burns for
proof-reading the stencils.
TABLE OF CONTENTS
PAGE
ABSTRACT 2
ACKNOWLEDUM-NTS 4
TABLE OF CONTENTS 6
CHAPTER 1 INTRODUCTION TO THESIS
1.1 Introduction 11
1.2 The Stron-Tth of Cehcrote 11
1.3 Levels of Study 11
1.4 Levels of Stu:'.y User'_ for Concrete 12
1.5 Objective of Thesis 13
1.6 Outline of Thesis 14
1.7 Notation 15
PART I
COMPARISON OF AVAILABLE DATA SITE THE EXISTING FAILURE
THEORIES AT THE PHENOMENOLOGICAL LEVEL
PAGE
5 CONCLUST(VS TO HART I
CHAPTER
5.1 Introduction to Chapter 5 95
5.2 Interrelation of the Theories 95
5.3 The Validity of the Failure Theories a Failure
Criterion for Concrete 96
5.4 Reasons for the Lock of A7reement between Theory
nd D-ate 96
5.5 Conclusions to Chapter 5 101
PART II
STRESS DISTRIBUTION IN SINGLE AND MULTI-PARTICLE
NON-HOMOGENEOUS MATERIALS
PAGE
PART III
EXPERIMENTAL METHODS FOR MEASURING THE FORMATION
AND PROPAGATION OF MICROCRACKS
PAGE
PART IV
THE DateURPLLTIUN AND FAILURE CHARACTERISTICS OF
CONCRETE IN BIAXIAL COMPRESSION
PAGE
PART V
EI, FECT OF THE MACROSCOPIC STRUCTURE OF CONCRETE ON
ITS PHENOMENOLOGICAL BEHAVIOUR
1.1 INTRODUCTION
1.5 13
1.7 NOTATIUR,
principal strains
81' 02, 0'3
E Young's modulus
K K
I 1 7 etc,
K2 - constants
ass infinity
CA coarse aggregate
of degrees Fahrenheit
o
A Angstrom unit
16
PART I
PHENOMENOLOGICAL LEVEL
2.1 17
CHAPTER 2
INTRODUCTION TO PART I
2.2 18
3.1 19
CHAPTER 3
THE PHENOMENOLOGICAL BEHAVIOUR OF CONCRETE
CONPARED TO THE CLASSICAL FAILURE THEORIES
3.1 INTRODUCTION
3.1.1 The Classical Failure Theories
TABLE 3.1.1.1
The Classical Failure Theories
SECTION OF
CHAPTER 3
NAME OF THEORY
3.1.5 Conclusions
TABLE 3.1.6.1
Standard Deviation and Coefficient of Variation
for Different Degrees Of Control
With these data the K values for any given specimen may
be computed and from this calculate the statistical variations of K
for any mix. That the calculations must be kept within a mix is
fortunate, since it is unnecessary to apply correction. factors for
different shapes and sizes of specimens, different aggregate shapes
and sizes, etc. In spite of this, the analysis of the data from
the 516 specimens for each of the theories listed in Table 3.1.1.1
would have required a vast number of repetitive calculations. To
save the author time, the entire calculation was programmed on the
University of London Mercury Computer.
Sect ion A
V
UNIAXIAL COMPRESSION UNIAXIAL
TENSION
BIAXIAL TENSION-
_ COMPRESSION
TRIAXIAL COMPRESSION
Section B
BIAXIAL COMPRESSION
3.2 26
Ultimate load
-Yield point
STEEL 4-Ultimate load
Proportional
limit
CONCRETE
(not to scale)
STRAIN
p 0 K or
1 t
0 = = K, ... 3.2.1.1
3
where 5 and (5
are the uniaxial tensile and compressive strengths
( t c
respectively. The failure surface (Section 3.1.1) is shown graphically
from these equations in Figure 3.2.1.1.
maximum stress • /
eq. 3.2.1.1
maximum
2-shear
eq. 3.4.1.2
/oc-F,ealedral ..
410
shear ./ */
eq. 3.71.2
and energyl
eq 3.8.1.3 1
-1.0 / maximum
/ strain
eq. a3.1.2
maximum strain
eq. 3.3.1.3
K =1.0
COMPARISON
of
CLASSICAL FAILURE THEORIES
(from JM)
FIGURE 3.2.1.1
3.2 30
and crushing also occurred (R,B,i3). With the exception of the crushing,
none of the other modes of failure arc in agreement with the maximum
stress theory.
Table 3.2.3.1
MAXIMUM STRESS THEORY
(Maximum Stress)
= K
G-t
Triaxial Richart, 1 43 40" x 8" cylinder 1:2.1:2.5 0.75 0.88 -1403 1187 85
Compression Brandtzae,
and Brown 2 25 1:1:2 0.75 o.64 -1426 1251 88
Pt
3 27 PP 1:3:5 0.75 1.25 -1414 1257 89
it
Balmer
(G,GB) 4 9 60" x 12" cylinder 1:2.86:4.47 1.5 0.58 -9555 8730 92
it
Bellamy 5 6 6,10 x 12" cylinder 1:2.46:0 sand 0.55 -3467 2411 70
(C,J,B)
IT
Biaxial ** 6 8 6"0 x 12" hollow sand -4388 2645 60
Compression cylinder
I?
Wastland 7 64 5.9" cube 1:7:0 sand 1.0 -2490 1831 74
(S,W) 4000 psi
uniaxial strength
PP
Weigler & 8 30 3.9" x 3.9" x 1" 1:8.61 * 0.275 0.91 -294o 2550 87
Becker slab
(W,B)
If 11
30 If
1:6.07 * 0.275 0.67 -3930 3410 87
MAXIMUM STRESS THEORY (Cont.)
(Maximum Stress)
Biaxial** Weiler & 10 30 3.9u x 3.9" x 1" 1:2.93 * 0.275 0.36 -6470 5570 86
Compression Becker slab
(W,B)
**
value is for total a.:,;greate by volume include for completeness
mximum size 7 mm. diameter - 0- 2 = k as T = 0 for
1
this case
3.2 35
Table 3.2.3.2
= K
3 =
Triaxial Compression-
Compression 1 -7944 4419 56 tension or
ii 2 -8731 4870 56
shear 22 -1063 813 76
PT
23 -719 855 119
3 -6878 5019 73
it
4 -36,170 25,380 24 -979 828 85
70
5 -15,420 25 -1040 85o 82
6760 44
26 -1001 771 77
Biaxial
Compression 6 -8612 27 -3030 2390 79
1292 15
ii 28 -1811 1557 86
7 -4930 627 13
8 29 -2325 1408 61
Pt
-6070 714 12
9 30 -1034 595 58
fi
-8160 875 11
ft
10 -13,500 1340 31 -1016 546 54
10
iP
32 -716 398 56
11 -7530 1020 14
33 -996 682 68
Pi
12 -7390 776 11
St 34 -940 609 65
13 -10,800 1360 13
35 -1089 667 61
Compression- ii 36 -768 509 66
tension or
shear 14 -1222 953 78 37 -933 463 50
15 -2125 1648 78 it
38 -520 476 92
16 -2544 2184 86 I?
39 -815 577 71
17 -1530 792 40 -1403 744 53
18 -1680 1000 60 41 -874 393 45
ii 19 -2270 1414 62 42 -824 433 53
iP 20 -3200 2030 63 43 -85o 739 87
21 -954 754 79
3.2 36
3.2.4 Conclusions
Although the maximum stress theory has often been used for
brittle materials in uniaxiul states of stress (JCJ; EHS; AH), it is
evident that this theory is not a general failure criterion for concrete.
In uniaxial tension there appears to be excellent agreement between
theory and data, but in uniaxial compression and biaxial compression,
alt:_ough the quantitative data provide good correlation, the qualitative
information on the failure of concrete cannot be explained by the theory.
In biaxial compression-tension and triaxial compression, and even to a
certain extent in biaxial compression, it is evident that G and 0
t
and hence K are not independent of the other principal stressesand
thus violate the assumption behind the maximum stress theory.
e
l
= e
t'
... 3.3.1.1
or, since the theory assumes that the law of elasticity applies up to
failure (JM-1 ; ST; HJC-1),
- v(c.; + ) = Gt = K. ... 3.3.1.2
2 3
3.3 38
= 0t . K. 3.3.2.1
3
Hence the vertical cracks may be caused by the induced lateral tensile
strains exceeding the critical limit, G t.
3.3 40
Table 3.3.3.1
0 - v(0 + 0,) =
1 2
Triaxial Compression-
Compression 1 +665 83 13 tension or
shear 22 +196 53 27
2 +878 105 12
3 +316 100 32 23 +99 29 29
24 +171 62 37
4 +2050 702 34
25 +19 7 39
5 +1773 167 9
If +284 30 11
26
Biaxial 27 +739 86 12
Compression 6 did not convere
Pr 28 +488 118 24
7 +1480 422 28
29 did not converge
8 +1800 627 35
30 +319 59 18
9 +2420 825 34
17 31 +342 42 13
10 +4000 1320 33
If
32 +212 101 48
11 +2250 827 37
33 +116 31 27
12 +2180 720 33
ft
.54 +186 62 33
13 +3220 1160 37
35 +356 74 21
Compression- 36 +479 81 17
tension or
shear 14 +292 23 11 37 +372 78 21
15 +305 41 13 38 +193 110 57
if
16 +417 81 19 39 +317 94 30
17 +368 41 11 4o +559 79 14
vr 18 +322 85 26 41 did not converge
19 +394' 13E 34 42 +402 64 17
it 43 +384 41 11
20 +435 105 24
21 +272 43 16
3.3 41
3.3.4 Conclusions
maximum stross th:, ory. Furthermore, the maxi; um strain theory has the
practical advante that the strain can usually b. measured directly,
oven in complex structures .
0 —0 .20
1 3 S
K ... 3.4.1.2
This is the generally used equation of the maximum shear stress theory
and states that when the maximum shear stress exceeds a critical value,
a constant for that materi7.1, failure wil . occur (JCJ; JM; AN; JM-2;
EHS; ST; LBT). The failure surface is shown is Figure 3.2.1.1.
Th, maximum shear stresstheory has been widely used
(FBS; LET; Ail; to 5; WAS) for ductile matorials, but it
must be more closely examined for concrete, a brittle matorial.
K 0 + K 0 K, ... 3.4.3.1
1 1 23
where K, K1,
and K are constants for any mix. Hence for the
2
majority of the data available the two theories are identical; the
data in Table 3.3.3.1 apply to the Guest theory as well, except in
states of biaxial compression.
3.1+ 46
Table 3.4.3.1
MAXIMUM SHEAR THEORY
01 - 0 = 20S K
3
Triaxial 1Compression-
Compression 1 -6541 3263 50 {tension or
P. ! shear 22 -1169 723 62
2 -7305 3092 51
23 -765 819 107
3 -5464 3778 69
24 -1097 767 70
4 -26,230 17,800 68 I
25 -1111 843 76
5 -11,950 4359 36
PP
26 -1275 733 57
Biaxial 27 -3421 2098 61
Compression 6 -8612 1292 15
28 -1996 1391 70
-4930 627 13
29 -2472 1281 52
8 -6070 714 12
tt
3o -1171 496 42
9 -8160 875 11
PP
31 -1206 470 39
10 -13,500 1340 10
PP
32 -928 373 40
11 -753o 1020 14
PP 33 -1112 670 60
tt 12 -7390 776 11
34 -1119 597 53
13 -10,800 1360 13
PP
35 -1305 607 46
Compression- it
36 -1061 418 39
tension or 37 -1101
shear 14 -1388 872 63 399 36
it
38 -639 469 73
15 -2294 1547 67
39 -1021 542 53
16 -2767 208375
Pt
40 -1611 685 43
17 -1716 712 42
41 -1157 281 24
18 -1875 943 50
42 -1043 376 36
19 -2540 1370 54
43 -1426 673 47
PP
20 -3470 1892 55
21 -1105 655 59
3.4 1+7
3.4.4 Conclusions
Although from quantitative considerations there is evidence
to support Lie use of the: maximum shear stress theory in biaxial
states of stress, when the qualitative information is included, it is
evident that the maximum shear stress theory is not a criterion for
the phenomenological failure of concrete.
tmaxi-mum shear
stress theory
(a)
internal friction theory
•Cr) rz
z
zr
Cf) / r
// z
z
/ (b)
'RICHART et al.
Mix designator 1
Internal friction theory —
c)
TSUBOI AND SUENAGA
Mix designator 33
Internal friction theory--
I
3.5.4 Conclusions
—Mohr circle
Since the Mohr theory dces not specify the equation for
its failure criterion, the approach previously used in the
quantitative section of each discussion is not applicable. Instead
the theory must be studied on a more general level. Even so, two
significant points arise.
3.6.4 Conclusions
and
toct 4 /01-c2)2 +(°2-c3)2 +(C 3 1)2 - K
respectively.
3.7 57
Table 3.7.3.1
Triaxial Compression-
Compression 1 3084 1539 50 tension or 22 540 338 62
shear
2 3444 1740 51
23 358 386 109
if
3 2576 1781 69
24 498 365 73
4 12,550 7860 63
25 510 396 78
5 5630 2050 36
It
26 521 315 61
Biaxial 27 1560 1000 64
Compression 6 3620 548 15
28 920 665 72
7 2150 160 7
29 1136 630 55
8 2660 262 10
11 30 529 246 47
9 3580 322 9
31 535 226 42
10 5930 464 8
It
32 401 174 43
11 3300 400 12
33 500 316 63
12 3240 253 8
ti
34 494 278 56
13 4740 520 11
35 578 290 50
Compression- 36 456 202 44
tension or
shear 14 630 414 66 II
37 490 193 39
tt 15 1057 733 69 It
38 282 218 77
tt
16 1270 990 78 39 446 254 57
17 773 345 45 40 721 324 45
II 18 850 450 53 41
41 500 140 28
19 1150 650 56 it
42 454 179 40
20 1590 910 57 43 629 322 51
11 21 502 311 62
3.8 6o
3.7.4 Conclusions
W
... 3.8.1.1
= 1((3161 "262 363) = K
where W, the total strain energy at the elastic limit, can be
determined from a uniaxial tension or compression test (JM).
1 v, 2 2
Wd +0 +0 - I) 0 0 - 0 ) = K.
3E 1 2 32 1 02 - 2 3 13
000 3.8 .1.3
+v 2
Wd 1 6((0 -0 ) + (0 -r0 + )2) = K.
E 1 2 3 1
... 3.8.3.1
With the exception of the constants, this equation is identical to
that for the octahedral shear stress theory (equation 3.7.1.2)(ST; AMF).
3.9 63
Hence the v. Mises criterion and the octahedral shear stress theory are
quantitatively similar. Therefore the quantitative analysis of the
octahedral shear stress theory (Section 3.7.3) will also apply to the
v. Misus energy theory. This previous discussion concluded that the
octahedral shear str equation did not predict the behaviour of
concrete in triaxial compression or in biaxial compression-tension,
nor the value of approximately 10% for the uniaxial tension to compression
ratio. However, in biaxial compression a good quantitative comparison
was obtained.
3.8.4 Conclusions
2c — ... 3.9.1.1
t p
where 2c is the length of the crack and p the radius of curvature
of the leading •
then
(01 - 03)2
80 = K. ... 3.9.1.4
t (01 +03)
Note that equation 5.9.1.3 is the same as equation 3.2.1.1 for the
maximum stress theory.
flaws were 104 times tie molecular Spacing. As with glass, concrete
is particularly susceptible to flaws. The inevitable air holes, sand,
and coarse aggregate particles are all stress raisers, regardless of
whether they are microscopic or a half-inch in diameter. Thus it
would seam that the Griffith theory is particularly suitable for concrete.
(0 ) ( V, )1/m
c 1
) ( V ) ... 3.9.3.1
c 2 1
3.9 67
3.9.4 Conclusions
3.9 68
Table 3.9.3.1
(0
1 3
c0
1 t or 8 0, (c5 1 + 03) K
State Avg. State Avg. D.
of
Mix S.D.
K C. Mix
V.f S.
K . C.V.
No. (psi) No. (psi)
Stress (psi) Stress (psi)
Triaxial Compression-
Compression 1 -4645 1855 40 tension or +246 31 13*
shear 14 -2295 309 13
2 -5340 2250 42
Pt
3 -3633 2260 62 15 -3580 580 16
PP
5 -7649 1803 24 17 -2421 309 13
+249 56 22*
Biaxial 18 -2763 502 18
Compression 6 -8612 1292 15
+308 81 26*
7 -4930 627 13
IT
19 -3796 925 24
8 -6070 714 12
+383 94 25*
9 -816o 875 11
PP
20 -4825 962 20
10 -13,500 1340 10
Note:
11 -7530 1020 14
Tsuboi and Suenaga had insufficient
12 -7390 776 11
PP
but although the theory has been applied to rocks (RMSR) the
application to concrete (N K-3; JG) is difficult, especially to
account for vortical cracking.
0 + K0 K ,.. 3.10.1.1
1 2 3 1'
TABLE 3.10.1.1
SUMMARY; EQUATIONS OF THE CLASSICAL FilILUIRE THEORIES
a)
a b c d Le
1
1
Classical i
failure Equation with Value of Equation
General Equation
theory o--) = 0 Theoretical with
‘_
Constants * 0- = 0- = 0
in 0- = K + K a 2 3
1 2 1
K K-) K
2 2
min prob max.
. 0-
Maximum stross 61(3-2 1 = 6t
1 = j't
6 0 0 0 0-
o--.5 =o-c =0"C
3
+4 1
Maximum strain T1 -V + o-- ) ,
(3rt Ti - v 63 = 0- +1 +00 o- 1 = CT"
2 3 t to t
10
Maximum shear 61 - 0- = ay- T1 -6 = 2:5- +1 61 = 2.7-s
stress 3 s '3
0-1 - a- + “T + 0-3) = 2a- +4 ,
3 1 to 2T
Guest's theory or 0r - 6 +A(0- +0- )= +1 00 61 = 1+A
1 '3 1 3 +19
( 1 + X )cr1 - ( 1 - X )or = 20 s
3
aT
s
11-- - . 4
Internal sin 0 + 1 sin 0 - 1 00
K coo 0
same +1 +2
1
Friction COS 0 0-1 + cos 0 a3 - K °F-1 = sin0+ 1
I_ .1........
d e 0 _11
a b c
I
Value of 1 ,
Classical Equation with Theoretical . Equation
failure General Equation = 0
a2 Constants * with
theory in a = K + K i = 0- = 0
3 1 2a 1 3
K T.K.
2 , -'
min. prob. max. i
t___
General
Mohr's Theory
L —
6 - 6 )2 + (a - J 2 a - aa
2 2 1
- _ _
Octahedral 1 2 2 3) 1 1 3 IS = o- = 3
shear stress 2
+ (a - a ) = 9 t 0
0 2 1 y 2 toot
3 -1 oot
oct t act
_
62 + 0 2 + 0 2 -
1 2 3
cr2 2 vao-_,- a2 =
Strain energy 1 1 .) 3
(Haigh) . 2v (a a + cr o- + a 7 ).2E;1
_ _ = 11/2EW
1 _
1 2 23 1 3 2EW
i o
2 6'1 = .7-6-
(0.. .. a.
)
't
Griffith's 1 3 + 89-t(cri + u-3 ) = 0 same - -
theory a-1 . 0-
cr = 0- - t
1 t
-- i
* Note: of the 5 equations in column d, if o'2 = g3 only
maximum strain is affected 1 2
- 2) = a-
v(0- 0.5 1 to 00 1
61 t
1 3.9
3
--I
3.10 72
0 = K or (1 = K. ... 3.10.1.2
1 - 3
Thus regardless of the theory used, the failure stress under uniaxial
conditions is a constant, although as noted in (a) the constant varies
from theory to theory. It was for this reason that data from biaxial
and triaxial tests were used in the quantitative anal -ses in this
chapter (Refer to Sections 3.1.6 and. 5.2.3).
3.10 73
= K + K 0 ... 3.10.4.1
3 1 2 1
for biaxial and triaxial compression, and as
... 3.10.4.2
01 '3
= K1 + "23
for compression-tension states of stress. These equations were
programmed into the Manchester-Atlas computer along with the data
previously used in this chapter (except for Tsuboi and Suenaga's data
where there were too few specimens in a given mix for adequate analysis).
With the data the computer performed a regression. on these equations
and computed the values for K and K for each mix. The results
1 2
(column c), includin,7 the error of the regression (column d), are
presented in Table 3.10.4.1. As the value of K in equation
2
3.10.4.2 is the reciprocal of K in equation 3.10.4.1, the
2
reciprocals of the values of K for the compression-tension data
2
are shown in brackets to facilitate comparison with the actual
theories.
3.10 75
TABLE 3.10.4.1
RESULTS OF A REGRESSION ANALYSIS WITH THE DATA
OF CHAPTER 3 ON EQUATIONS 3.10.4.1, 3.10.4.2,
3.10.4.3 ,dip 3.10.4.4
b -----'-h
----
Richart
Brandtzaeg 1 -3173 +3.657 4.87 -952 +0.572 2.52 +1646 +1.744 3.11
& Brown
It
2 -4356 +3.692 5.49 -1299 +0.575 2.76 +2234 +1.734 3.51
VI
3 -1684 +3.944 7.28 -472 +0.597 3.67 +777 +1.672 4.10
Balmer 4 -8710 +2.87 8.00 -3075 +0.487 5.59 -6o45 -2.04 6.66
Bellamy 5 -5712 +2.799 3.03 -2117 +0.474 2.06 +4436 +2.105 2.74
McHenry
and
+0.085
f'+269 t+11.76) 16.9 +352 -0.843 3.73 +414 -1.18 5,8
Karni
PI
15 +300 +0.061
.
( +16.39)
25.9 +401 -0.883 3,60 +449 -1.13 4.77
L 1
* values in brackets arc rcciprocals of the
value of K
2
3.10 76
\v,<
\ 0\
• 7...) 45°
•
\ 11
r\
/
/ I \
/I \
/ I - \
I
-1
!„,,estimated
r failure envelope
1 for concrete
1
4.1 80
CHAPTER 4
THE PHENOMELOGICAL BEHAVIOUR OF CONCRETE
COMPARM TO THE THEORIES SPECIFICALLY
PROPOSED AS FAILURE CRITERIA FOR COECRETE
4.1 =TRODUCTION
4.1.1 Objective
4.2.3 Cnclusipns
or more simply
O
toot oct ...
K + K
f' 1 4 f'
4.3.1.3
RICHART et al.
a.
-+4000
Mix Desig. Symbol
1
2
3 -X--
•
cr)
-+3000
BRESLER & PISTER,
MCHENRY & KARNI
hence the: main support for this theory is that the data
r m
ebresont a sooth curve cn tin: t c )1L:ne. Si-Nlarly, Duch
t, - Ljoct
of the data discussed. in Chapter 3 plotted smoAhly on the Mohr
:ram yet it was shown that in situ of this the Mohr theory was
not a failure critorion fur concrete ),,ctin 5.6).
3
McHenry Karni replette the data or. the 01 - plane
and with cquati,- n 4.3.1.3 rewritten in terms of (31 and 03 snowed
(Fi. -,ure 4.3.2.2) that with these 1-dinates tb(_e criterion did not
adequately express the data. They concluded that when equation
4.3.1.3 was used as a failure criterion it ''night permit as a close
approoximation some stress combinations which in terms of principal
stresses wou12: be found t- b, strictly inadmissible'c (M,K). Bellamy
reached sinlilar c:nclusion (CJB).
2
toot
= K(01 + 0 - 0 0
1 3
,
... 4.3.2.1
and assume reasonable values for 1 and 03, say +200 psi and
-3000 psi respectively. Then the effect no. tact of chanin!T the
principal stresses can be studied. If, for example, o is doubled,
1
t will increase by 3.7%; if 0 is halved, the octahedral shear
oot 1
stress will decrease by 1.7%. In contrast, if 0 is doubled, t
oot
will increase by 96.5%. nonce any criterion of the form 4.3.2.1 will
be insensitive to changes in the tensile stress. Yet the work done by
McHenry and Karni (M,K), Berg (OYB), Jones (RJ-1; RJ-2), and
Blakey and Berosford (B,B-1), the tonsil.. .,:eodes of failure observed
by Richart, Brandtzacj and Brown (R,B,3), Weirfler and Becker (v;,B),
and cthars, as well as the conclusions to Chapter 3 (Suction 3.10.5)
concerning the qualitative a!!;rL munt of the maximum strain theory
with the piv,
nomonological behaviour of concrete all show that the
tensile strain is important to the failure of chcrete. Thus the
tonsile strain (or str,,ss) shoulJ, be emjiasized, not overshadowed,
and this would oxplain why equation 4.3.1.3, which sums the square of
87
FIGURE 4.3.2.2
COMPARISON OF DATA
WITH THE
OCTAHEDRAL SHEAR-
NORMAL STRESS THEORY
(from MI K)
-2500:-
i
7a-) \\
.. 4.-- data
\
i' \ \
0-2000-
Li \
1--- \
\\
(f)
\\
Li i
>-1500i-
(7)
(f)
Li \ octahedral shear-
4, sC- normal stress
a_
-1000- \\ theory
\
0
U - \
0 \
-500- \\
\/ 1
0 1 00 200 300 400
61, TENSILE STRESS, psi
4.4 88
a large compressive stress with a small tensile stress, does not fit
the more sensitive 01-O plane (Figure 4.3.2.2).
3
Expressing their equation in still a third way, Bresler
and Pister rewrote the criterion in terms of the invariants of the
stress tensor. They then noted that this is the equation of the mean
shearing; stress as a function of the mean normal stress. Of more
interest is a comment made by Blakey and Beresford (Discussion to
B,P-1) that, using; the invariants, Bresler and Pister's criterion may
be written as a function of the volumetric and deviatoric strains and
hence is of the form suggested by the conclusions to Chapter 3 for a
failure criterion of concrete.
4.3.3 Conclusions
4.4.3 Conclusions
CHAPTER 5
CONCLUSIONS TO PART I
5.3 96
5.4 REASONS FOR THE L.:CK OF AGREEMENT BET= THEORY AND DATA
The theories and the data used to test them may be in error
due to any one of the following:
specimens have been tested with care, and that consideration has been
given to ensure that the results obtained were valid. The greatest
scatter of coefficients of variation has been for the research of
Tsuboi and Suenaga, but this may have been caused by the srmll number
of specimens used for each mix and the wide range of mixes used.
CONCLUSIONS TO CHAPTER 5
Items (a) and (b) and part of item (c) will form the basis
of the remainder of this thesis. Before considerinG this, it must be
realized that this thesis is a comprehensive report of a coordinated
programme of research being carried out on the mechanical properties
of concrete at Imperial College, Newman (KN) is preparing a thesis
that considers on a general, overall level the structure and
properties of concrete. A summary of some of the major portions of
his thesis has been published at the 1964 Arch Dam Symposium in
Southampton (KN-1). More specifically, L2chance (LL) has studied the
uniaxial compressive test; Ward (MW) has developed a uniaxial tensile
test; and Sigvaldason (OTS-1) is at prosent studying machine effects
on the behaviour of concrete (item c), as well as biaxial tension-
compression and tension-tension in flexure. Now that Newman has
presented a coherent picture of concrete on the microscopic, macroscopic,
and phenomenological levels, and Lachance and Ward have studied and
102
PART II
CHAPTER 6
INTRODUCTION TO PART II
CHAPTER?
EFFECT OF A SPHERICAL INCLUSION
ON THE STRESS DISTRIBUTION AND C~\CKING
IN AN OTHERWISE UNIFORM CEMENT PASTE
INTRODUCTION
Introduction
on the possible mode of failure. This chapter \vill extend the analysis
to other complex states of stress and to a study of the onset and modo
of cracking. This will be compared with the observed mode of failure
in concrete (Part I). It will then be possible to observe how well
such an analysis describes the failure of concrete and whether
consideration of tho non-homogeneous nature of concrete may explain the
failures observed.
Mathematical Analysis
Ox=o
'.
, .L
F1GURE 7·2·2·1'-
EFFECT OF rj a ON CJe
j~)j=0.3
2.01~t~g:f "
~olo I g~=1-4(7~~5) t10-8m)(~3_18m(~J
'where m = 1/J)
a = radius of inclusion
I. e /0;
cavit ~~'c/
(E = 0 ' ,l._ _ _-l_~=~.r:...-·1_°/_0_ - ! ,
1.0 ,- -:::>
,, I . rIa
, .1 .
7.1 108
0*+0'
r r
= 0
r
,.
°e* i;·
+ 0e. °e ••• 7.1.2.1
f
°cp +
°<P = a cp
* + s r'
s r6 = s
r e'
where °r is the total stress in the radial direction, 0e normal
to Or in the e direction, 0cp norma~ to Or in the <p direction,
and sre is the shear stress (Figuro 7.1.2.1).
0* = ~O (1 + cos 26)
r z
0*
{1 = ~z (1 - cos 26)
0 ••• 7.1.2.2
0* =
<p
*
s r6 = --a-a z (sin 26).
a'
r =
06 = -2G r~
A
1 + 2 A2
A
+ 3 3 _
r5
(A2 _ 21
~
A3 )cos 2
r5
JJ
1 2(m-1)A2
(m-2)r3
~ + 9 -5 -
~.~ A2 A ~ J.... 7.1.2.3
a' = -2G
cP G-3 +
r (m-2) ? r
3 -3 - 15 3 cos 26
r r5
Elastic Parameters
TABLE 7.2.1.1
INCLUSION E (psi) v
Air void o not applicable
Sand particle infinite not applicable
Coarso aggregate (Thames Valley
10.2 x 10
6 (pc-MW)
River Gravel) 0.27
The use of such a range for the values of the variables will provide
information on the two extreme possibilities as well as realistic
values for average coarse aggregates.
The influence of the Poisson's ratio of the inclusion and the modulus
of the cement paste are not included since for the cases of an air void
and sand particle they cancel from the equation.
will then be possible to consider \'lhere and at what loads cracks will
form. These loads may then be plotted to form initial cracking
envelopes, an analogous concept to the failure envelopes of Part I.
However, it must be remembered that the value of the quantitative
rosults of this analysis are limited by the severe assumptions used.
7.3.1 Analysis
Conclusions
FIGURE 7· 3·1·1
STRESS D1STRl BUTION
ABOUT,A CAVITY (E = 0)
» ~ 0.1 0
1) :: 0.2 ---;l<---
)J :: 0.3 ·······0··· .. ···
1.0 '
o
(c) ~
I .-:----
Ur=O
Sre=O a zT
1+1.0
~
!\ '
7.5.1 Analysis
-10"
(r
~B
7.5.2 Conclusions
d) (] e is virtually independent of e.
e) The stresses at e = 00 and 900 are principal stresses.
CEMENT PAS TE
. 6·
E =2·5x10, psi
)J =0·25
I
!
-1.0-
. \
119
°e = K. ••• 3.2.1.1
From Figure 7.3.1.1,
°e = +2.02 0
1 ••• 7.6.2.1
which, when substituted into equation 3.2.1.1, permits an evaluation
of a 1 for initial cracking, or
a = ••• 7.6.2.2
1
Such an analysis, within the severe limitations of the
assumptions, provides an indication of where cracks are likely to
form and at what applied loads. In the notation of Figure 7.1.2.1,
Table 7.6.2.1 summarizes the relationships between the npplied load
that will cause cracking. By assuming an arbitrary 100 units
(i.e., Ie = 100 in equation 7.6.2.2) as a limiting stress or strain,
these relationships may be plotted in the form of initial cracking
envelopes (Figures 7.6.2.1 nnd 7.6.2.2) analogous to the failure
curves of Part I, Figure 3.2.1.1. That the envelopes are linear is a
reflection of tho linear analysis used.
TABLE 7.6.2.1
VALUE OF THE CRITICAL STRESSES i-u"\JD STRAINS ABOUT
AN INCLUSION AS A FUNCTION OF THE APPLI~ LOADS
Uninxial o kJ yz * = 0 xz = +2.0201 a
yz
=0
xz
= 2.01 0
1
tension
°a z == a0 1 10.2 x 10
6
Ozz = + 1 ~59 01 G'
zz == 1.4001
y
Ox =0
infinity ° zz :: +1.96° 1
Biaxial o
tension
z
('j == °1
ox =0
2
°y =0
ozz = +1 .6401 - 0.0680 2
Uniaxial
compression
o 0zx = azy = -0.5907-:; °yy =°xx == -0.5 2 2°3
0z = 03 6
ox =0 10.2 x 10 a
yy =G xx = -0.033° 3
(5
yy
= G
xx
= -0.122°3
0y =0
infinity 0
yy = 0 xx = -0.082°3 °yy = °xx = -0.068°3
-
Biaxial
compression
0
°zy = -0.587°3 + 0.065°2 0yy = -0.52~3 - 0.5 22 2 °
= °3 6
(j z 10.2 x 10 0yy = -0.03303 - 0.03302 0yy =-0.12203 - 0.1220 2
Ox = 02
0=0
y
infinity a
yy = -0.08203 - 0.082° 2
(5yy = -0.06803 - 0.0680 2
Biaxial
azx = 0 yx = +2.01 01 - O.44 0
compression
tension
0
°zx = +2.02°1 - 0.587 0 "7
:J B
az = 0
3
ex = 01
oy == 0
infinity 0xx = +1.9 6°1 - 0.082°
3
a xx = +1.64°1 - 0.0680
3
* 0 is the stress at y in the z direction
z
122
FIGURE 7·6·2·1
I Nrl-IAL CRACKING ENVELOPE,
(rnaximum- critical tensile stress )
.................... 'E =0
------ E '= 1q·,2 x 10 6 psi'
_._.-._' E = infinity .
OC
,,= -48 _-- -
+1~:I~ ~
.,..-ij::.-::::::
__
vt / _-~---- .. ,;.. ........... /'
/Y'
-3000' _--::z-o.oo
. A,- ~-o':': I
-
-1000·....
;J;~!
. ' 0
i
~,,/'i;--
' +100
--:--...:oi;;- .
, ) . : ! v"..
, ,
3 'f"'")
Vc
'.
"-.
" .
Vr
2
" -=-24 '. _ c =-3.4
" ,, Ut' at.
,,
,, -1000 -
,,
/' -2000
FIGURE 76·2· 2
I NITIAL CRACKING ENVELOPE
. (maximum critical t,ensile stral n )
~
'. . + °1?J__--,
1
-=-2.4 .. .,../t--~i-'-'-:::'';'::'"''
7:) .••••••••
01
~
..........
..,..
.,.~/.:
/. :" 0
~'.. -20PO ;~(;.:tO~~/ d"- ',+190
'U3 '''.'.'
'.
'y'
.', . "CJC '02
"0 . at =-3.8
1~1000 ·
CJC
ot-=-11.5
, ~2
-2000-- .
123
Orientation Dia.
State of Tangent or Plane of
Modulus of Location of crack with in
Stress inclusion of crack normal to propaga- respoct to the Fig~
inclusion tion applied loads 7.6.2.3
•.•
Uniaxial Inormal to applied a
tension 0 x,y nornnl xy
tension,O 1
a
z = °1
a :: 0 10.2 x 106 z tnncscnt II xy rt b
y
a 0
x =
infinity z [tangont /1 xy It
b
Biaxio..l
0 x,y normal x:y Normal to c
tension
rn-3.ximum tension,
= a1
°z :: 6
1
° x °2 10.2 x 10 z [tangent II xy " d
°y = 0
infinity z ~angent II xy If d
Uniaxial ,
comprossion 0 x,y ~angent II xz, yz parallel to e
applied com-
O
z
:: 0 pression, °
3 3
ax =0 10.2 x 106 x,y tangent II xz, yz Il e
Oy =0
infinity x,y tangent II xz, yz " e
-
Biaxial
compression
0 y tangent VI xz parallel to f
plane of applied
Oz = °3 loads
= 02 6
Ox 10.2 x 10 y tangent II xz " f
Oy =0
infinity y ftangent VI xz " f
124
~--------~-----------?---------T----------T-------~--------'----.--~----'-----~--
Orientation Dia.
State of Modulus of Lofcatiokn ~~~:~tt~r Plane of of crack with in
Stress inclusion crac propaga- respect to the Fig.
o inclusion tion
applied loads 7.6.2.3
Biaxial
compressior 0 y,z normal yz normal to applied
tension tension, 0 1 g
az --0 3
ax =0 1 6
10.2 x 10 x tangent II yz " h
a =0
y
infinity x tangent II zy " h
.
Initial Crack Location and Orientation for r1aximum stress Criterion
where Different from Maximum Strain Criterion
Uniaxial
compression
o z normal yz, xz parallel to j
applied com-
pression, a 3
Biaxial
compression o z normal xz in plane of k
applied loads
Biaxial
compression- o z normal zy normal to h
tension applied
tension, a 1
~----------~--------~--------~----------~.~-----~---------------------------
Conclusions
~01 ~01
!
~~ct
'""') vO-j b) c)
a v
I ~
101 ~o:
,~ 3 -0"2 IU3
normal toYi
page
~
iJ
~l02 C)2 I;
i<- -~>
I ~
c
I d) la1
\j e) t u3 f)
A
103
i . y~. G 3 i 0-3 t 01
I II 'V +02
normal to
page
'. <---
01 01
_Ir:;;::;a
0j
<-
I 10;
->
H
I
~ -r
Ig) 1U 3 . h) t U3 i) to;
,
I- U2 ~03
I VU3
normal to
page
'\J U3
Ii .1 0; 01
. rI
<-. .-::;:::=-
, ~
. \ !
l')
1 j) .~\0"3 i
!.fJ t<' "\ 03
I) r03
7.6 127
CHAPTER 8
A STRUCTURAL MODEL TO SIMULATE THE INTERACTION
OF COARSE AGGREGATE PARTICLES IN CONCRETE
8.1 INTRODUCTION
8.1.1 Introduction
8.1.4 Objective
&&&&&o~O&
00000000
88888888
00000000
134
(a) (b)
Idealized Packihg , . Ideal i zed Pacl<.i ng
Used by Baker Used by Roy & Sozen
1- ~,~
Ii
~
_I
, I
I " A" \
" ' I\ ~\, ,
, / /\". .
,/'I
/', I,' \,
'///;:' i/-!- <f ~~,', '
, --<./ / /I \
'cr' I""
I "">---..,
/ ," / I I' ' , , " "
<
I B~'
/\-I--+-'-T-~ ~(" \
. "~-.>--'-:.::::::
I ~ I '\'
. C I\
\ '~ ~, "
---1--'-----/\ I
, '.~ I ! 1 / .'
' - '......
-/, " !, I / , / ......, _/
' , " 1 ,~\ / ,
/ /~ .
......): ;-
""
, ......"" h- ij-.{ /.
...... ~ ";,,, d'J /
" ......
"~NY'
~. ~ ./' " Restraint Effect at the
I U I Aggregate-Mortar I nterfac e
,
\ '- D
-_/ /
I
8.2 135
Examining Figure 8.2.1.1a in more detail (Figure 8.2.1.2),
Baker noted that because the coarse aggregate particles had a higher
Young's modulus than the mortar (in Chapter 7 the ratio of the moduli
of the aggregate to the mortar was assumed to be 4 to 1), the stress
distribution would not be even throughout the specimen. Dantu's
(PD-1; PD-2) data confirm this. As a result, the paths ABb and ACD
(Figure 8.2.1.2) share part of the load P with AD. As the stiffness of
ABD and ACD increases in comparison with AD, they will assume a larger
portion of the applied load P. This is referred to as the thrust ring
effect and was mentioned in Section 8.1.2. Anson (MA) has shown that
the thrust ring does, in fact, exist.
8.3.1 Introduction
= •.. 8.3.2.1
P =
( L )
(AE) fiB
J-i.
... 8.3.2.2
where L is the length of member AB, :lnd f.
.Ii. and E are its area
and Young's modulus repsectively, and hence the force in AB is
dAB
= = cos €}
PAB AB
•••
At each joint there are five pin-ended members which must be in
equilibrium with the applied lO2.d at that joint, or
Wi + L!ijcos 8 ij = 0,
i j
where the unknowns nrc: tho deflections in F ..•
lJ
If joint Q (Figure 8.2.1.6) is now assumed to be the axis
contre and fixed, then five joints are able to move in any of three
directions. Thus there nrc fifteen urucnown deflections, but from
FIGURE 8·2·1·6
138
THREE DlIViEI\lSIONAL
PI N-JOINTED MODEL.
·,~rW1
,1
V
~~2~
h '
II \)FW3
o. f' P
. "~ :.j/1
J./ ~
~O/
[V] ~
,~-=--,:"""""""r--~----'-""
~
"1 \ ;~\i Y
,-W3 \, I //-:1
~\~ .
Q''3--__ .
--~v
f,~-W1 ' A
"0 ,
DEFINITION OF .TERfvlS
USED TO DESCRIBE MODEL
LETTERS CODE DESCRIP-I COMPUTER
DESCRIPTION IN FIGURE TION USED FOR' CODE
8·2·1·6 DISCUSSION
.
centre diagonal (X axis) . RS C2"" 1
" (y " )
II
MP C3 4
II 11
(Z II
) NQ C 1 0
diagonal (XY plane) RP , D23 5
II II II
, .. ~
PS ' D23 5
II II
" S l\~ D23 5
II II
" N1R D23 5
diagonal' (YZ plane) NP D13 3
II 11 II
PO D 13 3
II II II
OM D13 3
11 II II
MN D13 3
diagonal (XZ plane) NS 012 2
" II 11
SO D12 2
II II II
'OR 012 2
II ' II
" RN 012 2
'-;;'C2 means Centre member' in VV2 direction
139
. F-IGURE 8·3·2·1
i
I ., -
.________________~~
I~
Cfl"omJALM)
140
= .~. 8.3.2.5
P RN = P RQ = P
SQ = P
SN = P2
P = = P = P = P3
MN P HQ PQ PN 8.3.3.1
•••
P RP = P ::= P = P ::= P
RM MS SP 5
(for definition of numerical subscripts refer to Figure 8.2.1.6).
J'
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
. J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
TABLE 8·3·2·1 (Cant.) 142
..
- r
0 +Srp .+Srp 0 +Srri 0 . +Srn Arx =+W2
0 +Srp·. +Srp 0 ·0 0 0 D. ry Wry
0 0 '·0 0 0 0 0 D.m
..
..
0 -Srp-Ssp Ssp-Srp O. 0 0 0 1:::. px Wpx
.
0- Ssp-Srp D Spn-Spq' ·0 +Spn -Spn ~ py -W3
-- .' I
These in turn infer symmetry of some of the jeflections. For example,
at Node N the stiffnesses in the X direction nre equo.l and hence
.0. NX = 0 Clnd .0. NY = O. By similar re['.soning,
6 RX = -6SX
6 MY = -~Y
... 8.3.3.2
6 RZ = .0.SZ = 6 MZ = 6 PZ = ~~'JZ = 6
Z
Therefore by substituting equations 8.3.3.1 and 8.3.3.2 into
equation 8.3.2.5 nnd by considering only W' and the last
1
equation reduces to
(..£. + -1 1
+-) ..1 J.. .0. sx ~V2
P1 P2 P P5 P2 =
5
1 (..s.P + -P1 1
+-)
P
1
P3
6 py W3
"5 4 5 3
1 1 2 1 1 6 W
(- + - +-) Z 1
P P Po P2 P
2 3 3
.. ~
After the deflections have been determined from the nbove equations,
the forces in each member may be computed by
F = 2~ =
26
SX
o
b ~
1 1
= 1'21> 2 (6 Z +.0. SX) = ffp3 ~ Z + .0. PY )
1
= = ffP5 (6py +.0. sx)
••• 8.3.3.4
when the forces nre known, it is possible to calculate the stresses and
strains for each member as well as the over:tll stress and strnin, from
which the Poisson's ratio and modulus of elnsticity of the model m~y be
computed. The last two values can be compared with those measured in
concrete.
8.4 144
Am = An • ••• 8.4.3.1
.. FIG,URE 8'·4·1·1
...
-.
',
'I
I
__p~~~ggr_ega.te ~ort.ar ~ggr~gat~ . P . _,__
.A.aJEa AmJ Em Aa, Ea 'pin-joint'
I II . . .... ,I
f' lf- +-Im , 1f -~i
~
(J1
146
.... 8.4.3.2
L La Lim
EA: = - - +EmAm
EaAa - - •• t 8.4.6.1
and for members 0, 1, and 4
L la Lim
EA = -+~~~-
Ea.Aa x t AaEm •
••• 8.4.6.2
8.5 148
8.5.1 Introduction
=-----~~-
0.667 itLa3
••• 8.5.2.3
(t2CLa + Lm~ 3 - o.6671t La3 •
This in turn mn.y be equated to the aggregate to mortar ratio
of the concrete. Since the mix parameters are known, the only unknown
8.6
Lm=La. [ 12
. 1 [23 B'
TI: (' (1+E') + m(1+Er) + Ds(1+E') +
c'
~ .+
1TI =~ 6
2.65
where ••• 8.502 e 4
B' = water-cement ratio by weight
C' = coarse aggregate-cement ratio by weight
D' = sand-cement ratio by weight
E' = air voids by per cent volume.
The coarse aggregate and sand have a density of 2.65 while the cement
has a density of 3.15 of coarse, other values can be used for different
aggregates or cements.
To use the model, the equations are solved in nearly the reverse
of the order in which the:y he,ve been presented during the previous
discussion. The procedure is outlined below.
b) The values of 1m, La, Em, and Ea., which are now known
or are measurable, and the assumed values for Xl and Aaare substituted
into equations 8.4.6.1 and 8.4.6.2 to determine the member stiffnesses.
8.6 150
Em 0 = Em + xO , ••• 8.6.1.1
where Ema is the modulus of elasticity of the mortar segment at any
stress a in the members and x is a constant determined, at least for
uniaxial states of stress (Chapter 18 will Dodify this for other states
of stress), from tests on the mortar.
mortar segment and hence of the entire member (equations .8.4.6.1 and
8.4.6.2) becomes zero. Thus the member can no longer transmit any
load, which is therefore transferred to other members.
CONPUTER PROGRAMME
The inputs, with the exception of the loads, are at the beginning
of the programme. Their meaning and the equivalent thesis notation are
shown on the print out (Table 8.7.2.1). This explains some of the
unusual notation used in the above derivations. In the above discussion
the following four inputs were not explained.
8.7 152
TABLE 8.7.2.1
COMPUTER PROGRAMME TO ANALYZE MODEL
UNIVERSITY OF LONDON COMPUTER UNIT ROUTINE 905 REQUIRED
WRITTEN IN ACCORDANCE WITH (CHLF/3)
CHAPTERo
VARLlBLES1
HALT
JUMP25
0)1071,0,0,0
25) NEWLINE
NE\rJLINE
~
Young's Modulus of mortar (Em)
READ(F) Y. oung' s modulus of the aggreGate (E:'l)
'" inputs
i
READ(Q) (See Chapter 18)
READ(X) Factor in equation 8.6.1.1
READ(Y) Iteration error
READ(Xn) Area corrections for C1, C2, and C3
RE.i'lD(Bn) water_cement ratio
READ(Cn)
READ (Dn)
READ(En)
Fn=En+1
Dn=0.37736Dn
lCoarse aggregate-cement r~tio
Sand-cement ratio
Air void content
Cn=0.37736Cn
Gn=¢DIVIDE(BnF'nTDnFn+o.31847Fn+CnE'n,Cn) Step (a), Section 8.6.1
Gn=Gn+1 computes 1m from mix
Gn=o.66666666667£Gn !parameters, equation 8.5.2.4
Gn=¢LOG(Gn)
Gn=o.3333333333Gn
Gn=¢EXP(Gn)
C=0.707107GnB-B?
H=B+C
u=1.414214c+o.414214B
w=1.414214H Step (b), Section 8.6.1
Z=¢COS(D£/180) computes member length
Z=1-Z and urea.
D=0.5£BBZ?
Xn=XnD J
8.7 153
1=0(1)5
EI=E
VI=EI Step (b)
REPEAT sets member Young's
G=1/H modulus
Z=1/W J
9) NEWLINE
CAPTION
itPPLIED LOADS W2 W3 W1 Step (c) reads in
NE\vLINE applied loads, and prints
1=0(1)2 thorn on output
READ(UI)
PRINT(UI)01 4
REPEII.T
1)Xo=¢DIV1DE(FVoW,VoB+FU)
X1=¢D1V1DE(FV1W,V1B+FU)
X2=0DIVIDE(FV2H,V2B+FC) Step (b) computew Young's
X3=0D1VIDE(FV3H,V3B+FC) modulus of each member
x4=0D1V1DE(FV4I.rJ, V4B+FU)
X5=0DIVIDE(FV5H,V5B+FC)
Fo=XoZXn -1
F1=X1ZXn Step (b) computes member
F2=X2GD stiffness (equation 8.4.6.1
F3=X3GD
F4=x4zXn
F5=X5GD
1=0(1)2
J or 8.4.6.2)
B1=U1
REPEAT
Ao=2F1+F2+F5
A1=F5
1
A2=F2 Step (c) forms matrix
li.3=2F4+F5+F3 equ~tion 8.3.3.3
A4=F3
A5=2Fo+F2+F3
JUMPDOWN(R905) Step (c) solves
J
_. equation 8.3.3.3
8.7 154
Ho=¢DIV1DE(2FoB2,Xn)
H1=¢DIVIDE(2F1Bo,Xn) Step (d) computes stress
H2=¢DIVIDE(0.707107 F2B2+0.707107F2Bo,D) in each member,
H3=¢DIVIDE(0.707107F3B2+0.707107F3B1,D) equation 8.3.3.4
H4=¢DIVIDE(2F4B1,Xn)
H5=¢DIVIDE(o.707107F5B1+0.7o7107F5Bo,D)
Equation 8.6.1.1
JUMP27
26)WI;::EI
JUHP5,Q=O
Yn=X/Q
WI=EI+YnGI
1
27)JUMP18, WI>o Stops analysis if
NE\VLINE Young's modulus is
CAPTION reduced to zero by
E=o, FAILED J equation 8.6.1.1
JUMP17
18)v=¢D1VIDE(WI-V1,VI) Checks error of con-
V=¢HOD(V) vergence, if too large
JUHP4, V>Y computer resolves matrix
JUHP5 using new Em values
4)K=1
5)REPEAT
JUHP6,K=0
1=0(1)5
V1=W1
REPE.3..T
JUMP1
6)ACROSS11/1
17 )READ(H)
JUMP17
8.7. 155
END
¢COS
¢LOG
¢EXP
CLOSE
***T
TITLE
LATTICE THEORY ANALYSIS G ROBINSON U-797
CHiiPTER1
J~I.+5
B·,.2
U~
E'!o-5
F ..··5
G~5
H~5
V..,.5
W~5
X""'5
y.~
C~5
D·-:-2
Z-"">:5
l
11)S=0
1=0(1)5
Yn=VI
Yn=¢JvIOD(Yn) checks if critical strain
JUMP8, o. 01>Yn has been ~xceeded; if so,
YI=¢DIVIDE(HI,VI) sets Young's modulus to zero
JUl'1P8 ,A';" YI
VI=o
EI=o
vJI=o
YI=o
NE~vLINE
J
-, If a member fails in
CAPTION tension, number of momber
CRITICAL STRAIN, TENS ION F. :~I1URE is printed on output and
NEVJLINE computer resolvos matrix
PRINT(I)1,0 to redistribute the stress
S=1 in the model
8)REPE.'.T
JUMP14,S=0
ACROSS1/0
8.7 156
~step
14) NElrJLINE
CAPTION
JVIORTAR E, STRESS, STRAIN
NEvJLINE (d) prints out value
1=0(1)5 of Emo (equation 8.6.1.1)
NE\VLINE nnd th0 stress nnd strain
PRINT(VI)0,4 in e2ch mortar segment
PRINT(HI)0,4
PRIN'r( YI) 0,4
REPEAT
NE~ILINE
CAPTION
MEMBER E,
NEWLINE
STIFFNESS, STRI~IN l I!
1=0(1)5
NEWLINE
PRINT(XI)o,4 Istep (d) prints out the
PRINT(FI)0,4 YounG's modulus, stiffness,
ZI=XI and strain in each member
ZI=¢IvIOD(ZI)
JUMP10, .001~ZI
HI=HI/XI
PRINT(HI)0,4
JUMP12 I
10)620,27
620,6 I
620,0
12) REPEAT j
NEtvLI:NE
CAPTION
APPLIED STRESSES, DEFLECTIONS, OVE~lLL E,
OVERALL STRAINS
NE\VLINE
1=0(1)2 Stop (0) prints out
NE~VLINE applied stress, deflection,
DI=GGUI Young's modulus and overall
PRINT(DI)0,4 strains for model as a whole
BI=2BI
PRINT(BI)o, ~.
ZI=BIZ
DI=DI/ZI
PRINT(DI)0,4
PRINT(ZI)0,4
REPEAT J
8.7 157
Do=Bo/B2
D1=B1/B2
D2=B1/Bo
NEWLINE
C.L"l.PTION
POISSONS RATIO X/Z, y/Z Y/X Step (e) prints out
NElrlLINE Poisson's ratio
1=0(1)2
PRINT(DI)0,4
REPEAT
NEll/LINE
CAPTION
VOLUIvIE CHANGES DV/V , Vn-V
NEWLINE J
Un=o • 1666667ttJiiVW
Vn=o .5WW+o. 5B1 W+o .5Botv+o .5B1 Bo
Vn=o •3333333Vntv+o .3333333VnB2
h'n=Vn-Un
Vn=Wn/Un
l Step (e) prints out
Ivolume changes in model
PRINT(Vn)o,4
PRINT(Wn)o,4 J
I
l
R=o
.I=o( 1)5
JUMP15 ,HI~ Hn
NEWLINE Checks if limi tine;
CAPTION jcompreSSion is exceeded
SPECIMEI\T FAILED IN COHPRESSION DUE TO LOAD,
I
NEWLINE
PRINT(I)1,o
R=1
15) REPEAT II
JUMP16,R=o i
-...>
For any load stage, the output from the computer consists
of:
8.8.2 Conclusions
. / "4
.' /
/
t:J
/
X//
/" . ~
/' . "... ..
W
/" E~ .. ~··
I
r- /
.
.
,~".
~...,....--
~--
.
LL ./. ~.
.~..
~",rc--
-,.-
-----E a
.
o , / . . :.,..;""'"
>- ... ,. .................~.............. I( /~'?:;::::-
r- ~ , ,,~ . .. ...............';;;t.':;':;'' ~ .......:.. ~~ .v. '........... ~ ...
. .... ·.... · .... · .... · .. · ..x
. f,\ .. 4 .....
/
FIGURE 8·8·1·2 162
0.26
, \1.
0.24 \
'\ '
'\
\
o .\'XI
,
""
0.18
0.16', ",
, ,
T"
0 ,0001 A, .0002 inlin .0003 ~~ ..0004
.0 30 D 40 degree 50, 0 60
",0 3x10 6 Em 4X10 6 psi 5x10 6 , _ .. -A-.. - " 6X1'06
0 8x10 6 Ea 10X10~ psi '12x10o -~~-- 14x10o
0',+1000 X 0 -1000 ....... *....... I
-2000
'0 1.0 ....:. . ...g._.-
X' 2.0 3.0 4.0
FIGURE 8·8·1·3 163 j
If)
0. -5000
)c••••
' .•...X
.
' . ... ,
.
'. '.
'
I-
It.;
' .....
(f)
a: -3000
"
' .. '.
"
"
LL ,
".
I
U
-
I
3
o
« , .
9 -~ .
NOTE: see taxt for @xpJanaiion
of symbols
, I
I
0 .0001 A .0002 inlin .0003 ~ .0004
0 30 D 40 degree, 50 G 60
-.0 3x10 6 Em 4X10 6 psi 5x10 6 _.-A- .. - 6x106.
0 8x10 6 'Ea 10Xl0 6 psi 12x106 -...;..~--- 14x106
0 +1000 X 0 -1000 ........>'...... ~ -2000 '
0' 1.0 e. X' 2.0 3.0 ~'-lii1'-'- 4.0
8.8 164
8.9.2 Results
--6000
-- --
tf) -4000
0.
<f)~
<f)
W
0::
I-
(j)
--2000
~ _ _- . I I_ _ _ _ J ,
-500 '-1000 -1500
~TRAINJ in/in ' .
8.10 168
....
(a) (b)
I u- .
~ . 3
(c) I (d)
If a compressive force (W
3
= -03) is applied normal to the
tensile load (W = +01) discussed in the above paragraph, it will
1
increase the tensilv strain in members C1 and D12, but decrease it
in member D13. The eff~ct will be to cnuse the fnilure of member C1
at a correspondingly lower tensile load than in the uniaxial states of
stress described above. Again, after the failure of member C1 the
stresses in the model \vill be rodistributed and member D12 can be
expected to fail immediately thereafter. Thus thrust rings PNM and
PQM will become mechnnisms and the compressive load W3 must be carried
entirely by members C2, C3, and D23. As a mechanism, the compressive
strain in mmnber D13 is zero due to , but for W W member D13 is not
3 1
a mechanism and hence ".Jill fail in tension due to W • Thus for biaxial
1
compression-tension, the failure of members C1, D12 and D13 will create
a plane of failure normal to the applied tensile lond (Figure 8.10.1.1c).
This is in agreement with the modes of failure observed by McHenry and
Karni (M,K) for concrete subjected to biaxi2;.1 compression-tension.
8.10.3 Conclusions
, "CJc . , ',' ,° 1
r o-3~. . . L./ ' ,
_._._._._._.·.·,. t.01 11 /
_.-.-. -.~~-.-. ..-.-.--:;~:......................... ~....
.....
r I . _ _ .. - ·
'.
-'-'r
/
///
•.",' ~ ~ .~
~. .
103 aO> ..
..•
• •
I :f
•• • • • •
t : /
)i
,.02
I
/
//
/ /'
.
ULTIMATE LOAD CURVES
BASED ON A GENERAL
'
. /
I // DISCUSSION OF tHE MODEL
"I). // ~
"'-.J
\ SECTION A // (Jl
L
8.11 176
8.11.3 Conclusions
This chapter hEls shown th2.t from n i2,imple model of the structure
of concrete, proposed originally by Baker, it is possib10 to derive
equations based on tho concept that aggre~~tes will naturally pack in
3. diamond formation and when subject to load will create thrust rings.
To maintain equilibrium within a given ring tensile forces must develop.
At a critical load these will fail, signifying a crack formation in the
actual concrete. Such a concept has been expressed mnthermticnlly and
tho size of the model related to the mix proportions.
CHAPTER ,9
CONCLUSION TO PART II
PART III
CHAPTER 10
GENERAL INFORMATION AND DESCRIPTION OF
EQUIPHENT FOR THE TESTS OF PART III
10.1 INTRODUCTION
10.1.1 Introduction
the energy loss due to sound can often be quite audible even at load
stages below ultimate. Both Rusch and L'Hermite have plotted this
noise or its integral as a function of the load and concluded that
above a certain load internal cracking sounds are heard and that
these increase until failure occurs.
a) it is too time-consuming,
b) it is limited to the surface, and
c) it repeats information that can be obtained through
the X-ray tests, since X-r~ys detect surface cracks as readily as
interior cracks.
10.2.1 Materials
Thames Valley river sand from the Stone Court and Ballast.
Company (Rickmansworth Pit) was used. Before the sand was batched,
it was slowly dried on a 2 kilowatt pan at a temperature less than
100°C (KN-2) and then passed through a 3/16-inch sieve. A represent-
ative sieve analysis is shown in Table 10.2.1.1. This sand conforms
to the Zone 2 limits of British Standard 882: 1954.
10.2 186
TABLE 10.2.1.1
SIEVE ANAIJ YS IS
Secondly, each mix was poured in one batch, but the batch
was of sufficient size to supply at least two specimens for each test.
This eliminated batch-to-batch variation when comparing the techniques
and ~oreover, by providing two specimens for each test, the average of
the results was more significant.
To attach the string one end was glued to the stone and the
other to a metal wire cross (Photograph 10.2.4.1). A wooden jig was
used to ensure that the strings were of the proper length. Also, two
minute pieces of metal were glued to the aggregate surface to assist
in identification of the aggr8gate with X-rays (Chapter 13). This
technique proved quite successful (Photograph 10.2.4.2), although in
one or two instances the string became so unravelled that it parted
188
~. :!
I
I
Photograph 10.30301 \
As u result, all Poisson1s ratio values have been divided by the value
of Poisson's ratio at the first load stage to eliminate calibration
errors. The value of the volumetric strains is also affected. The
lateral gauges were removed just prior to failure in order to prevent
damage.
CHAPTER 11
11.1 INTRODUCTION
11.1.1 Introduction
E = 4(fr)212p
••• 11.1.1.1
g
FIGURE 11·2·3·1
. EQUIPMENT SET-UP
FOR THE TESTS OF CHAPTER 11
20 A.F. Transducers
(l\ Gener- emitter . receiver Voltmeter
V ator (valve)
volts II! ' ,)
t--------j
f\)
.0
Advance Type TC1 Solarscope CD 1014.2 w
0~6'" ) ~O.225
RESPONSE FRO[vfJ VAR lOUS 0.204
TRANSDUCERS A,S A FUNCTIO I
FIGURE 1"1·2:3·2
0.14r,. OFFREOUENCY ..
......)':....... ....... ~
0~2!..
"
DRIVER
bari urn ti tin ate,'
'
. ,'
. "
. ...... ...
.. ~ ... :........... ~--,.... ~ ...1-.............. ..'. " . .
.
: I
\
~..... : ........ ~ ......... ···t····· ... ···x· ...................... .
,
. •••.••• '" ...... ~ ••••••••••••••••• ;?Co
"<
~~
'1/ "'yf k . \,
~, .
.J\ ,l tl> ! \\ (\)
f
0,21" . \
: I
,
" \"
"
:"'," q,""
I
A.
~
.
r.l
In
,,,,\
/'\.t
(,!J
/
/
'
•
\bR 0
~
C>\\
',Q
~
•ru
r--'-'
I
-----l·~-···'---'·---'!'--
~-.
RECEIVERS
·'~---~----·-·------~---·---·--------~'·--'---I
,-,-"~--"~-",---.--.,,,,,--,-,,~,"'.--,,"---'---'.-.---,---.-,.----~~~
I
. ,'--.. -.. '.. . ,...."'.... _',- '''·---'''-1
~
I
I
i
j
I
t
Barium
Ti tina te
Lead
(0.5 me)
Lead
(1.0 me)
Rochelle
Sal t
Unknown
e~_ys__t~_l__f
I
I ; I ---~.... --~,.'-----'--~-.... ,
i 1 En. r Very' good poor poor good I
T·t:~U~ f 23,031 28,888 29,276 No response 22,563 t
" ~ lllU e , 0.225 V 0.00049 V 0.00008 V 0.145 V I
f
}
Lead
(0.5 me) No response Not tested No response
~
lJO response
.
No response
I
i
I
:
Lead
( 1.0 me) 1-
poor
28,808
No response Not tested
poor
29,273
0.0098 V No response
!EHITTERS t 0.0038 V
Ii Rochelle
Ii poor poor
28,676
! Salt I 0.0063 V
No response No response Not tested 20,108
0.0035 V
I I
I i fair fair
! Magnetic 21,701 No response No response No response 20,135
, ' 0.0109 V 0.035 V i
I r '1
i s 't .
i! ensl 1. V1.. t
Yf i d
goo f
high . poor N0 response good I
i
i requenc1.es 1
IL____J i only i
I I .___. .,,~__~._.~~.. . . _. _,_____________ .~_,,______~.___,___~_____._,.__,~ __._._.
____
I\.)
o
\J1
11.2 206
When this had been decided, the barium titinate transducers were
checked for their own resonant peaks. Their response to different input
frequencies was not linear, but there were no significant discontinuities
in the range of 500 to 30,000 cps. Thus the peaks in Figure 11.2.3.2 are
not caused by the transducers, but either by resonances in the concrete
or mutual resonances between concrete and transducer. This will bo
considered in more detail in Section 11.2.5.
It was shown in Section 11.2.4 that the peaks adjacent to the resonance
p0ak arc not caused by resonanee of thG transducers, but might be due to
inter.actions (chattering) of the transducers with the concrete. Tho use
of glues as bonding agents prevents this chatter yet the peaks still
exist. Therefore these pe.Cl.ks adjacent to the r0sonant frequency are
probably causod by internal reflection or interferences within tho cube.
I'
ototrash 1 1.2.7.
Acoustic Transducers and Holder i Pssiis
11.3 210
11.3.1 Theory
Q =
~kH
~., and .•• 11.3.4.1
11.3 212
f
r = ••• 11.3.4.2
ltlith peaks of this size, the shape of one peak will often
overlap and influence the shape of the adjacent peak. As a result
Q measurements are inaccurate.
Although theoretically the technique has promise, it w~s
1'\ \ \
o
.1
/
/
'I
,; ILl
;1
I \ \
I ,\.
1\
.,
! :1 \
J / I
~\ h-
o~o A---;· I A- :' I \ I \ \ \ I \ ' :I ' :
~
I /X'
/! \1 ,\
I ~ .j. \ ' ! 1/ \ \.
\ \~,
til //
'J I I tJ ' .
"I J:I \. :\
,+-'
o foam rubbelr
. // : 'I . \ \! \
>
.........
c==::J
//
: 1/
/ !I \ \ ',I
. \ \'
\. A ! '/.1l '\ b
f
O.B -CJ transducer __ x / /
I '. \ \ ,!S
/y,
\' ' :/ . \.
w : [I • \ 'N'P .. /1
o . \~ .~
' : :J
J
1- I .A wedge
--~- ",-",--
.. /
/
: ./1II
,I, \
~"
\
.\ \ I.r\:\\,
\ ' , . . .
\
/
AI "I
i
:
:.
!
j
J I
Ii,.
~
· \iI! \,\
. , ~,\'.
\ \ '1:\\\
\
. ~ ., :
\
\
:.... ..""
. '.
....
tii'
~.
\ J ~[ :
,:
tl ::
\'. •
..\ \"
\.v: \i
Ii:
~);I' ......~ :. .
: \/
~.
/\...
"
.
h
I
.. *..~.~ . .' .... :\;
~.I:!>' ..
. '.\ .
\ I •' er ~
,
:
•
~t-- ,,' .
• ~" ,.' ......... / ..... -..1-~_ ,.' ': I /:' ... ~'j \~:
". /' /' ........~. ... :" / : '. tv· : ~/:'
........ /' ~~ ..···h· '. .' /
02 ..- , /:r'" /' ....;....
,:.~
- ~"
. 'r" "
/ .', '"
'. 'H1-t
~Ill •
•/ , . / ; l - _ "'::~ : . '\. I: :
_/-- 'il_ - - __ ~~')'!A. ._
'J~~" :
:' /'
..
p!
./ \
.,j[ \.1.",..""
Y ~ ,. .
.~~
,>(?(' .' a.:/
~J"\ . ~ \ ). f\)
~ *'
,,~r\:
' • •~ :",-y
i;J> \...,/ ~
.)!-~ f:I
o ~ =a~t...._ _ _ ...J. , _' ~(! L----------'
'14000 16(000 18000 20000 22000
FREQUENCY (cps)
1...
11.3 215
Even so, with thc knife edges, Obert used a pulse regeneration
technique that would determine the resonant frequency. .l similar but
greatly modified set up was tested with spare pre-placed agsregate
sp(~cimens. A Cintel delayed pulse and sweep genera tor was connected
to both the emitter transducer and the output of a Cossor Hodel 14ll-9
amplifier attached to the receiver transducer. Thus the generator
pulsed the emitter transducer, the resulting acoustic energy was
trnnsmitted through the concrete, r:.;ceived, amplified, and returned
to the generator to act as' a trigger for the next pulse to the
emitter trnnsducer. In this 'vvD.y the system regenerated itself and the
frequency of pulsing was measured by a counter. Considerable difficulty
was encountered bocause of instability, but a preliminary test was
conducted. i{ mortar specimen with six pieces of nggrega te ',was loaded
to failure and the frequency reading taken with the pulse regeneration
apparatus. The results are shown in Figure 11.3.6.1. No strain measure-
ments wore taken as this w~s only a preliminary test. The results agree
well with rGsearch by Obert on tho effects of pressure.
11.3.7 Conclusions
The curve in Figure 11.3.6.1 may also indicate that the resonant
frequency technique is insensitive or that the specimen is failing by a
series of small cracks or bond failures (Section 11.3.4). To determine
216
FIGURE 11·3·6·1
PULSE REGENERATION TESTS
. .
EQUIPMENT SET-UP
freqUeil:-JY Advance I 0,
counter TC1... V 3
(~l . \l
I ~~~iX F
IbJ .d r _0>- I [r-,~-~
I
'PUlse r
genera:or
'--' .
'
! amplifier
. r
I
U/'
Ci ntel
'1\ 1"""'\:3
V~
iI
I Cossor
1449
!.-
' - - - - - -- -- -- -- -- - -- -- -- -- - _-_-_-_-_-=--_-=--=-~I
--
TEST RESULTS
2L\ OOO r
':\
~X~:~~~~~
>-
U 22000-
z
w
::J
'0
W
0:::
LL' ,
20 00 0 l l . - . , !_ _ _ D g , r,
-'
11.4 217
11.4.1 Introduction
The test set up for the sound emission tests was similar to
that shown in Photograph 11.2.7.1 except that only one transducer was
used. The other was in position solely to balance the trans~ucer
0-1
oscilloscope
4.5v _ 3 t I
I ') : time base circuit (1"/lllin) :
I =!- ~
'I
~' I
I
45v !,
-
l ' -< ;
I
A '
-~ ~I f\)
T ... -~
ILcaa::»~CIt;IId_-=-C:Oasllb~_&St»~~:Jtl'Ea""'fIW,D~~~C;G;;3I>e:.-u_~
" ...... ~-o .... _ . . , . . ~~ cr.:.:a J
I
f\)
'0
11.4 221
a) The small pips have been lost due to biasing and subsequent
rectification.
o Q
o
";t
<X)
.
LOI
o
o
$2I
-490 -1460
-240
(j)
.
-970 ~
(PSI)
j I
Q I
I~ ;f
(fj FIGURE 11.4.4·3 /
2 I.
. 0
z
<to
O:::C\I
I-
(j)
0:::
on...
0-
LOre
In...
....io
O·
>'-
80
,-I
(GSR. '64) .......o...... ~.--sr. :
-970 -1940 -2890 -3860 -4840 -5840 -6830 -7790
-214 -L\~,O '_' -14,60 ._.-21?O.,. -33.70 ,_, -4~30 Of. -5~<30 -6 0 -.7~90
0-)- LOAD (PSI)
INTEGRATION OF E1ViITTED SOUNr)/223
MIX 'D'
MORTAR A/C=2.o W/C=0.42
MAXIMUM LOAD= -8710
PLACED AGGREGATE'
01
~rn
o ?
Lfl·
J,
i'
c)
io
~M
ZI
o
<.9r-;V)
6<r~
-1....JO
>
Z
00-
Z
o
i=;'
<r0::
(G.SR '64) <.9
-490 -1460 . -2420 -3370 -4350 -5360 -G310 -7300
w
f-
?44 -97~, -1~40 -2890 -386,0 t -4840 -5840 -6830 Z
(0 LOAD (PSI)
[~l ~
I
I
I
I
I
FIGURE 11·4·4·5 /
I
I
I
I
/.
o o
II ..........
Q
Pi
.........."lS.............C( ;1
c~ I
(G.SR '64)
-490 -1460 -2420 -3370 :-4350 -53GO -6310 -7300
-1~40 ~_ .-2?~0 • -3?6,o • ;484? . _.-5?40
'w .
-244 -970
LOAD . (PS/)
-6830
FIGURE 11. 4·4·6 224
INTEGRATION OF EMITTED SOUNDS
MIX 'G' /
CONCRETE A/C=4.5 W/C=0.45 /
MAXIMUM LOAD = -7060 /
l-
"i'
I
8~
LO· /
I
I
7-1.. /
I /
to
~M
I
/' /
.l
/
,
./
Z / ./
o ./
o 8C\I '/
/ /
0-............ Q .. .
........................ . / /
/
z ..... ............. ,,/ /
«0 o IV"
" ./>.
/
• •• CJ ..........................._ ..
O:::C\J //"" /
I- .... ~ ..
l/). ........... g,.••••••••••••••••• Q ••••••••• , •••.
_ .,,/'
8
o
'I
(G.SR '64)
241 -490, -970, -1460 -1,940 -2420, -2890' -3370 -38GO -;B50 -484q -53 GO -5840
(i) . --~~LOA~D~--~~~~(P~S~I~)--~----~~~--~~--~
FIGU.RE 11.4·4·7
INTEGRATION OF EMITTED SOUNDS ,
! t
MIX 'H' d
I I
/ '
CONCRETE A/C=6.0 WiC=o.50 i I
MAXIMUM LOAD: -5550 ; /
i / ......i
j
i I
I
.
....
/ I~...·······/
I /..""
8 'I I"
I'- i ..../
:,
1'1 I
/ .....
.. ~"'/
I
a Do ..... ; .................................... ~ 0 / D ••• •••••• ",'
z ................................".....................,,/:......... //' ,
« 0
CC N .I
/'
/
/
~ / /A
ci::: / /'
~~T
I"
,I
Ii
I .
I I ZZ.
/
«
"
/
o /
/
. <t OJ
o co ". .' 0:::0:::-
(j) / I f-f- ~
z
D""" ••, •• ."., •••• , •• ".,.
"&""""
0
~
It
II
//
/ //
./
/
000
°oe
U)
recorder. Tho sound pips are dark lines extending above the background
noise. The first one is adjacent to the symbol a in Figure 11.4.5.1,
although Figures 11.4.5.6 and 11.4.5.7 show the pips more clearly.
Tho load stages during which the loads were held ccmstant for
strQ..in readings are nnrked.
Since theso r0cordings are from the same specimens as thoso used
in Section 11.4.4, the same stress-strain curves are applicablo.
EMITTED SOUNDS FIGURE 11 ·4·5·1 MIX ;t\
-2420
I .
-- - -- - - .. _- _.- .__ ._. - "--- -
'. ,
EMITTED SOUNDS FIGURE 11.4.5.3 MIX 'D'
-970
40001mat
) 14 -1460
I,141.11....6111.10•111010.' • • ."-J
-5840
EMITTED SOUNDS FIGURE 11 4.5-4 MIX 'F'
EMITTED SOUNDS FIGURE 11.4.5.5 MIX 'G'
-2890
°A*011011111.1411010337°Fw
k -5360
EMITTED SOUNDS FIGURE 11 4.5 6 MIX 'H'
-490 -970
„6011,140kulitt.0.144)41
,,,s#4,1
-4840
I 1,
. ;
EMITTED SOUNDS FIGURE 11.45.7 MIX 'J'
-970
-1940
pr
11.4 234
Photograph 11.4.6.1
Table 11.4.6.1
FREQUENCY ANALYSIS OF EMITTED SOUNDS
HIX A
Load
(psi)
AVG.
f*
S.D. C.V.
Load
(psi)
Avg.
f * S.D. c.vl
---
-350 3.2 1.1 34 -350 22 2.5 11
-700 3.2 1.0 33 -700 22 2.7 12
-1200 3.1 0.5 17 -1200 20 2.5 12
-2200 3.2 1.2 37 -1950 20 1.5 8
-3100 3.2 1.4 43 -2200 21 1.4 7
-3650 2.7 0.3 11 -2700 18 3.5 12
-4050 3.4 0.2 6 -3100 19 0 0
-4350 4.8 - - -3400 18 1.7 10
-5100 2.6 0.6 23 -3650 21 6.2 29
-5750 2.5 0.7 28 -4050 16 2.1 13
-6100 3.4 1.4 41 -4350 80 - -
-6300 2.6 0.2 7 -4600 16 1.9 12
-6600 2.8 0.9 3 -5650 20 0.8 4
-7050 2.4 0.3 12 -6100 20 11.5 58
-7300 4.3 0.8 19 -6600 17 2.4 14
-7500 2.9 1.1 37 -7050 28 10.7 39
-8000 2.8 0.9 34 -7500 16 1.7 11
-8000 28 10.7 75
-" ~
MIX C
HIX F
----,.... ~-
Load -.----------··-r-----~--~~---~·--·-~·-·
( ps i)
Avg.
f * S •D•
. .......-
C. V
.........,-.... -
-~ --~- .....
Load
(ps i) . Ii
Avg.
f *
--~.-~
S•D•
..
.
C. V•
.....-.....--.-,----..........-.....--.........-- -
MIX J
The Dxperiments in this ch.::tptcr showed that the testing frame so damponed
tho rosonant peaks that tho value of Q might be difficult to measure, if
nDt moaningless.
~:.
I ~
-======~.
71 ~~.l-,
If ~.::
lr_ <=t ~.:.
=====:::I~
~'
r'
;)
~~I
[ .c-:=.
.I
-~ ..J 1.
~i
::z=;··r;·
-=7.: . . ]":. . ".·
. ~i
-I
I
[
~ .. I I
~\
--;l;.,.
o 1
I
0 o
(a} ((:? ) (c) f\)
_~.
applied compr'S'ssive
.. load increasing froDl
~,_~_~~"".",>,"""",~ ~~,_:"-~."...r:~
(2)
_ _ _.............;:cr",,,,,,,_ _~_~,,,~ _ _~~~-=
to (c)
_______.~'
~-.,.:.:-.~
.... , .~
CJ)
high stress ccmcentra tions. \"lhen this occurs the local stre3SGS a.ro
rodistributed, causing an increase in tho stress in ndjacent aroas.
If this crnck is of sufficient size, its formation and the subsequent
stress redistribution mo.y be reflected in tho ol-:.stic behaviour of the
spocimen.
hiGher volum;,.; cc;ncentr:::t tion h:J. V0 D. hi,~her scn tter of the emitted fre(].uencies.
j) J, only
With tho exception of Mix ~n insiGnificant amount
12.1 INTRODUCTION
Introduction
12.1.2 Theory
TESTING
12.2.1 Equipment
The specimen was thon checked for strain grmgo (lrift, and
initial pulse velocity and strain gaug'') readings were taken. A stop
watch was started and tho land increased to the first load stage,
where the strains wore agnin rond with the Sol~rtron dGtn logger.
Three readings were taken on tho pulse velocity apP1.ratus, another
set of strain rendings were takon, and tho timo recordod. Tho lond
was then increased to the second lond stago, etc. By measuring the
strains at the start and finish of each lond stage, it w~s possible
to determine creep during the pulse velocity readings.
12.2 253
both these specimens wore generally within ~ 5%, especially for Mixes
A to F. However, for tho concrete Mixes, G, H, and J, the variation
in some instances was as high as ! 10.% or, at high loads noar ultimate,
even greater. One specimen in Mix J produced erro..tic results and
another specimen wns substituted. The compressive strengths were
calculated by dividing the true loncL by tho cross soc~ionnl area of
the specimen; and tho results were averaged. The agreoment was good,
genern.lly within ~ 3%. The pulse velocity was calculated by dividing
the distance between the transducers by tho time required for the pulse
to traverse this distance, as determined by tho Cnwkell apparatus. To
facilitate plotting and comparing the data, the pulse velocity at any
load was dividod by tho pulse velocity at the initial load stage,
thus reducing the plot to relative values. Tho results are shown in
Figures 12.2.3.1 to 12.2.3.8.
i
:0 \ .
\
\
~\ r1/
.
.'
....
rol/ Pi-, . _),1.
--' \
\
I .
..'
~/
\ ! / \ I . .-
~ '?\"J .¢ 0 \ i ./
'~:.
I .... I ' ...
:-0 . \ i ...... r0 \ ! ./ .
o u ~~ q ':,< fI' TI a." ~ } . P
g \~!;.\ I
J,' \ \
./
i/
~
C 8·
0 ~\ f~~\ ."" ~7
\i ./ 2( ~ r \ :' <:::-...
o
I
t:.'l),~ ..\ :d (j) O. y 0 C,~)'O ~ ...0 G> 0
»~ \1 ,':'''''"1)' ~ 0» \J . ./.'~
o
k..~t»
o :r-----+-d.--
?
b ,,-.\
.---~.\
9"PR=-38GO./---
-~
CJ)
C
\ CTpV~// -3980 JJ
.p' TJ r ~
G f; » e
0 f1l ><
»
:::0
~
0 i.
L..a
-
0
~
f(\,)
0 n ,::
i\\ 0 r
; \I~;--\ \ \XP-J 0 '\
.~m
,./
.)
'~,9 J
~ :l f' ;
.
<;
~~ ~\ l:>
\
\ "
........
.....
.p...r./·
fll
r-:\
~
~
f\)
.
G 0
~
(j) 0
;0
f1l
0 () ---
» ><
11
II
fj ~ (f)
r 'iJ
(J) ~
()
--
LJ
C
f1l ---,
0 b
! \ .
~ I;l~
i
\
...
\
fll »
~
.'
. ""
r- i ' f\)
.
[\)
0
II
II
~ -~
/'
).<
f tt····p-·-cl· . ~\ <:J (j) I ~
~
1" bf . .p. d M-\ W ~ I
(X)
8
I
Q
!
.
\
n.............. "........ ~.-.~ ....... ,..............................~~, ..... Cl...
~
" ,0)
~~
' -
-I
::tJ»'
I
(X)
8~
,.1
~7
1~
I;
.
\
\
b-..-4
.
\ .
. -1'\ tv
(JJ
~.
o 0 l:.~o b..---t> ~{:
'00 STRAIN -20'00 (L!S.~-) -3doo Z 0 Lo'~~. -10~O STRAIN -2doo (l.J.S.~-:3~OO
o LAT 400 STRAIN 800 (qs.--..".--) 1200 0 LAT. 400 STRAIN 800 (4S-..~-) 1200
1.025 P.VJp'Vo 0.975 0925 (~---~~) 0875 1.0?5 P.V/P.Vo 0.975 0.925 . ~""""""~=-) 0.875
I
I
:I
·~I STRAIN STRAIN -500 . (L!.s....... O .••••. ) 0
4.0 ~'Q'-1
J
/~
. ./,s
//'" ..
.:
:"".
/r6'
....../,6
ON ..-..
1 . / ...... TIg /:
-·1 0 !J./l
L 0"00= -2490
=_
~5
a ..../Dl·
I>~~ 28~O ~~ O"pv= -2550 .../ ~~
1 ,\ :'
:/~ ~0
\J~~
~--.~ ~~------
.
O~ ___ /~
l -0-
r :s: ~
;0
J,. /'. rs:jj
1
o. /'-3860 I~ 1f)>::O . Ok.
Zo if.
/ ~f)>:U
000' fTl X
8!\ . '-W
°5~~
1
fll« .. O· /"
i 0
,/ .' °r'»~
.1\ ·
40
0
1
~ b..t
I] -U
~ OOX-
cl ~ I
CYpR- /48 .......
(j) l;)> " X C-u .....-l- EJh»~
I' !
~.
\
~\ '-W::o
G> \? b IJ~ r(f)
(f)v ~ '\
lot o;PV=/ -4890
/ . T]
- QO·0 0...
::0"
rYV"\
I'i ::::0 T
~
0- '-AA.)
!
'm r;i:
."" \\. . -~ ,
~
I\) ~.
I \
\~ ~
r:.:> 0 :t>.
1
,..
I. ./
p.-~
\ G) 90 .
!
A
! ,
I\)
I ~ '~-...
! ,
e., J!i.·_···o ~ (f) III
i ~ (f) 'I. i-"
! :\, £\)
i
I
CO a) -;00 ~, ~.. -..g , ....
p--~
(Jl
o :::00 " (Jl
o ,~
'I o »8
j
i
o LONG. -1000 STRA.lN -2000 (~.S. -.-) -3000 z 0 LONG. -1000 STRAIN -2000 (~.S.-H-) -3000
J o LAT 400 STRAIN 800 (qS.--&--) 1200 0 LAT. 400 STRAI N 800 (~.S.--~--) 1200
1.025 P.V/P.V O Q975 0925 (-"-) 0875 1.025 P.V/P.Vo 0.975 0.925 (~) Q875
STRAIN STRAIN
f"
III
ir---1' _j I
~
b
ai
/ f-
I
//
! I
. A
k\ 0'
i
f• I
I i
o tJ
~ ~n~
f\)'
g
I ~, rj J (JPR=-2100 I I / ()
o \ \ \ ! (J,...,,=-2420
8 'l]
r- o
i j =-2720 .
L! ~R .-
Or
0 '-
!'
' .
!
!
_'
TJ u'm
-1
I"'i ::0
T]0~ » .t.
~~
m
i
9-e 0 (J) .:..
i 0·» / c 7
(j)~--f O0
~ fT1L~X
I
_ : : : 0
~ h\ c»nl
::o?< ~ 00 ~
o
o \
b~ rrlr»- -"
Zo
· i
--'»11
~ooX
\ L
\
(\)
. »0 II
(j) ...
8 l\> O:l'- (j)
\ ( \ ) .0 I' (\) 1IU1 ...
n' II ... • I'
Wi
• U1
w(j)
• . ()) <
())O>~
1\)-";;
(j)
o U1 ex>:::;
() o
011 o oQ
'0
U1
o
~
0Q 90 U1
fj)
;:0 (f)
~I I\)
- -n CD
(j)
~
AJ 0
».0
........0
(Jl
.0)
o . LONG. -1000 STRAIN -2000 (~.S.~)-3000 Z 0 LONG. -1000 STRAIN -2000 (qS.----)-3000
o L. AT. 400 STRAIN . 800 (Lj.S.--A--) 1200 o LAT. 400 STRAIN 800 (4.S:-~-) 1200
1.025 PVlPVO 0.975 . 0925 (~) 0.875 t025 P.V/P.Vo 0.975 .0.925 (->-) 0875
J
o ,MIX 'J' 257
OOLO
o q,.. CONCRETE A/C=7.5 W/C=O.55 o Of'.
l!)<D ON(X)'
AVG. MA'X.' LOAD= - 5200 C( ~d '
,....,....
FIGURE 12.2·3·7
1 iT
o 0
T'
....6
1.
(f)(f)
I, '( .
ro M I: :j-~
'I\t LO
~
'--'--
-8° oO'LOC\J
C\J N
i fi'
i I
~(X)~
••'Q. .... '
:I
/~t.o
••••••••••• Q
.,.... (.
"11 ......, : ! zz
... .........
~ i I
..... ~ ......... .G)
! 1i
~~
'i}.........-n...
'"'" ...... 1;1
..........-'"
-...
1"('/ I
T
l-I-
(/)(/)
j
il
°°OLO
./
°LON
OQ
I
Of'.
O~(j)
'7 a
. 0
(9~>
z D..;
O«~
~-.J~
o lD
o OON
$f 0 0 (j) -2000 LoAD -4000 (PS I) - 6000 -8000 0
~
G / (j)
d
H --'--J--'-- 'I
/
Ii-
J I
i
V
LO
~ _ _ _ _~_ _~~~~~_ _~_ _ _ _~~_ _~_ _ _ _~_ _ _ _ _ _~C\J
, 0 (j) -2000 -<3000 -8000 $
12.3 258
Item (b) is not valid. A simple calculation shows that for a lateral
strain of, say, 400 microstrain, the pulse velocity would docrease a
negligible 0.04 per cent. Hence the decrease in pulse velocity which
occurs in concretes must be caused by an internal disruption. From
the work previously conducted by Jones, it cnn be concluded that this
internal disruption is of the form of a tensile crack or cracks.
(a) MEASURING
TRUE PULSE VELOCITY
·time
markers,
\'
front of ~
return pulse ,
,(b) RECOMMENDED
FOR DE!ERMING ()pv
., ."'
, .
261
CHAPTER 13
X-RAY DETECTION OF MICROCRACKS
-~I-----N-C~O--N-C~RET='=E....
· --";"UN~ID~E~R LOAD
13.1 INTRODUCTION
13.1 .1 Introduction
13.1.2 X-radiography
For example, a crack is nothing more than a thin air or water pocket.
Since air or water is less dense than the surrounding concrete, a crack,
if properly oriented, will appear as a line of high intensity in the
emerging beam.
I = •••
x
Total AttenuQtion
13.3.3 Absorption
Lp = K}..3 + ~
p
where A. is the wavelength of the X-rays and K is a constant over
certain· wavelengths. Absorption will then be greater for long wavelength
X-rays.
13.3.4 Scattering
cu From' (NB~)
L
. en
~l"
E
u
-."
z
o
·fc
<:(
::)
fo/t
z
~a2
~ Coherent
. scattering
rv
(j)
<D
01 • . -J -. I
.. ] -. I I
~:Jr::It~
I . ~
(59 0 )
(TrJ
0.01 0.02 0.04 0.06 0.08 0.1 0.2 0.4
PHOTON ENERGY'
---
270
13.4.1 Introduction
The previous section discussed tho n~ed to use 10n8 wave length
X-rays. These can be produced by most commercio.lly availablo machines,
but a number of other factors must be considered before the most suitable
machine can be cho"sen.
I
!
60 I
/
I
/
I !
I
/
/
/
/
/
40 / .
/
-
'. E /
/
-
~
u
/
/ .
/~uc -Uw:
"'-'
z
o- /
f- /
«::::> /
z /
~20 /
~ 7:
'/ '
/
/
/
/
/
.~
/'
./
;,;-
:;..":;...'
o~~ ! ,
o 0.5 . 01.0
WAVE LENGHT .(A)
13.5 272
The X-ray unit used should hnve low inherent filtration and a
polychromatic (multi-wave longth) beam. Since the X-ray tube itself
is encased in n cooling liquid, the Iiwindm·Jn of the case and the liquid
should be nnde from such materials that the ,desirable long wavelengths
are not removed by excessive absorption. For a similar reason filters,
which permit only a given wave length to pass, will allow only short
wave lengths to reach the concrete. Therefore a polychromatic beam,
such as that from a tungsten target, is bost.
curve for the available films must be studied. From these the contrast,
latitude, speed, and some idea of the sensitivity may be determined.
D = Log E - log K,
2.0
contrast
From Kodak,ltd and (GLC).
>-
f-
under-exposure . correct·
(/) exposure
z
W1·0
o
1«0 latitude - - - - - -
!!-If
fog
Oi..J .
Inertia f\)
, , ! ___ l
-.....J
~
100 200 600 800 1000 2000
-. --._- (milli -ampere seconds)
----------
13.5 275
curve. High contrast films generally h~ve a large inertia and thus
require long exposures. Tho fourth and last important characteristic
of X-ray films is sensitivity, but this is difficult to define from the
characteristic curve. Basically, it is the amount of detail revealed
in a photograph and is generally measured by a small step-like instrument
called a penetrometer. This permits [l det~rmination of the smallest
I-
detectable change in thickness of a specimen. Yet Pollitt and Durant
(P,D) have shown that p~netrometers are not a valid measure of crack
sensitivity and thus of little signific~nce to this work.
Exposure Time
13.5.8 Developine
The distances between the film, specimen, and X-ray unit are
particularly important. As previously mentioned, the eye can detect a
change in density of 0.01 providing the change is abrupt. To reduce
any fuzziness or penumbra, Ug (Figure 13.6.1.1), the X-ray unit should
have a small focus, and be placed as far from the specimen as is
compatible with the decrease in intensity with the square of the distance.
Also, the specimen should bo thin so that defects will be nearer the
film, and the film should be placed in contact with the specimen.
Crack Size
, Ug = Fe
8,
Specimen
r------:- -------- ...,
I I
I I '
~ -1--1
" -li-A r - - - Jo--lI I
.I I..
I
I'
I
I
C : I
I
I
I I
I , I
-"'--------'"'----i---.l"----'------'-- Fi 1m
13.8 278
Specimen Size
13.9.3 Film
the Ilford ex film required only 6 minutes. Both films were developed
in Phenisol developer for 9 minutes at 68°F and then fixed with an
acid hardening fixer.
The control panel for the tosting frame and the X-ray controls
were all located in an adjacent room. In this way no one was in the
testing room while the X-ray machine was working. A flashing red light,
part of the Muller equipment, and rndintion signs served as warnings of
danger.
For any Given mix the two X-ray and two strGin gauge specimens
were prep:tred in Dc. similnr manner to tho sound and acoustic speciI!1cns, the
only difference \vo.s that tho X-ray tests were conducted the day aftor the
sound and acoustic tests.
placed on tho specimen and then relensod. The load was then re-applied
to hold tho ali3nmont of the specimen.
The X-ray film, '\,'I'hich hCtd beon stored in Do cool, dark room
somo distance from the laboratory, '\,'/0..8 then prepared for the two
initial X-ray photo~rnphs (i.a., \vith virtual zero load on tho specimen).
The cassetto, smooth face toward tho X-ray unit, was held
flush against the specimen by two olastic bands. This is shown in
Photograph 13.10.2.1, exco)t that for actual testing tho X-ray unit
was on a trolley o.nd the testing machine control panel was in the
adjacent room. Lead putty was then forced into the corner formed by
the specimen and cassette- (Photoe;raph 13.10.2.2) to absorb the intense
radiation passing near the specimen but not through it. This, in
effect, prevented burning the film Gnd blurrinG the edgos of the
picture of the spocimen.
Tho X-ray control panel wns set at 1OOkV , 5 ~~, and 15 minutes,
and the unit started. ~Vhile the exposure was being taken, the next
cassette was prepared.
~Vhile the second exposure was beinG taken, the exposed film
from the previous exposure was romovod from the CQssctto Qnd inserted
in its folder. Tho folder 'v'IGS then lnbelled. Fresh film was placed
in the: cassette, in this case since both initinl photographs had been,
or were in the process of being t~ken, the faster Ilford film was used.
For each spccim011 there were: two views (north and west) and
hence two sets of films taken at progressively hicsher loads. Therefore,
for either the north or west direction it w~s possible to locate a crack
and follow it to failure via the lond stages used.
13.11 287
13.11 .3 Presentation
The proportions wero such that when the tracings were photographically
reduced to thvsis size, the crack pn:ttorns \voulc1 bo nctu8.l size. The
results are shown in Figuros 13.11.3.1 to 13.11.3.8.
a) the load stage at which each X-ray exposure was taken, and
For example, the length of the crack nt :::tny :-;iven load stagG is the sum
of all the types of lines shown on the key below that load stage. The
origin of the crack is tho location shown by the line representing the
smallest load stage along that crack. Shrinknge cracks arG shown by a
series of xxxx's.
0.) the lines form and spread with increasing load in a manner
expected of microcrncks;
FIGURE 13.11. 3.1
X-RAY DETECTED CRACK PATTERN
AND 289
COMPANION STRESS-STRAIN CURVES
IP IP
(ACT SIZE)
WEST VIEW NORTH VIEW (G.S.R.'64)
tP tp
MIX 'A
MORTAR A /C=2.0 W/C= 0.42
AVG. MA X. LOAD = (X-RAY) -8430 (STRAIN) -8200
I
KEY TO X-RAY DRAWING
>- IIimo
<
o 0
08000
c‘, (di ,r)
r,) —
<
2
0
O —2000 LOAD —4000 (PSI) —6000 —8
0008
X-RAY DETECTED CRACK PATTERN
AND 290
COMPANION STRESS-STRAIN CURVES
1
(ACT SIZE) NORTH VIEW GSR'64
tp
MIX 'B' FIGURE 1311-3.2
MORTAR A/C=2.0 W/C=0.42
PLACED AGGREGATE
AVG. MAX. LOAD= (X-RAY)-8180 (STRAIN)-8180
I
KEY TO X-RAY DRAWING CL 1 I I I
CC Cr CC x cf. co 0
>k x_J _ds8
,- 0
a
2 1".n -Jri Cr)
D- 7'
0
0°
0 0 0
CV CO
Z Z Z
74 ( 7
CC <7
Et CC CC
I- 1-
V) V) (r)
0
8 008
o tE___j.
z Vzo
a_
0 00
0 (I) -2000 LOAD -4000 (PSI) -8000 7
X-RAY DETECTED CRACK PATTERN
AND
COMPANION STRESS-STRAIN CURVES 291
P
/P
( // 11
WEST VIEW (ACT SIZE) NORTH VIEW (G.S.R. '6 4
fP
MIX 'C' FIGURE 13.11.3.3
MORTAR A/C=2.0 W/C=0.42
PL ACED AGGREGATE
AVG. MAX. LOAD= (X-RAY)-8420; (STRAIN)-8180
I
KEY TO X-RAY DRAWING CL I I
.1 , i 4 ..i
.1111111111111111111111011/1111410111111.11.1141.111.1
< f 4
‘4,
_rno
,L..
CC CC
0 0oo 0
X k k k --I _ o0o 0
o N (.0 0
k re)
<
2
D- a 7
0 0,
; n
O ny
0 co' '7 '-'
C:4
0 zz z
0 1
r, ...it ZI .71 ry 'Tr
in —
. i .... Ic H
r". cno Lc)
.. ,,
ix I °-1 0 (---)
(_ - 0
61- ------- OvNin
0000
(I) -2000 LOAD -4000 (PSI) -60010 -8000
X—RAY DETECTED CRACK PATTERN
AND 292
COMPANION STRESS—STRAIN CURVES
<
cc! cr
>l< X
()Ui Tr>
o-
0
0
1
•••• •
.••••"'"*"
4
(i) -2000 LOAD —woo (PSI) -6000 -8000
or, ••• .....
m ...........
H
0 X-RAY
m m .....
C ...... ""`
O •••••—^ 1,40, ,,,,, "
..
,
/,\ •••••••-.../
N.)
0
0
0
r_ X-RAY :u ••••• nnf .
......
ID
ono
..... .....
0 ••• ••••••••=•~•••••••••- .......
rrl
i
0 —
etc
10
g •E 'L L •C LEelneH Fri
..
.... \OM .............................
\ X-RAY
,•\ • ...........
... C
m
••••
-o < 27
O
0 X-RAi z
0
MAX. LOAD (r)
GRAPH
0 LONG. -1000 STRAIN -2000 (US -3000 -
0 LAT. 400 STRAIN 800 (US) 1200
0 PR/PRi 20 40 GO
-1000 VOL. -500 STRAIN 0 (.45) 500 4
X-RAY DETECTED CRACK PATTERN
AND
294
COMPANION STRESS-STRAIN CURVES
P
1)
.
!
WEST VIEW (ACT. SIZE) NORTH VIEW (G.SR:64)
IF
MIX 'F' FIGURE 13.11.3.6
MORTAR A/C=2.0 W/C=0.42
PLACED AGGREGATE
AVG. MAX. LOAD= (X-RAY)-9500; (STRAIN)-9260
KEY TO X-RAY DRAWING a_ I
)75- < 1,
cc o
>k >> O c: Q
0
.
-. k
2
8
0
0 or' zz z
cc cc
I-- I— I—
(1 U)
9 0
-- 0-
000§
0 (i) —2000 LOAD —4000 ( P SI) -6000 —8000
•..... _
,.
"..........,„. :...:
.• ,,,..‹...
>"
......i,..:.
.. .
.........:::.
SO. ............„... • • ••,,.„ .... .'.. .
... •
.
V) ..a•Z::
• ......
X-RAY1
H
:4:a....
." ... '..,.. .....,,..
• ••••• .
.
-<
. ...................
......
--I '..........................
. ••••••,•••••••....••••
•••• .. .......
o
7•110".. ••
0 •••••
....
......... • ••I
010
>
0
<
0
> 1:) .. , .. ....... t•I ••• 141 >i<
.
ti K>n --.7.••• ..... •• •••_•.......17:
. _______ ....::• ...,:,:.
> >
0 1 e....
.... 0,0, ........ •
GRAPH
0 LONG -loco STRAIN - 2000 (q - 3000
. • . ..... .......... • •. ..
. **•1
....... .......
.. ..
,
0 LAT. 400 STRAIN BOO (is) 1200 ..........
....... ........
....................... ............
o PR/PRI 2.0 4.0 6.0 o5
4000 VOL. -500 STRAIN o (y5) 500 4-% ...... ••
X—RAY DETECTED CRACK PATTERN
AND
296
COMPANION STRESS—STRAIN CURVES
r.
8 8 o8
M
(C3
0
1•C‘l o
of,
101
.e.
Aggregates
Photograph 13.12.1.1
Mix E
Print of X-ray Through 2.75" Thick Specimen
(White Dots Are Metal Wires Glued to Coarse Aggregates;
Thin Blaok Lines Are Cracks)
13.13 298
The fact thQt a specimen only 0.15 inch thick has been X-rayed
in the Cornell studies means thnt D. much lower kV is necessary. Such a
technique can, on the basis of the comments at the beginning of this
chapter, be expected to produce good resolution.
Photograph 13.13.2.1
Print of X-ray Film of a 0.15 -inch Thick Slice
Through
Mix F
Mix H
13014 301
DISCUSSION OF DATA
discussed.
Photograph 13.14.1.1
14.2.1 __________________________________________
BGh3. viour of Concrete n t Loads Less thnn (J ~pv
( ( I 1 I
I " ~
I'p p. II
z
pO?~lble .....-'7[t-'-;:
1/
iJ
~"~
·1-
« '. I'
crltl.ca! . tr-internal ~tra~n
-1J~
..;r~
a: ~
.,-
1/
T· l- straIn lImit ii= distribution' 1/
'r (/) , I I·
1/
'9"
-
. ....-.
/./
._._'~oi?s<?Jl.Ra~ io._._.--!-._. _. _ . /
" /
//
..",
STRESS w
-0+-- o
<0
0pv ()c
310
G - 41.3 2500 7060 - - 2720 2720 6880 2050 4100 2500 6810
H - 44.1 1700 5550 - - 2420 2100 5820 2050 4100 2420 5920
J - 45.9
.----~---.
I 1980 5260 - - 2480 2530 5200 NT NT NT NT
\.N
* correction factor accounts for variation in mortar strengths as shown in Table 10.2.8.1 -lo
-lo
** value in bracket has been corrected for variation in mortar strength
*** NT means not tested
14.2 312
TABLE 14.2.2.2
SUMHARY OF LATERAL AND LONGITUDINAL STRAINS AT a PV FOR
THE SPECIMENS TESTED IN PART III
14.2.4 Conclusions
From both a theoreticnl analysis and experimental results
the substantial influence or the non-homogeneous nature of concrete
has been shown. The high stress concentrations caused by the inclusions
in the mortar matrix nre a major cause of microcracking. It is these
and other cracl<::s (shrinkage, etc.) which eventually cause the failurG
of concrete.
PART IV
CHAPTER 15
INTRODUCTION TO PART IV
15.1 INTRODUCTION
numerous complications.
16.1 INTRODUCTION
(A)
!
I
l/~~_:~_~
.•..
/ f-------rr
I
.\
\
spec. ram!:
f-o----..J,)
I,
I I i
~
I I
I'ramiI
to
L ___ J pumps
L ___ ..F- :=!.~>
320
a) ns only the '0' ring nnd the hydrC1.ulic fluid touch the
specimen, platen restrnint cQused by friction between concreto and
steel platens is elimin~;.ted; .
c) it is very compnct.
largo fro.me of which none wns :lv:til.3.ble at Imperial College, but there
were two smaller testing frClmes, one of which c IJnveniently fitted botween
the columns of the other. Therefore it was decided to design'n double
rnm machine provided the platen restrnint problem was found to be
negli~ible or could be overcome.
f~'\)
a::.
'4
~
~
't
~
~
~
~
~ ~~
~
~ rJ
;;"
§
~
I I
325
Photograph 16.4.1.1
testing frame. Those jacks had been specially purchased so that they
would work in parallel and exhibit similar characteristics. To ensure
this, both rams were firmly bolted to a 4 inch thick platen, thus
requiring the rams to extend at the S2mo rate. The hydraulic system
was kept symmetrical and attachod to a common mnnifold at n distance
of 6 inches from the ram.
The Epco ram was not lapped in to thv cylinder, but only a
loose fit, and the oil pressure was contained by IV' glands. This
meo.nt thnt when used in a horizontc1.l position the ram might, due to
gravity, emerge at n. sliGht o.n31e to the axis of its cylinder. To
overcome nny possible effects of this, a complicated letppod piston
was required to fit over the entire jack. 'rhis is shown in
Figure 16.4.3 .. 1. The oil pr0ssure from the pump forced the Epco rnm
out, applying pressure on the column. As the column hets n betll joint
at either end, nn axial lo~d must be transferred through tho column
regardless of whether the ram w~s slightly out of alignment in its
cylinder. The balls were lubricated with molybdenum disulphide and
the entire area around the column packed with a heo.vy grease. Thus
the column transferred an axial load to the face plate of the lapped
piston. Since the steel piston was lapped to tho cast iron sleeve
over a 12-inch length and the sleeve in turn wets rigidly clnmped to
tho nnchine columns, the piston was forced to travel parallel to the
axis of the mQchine reGardless of the alignment of the internal Epco
ram within its cylinder. This lnd the o..c1.:lec1 advantage of greatly
stiffening the entire frame. The lapped piston was also lubricated
with molybdenum disulphide, but despite this the friction was still
significant. In order to retract the piston, screw returns were
required in addition to the: smClll springs at the bottom of the lapped
piston.
The hydraulic systems for the 100- and 200-ton mnchines were
identical. The pressure for each system was supplied by Cln Oswnld and
Rid~ewny centrifugal pump, type SW01. This fed to an Osw:11cl and Ridge''lay
relief VC',lve set at the m!J.ximum opero.tinc- pressures of the 100-ton and
200-ton rams (8000 o.nd 5000 psi respectively). A spool valve was used
as n control to the rams. Betwoen tho spool valve ,2nd thG ram aT-joint
led to two Burdenberg 10-inch testine; SO-uges, one from 0 to 2000 psi :md
331
c:
the other from 0 to 10,000 psi. At high pressures an Ermeto stop valve
isol::tted the 2000 psi gauge. The ontire system 1.tJ~s interconnected by
i inch OD Ermeto high pressure steel tubinG and fittings. Steel tubinG
is preferable to thc flexible tubing required for the X-ray ~nd acoustic
tests (Part III) since the stiffer steel tubing increLlscs the elasticity
of the testing m~chine (pc-OTS). The tubu lending to the 100-ton capacity
ram WD.S coiled to pe rmi t free movement of the In ter::ll machine (the coil
cnu be seen in PhotoGra.ph 16.4.1.1 b(;t"wcn the 100-ton nr. chine o.nd the
control panel).
16.5.3 Cruciform
Cruciform-shaped specimen
Fisure 16.5.3.1
16.5.4 Slab
This size comp~rGs most favourably with the 3.94 x 3.94 x 0.98
inch specimen used by Weigler and Becker, since it:
All the steel platens of the bio.xial testinG rrnchine were only
9.75 inches long in order to prevent intErrferenc e between the pIc.. tens
at the corners of the specimen.
16.6.2 Acknowledp.;ement
16.6.5 Conclusions
I
L' I .
J
~
( i
.r
: It
. : il .
~
: '\
~ I
!
;"~
I . ~. i, \\
: 'I
. , \
I \
:i \
Zo 3 .~i~ ~\
1--~c.L.f .'
,1 \
.I- 1\
r
\
(3 i\. \
~ \
U)
Z
W
5 r ....
.. '
.ff.
/.
t~ //r;]
\./
.... ~/ \
. I--
.X
6 ~1O
I
W 7 ..,.
_-i--
_-s------~.
~..s....
/.
~ .
.J
16.6 338
TABLE 16.6.4.1
EFFECT OF VARIOUS PACKINGS AT THE CONCRETE-STEEL
INTERFACE ON THE LATERAL EXPANSION OF THE SPECIMEN
- -.---
The distance from the
interface to which
. No. the lateral strain is
Packing Material affected by
of
Spec. ~), 3% 5% 10%
Series I Mortar specimens
.-
1) Febcure super curing compound 3 2.3 2.0' 1.3 <0.2
2) None,concrete surface dry and
steel platen 3 4.0 3.7 3.1 2.2
3) Molybdenum disulfide (Rocol) 3 2.8 2.6 2.1 0.7
4) PTFE baked on platen 3 4.2 3.8 3.1 1.3
5) PTFE sprayed on surface 3 2.7 2.4 1.8 -
Series II Mortar specimens
1) Febcure super curing compound 3 3.1 2.5 1.8 1.1
2) Ship's wax 3 3.8 3.5 3.2 2.6
3) Acetate Durafixed to concrete 3 3.2 2.8 0.9 0.6
4)
5)
Restrained rubber
Restrained rubber
t
t
inch thick
inch thick
3
3
2.8
3.6
1.8
3.4
-
3.2
-
2.8
Series III Mortar specimens
1) Febcure super applied on damp concrete 3 1.0 0.8 0.6 ...
2) Febcure super applied on concrete
3)
left to dry for 20 minutes
PTFE sheet and PTFE platen
3
3
2.1
4.0
1.9
3.4
1.5
2.8
-
2.0
4)
5)
Sheet of acetate 0.003 inches thick
Three sheets of acetate
2 1.2
2.4
0.9
2.2
-
0.8
-
0.6
3
Series IV Mortar specimens
Hours after applying Febcure super
a) 20 to 22 hours 2 1.8 1.6 1.3 0.7
b) 68 to 70 hours 2 3.0 2.4 1·7 0.8
c) 116 to 118 hours 2 2.5 2.1 1.5 0.1
Series V Aluminium Specimen
--
1) Steel platen 1.8 1.6 1.3 ·0.9
2) Sheet of acetate 0.003 inches thick 1.7 1.5 1.2 0.8
3) Molybdenum disulfide - 0.1 - - -
16.6 339
Machine Repeatability
The specimen was insorted into tho rW:l.chine and various loa.ds
a.pplied in both directions. The specimen was then removed and the
entire process repeated. This was done a total of 12 times, 6 with the
gauges vertical, 6 with the gauges horizontal.
From theso tests it was found that for any given set up of
the specimen tho strain measured by tho g.:luges was repeatable to +
- 1%,
the accuracy of the strain gauges. The maximum difference shown by
any single compression gD.uge wns :t 1.896 and this was near the edge of
the specimen.
State of Stress
11' II 11'< 1/ 11 03
...
/I " II II II
+-'
cu
(})
U1
, ,
16.7 344
a) Failure originated :in the centre of the specimen and not :in
a corner or particul~rly near one platen.
results agree with those obtained from the unchine described in this
ch~pter and presented in Chapter 17.
~I
0'
I
'>t " . 0 o. ".0. N : o· -or ·0.·· ill)" (Y) "(Xl
o f). . (;\
0 .;
.Q
. 01-
1
-: Ci.' . Q\ <l?
(f)~>l-7"t-= or-. -,
o
" -~. -,
0
I
0 - ,_ .
E
I --
co
L
. " ,I':
'<'I '
o·
. .
r
'~~(--~I.~---J~I\--'~+---n~~J~--.-J~\--~J\--'-~rt~J.\--~1l~
: . If 'I Ii . I . T ". . f' f T I r'
0-3 ram Side 03
16.8 347
17.1 INTRODUCTION
With the biaxial compression testing machine described in the
previous chapter it is now possible to conduct a series of tests that
will extend the study of Parts II and III into biaxial compression.
The data may also be used to confirm the conclusions of Part I. Of
particular interest will be the influence of the aggregate-cement
ratio, the water-cement ratio, coarse aggregate, and especially the
lateral load on the ultimate strength and ° pv .
17.2 PREVIOUS RESEARCH ON THE BIAXIAL CONPRESSIVE STRENGTH OF
CONCRETE
\veigler and Becker did not correct for any stress losses due to
friction (Section 16.7.4).
17.2 352
17.3.1 Haterials
"As only one batch was prepared for each mix, a sufficient number
of specimens within any batch was prepared so that the results, when
averaged, were significant (M\rJ).
354
TABLE 17.3.3.1
Number and Type of Specimens Used in the Biaxial Tests
------------------ ..-.--------
Type of Number of Specimens
Specimen in each mix
Standard Tests
1) Uniaxial tension waisted (Ml\h) 2
20" x 4" X 4ii
2) Uniaxial com- 12" x 4" x 4i1 2
pression
3) Flexure 20" X 4" X 4" 2
Biaxial Tests
4) Biaxial com- 10 B X 10 11 X 4" 6
pression
17.3.5 Casting
Casting \'Jas similar to that described in Section 10.2.7
except that the moulds for the biaxial specimens have not been described.
17.3.6 Curing
The curing procedure up to and including stripping was similar,
to that described in Section 10.2.7. After the moulds h:tve been stripped
the specimens were placed in a special curing tank at a temperature of
0
70 0 ±0.5 F. During curing tho tanks developed faults and twice the
temperature fell to 60 0 F and twice rose to a:?proximately BOoF. The
dur~tion of these temperatures was only a few hours, insignificant
when compared to the remainder of the 28 days cured at 70 oF,
The only part of the specimen free from curing compound was
the area where the strain gauges were to be placed. This area was
allo\ved to dry, since;; the strain gauge cement would not adhero to a
damp surface, a.nd then the strain gauges were applied. The gauges,
the point of contact with the lateral gauge, and tho surrounding
exposed concrete were covered with a light grease to prevent further
evaporation.
FIGURE 17·3·5·1 356
,
I.
,
14-......
,--- 12±001" -----iH+-+---+----;--~t__i
I I
I, .. i
1.5'~ 1.5' .j. 3' .L . 3' --1-1.5'+-15
. 'JO.01·-------------I
PLAN 'VIEW
STEEL MOIuLDS FOR
BIAXIAL' COMREtSION SPECI MENS
close tolerances are
required on angles (9(1)
I
marked
I
I
I
~~--------------12~001------------~----~~~
I
SECTiON A-A
357
17 .4 TESTING EQUIPHliJJT
TABLE 17.3.7.1
SUMHARY OF MIXES 1 TO 6
02 -~ 1--._.- of-'
-I
. I ,
I
,\02 :. :.
front view side view _ .
BIAXIAL COMPRESSION ....... '
{1 U3 1 U3·
- I--~.-;-
, . I
\
I
..
i
:1
, ~~
,I trCTtI
. .. 11\ 03 . Ii 0"3 UNIA XIAL TENSION
UNIAXIAL COMPRESSION
~! ~,
-I
FLEXURE
17.4 361
r(2 •1 11
!M.
Photograph 17.4.3.1
Lateral Extensometer and Transducer
Photograph 17.4.3.2
Lateral Extensometer in Position
17.4
17.5.1 Set up
After tho spocimens had been prepared on the 28th day, they
o
were moved into the testing room (temperature 68 0 : 0.5 F) and left
overnight. On the 29th day a biaxial specimen with strain gauges
was inserted between the columns and into the machine with its cast
fnce away from the operator. The specimen was aligned within the
verticnl machine and a slight load (-200 psi) applied to hold the
specimen in position.
Photograph 17.4.4.1
The lateral load was next increased to -650 psi and then
exercised three times over the range -22 to -650 psi while a 3 was
varied from -785 to -1720 psi. Three sets of strain readings were
recorded and averaged for each of the following methods of loading:
a)
°7
:J
= -785 °2 varied from -22 to -650 psi,
psi;
If the specimen behaves elastically and the frictional force doos not
vary significantly, the sum of the strains due to (n) and (b) should
equal those caused by (c).
consto.nt while 03 was incren.sed at D., :rate of -700 psi por minute
17.5 368
The other four specimens did not have strain gauges fixed to
their surface and were tested for 0pv and ultimate load only, for
various values of O • The lateral loads for each specimen were
2
-1980, -1290, -650, and ° psi. This corresponds to the laternl loads
used for the elastic measurements on specimens No.5 and No.6. For
these four specimens the pulse velocity transducers were placed in the
centre of tho specimen.
17.6
TABLE 17.5.4.1
SU~ffiRY OF THE LATERAL LOADS USED FOR EACH SPECIMEN
.---~---~.~--------------.----.-~--~----------
17.7 DATA
The data from each of the six mixes are presented in Tables
17.7.1.1 to 17.7.1.6. These tables are briefly explained in the next
two sections.
TABLE 17.7.1.1
SUMMARY OF DATA FROM BIAXIAL COIvIPRESSION TESTS
PR e
e2 :IT: PR 1/ S2
2 ~ &1
(J
s 23 13
x10+6 x10+6 x10+6 x10+6
Specimen No. 5
0 -198 +38 +32 4.72 0.19 0.16 0.84
-22 -199 +34 +24 4.70 0.17 0.12 0.71
-1290 -195 +28 +29 4.79 0.14 0.15 1.04
-2700 -188 +34.5 +30 ·4.97 0.18 0.16 0.87
Specimen No. 6
0 -198 +39 +24 4.72 0.20 0.12 0.62
-22 -200 +36.5 +58;5 4.68 0.18 0.29 1.60
-650 -193 +26 +51 4.84 0.14 0.26 1.96
-1980 -191 +29.5 +45 4.90 0.15 0.24 1.53
-3420 -188 +34 +40 4.99 0.18 0.21 1.18
Uniaxial
0 -201 +40 - 4.87 0.20 - -
Ultimate (0
c and Cracking Loads
) (0 )
pV - Ev
Corr. ac Corr. u c. Corr. v u....E!
°2 factor 1b c 5 o pv a pv
ac 0
(OpV) 0 0
c
Uniaxial
0 1.000 -7320 -7320 1.05 -4350 ..;4350 1.06 0.59
TABLE 17.7.1.2
SUMMl-\.RY OF DATA FROM BIAXIAL COHPRESS ION TESTS
Mix No.2 Concrete (W/C = 0.45; SiC = 1.8; CAlc = 2.7; Alc = 4.5)
Elastic Measurements "-
-
a
2
e
3 E2 (;1 E
3
PR
23
PR
13
CJ 11 &
2
x10+6 x10+6 x10+6 x10+6
-- .
Specimen No. 5
0 -122.5 +16 +19.3 7.63 0.13 0.16 1.21
-22 -123 +16 +21.8 7.60 0,13 0.18 1.36
-1290 -118.5 +14 +17.4 7.89 0.12 0.15 1.24
-2700 -114 +21 +16.2 8.20 0.18 0.14 0.77
Specimen No. 6
0 -131 +19.5 +15.6 7.14 0.15 0.12 0.80
-22 -131 +19 +30.5 7.14 0.15 0.23 1.61
-650 -126 +16 +30.5 7.42 0.13 0.24 1.91
-1980 -123.5 +17 +30 7.57 0.14 0.24 1.76
-3420 -120 +22.5 +26.8 7.79 0.19 0.22 1.19
Uniaxial
Uniaxial
0 1.000 -6440 -6440 0.97 -4830 -4830 0.97 0.75
TABLE 17.7.1.3
SUMMARY OF DATA
FROM BIAXIAL COMPRESSION TESTS
Mix No. 3 Mortar (wjc = 0.45; sjc = 2.4)
Elastic I1easurements
PR PR B1je 2
O
2 £3 £2 £1 ~ 23 13
x10+6 x10+6 . x10+ 6 x10+6
Specimen No.5
8 e 2 e 1 E3 PR PR t 1/e
°2 3 23 13 2
x10+6 x10+6 x10+6 x10
+6
~-#---
Specimen No. 5
--- ..
e e 2 e 1 E3 PR PR e 1/ e 2
°2 3 23 13
c
x10+6 xi 0+ 6 x10+ o xi 0+6
Specimen No. 5
Specimen No. 6
Uniaxial
0 -233 +52 - 4.20 0.22 - -
Ultimate (0c ) and Cracking
' Load (0 )
pv
Specimen No. 6
Uniaxial
Uniaxial
~.--:
0·25(· ~.~~
.---.-.~
.---.--.--
('t")
(\J • 1 "
-00.20";"
I-:
<{ i "
"~ .
.............. 4
......
.
(/) . . . . ......... .
'.' ··········Mix
....... .
W/C _ SIC
) . 045 J·8
CALC
-" x
o "'". . . . . ......fl···· .2. 0.45 ).B . 2.7 _.0--
0.. ~----- 3 .. ' ..... 0.452.4 --t:,.--
.'-' 0·10 ........... ..
4. 045 2.4 · .. ··0· .. ··
3·6
..... :..... 5 0·60 2·4 _.-4--
. ............... 6 0·60 "2·4 w
"'~" ................ [!J ••••• e" 36 --~-
CO
o
0. 05 6 L -10'00 ~2000 -30'00 -4000
LATERAL .LOAD, 02J psi
FIGUR·E 17· 8·2·2
:. P R.13 AS' A FUNCTION OF LATERAL LOAD
(SPECIMEN. NO. 6~
MIX W/C SIC CAlC
1 0.45 1.8 -1(,-
O"25~ "
~.~.-._ . _ - . -'-'-f/)-.-.~. _ _
~
o .
(Y) ........
....................
~'-...
' . .
. .
(
.----. ---~ mix 2
• _____ •
.............
Z . ····················[})JXA .. ----_"
. 0 ··· . . . -;-r.-Sn --
tf).1<----~_. ········7iJ~~i~e·
.!.I/ n n
. -.- - -6 mix 3
tf) 0'15-'
o
. ~
.
.
. ' - - ---.. --... ---------
............ 0 5
-- ..-.-... -.. .
. -:-:-:-:-:-. ~ oL.:'...... • • : :
~ .a..
.
. . . . . . . . . . . . . . .; . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
m IX 4
n . .................. ...:..~.:-- _ . .'.
!::;.~ ........... -......... -····13 ... ::::_~_:.-.;- m~x 6 w
~......................... f1.. - .. :.... mix 4 (NO. 5) ~
0·1 0 . I . • I ~
o -1000 -2000 -3000 -4000
L
LATERA L LOAD, 021 psi.
FIGURE 17· 8·2·3
.e1/e2 AS A FUNCTION OF LATERAL LOAD
. . :(SP'ECIMEN NO.6)
MIX W/C SIC CAlC
1 0.45 1.8
-.e--.-
"
2 0.45 1.8 2.7
3 0.45 2.4 - .- - , A - - -
...... -------.....
, , ........
........
__
.
.--- ---- ,
.
~
" ,-,
. ; .
£!l'" ". ~'. ,"' "-......
C\J'
"
.~
(})
. "'- 1·4 .....-'. '"
'.~ ............. .
........... .
. "........, ""
.....
~
<l)'
-"-" .....
I,
"
~,
~---
•
• -. -.
. -.-.of-- '_._
-2000
! .
.
' -I : - ' - . '-'-'-'-7:fDIX
t
-3000
-
~_ .'
5 !I
4000
co
(\)
The amount of the decrease in Poisson's ratio becomes less with higher
values of the lateral load. The influence appears to be the same for
both &1 and &2 and hence the plot of &1 / &2 is comparatively
insensitive to changes in coarse aggregate volume (or size). For
uniaxial specimens, a similar decrease was observed in Part III
(Table 14.2.2.2).
17.9 THE EFFECT OF0 , Ale, wle, AND COARSE AGGREGATE ON THE
2
ULTIMATE AND CRACKING LOADS OF CONCRt"'TE
17.9.1 General
1
1 0.45 1.8 x '. ' . . " + 1000
2
3.
0.45
0.45 ~:~ 2.7 :_ estimated· 0;-0"3 envelo es Oil
4 0.45 2 .4 3.6 i
...... ~......
r.\
. . . : - - : - : -. . . . . ~:::-• • • ~-.-.~~.
__ - - - - -
' -- -~
:-~.~:
- - --=-----=-------
---:-.-:::::-:.....................
• • •
~
··_·······1'
~
0"3 -80pO . (p s i) . .'~ 0.· .... ··:;.: ~...-.....-.. ~ . -:::Lf<5O" ... ;: /,;,;,;':-:- . . - 2 000 0
. ,.-
.
,no
.....
.-~C
I
y""--- , ."pi
'
,r~~ • , It
•
...... / :/
. ~ . .'. . . ,/ . : ./
'
// . , '/ :;
I
,
/'.' I .j
. (:1' A X'o 0 : ~
td G . ~ ;~
:it
I
I
I I. ' 1
,1
a J~ -2000-
i
61 1
, I II .'
,
I .. l
I
X : Cl ., A
~ If\:.
I
I ' \
\.
~U pV
'\
r;1
O"pv
bl
I
6c S/
O"pV·....
0"2
(psi).
I~ ~
oc
-4000
t:-
. FIGURE 17·9·1·2
.... EFFECT OF W/CON 0c AND 0pv
MIX W/C SIC CAlC
3
',4
0.45
0.45
2.4
2.4 , 3.6
;f.,
-'.-9-'.-
. .',
. .
. ,', + 1000
~
1
5 0.60 2.4 '-- ~--A-- estimated 01-0-3 envelopes . 1 <
6 0.60 2.4 3.6 ······0······ _~~=-=-=.~.~=-........--·i)'
. ~. ---.-:-~ ~:.~:-- ··-if r-:-:":--;=: •• ::::::: ••••••••••••• , ••
=-- - ___ __ ~: ........ ___ e.····
()3 -8000 ,(psi) - 4QOO.~· ...... /~-=-200C)"e .. ' 0
I---.;;----~! .. ..;.~~~-~~.~--a-D~~-----~~--
........ / ' /
..... '/ /
. .... /.1) / ..
.~.
.,. II
'1. 0X '1
/ .. s
.: I/ . I.:
.: I ., I.
X P, if p ," I ..~
:' I. I,:
.. . . . , 1 J
!fil
.
Jj
1~
Fi"A
.. \ -2000
: Ii\ . . i: \\
1
·0 x .p' \ ' 'C1. A\ . ;
'I \...... \
----
. . '
14,- ~~
//
_._.~.---.~
/'
t) /
/ X'
omix 1
~ ./." )( __ ;- ,-;- J.. mix 3
u
t5 1-2--
I
.j
.
;:. ---
~
---
/ ·,' ;..--.-p- .--:-- - ---
J.
I
.
:-. _..------
/ ... :.:1 /'..--"--.- .-: : : '". .-:.--'--
I .~~ ' ·-
----
. r·- .- . - . -
",-:r--........
.
/f 1
1-0~'O' .//" /". -'-'-' -
.... . , / /AJi "':-'-'
/ ./ " . /J>
'1--'-
. £/ .,./
' - ' - '-'-':-mix 5 w
())
(J)
4
5
6
0.45
0.60
0.60
:2.4
2.4
2.4
3.6
3.6 '-~-
......p ......
_.-f-.-.
-- ----- -"";', .......
~.
oII
~
Q
~
>
1-6'"
'." "
. ~mix6'
~
0-> 1-4
I· b
_--- mix 3
. 1-2~ / __ . _"_. .-_-.---- I W
.a..
'"I; ."-
.- '
.,~ •• "
.p .......... ..::;::-:::> --.
-...-......-....
CD
~~( '-.J
. . --' '. .....
.~~·-f;}:.::~---=--..::::..._. .
__ • . . -
. 4 '~" .........
• •••••
FIGURE 17-·9·1·5
--- Opv/CJC - ASA -FUNCTION OF LAT"ERAL "LOAD
~ -- --
--
---..........">.- -il>m ix 3
"
- - "~mix 5
-- -- -- ---
fo
---
~--
--=--~
. . - e--- - - ___ ~r - - -. - _~
--
. - -
r _
. --';;;;;;;;'--- . .,'"
'--'~~'"Q--'~'-' :~·-·--.;>mix 2
- .
./
. .
~ .~, -
o , .r,' "'--
)Y . ,
..... /' ......................... l'l............. ..... ~. .
., ". '. '1<> .. " .-..... "':.;7 / .... .••1'1 '. •: .•.• "llt ....... . -.. '" miX
. 6 w
.
'1~ " , . / ....... '"_ ())
00
\
oN .-
1
\
. \ .
\
o
~ o
-If)
x-
2
0::::
0
LL ....
'7 (1)
--1
Z
- :::J «
<t (9- - Z
0::: - o
LL _..0 :::)
I- r-
-U)
I 0(9
U) -OZ
U) LfIO
W 1--1
0:::
I-
-If) .
0 0 0 0·. 0
0 ·0 0 0 -0
0 0 0 0 0
0 CO CD ~ NI
l' n
1
, I I
0 0LO
; C\J .
. tsd £D -SS3C11S l'v'NIOnl18NOl +
\
17.11 393
As these plates develop and break off from the rest of the
specimen, the strains in the plates relax, for the surface plates no
lon~er support part of the applied load. It is for this reason that
the strain drops off rapidly in some of the stress-strain curves
(Figure 7.10.2.1).
Mix 2 (biaxial)
Mix 2 (uniaxial)
Photograph 17.11.1.2
Close-up of Plates and Aggregate Pieces
from Specimens Failed in Biaxial Compression
396
PART V
CHAPTER 18
THE FAILURE MECHANISH OF CONCRf.!J"'TE
18. 1 SUIvjNARY
Both Rusch (HR) and Kaplan have suggested this point may be
the long-term strength of concrete.
18.2.3 Analysis
Using the sJ..me computer prosro..mme as was used to process
the data in Chapter 3, an analysis was made of the binxial data in
Chapter 17. By comparing these data, usin:s both a and a values,
pv c
with the variations listed in Table 3.1.6.1 it may be possible to
estimate the extent of the error due to the definition of failure and
whether the new definition for failure permits a better cOiilpromise.
TABLE 18.2.3.1
BL~XI.i-\.L DATil. APPLIED TO THE CLASSICAL FAILURE THEORIES
ultimate load
Mix Avg. s.d. C.V. Mix Avg. s.d. C.V. Mix Avg. s.d. C.V.
TABLE 18.2.3.2
BIAXIAL DATA A.PPLIED TO THE CLASSICAL FA.ILURE THEORIES
Mix I. . vg. s.d. C.V. Mix Avg. s.d. C.V. Mix Avg. s.d. C.V.
Maximum strain
(compression) Octahedral shear stress
Note:
Mix iivg. s.d. C.V. Mix Avg. s.d. C.V. same as table
18.2.3.1
1 no convergence 1 2680 453 16.9
2 -4970 236 L~. 8 2 2240 96 4.3
3 -3950 245 6.2 3 1820 153 8.4
4 -4110 209 5.1 4 1810 40 2.2
5 -3210 218 6.8 5 1440 75 5.2
6 -2520 389 15.4 6 1330 247 18.6
18.2 405
that they arc in excellent QGreement with the two theories, with the
one exception of Mix 5, which had ~ coefficient of variation of 29 for
K as com~utGd from the octallGdral shear stress theory.
However, S(;ction 3.3 showed that the maximum strain theory gavG
good quali ta tive agreement with the phenomenological be}1..aviour of
concrete. Yet this was for a limi tinc:~ t€:nsile str.::tin, whereas the dD. to.
of Chapter 17 (Tables 18.2.3.1 and 18.2.3.2) indicate that a limiting
compressive strain gives the best quantitative results. However, if the
maximum compressive strain were the criterion, failure would occur norma.l
to the comprc-ssivc: stro.m. In..stend, Photographs 17.11.1.1 sho\v that
failure was normal to the maximum tensile strain, B • Therefore there
1
is a basic conflict botweon the qualitative and quantitative agreement
of the theory and the data.
406
18.2.5 Conclusions
In sUlmilary, tho existinG failure theories are not failure
criteria for concrete. The generally poor aGreement between theory
and data is unlikely to be causod by machine effects nnd, unless
~nothor definition of failure is proposed, tho analysis of Chapter 3
is independent of whether a or 0 is used to define failure.
pv c
Therefore it is concluded that tho major C2use of the
discrep:tncy between tho theories discussed in Part I and the data
is their nssumption that concreto is homoeeneous. The thesis has
shown that the mncroscopic structure of concrete is important to an
understD..ndin~ of the behaviour of concrete and since tho thoories neGlect
this they may be neGlectin'~ one of tho primary cC\.usos of f3.ilure in
concrete.
FIGURE 18· 2·4·1,
5000
OCTAHEDRAL SHEAR-NORMAL STRESS THEORY
........
tf)
COMPARED WITH DATA OF CHAPTER 17
0.
~
Mix2 (;)
'01"4000 Mix 4 x
.1:1
o ivi i x 6
r-
eYe ~ .~ x
u)I' ,
(/)'
(Jpv -----
~1-3000
J-
.(/)
n::
«
w
Il"2000
(/)
::---....::----r---.~--~-~
-1 "
«
n::'
"0
w
.
_------~---~--s~-
_~-.- J!l .
I I.. 1000
~
U
o
~
l 9· • ,. ~u... l ' 2 .~" '""' -' 0
'+500 0 -1000 -2000 -3000 -4000 '-J
OCTAHEDRAL NORMAL STRESS, CYoet (psi)
18.3 408
= ••• 18.3.1.1
or extended to include the influence of a 2' this might be
= ••. 18.3.1.2
theory (tensile strain) was the only theory that could explain all the
modes of failure in uniaxial tension and compression, biaxial compression,
and biaxial compression-tension. The analysis of Part II supported this
conclusion by showing that the assumption of Q limiting tensile strnin
in the mortar could explain certain aspects of the failure of concrete.
18.4 409
The modes of cracking observed in Part III and the plo.te-liko failures
obtained in bin.xin.l compression and further evidence that a limiting
tensilo strain is import~nt.
TABLE 18.4.1.1
Designator
Computer Programme Inputs
in
Mix 2 Mix 4
(Section 8.9.1) programme
Elastic Constants
COMPAF~ISON OF UNIAXIAL
COMPRESSION DATA WI TH 'MODEL
o Actual
., data (Ch.17)
x ----- Data from model (Ch.8)
L-6000
j
/
/
/
/
/
-4000' /
(/)
(j)
W /
0:: /
I- /
(j)
I'
...J -3000' I
«
,Z /
I
I
o I
:::J I
I- '~
(9 , upv
z -2000
o
...J
/ ,
/.- / \
/' {0 '\'
. /
Mode~ -z-~/ / \
-1000
(x=f(cr~) .1
c:,.
I '\
~·.P ~\.
" I / '\4z-- Model (x =0)
-"
i·/ .".
\
: qI ". .
.
.
\.
~
\
~
\ Mix 4
-\ (corrected for
" .
'-"
,
--~'-2'000
"\ \0 friction)
. \
~ \ -3000
\. \ ~
~
w
\ \ ()2 psi
\ ~
\ 'Mix 4 (uncorrected)
FIGURE 18·4·2·2
414
COMPARISON OF MIX 4
WITH MODEL
<.D
YOUNGJS MODULUS
o
....-
x -
.~ ro rt
0. /'
/Model (x =f(G2 ))
.CJ) /
::J
...J /'
::J
o
o Mix 4
2 specimen No:6
. '){
,/ . Model (x=O)
----- ---- -{::l--------- --El- -..;... - - - - -
~-,-~ ' ;
o -1000 -2000 -3000 -4000
LA TERAL LOA-OJ psi
()2 J
18.4 415
The similarity betweon this equation and the equations for the
internal friction (Hohr's theory), Cow~n's theory, nnd Bro.ndtzaeg's
theory is obvious: Ema corresponds to the shear stress /s/ and a
to tho normal stress (Section 3.5). iilthouGh the pt.'l.rameter x
adjusts the moduli cf the me:mbers it c~n be interpreted as accounting
for the effect of the volumetric stress.
x
Em °= Em + Q 0, ••• 18.4.2.2
and which ~pplies only to the ccntr~l members, C1, C2, C3 (Figure
18.4.2.4). However, the progrnmmo is written so that if Q=0
then x/Q = 0 n.nd the effect of the volumetric stress on the central
members is entirely elimi:n.::ttod.
~-W3
V RESTRAINT' IN
LONG AND. SHORT
MEMBERS
-W-"'>
2
<--w
, 2
MODE-L
NOTATION
(FIGURE 8·2·1~6)
. 0'
\
, FIGURE '18·4·2.5
COMPARISON OF M1X 4
WITH MODEL
~v VALUES
U3 psi
,-60,00=--.-5Q{JO -~30 ~~-3o,P?===_ -29go -10:00
I
d,","'2-,
{\
'I .
/ \.
/><'0 \
I \
Mix 4 ~// . &~-1000
(uncor'rected) t b \
/ ..
I ~ Model
I "'-. X= fCo-2), Q=O'
1 '-.' '
\
r-Mix 4
". ",
..1,"'-2000
\ Q)
<\ U2 psi
L
FIGURE 18·4·2·6
419
COMPARISON OF Ml X 4
WITH MODEL
(,0
YOUNG/S, MODULUS
o
~
.U)
x
OJ
0..
Mix 4
/SpeCimen No 6
Mix 4
specimen No 5
~
---~Model '
--- -- -- -- -- ---
-
C''--
1-----
..
-- , X=f(02)~ Q=O
I
I ~--
.-',,.-;;..~-.
5J---- ". " .~:",., .
""-.
o--Model
x=f(02),0=0
~--:.~
(f) ~ , , Model
(f) d x=f«()2), Q=O
o
0...
f'.-
Q~o/ ______~p_ _ _ _ _ _~_ _ _ _ _ _~ft_ _ _ _ _ _~'
°0 -1000' -2000 -3000 -4000
LATERAL LOAD, 0"21 psi
18.5 420
18.5.1 Introduction
The: mixvs of Part III wer~ :J.ir-driod for 8 weeks after curine
for 2 weeks. In contr:J.st, tho specimens of Part IV woro cured for 28
days Gnd t...:sted on the 29th nnd 30th d:J.ys in a StJ.tur::.ted condition.
Mix G of Part III and lvlix 2 of Pc.rt IV aro the simplest to compare as
they hi':lve the so..mo \\1.:1. tor-cement ret tio .::.nd the same 0. ggregC'. to-cement
ratio. The o-vor.:J.C;8 ultimo.-ce lo.::td of tho X-ray nnd pulse velocity tests
w:ts -6840 psi and the r2tio of a to a was 38.2%. In contro.st,
pv c
the ultimate strength of tho sGtur:ttod spocimens \V2cS -6440 psi and
a pv/oc w'].s 75%. Hence drying the C'. ncrete 11.:'..s c;ro.::ltly lov/cred tho
v.J..lue of 0
• This cle:J.rly shows the effoct of drying the concrete
pv
ctnd tho influence it is likely to h::tve on shrink2[;o cr2cks 2nd their
prop2gntion into m1.jor cracks.
18.5 421
I
X~-
0
~l /
I
I
I·
I
/
.. 1/
/
~ /
0 /
~~
1
/ /
Xt'- /
<.01 I /
/
/ 4
/
.. j
I /
01 I
t-:..? 01
~
1/ /
CVj
S::!/ 1/ .
W $2~ 0/ ~/
--x / 0/
A tD.
/ /
/ /
/ /
/ tl
.- /
J /
/
/ /
~
I ~ /
0..- / . Jl../
/
X .- d
~ /
, I
/ EFFECT OF W/C ON ))/E
IN MORTARS
From (MA)
"0 1
xl
(Y)
03
~ ,
04 . 0·5
~
0·6
e
. WATER-CEMENT RATIO
"
18.5 424
That such nn effect miGht occur h2d boen sue;gestcd in Po..rt IV,
where it WetS concluded tho. t the :;'PP'-:lrcmt increase in Younr~' s modulus
was cD.used by either a change in the Young's modulus of the r.1nteriClI
or the loss of applied land duo to friction of the platens. It was
assumed in Chnpter 17 on tho information then nvnilable that the entire
influence wns due to friction. It is now evident that the increo.se in
the Youn:;'s modulus is most likely J combination of these two effects.
Howover, with the data at present o.vnilable, it is impossiblo to separate
tho two influences, therefore hoth the oriGin3.1 and the corrected datn
were prosented in Tables 17.7.1.1 to 17.7.1.6. Fortunately the corrections
were sm2-11, Generally in tho order of 5%, .::tlthouCh a few corrections were
ns hiGh ns 10% It is concluded thc..t nn increase in the volumetric
compressive stress will cause o..n increase in tho Youn3'S modulus.
Such an effect i,s most likely c:luscd by (b), but once the lateral machine
was in position the Gxtensometer waG not touched nCnin until the tests
were coft1plctc. ~\s.'J. result the dccreo.so of PR with n.n increase in
13
lateral load (FiGure 17.8.2.2) is a reflection of the bohaviour of the
cencrate. This influence of lateral lO:1d (volumetric stress) on PR
13
425
confirms the above conclusion that the volumetric stress stiffvns the
thrust ring and thus there is proportionally less strain in C1 (~).
Hence with increasing a 2' e 1 decroases and so dOGS PR •
13
In contrast, Fisure 17.8.2.1 shows that PR initially decreases
23
3nd then incrcCtscs as 0- increases. The subsequent incro.::tse hns been
2
duplicn ted in the L1oc1el by usinG a fnctor Q which reduces the influence
of tho restrainin~ effect (x) on the contrnl members C1, C2, and C3. As
the mortQ.r seements of the centrnl members nr.::; lonr;er thnn the mortnr
segmonts in tho thrust rings, the restraining influence caused by the
differenc0 in elnstic properties of the n~Sregnte and the mortar may be
expected to be less in these centro.l members (FiGure 18.4.2.4). Q
compensates for this. J.ls 3. result, the thrust rinES is more sensitive
to the influence of °2 than is C2. Hence for hic;hervalues of 02'
0. given chnnce in 03 will C2use n higher proportion of tho stress to
be supported by the thrust ring. Since equilibrium of the thrust ring
must be obtninod and the stiffness of C2 is not increasinG as rapidly
as tho thrust rinG, the strain in C2 must increase. This cnuses the
increGse in Poisson's r~tio observed in Figures 17.8.2.1 and 18.4.2.6.
The model has shown that those two concepts are compatible and do
provide a reasonable explanation of the behaviour of concrete, including
the elastic behaviour, fo.r the states of stress considered.
HICROCRACKING IN CONCRETE
The tests using the [lcoustic, pulse volocity, and X-ray techniques
produced considernblo informCl tion on microcro.cking in c·"ncrcte. From
these tests it was shovJn. that microcr.::lcks form 2nd propctga to throughout
loadinG, o.nc1 it was confirmed that some of tho cracks had formed in the
concreto prior to lOCldille;, probnbly due to strains c.C'..usod by shrinkage.
At 101,01 loads these cracks, \-'hich \-Jere primarily bond failures, "rere
small and had only n minor influence on the bc~~viour of concrete.
However, at ~ certain load defined by the pulse velocity or X-ray
techniques, or by an increase in Poisson's ratio, thos~ small cracks
spre.::ld through tho mortnr nnc1 interconnected to form mnjor cr3.cks.
These caused an increClse in Poisson's r:J.tio and a decrease in Young's
modulus.
The do.ta of Part IV, which studied the behaviour of the onset
of major cracking in binxi21 compression, were contrary to the predictions
of the model theory (Chnptcr 8) but Cmpter 18 showod that if restraint
437
REFERENCES
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference
Initials Reference