Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
--``,`,-`-`,,`,,`,`,,`---
EDITORS
D. J. Stevens, et. al.
Copyright American Concrete Institute
Provided by IHS under license with ACI
SP-155
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-L.55 95 W 0662749 0523086 622 =
Testing of Fiber
EDITORS --``,`,-`-`,,`,,`,`,,`---
The papers in this volume have been reviewed under Institute publication
procedures by individuals expert in the subject areas of the papers.
Copyright O 1995
AMERICAN CONCRETE INSTITUTE
P.O. Box 19150, Redford Station
Detroit, Michigan 48219
All rights reserved including rights of reproduction and use in any form or
by any means, including the making of copies by any photo process, or by any
electronic or mechanical device, printed or written or oral, or recording for
sound or visual reproduction or for use in any knowledge or retrieval system
or device, unless permission in writing is obtained from the copyright
proprietors.
PREFACE
--``,`,-`-`,,`,,`,`,,`---
CONTENTS
PRACTICAL APPLICATION
by G. Spadea, R. Cava, D. Gallo, and R. N. Swamy . . . . . . . . . . . . 233
SP 155-1
The range of test specimens for which comparative data are reported includes
a series of third-point-loaded 500x150~150mm beams with three different steel fibers
ranging in length from 18 mm to 63 mm, and a second smaller series of 350x100~100
mm beams that allows for assessment of the effects of beam size and fiber alignment.
Fiber contents vary from 20 to 75 kg/m3 (0.25to 0.94% by volume). Also included
is a series of 350x100~100mm beams with a single type of fibrillated polypropylene
fiber of length 38 to 64 mm in amounts of 0.5 to 0.75% by volume.
The results illustrate the extent to which the Cl018 parameters 15,.Iio, 120,
R5,10,and are effective in distinguishing the performance of the various FRC
mixtures in terms of fiber type, geometry and amount. The index I5 is found to be
least effective and a case is made for greater emphasis on use of residual strength
factors, especially when employing the test to specify and control the quality
of FRC.
--``,`,-`-`,,`,,`,`,,`---
2 Johnston
Colin Johnston is an AC1 Fellow, a past-president of the Alberta Chapter of AC1 and
a professor of civil engineering at The University of Calgary. He was awarded ACI's
Wason Medal for Materials Research in 1976. He was chairman of ASTM 0 . 4 2
on Fiber-Reinforced Concrete from 1980to 1990 and is a member of AC1 Committee
544 on Fiber-Reinforced Concrete.
INTRODUCTION
--``,`,-`-`,,`,,`,`,,`---
The issues of exactly how deflection should be measured and the possible
effects of different methods of determining deflection on the loaddeflection
relationship and the toughness parameters derived from it have been the subject of
much discussion and some controversy since ASTM C1018(1) was first introduced in
1984.
Since 1984 the Apparatus section of the standard has required deflection-
measuring equipmentto "accurately determine the net deflection of the specimen under
load exclusive of any effects due to seating or twisting of the specimen on its
supports, and the Procedure section of the standard has included the imperative
statement in mandatory language "Exercise care to ensure that the measured
deflections are the net values exclusive of any extraneous effects due to seating or
twistingof the specimenon its supportsor deformation of the support system". While
the intent of these statements was clear enough, the specifics of how it might be met
were contained in a non-mandatory note recommending the use of additional
deflection-measuring devices at each beam support, The 1984, 1985, 1989 and 1992
editions of the note also acknowledged that the increased number of deflection-
measuring devices makes the processing of data to obtain average net deflection more
complex and stated that a recommended correction procedure for drawing a tangent
to the initially concave upwards portion of the loaddeflection curve "allows the net
deflection to be obtained reasonably accurately. "
Unfortunately, the notion conveyed by the note in the 1984 and 1985 editions
of ASTM C1018 that nominal deflection measurement was reasonably accurate and
good enough for most testing was widely believed until well after 1985. However,
in the 1989 edition the standard was modified to delete reference to this notion and
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 75 0 6 6 2 9 4 9 0523094 î T 9
Testing of FRC 3
replace it with wording stating that "Nominal deflections based only on midspan
measurements may be much larger than corresponding net midspan deflections
obtained by subtracting the average of the deflections measured at the two supports
from the corresponding nominal deflection at the midspan". Also added was the
comment that "Toughness indices based on nominal midspan deflections may be less
than the equivalents calculated using net midspan deflection". This non-mandatory
language stating the desirability of deflection measurement at the supports as well as
the midspan and the possibly significant effect on toughness indices remained
unaltered in the 1992 edition of ASTM C1018.
In 1994 the standard was modified to delete ail reference to testing based on
nominal midspan deflection measurement thus making even more explicit the
requirement to determine net deflection and to compute toughness indices and residual
strength factors solely on that basis. Detailed descriptions with photographs of two
alternative deflection-measuring systems for doing so were added, and a formula for
estimating the first-crack deflection in terms of the size of the test specimen and the
modulus of the concrete was introduced to help users confirm the validity of deflection
measurements.
Like most ASTM Standards, ASTM C1018 has evolved through consensus
and compromise between those who advocated the need for net deflection
measurement despite increased experimental complexity and those who argued for
permitting nominal deflection measurement because of experimental simplicity and
practicality. Naturally, fewer laboratories had the more complex equipment needed
to determine net midspan deflection, and those that did not have it tended to test
specimens using nominal deflection measurement despite failing to comply with the
intent of the standard. For example, in an interlaboratory comparison of data(3),
supposedly obtained according to ASTM C1018-89, only four of six participants
measured net midspan deflection while the two others measured nominal deflection.
Despite much discussion and some controversy over the issue of deflection
measurement, there is little published data comparing results obtained using net
deflection measurement with those obtained using nominal deflection measurement.
This paper makes same-specimen comparisons for a variety of steel and polypropylene
fibers at different fiber contents. The results reflect the evolution of ASTM C1018
from 1984 when only toughness indices Is and Il0 and the ratio Ilo/Is were reported,
to 1989 when the residual strength factor R S , became
~ ~ a requirement and the index
120 was highlighted instead of 130 as a first option along with the residual strength
factor R10,20 This followed introduction of the concept of residual strength factor
in 1986(4).
4 Johnston
corresponding residual strength factors, such as R10,5@ The merits of low-end point
versus higher end-point deflections and the significance of toughness indices versus
residual strength factors can also depend on the response of the testing system. The
discussion also deals with the effect of the testing system response rate on the load-
deflection function immediately following first crack when there is sometimes a rapid
and unstable decrease in load and increase in deflection, particularly for low fiber
contents.
EXPERIMENTALPROGRAM
In the first set of tests reported, the fiber distributor who sponsored the tests
in 1987 wished to compare the performance of several types of steel fiber available
in Canada in the context primarily of the relatively low fiber contents and low to
medium strength concrete matrices associated with industrial floor slab applications.
In choosingthe specimen size it was recognized that stiff fibers of length 50 to 63 mm
would be subject to significant preferential fiber alignment if evaluated using
300x100~100mm beams and that the ASTM Cl018 requirementfor thicksectionsthat
the minimum specimen dimension be at least 3 times the fiber length would be
severely violated. In order to obtain results more representativeof thick sections and
meet more closely the specimen size/fiber length minimum of 3, the heavier less
convenient 450x150~150mm beam size was employed in the majority of the tests,
although additional 300x100~100mm beams were tested in two cases to get an idea
of the effect of specimen size and associated preferential fiber alignment on test
results.
In the second set of tests also in 1987 another fiber distributor who wished to
evaluate the performance of fibrillated polypropylene fibers of length 38 to 64 mm
chose the more economical 350x100~100mm specimen size recognizing that flexible
polypropylenefibers would probably be less subject to the effects of preferential fiber
alignment.
Testing of FRC 5
measurement are in bold print while those based on nominal deflection are in normal
print.
DISCUSSION OF RESULTS
The effect on I and R values of calculating them on the basis of net deflection
versus calculating them on the basis of nominal deflection varies with fiber type,
geometry and amount as illustrated by the specimen examples in Fig. 3 and 4.
Starting with 450~150x150mm beams, the three relationships for steel fibers
in Fig. 3 illustrate the main possibilities.
The case of gradual stable strainsoftening after first crack (Fig. 3a) is perhaps
the easiest to explain because of its approximately constant slope. In this case, values
of toughness indices denoted by I' are derived from the relationship obtained using
nominal midspan deflection (broken line in ali figures), and they are less than the
indices denoted by I determined in the proper way from the relationship obtained
using net midspan deflection (solid line in all figures). Clearly, this results from the
fact that the average vertical ordinate for the total area representing the numerator in
any index decreases as the end-point deflection moves to the right, making this
ordinate for any end point on the broken line less than the ordinate for the
corresponding end point on the solid line, while the horizontal abscissae for the
numerator representing total areas and the denominator representing first-crack areas
remain in the same proportion for both broken and solid lines. Thus, Is', Il,' and
Izo' are less respectively than 15: Il, and 120. Likewise, residual strength factors,
which are in fact the average ordinate between consecutive end-points divided by the
first-crack ordinate, decrease as the end points move to the right, so and
R'lo,20are less respectively than R5,10and R,,,,,
Considering the special case of elastic-plastic behaviour where both broken
and solid lines are horizontal from first crack, it should readily be understood that I'
and R' values will be the same as I and R values because the vertical ordinate remains
the same for all parameters. Extending the analysis to the case of gradual stable
strain-hardening where the ordinate actually increases as the end-point deflection
moves to the right (Fig. 3b), it is expected that I' and R ' values will exceed I and R
values. This situation is relatively uncommon, but is known to happen for certain
type-amount combinations of hooked-end fibers as in Fig. 3b.
--``,`,-`-`,,`,,`,`,,`---
6 Johnston
than for the solid line one, so I' values are less than I values. However, when the
portions of the relationships between consecutive end points used to calculate R' and
R values are horizontal (plastic) R' and R values may be nearly equal, for example
~ ~ Rg,10in Fig. 3c.
R ' S , and
Similar situations arise with the 300x100~100mm beams. For example, with
steel fibers (Fig. 4a) the extensive plastic portion of both relationships following first
crack accounts first for R10 2o almost equal to R S , and ~ ~ second for almost
equal to Rs Polypropyiene fibers (Fig. 4b) can also give rise to the situation
where I' vdues are much less I values while R' and R values are not substantially
different.
Toughness indices calculated in the normal way using AXY will obviously
decrease if recalculated using AZY. In Fig. 5a the effect will be greatest for 15,
derived from the area AXYCDB, and relatively less for Il0, derived from AXYEFB
in which the portion CEFD is unaffected by the position of AY, and of course less
still for Izo. Perhaps the most important point is that Rs,lo, which is based only on
the area CEFD, is unaffected by any uncertainty in the position of AY because C is
the right of Y, and likewise R10,20(CC63 fibres in Table 3). This applies also to
Fig. 4a. However, in Fig. 5b the effect of uncertainty regarding the position of AY
extends to R5,io because the area CEFD is affected since C is left of Y.
Nevertheless, which is based on the area to the right of EF is unaffected @E18
Testing of FRC 7
fibres in Table 3). This applies also to Fig. 3a, 3b and 4b, but Fig. 5b represents the
worst case observed for the data set in terms of C being furthest to the left of Y.
--``,`,-`-`,,`,,`,`,,`---
elastic-plastic or yield-like behavior which is the reference level against which actual
performance is usually compared (Appendix XI of ASTM ClOl8). Accordingly, the
1s scale is twice as large as the Il0 scale and four times as large as the I,, scale.
Naturally, R5,10 and R,,,,, are plotted to the same scale as they both have the same
range of O to 1 0 .
Since the graphs for toughness indices (Fig. 6) are scaled vertically to make
the 15, Il,, and 12, values graphically comparable, the steeper the slope the better the
distinction of performance in terms of fiber content. Clearly, Is is least effective and
8 Johnston
--``,`,-`-`,,`,,`,`,,`---
120 is the most effective with Il0 almost as effective as 120. In terms of residual
strength (Fig. 7), both R5.10 and R10,20 appear effective in making the distinction by
fiber content. However, is in fact better because the uncertainty associated
with the unstable AXY portion (Fig. 5) of the ioaddeflection relationship which
influences some R5, values at low fiber contents is without exception eliminated for
R10,20*
The importance of fiber geometry for steel fibers reflects the influence of
aspect ratio and improvements to pullout resistance by use of crimping (CC and CW)
and hooked ends (HE) or enlarged ends F E ) . For example, at 0.5% by volume or
40 kg/m3 of fibers values are about 88 for HE 50, 82 for CW 60,62 for CC
63 and 42 for EE 18 over a range of aspect ratio of 100 for HE 50 fibers to 38 for
EE 18 fibers.
Concrete matrix strength within the limited 25 to 30 MPa range examined has
little influence on I or R values.
Once again, the manner in which the graphs in Fig. 8 are scaled means that
slope is an indicator of the effectiveness of each toughness index in distinguishing
performance. Clearly, 120 and Il0 are again more effective for this purpose than 15
(Fig. 8 left), just as in Fig. 6, For residual strength, Rs,lo and are both
effective (Fig. 9 left), and illustrate the expected influence of fiber content and in the
case of the polypropylene fibers the effect of length or aspect ratio. For example, at
0.5% by volume of fibers the R10,20 values are about 80 for the CC 63 steel fiber,
37 to 39 the FP 38 and FP 64 polypropylene, and, by interpolation on Fig. 7, for the
EE 18 fiber about 42 plus a small amount attributable to specimen size effect (see next
section).
Testing of FRC 9
Precision of Results
Within-batch coefficients of variation for each set of four specimens are given
in Tables 4 and 5 for the two different specimen sizes.
The most meaningful numbers are the mean values which indicate the level
of precision that should be expected on average when testing multiple sets of
specimens. In this regard, there are no major differences for the two specimen sizes,
--``,`,-`-`,,`,,`,`,,`---
which if combined total 24 sets, and it is clear that the parameters which it has been
argued should be highlighted more in future testing, that is 120, RS,JO,and R10,20,
can be evaluated with reasonable precision. The highest mean within-batch coefficient
of variation is about 13%for and 19 of 24 values are less than 18%. It must
be recognized than the higher the end-point deflection the greater the variability of the
results, but in view of the arguments previously discussed it is pointless to rely on Is
determinations simply because of better precision.
The mean values reported in Tables 4 and 5 compare closely with values for
another recent data set (6) analyzed in the same way for 300x100~100mm beams
tested at various ages and machine stroke rates.
CONCLUSIONS
(i) Toughness indices and residual strength factors derived using nominal
deflection measurement are usually less, often considerably less, than
values derived using net deflection measurement as required by the
standard. However, there are exceptions where nominal deflection
measurement may produce values equal to or slightly more than
values obtained according to the standard.
10 Johnston
(iii) The use of the preferred 300x100~100mm beam with long rigid
fibers like steel tends to produce toughness indices and residual
strength factors greater than for otherwise comparable 450x150~150
mm beams.
The index I20 is least influenced by uncertainty regarding the position of the
portion of the loaddeflection immediately following first crack in cases where
rapid unstable behaviour occurs. The fact that the effect of this uncertainty
can be minimized or eliminated entirely by calculating residual strength
factors, coupled with the fact that these factors are more easily understood
than toughness indices and have more potential for direct application in
strength-based design (7), is an argument for less emphasis on I values and
more emphasis on R values, particularly R10,20, in using the test method to
specify and control the quality of FRC. Certainly, it is imperative that the
widely practised tendency to highlight I5 and ignore R5,10 and must
change.
--``,`,-`-`,,`,,`,`,,`---
3. The conclusion that parameters such as I?o. and R10,20, and to a slightly lesser
degree Il0 and Rg,lO> are effective in distinguishingperformance differences
by fiber type, geometry and amount is not unique to these two series of
specimens or to the particular equipment used. Similar results were reported
(8) in the same format as Fig. 6, 7, 8 and 9 for a series of 108 tests on 18
steel fiber-rnatrix combinations performance on different equipment using
750x150~100mm beams and supervised by the writer in Stockholm in 1989.
However, both investigations employed the 3-transducer deflection-measuring
system (Fig. 2) rather than the rectangular jig arrangement identified as an
alternative in the 1994 edition of ASTM ClOl8.
ACKNOWLEDGEMENT
Testing ofFRC 11
REFERENCES
1. ASTM C1018, Standard Test Method for "Flexural Toughness and First-
Crack Strength of Fiber-Reinforced Concrete (Using Beam with Third-Point
Loading)", ASTM Standards Vol. 04.02, Concrete and Aggregates, 1992.
2. Johnston, C.D., "Toughness of Steel Fiber-Reinforced Concrete", Proceedings
of a U.S.-Sweden Joint Seminar, Swedish Cement and Concrete Institute, Ed.
S.P. Shah and A. Skarendahl, Elsevier Applied Science, London, 1986, pp.
333-360.
3. Gopalaratnam, V.S., et. al., "Fracture Toughness of Fiber Reinforced
Concrete", AC1 Materials Journal, Vol. 88, July-August 1991, pp. 339-353
and Discussion by C.D. Johnston, May-June 1992, pp. 304-309.
4. Johnston, C.D. and Gray, R.J., "Flexural Toughness and First-Crack Strength
of Fiber-Reinforced Concrete Using ASTM Standard Cl018", Proceedings of
Third International RILEM Symposium of Developments in Fibre Reinforced
Cement and Concrete, Ed. R.N. Swamy, University of Sheffield, 1986, Paper
5.1.
5. Johnston, C.D., "Effects on Flexural Performance of Sawing Plain Concrete
and of Sawing and other Methods of Altering the Degree of Fiber Alignment
in Fiber-Reinforced Concrete", ASTM Cement, Concrete and Aggregates,
CCAGDP, Vol. 11, No. 1, Summer 1989, pp. 23-29.
6. Johnston, C.D., "Effects of Testing Rate and Age on ASTM Cl018
Toughness Parameters and Their Precision for Steel Fiber-Reinforced
Concrete", ASTM Cement, Concrete and Aggregates, CCAGDP, Vol. 15,
No. 1, Summer 1993, pp. 50-58.
7. Johnston, C.D., "Fiber-Reinforced Concrete" Chapter 5 1 of Significance of
Tests and Properties of Concrete", ASTM STP 169C, 1994, pp. 547-561.
8. Johnston, C.D. and Skarendahl, A., "Comparative Flexural Performance
Evaluation of Steel Fiber-Reinforced Concrete According to ASTM C1018
Shows Importance of Fiber Parameters", RILEM Materials and Structures,
Vol. 25, No. 148, May 1992, pp. 191-200.
--``,`,-`-`,,`,,`,`,,`---
12 Johnston
TABLE 1 - TEST RESULTS FOR 450 x 150 x 150 mm BEAMS (VALUES BASED ON NET
DEFLECTION SHOWN BOLD)
= - -
Matrix Fiber Amount Vol First-Cmck Toughness Indices Residual
Codeb kg/m3 x - -- -
Stren Factors
25
- --
---- -
O
---- Oft 6, '5 '10 -
120
-
Rs.ia p10.20
O 4.45 0.052
0.168
25 CC63 25 0.32 4.53 0.061 4.11 6.27 10.41 43.2 41.4
O. 179 3.64 5.67 8.38 40.6 27.1
25 CC63 40 0.51 4.68 0.060 4.38 7.67 13.90 65.8 62.3
0.181 3.95 6.91 11.16 59.2 42.5
25 CC63 55 0.70 4.88 0.060 4.66 8.63 15.95 79.4 73.2
0.193 4.29 7.80 14.09 70.2 62.9
25 EE18 25 0.32 4.44 0.060 4.40 6.7ï 10.08 47.4 33.1
0.174 3.61 5.13 7.01 30.4 18.8
25 EE18 50 0.64 4.80 0.064 4.54 7.66 12.45 62.4 47.9
0.186 3.71 5.87 8.64 43.2 27.7
25 EE18 62.5 0.80 5.15 0.060 4.60 8.12 14.54 70.4 64.2
0.189 3.97 6.94 11.16 59.4 42.2
25 @E18 75 0.96 5.11 0.059 4.64 8.39 15.39 75.0 70.0
0.193 4.10 7.34 12.36 64.8 50.2
30 EE18 62.5 0.80 5.33 0.062 4.64 8.24 14.27 72.0 60.3
0.193 3.95 6.37 9.07 48.4 27.0
30 EE18 75 0.96 5.62 0.065 4.64 8.33 15.07 73.8 67.4
0.216 4.07 7.03 10.79 59.2 37.6
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 13
TABLE 2 - TEST RESULTS FOR 300 x 100 x 100 mm BEAMS (VALUES BASED ON NET
DEFLECTION SHOWN IN BOLD)
Matrix
Typea
-
I
Fiber
Code
b
1
Amount
kglm3
x s t - C r a c k
-6fc
Toughness Indices
--
---
-
I20 Rs,io
Residual
Strength
- Factors
R10.20
TABLE 3 - COMPARATIVE I AND U VALUES CALCULATED FROM AREAS BOUNDED BY AXY AND
Alï (Fig. 5)
--``,`,-`-`,,`,,`,`,,`---
MatliX
Type*
Fiber
Codeb
Amount
kg/m3
Vol
96
--
Toughness Indices
~1 m
I
h
:e
t-
Residual
I
Factors
-- - - - -
- 6fc Ufc 15 I I10 I 120 I %,i0
-
R10,20
Mean
0.51
Values
-
3.5 8.0
7.3
1.8
2.1
3.3
4.9
4.4
8.3
6.0
9.5
6.2
13.5
4.7
=
* - Specified CompressiveStrength in MPa
- CC is Crimped Crescent-Shaped Steel, CW is Crimped Wire.
HE is Hooked-End Wire, EE is Enlarged-End Slit Sheet Metal.
Numbers are Nominal Length in mm.
Ufc *fC I
I
15 I
I
I10
-
120
--
& R10.20
25 CC63 25 0.32 1.5 14.5 18.1 23.5
25 CC63 40 0.51 1.7 7.2 10.6 8.7
30 EE18 75 0.96 3.5 13.0 5.9 7.8 7.5 11.0
30 FP38 4.55 0.50 10.7 18.6 5.4 11.5 6.3
30 FP38 5.91 0.65 6.1 10.3 15.0 17.6
30 FF’38 6.83 0.75 3.9 3.0 5.2 5.4
30 FP57 5.91 0.65 4.1 7.8 5.3 17.4
30 FP64 5.91 0.65 2.6 9.8 14.3 13.9
30 FP64 4.55 0.50 5.3 5.4
- 8.9 7.2
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 15
20
15
z
Y
I
o IO
a
3
5
I
O
DEFLECTION -mm
Fig. 1-Load-deflection relationships and associated toughness indices I, and I,, compared on
the basis of nominal versus net deflection in 1985 (2)
--``,`,-`-`,,`,,`,`,,`---
16 Johnston
DEFLECTION
Fig. LSingle-specimen load-deflection relationships for 450 x 150 x 150 mm beams showing
end-point deflections and derived I and R values for nominal deflection (broken) and net deflec-
tion (solid) with 75 kg/m3, 0.94 percent volume of EE 18 steel fibers (top), 40 kglm3, 0.5 percent
volume of HE50 steel fibers (center), and 25 kg/m3, 0.31 percent volume of CC63 steel fibers
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 17
F I BERS C C 63- 4O
, kg/m
Rs.10 = 88 R 10.20 = 87
15- 4.5
%,IO - 85 Río. 20 - 75
Iio * 16.3
I I
I I
4 0.1mm+
8’ 10.58’
-FIBERS FP38-4.6kglm3
R5.10t 4 5 R10.20 = 3 8
h!
1/20 9.6
1- 10.2 \
--``,`,-`-`,,`,,`,`,,`---
I I
I I
I l
n
- 8 318 5.523 10.58
8 38‘ 5.58’ 10.58‘
40.1
I
DE F LECT ION
Fig. &Single-specimen load-deflection relationships for 300 x 100 x 100 mm beams showing
end-point deflections and derived I and R values for nominal deflection (broken) and net
deflection (solid) with 40 kglm3, O5 percent volume of CX 63 steel fibers (top) and 4.55 kg/m3,
05 percent volume of FP 38 polypropylene fibers (bottom)
18 Johnston
FIBERS C C 6 3 - 25 kg/m3
1
.
1
.... \
'. 1
1-
\ Y'
--L--,--
E -
I
l
I
I
I F 1 -
DEFLECTION
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 19
Fig. &Effects of fiber geometry and amount on i values based on nominal deflection (broken)
--``,`,-`-`,,`,,`,`,,`---
and net deflection (solid) for 450 x 150 x 150 mm beams with four different steel fibers
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 0662949 0523111 608
20 Johnston
100
i.,25
FIBER-CC63
MATRIX
Oi5
FIBER VOLUME FRACTION
0.Y5,Oi25 0;5
FIBER-€EI8
MATRIX
0.75
-%
I.?
-
0.25
FIBER
HE-50-0.
0.5
J
8 90 25MPa- O 0 25MPa-0. cw-60- o
30MPO - A A 30 MPo - A A
.Pc"
8
R
--``,`,-`-`,,`,,`,`,,`---
K 70
d
a
O
I-
2 60
LL
E 50
a
z
h o
E /
/
30
d'
20
0
$90 P
I
8
o6 80
K
8 70
s2
60
I
I-
<3
50
K
!- c/
u)
-
40
30 'O
I_
1 I I 1 1
25 40 55 25 50 75 20 40
FIBER C O N T E N T - k g /m3
Fig. 7-Effects of fiber geometry and amount on R values based on nominal deflection (broken)
and net defledion (solid) for 450 x 150 x 150 mm beams with four different steel fibers
16
.aH
8
O
L
U 1
4
H
X
w
n
f 12
ia
8
--``,`,-`-`,,`,,`,`,,`---
X
w 7
n
Z
.m
H
8 4.5
ln
I
X
3.5
Fig. &Effects of fiber type and amount on I values for 300 x 100 x 100 mm beams with steel
and polypropylene fibers (left) with an indication of the importance of specimen size for rigid
Copyright American Concrete Institute
steel fibers
Provided by IHS under license (right)
with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 = 0662747 0523113 480
22 Johnston
100 I I I I I I I
FIBER FIBER FIBER FIBER
CC63-0. CC63-• 0 EE 18- AA
O 9 0
FP38-AA
FP64-vV -
d
ìE
a 80
O
O
I
Q 70
a
IY
O
5. 60
9
F
o
2
!i
50
40
OI :& SPECIMEN
30 - SPAN / CROSSECTION
0 A- 300/100mm
o A- 4 5 0 / t 5 0 m m
20
0
g-90
8
0 80
ui
0:
E
--``,`,-`-`,,`,,`,`,,`---
70
o
8 60
I
I-
(3
Ea 50
I-
V>
40
30
Fig. 9-Effects of fiber type and amount on R values for 300 x 100 x 100 mm beams with steel
and polypropylene fibers (left) with an indication of the importance of specimen size for rigid
Copyright American Concrete Institute
steel
Provided by IHS under fibers
license (right)
with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-1.55 95 m 0662749 0523114 317 m
SP 155-2
Toughness of Fiber-Reinforced
High-Strength Concrete from
Notched Beam Tests
by D. Jamet, R. Getîu, V. S. Gopalaratnam, and A. Aguado
24 Jamet et al
INTRODUCTION
Testing of FRC 25
uses the displacements associated with this plane. The size effects on the matrix-
and fiber-dominated regimes of the specimen response are studied by testing three
different sizes of geometrically-similar beams.
RESEARCH SIGNIFICANCE
In the present study, the three-point bend (3PB) specimen shown in Figure 1
has been adopted as the test geometry. Nevertheless, most of the discussion and
its implications also apply to other notched specimen geometries. The 3PB
specimen has several advantages: (i) it is relatively easy to fabricate and test;
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 W 0bb2949 0523117 0 2 6 W
26 Jamet et al
(ii) it deals with just one critical section and one propagating macro-crack, making
its analysis and modeling simpler; (iii) the notch localizes the crack along its
plane, reducing the variability of the crack path and non-symmetry of the
deformations,both of which can occur in unnotched third-point loaded beam tests;
(iv) a stable test can be performed, even for plain concrete, under crack-opening
displacement control and (v) it has been extensively studied and documented in
the fracture mechanics literature (see Reference (IO) for a review).
MATERIALS USED
Beams of three different sizes, three specimens in each size, were cast from
each of the concretes. The specimens were geometrically similar (fig. l), with
thickness = 90 mm and depth (d)= 90, 180 and 320 mm. The beams were cast
in two layers normal to the loading plane and vibrated with a finishing screed.
Notches of length a, = 0.275d were cut with a diamond disc before testing. The
beams were tested at the age of 25-28 days in a 100 T closed-loop
INSTRON 8505 servo-hydraulic system. The tests were conducted at constant
CMOD rates that were chosen to give peak loads in the HSC-0.0 specimens at
approximately 3 minutes. The CMOD and net mid-span deflection (fig. 2) were
monitored with an INSTRON clip-on crack-opening displacement gage and a
Schaevitz LVDT, respectively. These measurements, as well as the load and
piston displacement, were recorded electronically using the INSTRON FLAPS
computer program. The data sampling was done at 0.6 seconddpoint initially, and
after substantial crack propagation was observed, the acquisition was slowed to
30 secondsìpoint.
beam depth. The first-crack load corresponds to the point at which the load-
CMOD curve exhibited a significant change in slope or a kink. The identification
was made graphically and is admittedly subjective. The first-peak or bend-over-
point is more easily determined, at least when non-hardening behavior is
observed, and is taken to be the end of the matrix-dominated response. The
From Table 2, it can be observed that the scatter of the stresses is much less
than those of the displacements. Also, the variations in the stresses generally
increase with Vf . The first-peak stress exhibits much less scatter than the first-
crack stress in ail the concretes. Strong size effects are observed in both the
matrix- and fiber-dominated responses for HSC-0.5. The stresses and relative
displacements decrease with an increase in specimen size, the latter much more
significantly. This is similar to that seen in HSC-0.0 as expected ( 1 O). However,
this effect is observed only at the maximum loads of HSC-1.0 and not in the
matrix-dominated response, implying that higher fiber volume-fractions may make
the size effect much milder.
Another interesting aspect seen in Table 2 and Figure 6 is the relation between
CMOD and deflection. The rate of change of deflection during the test decreases
progressively (see inset of Figure 6). especially after first-cracking, to ultimately
reach a constant value. This demonstrates why the test can always be controlled
--``,`,-`-`,,`,,`,`,,`---
in a stable manner with CMOD control. Moreover, in the post-crack regime, the
CMOD-deflection curve is linear with a deflection/CMOD ratio of approximately
0.65, for most of the specimens tested. This behavior may be attributed to rigid-
body rotation of the beam halves, indicating the presence of a predominantly
planar energy dissipation mechanism due to localization of the deformations.
Another implication is that CMOD will be a more sensitive parameter to base
toughness on since it increases much more than deflection, for a given change in
load.
JCI (2) and RILEM (13) recommendations define toughness as the work done
@e., area under the load-deflection curve) to reach a prescribed deflection. A
similar measure of toughness, r', , is defined here as the area under the
load-CMOD curve. The CMOD limit is prescribed as a fraction of the beam depth
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-355 95 I0 6 6 2 9 4 9 0 5 2 3 3 2 0 b L 0 I
Testing of FRC 29
dln, where values of 250, 500, 1000 and 2000 have been chosen for n. The
corresponding values of TR are tabulated for the three concretes in Table 3.
Obviously, r", increases with V, , for all n and for all d. This type of toughness
measure can only be used for comparing the values obtained with a specific
specimen size and geometry.
For better comparison, r", cau be divided by the ligament area to get a more
fundamental quantity that is analogous to fracture energy. This, denoted as Tn ,
is the area under the load-CMOD curve, until the prescribed CMOD limit dln,
divided by the ligament area. These values are given in Table 3 for different d
and n. Also evident from the table is that the value of increases with the
specimen size. This effect increases with V f , especially during post-crack
hardening where the larger specimens exhibit more significant increases in load-
carrying capacity than the smaller specimens. Also, TE increases considerably
with CMOD limit (Le., n), which should, therefore, be based on practical crack
--``,`,-`-`,,`,,`,`,,`---
Figures 7 and 8 compare typical load-CMOD curves of the two FRCs with
those of HSC-0.0. The differencebetween the FRC and unreinforced concrete can
be attributed to the presence of the fibers, and can be used to quantify the fiber
performance. Such toughness measures, which explicitly deduct the toughness of
the unreinforced matrix, have been recommended by AC1 (3) and some German
standards (e.g., Reference 14). Following this approach, a relative toughness can
be computed as Tr = TE (for FRC) - r f f (for HSC-O.O), for the values of n used
earlier. These values presented in Table 4 show clearly the increase in toughening
with fiber effectiveness and the specimen size effects. This type of toughness
measure obviously has several advantages but would always require the testing
of a companion specimen of the base concrete. This could be considered as a
disadvantage.It can, however, be handled by defining an idealized curve for the
unremforced concrete based empirically on other tests (as in Reference 14) or by
determining it analytically using an approach such as nonlinear fracture
mechanics.
The ASTM Standard C 1018 (4) defines toughness indices that are equal to
the areas under the load-deflection curve of an unnotched beam until prescribed
multiples of the first-crack deflection divided by the area until first-crack. A
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 M 0662949 052312L 557 M
30 Jamet et al
similar approach was used by Barr and Hasso ( 5 ) for defining toughness based on
notched and unnotched specimens.
Bryars et al. (9) recently used a toughness index based on the load-CMOD
area of a notched beam defined in terms of the first-peak, instead of the first-
crack. They found that the value of such an index increases with the CMOD limit,
is practically size-independent and reflects the effectiveness of the fibers
satisfactorily. Two similar indices are evaluated here: Ti" defined as the ratio of
the area of the load-CMOD curve until rn times the first-crack CMOD and the
area until the first-crack; and ';"7 is defined similarly in terms of the first-peak,
instead of the first-crack. The results for rn = 5, 10 and 20 are given in Table 5.
From Table 5 it can be seen that all the indices reflect the fiber effectiveness
and are practically size-independent. Further testing is required to confirm whether
these indices are uifluenced by specimen geometry and dimensions. However, as
concluded by other researchers (1, S), the indices are not very sensitive to the
reinforcement parameters at small values of rn. This implies that tests have to be
continued until larger displacements are attained. Moreover, comparisons with
previous results of Bryars et al. (9) suggest that the sensitivity or the dependence
on the factor rn may depend on the material. In their study, conducted on a HSC
with shorter straight fibers, use of rn = 5 yielded 7'; values that were reasonably
sensitive to the increase in fiber length (i.e., fiber effectiveness).
--``,`,-`-`,,`,,`,`,,`---
CONCLUSIONS
2. Several toughness measures have been analyzed in this study. They have been
based on (i) the absolute area under the load-CMOD curve, (ii) area between
the curves of unreinforced and fiber reinforced concretes and (iii)
nondunasional indices defined as ratios of the post- and pre-cracking areas.
All such quantities reflect the effectiveness of the fibers adequately and
warrant further research.
Acknowledgements
REFERENCES
2. JCI Standard SF-4, "Method of Tests for Flexural Strength and Flexural
Toughness of Fiber Reinforced Concrete", J q a n Concrete Institute Standards
f o r Test Methods of Fiber Reinforced Concrete, Tokyo, pp. 45-51, 1984.
4. ASTM C 1018-92, "Standard Test Method for Flexural Toughness and First-
Crack Strength of Fiber-Reinforced Concrete Using Beam With Third-Point
Loading", Annual Book of Standards, ASTM, Philadelphia, V. 04.02,
pp. 510-516, 1992.
32 Jamet et al
9. Bryars, L., Gettu, R., Barr, B., and Ariño, A., "Size Effect on the Fracture of
Fiber-Reinforced High-Strength Concrete",Proc., Europe-US. Workshop on
--``,`,-`-`,,`,,`,`,,`---
Fracture and Damage in Quasibrittle Structures (Prague), E&FN Spon,
London, 1994.
12. Mindess, S., Chen, L., and Morgan, D.R., "Determination of the First-Crack
Strength and Flexural Toughness of Steel Fiber-Reinforced Concrete", Advn.
Cem. Bas. Mat., V.l, No.5, pp. 201-208, 1994.
13. RILEM Committee 49-TFR, "Testing Method for Fibre Reinforced Cement-
Based Composites", Mater. Struct., No. 102, pp. 441-456, 1984.
Unit Conversions
Stress Relative Relative Stress Rel. I Rel. defl. Stress Rel. Re1 defl.
CMOD deflection CMOD CMOD
(MPd (W-) (P&-) (ma) (P&-) (rd-) (MPa) (Y&-) (cid-)
HSC-0.0 90 4.53 0.156 0.156 5.28 0.252 0.237 5.28 0.252 0.237
f4.7K iL2.4% 112.4% *0.7% i6.7'h ì2.7% ì0.6% *6.7% +2.7%
I I 412
i6.6% 1
~ 0.120
19.6%
I 0.136
*20.2%
I 440
ì2.5%
1 0.167
ì20.0%
1 0.156 I 4.40
*2.5%
I O I67
*20.0%
1 0.156 I
320 3.75 0.138 0.117 6.09 0.675 ~ 0.514 6.94 2.81 I91
f22.2% +31.5% ì29.6% ì26.3% *%.i% 1I *51.2% ì24.7% *l4.0% *l4.8%
34 Jamet et al
--``,`,-`-`,,`,,`,`,,`---
Note that values are omitted when the experimental CMOD range is smaller than the limit
Material T r (N/mm)
n = 500
0.085 0.253
Testing of FRC 35
b
S=2.5d
Fig. 1-Three-point bend geometry
36 Jamet et al
2oooo
High strength concrete, HSC-0.0
15OOO
E
n
loo00
75
m
O
4
Large
d = 320 rnm
5000
Medium
d=l80mrn
Small
d=90mm
O
O 100 200 300 400 500
CMOD, A (microns)
25000
High strength fiber concrete, HSC-0.5
Hookedend steel fibers (30 mm, V, = 0.5%)
2oooo
t
n
U- Medium ,
8
J
loo00 -
-----.- d=180mm I
32000
28000
-k 2 m
Q
2oooo
U-
m
2 16Ooo
12000
8000
- d-9ümm
4Mxi
O-
O 400 800 1200 16W
CMOD, A (microns)
--``,`,-`-`,,`,,`,`,,`---
Fig. &Typical load-CMOD curves for concrete with fiber V, = 1.0 percent
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 I0 6 6 2 9 4 9 0523129 8 4 8
38 Jamet et al
1000 I
n
800 /
cn
c
g
.-
E 600
v
Lo
t
.=
O
o
400
a>
=Q)
n
200
O
O 400 800 1200 1600 2000
CMOD, A (microns)
Fig. &Typical deflection-CMOD curves for all concretes and all specimen sizes
25000
HSC-0.5-Small
Fig. 7-Comparison of iypical load-CMOD curves of the small and large specimens of
Copyright American Concrete Institute
unreinforced concrete and FRC with V, =Not 0.5
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS
percent
for Resale
A C 1 SP-155 9 5 I0 6 6 2 9 4 9 0523330 5 b T
Testing of FRC 39
Small: d = 90 mm
t
n
O
4
3
HSC-1 .O-Small
--``,`,-`-`,,`,,`,`,,`---
SP 155-3
It was found that the ASTM C1018 toughness indices, particularly Ili
and Il0, did not discriminate very well amongst the different fibre contents or
different fibre types; the JSCE parameters were rather more successkl in this
regard.
--``,`,-`-`,,`,,`,`,,`---
42 Chen et al
BIOGRAPHIES
AC1 Fellow, Dudley R. Morgan, is chief materials engineer for the AGRA
Earth & Environmental Group in North America. He is a member of AC1
Committee 544, Fibre Reinforced Concrete, AC1 Committee 506, Shotcrete
and AC1 Committee 234, Silica Fume, and has been extensively involved in
specifjing and testing toughness for fibre reinforced concrete and shotcrete
projects.
Testing of FRC 43
INTRODUCTION
--``,`,-`-`,,`,,`,`,,`---
44 Chen et al
applicability of this test method for speciQing FñC,or for carrying out quality
control in the field.
the test specifications.) However, this method is less sensitive than that
prescribed in ASTM Cl018 as to whether the extraneous deformations have
been properly eliminated (and completely insensitive as to whether the first
crack deflection has been properly chosen). For instance, Fig. 3 shows the
load vs. deflection curves for three different FRC mixes, both including and
excluding extraneous deformations. These three sets of curves show quite
different behaviours. The calculated toughness parameters for these curves are
shown in Table 1, for the mixes described in Table 2. It may be seen that T,sc,
and FJSCEare essentially independent of how the deflections were measured;
the various ASTM C1018 parameters, on the other hand, can vary
considerably.
It has been suggested (1,3) that JSCE-SF4 is better able than ASTM
Cl018 to distinguish amongst FRC’s containing different fibre types and
different fibre volume fractions. This view, however, remains a controversial
one (2,7).
EXPERIMENTAL PROCEDURES
SDecimen PreDarrtion
Testing of FRC 45
Test Protocol
After moist curing, four specimens of each type were shipped to each
of the participating laboratories, where they were tested at an age of 42 f2
--``,`,-`-`,,`,,`,`,,`---
days after casting. The tests were camed out in accordance with ASTM
C1018, but taking into account the proposed revisions to this standard
circulated to the members of ASTM Committee C09.03.04 on Fibre
Reinforced Concrete in March, 1993.
ExRerimental Arranpements
ExDerimentri Results
--``,`,-`-`,,`,,`,`,,`---
46 Chenetal
laboratory, except that the results obtained with the closed-loop testing system
are shown by thicker lines. Figures 13-18 show typical load-deflection curves
for each mix as obtained in each laboratory, so that the differences between the
various loading systems can be seen more clearly.
flexural shear
component component
where
mid-span deflection
load at first crack
span length
modulus of elasticity (assumed to be 35,000 MPa)
Poisson’s ratio (assumed to be 0.2)
moment of inertia
beam depth
These values are also given in Table 7. It may be seen that these theoretical
first crack deflections agree within k20% with the measured values, except for
four of the thirty-six results (see Tables 4 & 7), which fall outside this range.
Of these four results, two are from Northwestern University, the other two
from the UBC MTS machine; in these cases, the measured first crack
deflections all appear to be low. These differences are small when contrasted
with the values that are obtained when the extraneous deformations are not
properly eliminated (Table i). However, the measured values are much more
variable than the theoretical values, probably because of the difficulties in
accurately measuring the very small deflections at first crack (4), even when
extraneous deformations are eliminated.
Testing of FRC 47
DISCUSSION
However, fiom Fig. 19, for plain concrete (Mix 1) and for the relatively
low fibre content mixes (Mixes 2, 5 and 6), the load falls very rapidly from the
maximum load to quite a low level immediately after cracking has been
initiated; this represents a region of instability. Because of this, the recorded
lines B-C, even with properly designed testing systems, do not represent the
true load-deflection response of the beams in the region B-C;they are largely
an artifact of the particular system. In fact, it is very difficult to measure the
‘ h e ” load-deflection curve in this region unless a closed-loop servo-
controlled testing system is used, such as the one at Northwestern University.
This may be the reason that the load-deflection curves for these four low
toughness mixes (Figs. 7, 8, 11, 12) exhibit more variability. This is manifested
in Tables 4-6, where the greatest variability amongst the different laboratories
in the ASTM toughness indices 13,, and Ise, and in the Japanese Toughness
TJSCE,occurs for these four mixes. It therefore may not be appropriate to
evaluate low toughness mixes on the basis of their load-deflection curves
--``,`,-`-`,,`,,`,`,,`---
48 Chen et al
unless this instability is taken into account. This is because, for the low
toughness mixes such as Mix 5 (0.20% polypropylene fibres), the end point for
130for some specimens fell in the unstable part of the curve. Thus 15, Il0, bo,
and R5,i0,Rlqz0and even &o,3o may be affected by the instability.
volume of steel fibres) did tend to stand out. Only when using Izo, and more
particularly 130 and 150, did the ASTM toughness indices begin to discriminate
sensibly amongst the different mixes, even though the results presented in
Table 4 are not entirely consistent. From Table 4, it may be seen that, for the
UBC-Instron machine, the Izo, 130, and Is0 values for Mix 5 (0.20%
polypropylene fibres) are considerably higher than those for Mix 6 (0.50%
polypropylene fibres), which can not be correct. The reason for this is that
these values were calculated from the unmodified P-6 curves, which contain
irregularly shaped unstable portions. For such curves, the calculated toughness
indices are simply artifacts of the trace of the dropping pen (the recorded curve
in Fig. 21). The "modified curve in Fig. 21 is an extrapolation of the initial
slope of the line &er first crack, and is a better approximation of the real
behaviour. If the curves are modified (Fig. 21), the values for the two mixes
fall much closer together (values shown in parentheses in Tables 4-6). In
addition, some of this inconsistency may also have been due to differences in
estimating the location of the first crack.
Testing of FRC 49
CONCLUSIONS
The ASTM toughness indices Is, Il0, and the corresponding Rs,lo, and
to a lesser extent Izo, are not particularly sensitive to either fibre
addition rate or fibre type; 130,Iso, and particularly are more
useful in this regard.
50 Chenet al
even when every effort was made to test rigorously in accordance with
the testing procedures prescribed in ASTM C1O18.
ACKNOWLEDGEMENTS
REFERENCES
4. Mindess, S., Chen, L. and Morgan, D.R., First Crack Strength and
Flexural Toughness of Steel Fibre Reinforced Concrete, Journal of
Advanced Cement Based Materials, Vol. 1, No. 5, pp. 201-208, 1994.
Testing of FRC 51
--``,`,-`-`,,`,,`,`,,`---
52 Chen e t a l
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 53
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 55
1 2 1 58 I 82
3 82 90 1O0 65 94 74
R10,20 4 100 113 110 111 111 101
5 26 39 77(65) 51 16 55
6 34 68 53 50 39 71
3 78 85 96 66 92 82
R20,30 90 105 103 104 104 96
5 17 33 72(57) 29 6 53
--``,`,-`-`,,`,,`,`,,`---
Chenet al
95 = 0662949 0523346 9 2 7 -
TABLE 6 - JSCE FLEXURAL TOUGHNESS PARAMETERS - AVERAGE VALUES
Testing of FRC 57
U
E
P-
rn
æ
U
U
E
vi
--``,`,-`-`,,`,,`,`,,`---
Ym
W
z
æ
c"
a
W
z
!z
n
5
WY
m
w
32
v,
Y
æ
æ
W
a
c
O
LL
O
E2
t
vi
Y
vi
Lu
z
c
2
00
I
w
2
m
2
58 Chen etal
Flexural Load
25.56 -- - --
~:
; ;s.SS -j R20.30 = 101130- lm) I
.- :36 -4
I
R30.50 = 5 ( l ~ 130)
-
I
SJ
fiB 10 ,I F I /H
I IJ I I L i
o \s is the first crack deflection. Net Midspan Deflection
Bending Load
I L. B and H are the span, width and height of the beam respectively.
L!tb is deflection of II150 of span (2mm when span is
6 t b = LI150
Deflection
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 59
O 0.5 1 1.5
Mid-Span Deflection (mm)
+
b
60 Chen et al
-L
r!
--``,`,-`-`,,`,,`,`,,`---
Specimen
I
Specimen
I
Base of Testing Machine Base of Testing Machine
I
#
Leads -LVDT
LVDT
Pins Leads
Testing of FRC 61
Fig. i-load versus deflection curves from all laboratories for Mix 1 (plain concrete); thicker
lines denote tests carried out with closed-loop testing system
--``,`,-`-`,,`,,`,`,,`---
"T i
Fig. L L o a d versus deflection curves from all laboratories for Mix 2 (0.25 percent steel fibers);
thicker lines denote tests carried out with closed-loop testing system
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-355 95 m 0662949 0523352 320 m
62 Chen et al
Fig. L l o a d versus deflection curves from all laboratories for Mix 3 (0.75 percent steel fibers);
thicker lines denote tesis carried oui with closed-loop testing system
--``,`,-`-`,,`,,`,`,,`---
Fig. lû-load versus deflection curves from all laboratories for Mix 4 (1.27 percent steel fibers);
thicker lines denote tests carried out with closed-loop testing system
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 0 6 6 2 7 4 7 0 5 2 3 3 5 3 O67
Testing of FRC 63
Fig. 11-Load versus deflection curves from all laboratories for Mix 5 (0.2 percent polypropylene
fiben)
25
20
15
Fig. 12-Load versus deflection curves from all laboratories for Mix 6 (0.5 percent polypropylene
fibers)
--``,`,-`-`,,`,,`,`,,`---
64 Chen et al
25
UC-MTS UBC-MTS UBC-lnstron UL-INstron HBT-MTS NU-MTS
'
--.- - -
- - Top LVDT'S Top LVDT'S
-- --
-----I
-- - -
........... --- ...
Yoke
--- --
Yoke Yoke
--
Side LVDT'S
-- --- --
-.-.-
--A--- --
20 2 L _I_ .
I L -0- _I_ a L -I- I
i -- -0- --
--``,`,-`-`,,`,,`,`,,`---
Fig. 13-Typical load deflection curves from each laboratory for Mix 1 (plain concrete)
25
UC-MTS UBC-MTS UBC-lnstron UL-INstron HBT-MTS NU-MTS
- --- -..-..- -.-.-
_.
Top LVDT'S Top LVDT'S Yoke
-----. Yoke Yoke
*.........
Side LVDT'S
20
- _j_ Mix 2: SFRC (Fibre Volume Content: 0.25%)
I
; 15
All curves end at 1 /150of span.
..........................................
O
;- -;. -i
---- -- ----- -i. --- -i- -- -i-
-: :
-I - .~.
2
U
,o -- --- - ------ ------ ----- j -[- - - - - - -;- - - - - - -:.
C
m"
5
......................................................
a I I I I
O 0.5 1 1.5 2
Midspan Deflection (mm)
Fig. 14-Typical load deflection curves from each laboratory for Mix 2 (0.25 percent steel fibers)
Testing of FRC 65
25
UC-MTS UBC-MTS UBC-lnstron UL-INstron HBT-MTS NU-MTS ;
Top LVDT'S Top LVDT'S Yoke
...........
Yoke Yoke
--- -..-..-
Side LVDT'S
-.-.- -a-
I
--``,`,-`-`,,`,,`,`,,`---
20
A
5z 15
O
J
CD
C
10
C
Ill
m
'
5 _ _ . Mix
_ -3: _SFRC (Fibre Volume Content: 0.75%) -- -:- - - 7 .
O 1 I l I
O 0.5 1 1.5 2
Midspan Deflection (mm)
Fig. 15-Typical load deflection curves from each laboratory for Mix 3 (0.75 percent steel fibers)
25
20
o,
C
5 10
C
9>
m
5
a
0.5 1 1.5 2
Midspan Deflection (mm)
Fig. 1 b T y p i c a l load deflection curves from each loboratory for Mix 4 (1.27 percent steel fibers)
66 Chen etal
-
Top LVDT'S Top LVDT'S Yoke Yoke
- - 1 - - 1
Yoke
......
..
Side LVDT'S ;
I.I.8 --- -o-..- -.-.-
Fig. 17-Typical load deflection curves from each laboratory for Mix 5 (0.2 percent polypropylene
fibers)
25
UC-MTS UBC-MTS UBC-lnstron UL-INstron HBT-MTS NU-MTS {
.--
-
Top LVDT'S Top LVDT'S Yoke
-----. Yoke Yoke Side LVDT'S -;-
.......t. II --- -..-I.- -a-.-
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
O I I I I
O 0.5 1 1.5 2
Midspan Deflection (mm)
--``,`,-`-`,,`,,`,`,,`---
Fig. 18-Typical load deflection curves from each laboratory for Mix 6 (0.5 percent polypropylene
fibers)
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 '75 0 6 6 2 ï Y î 0523357 702 W
Testing of FRC 67
25
20
TI
15
10
Mix 2
5
Mix 5
O I
Fig. 19-Typical load versus deflection curves for low toughness FRC mixes
20
15
z
Y
U
m
O
J 10
o
O 2
Fig. 2û-Load versus deflection curves for same specimen recorded by an X-Y plotter and by a
digital data acquisition system
--``,`,-`-`,,`,,`,`,,`---
68 Chen et al
20
- 15
5-
U
m
O
-
J 10
3
X
o,
h
5
O
O 0.2 0.4 0.6 0.8 1
Mid-Span Deflection (mm)
Fig. Ill-Effect of irregularly shaped unstable portions of the P-6 curves on calculated toughness
parameters
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 69
c
T - UBC-lnatron
16
14
O
Y 12
X
Q 10
-E
U
8
E 0
2 4
O
Mix 2 ISFRC 0.25%) Mix 5 (PFR C 0.20%)
14
u)
u)
QI 12
E
c
cn 10
a
O
c 8
Q
u)
e 0
E
m
Q 4
m
7
2
O
Mix 2 (CFRC 0.25%) Mix 6 (PFRC 0.20%)
@)
Fig. 22-Comparison of (a) ASTM C 1018 I, values, and (b) Japanese toughness values for: Mix
2 (0.25 percent steel fibers), and Mix 5 (0.20 percent polypropylene fibers)
--``,`,-`-`,,`,,`,`,,`---
by Colin D.Johnston
The first three authors are sincerely commended for initiating and
coordinating this interlaboratory comparative testing program, which the writer
considers to have been a worthwhile exercise despite his reservations about
some of the results and conclusions. These reservations are based on an
analysis of the first-crack deflection data (Tables 4 and 7 in the paper) which
reveals characteristics that suggest significant inaccuracy or inconsistent
performances for at least two of the six deflection-measuring systems, and
consequently raises the possibility that conclusions regarding the ability of the
various toughness parameters to distinguish performance differences between
mixtures may have been reached using unreliable deflection measurements.
Obviously, the reliability of the data for toughness parameters and the validity
of the conclusions reached about them depends significantly on the quality of
the deflection measurements and the consistency achieved in their analysis.
Since the six concrete matrices have essentially the same mixture
proportions, and are in any case nominally identical for any set of specimens
from a given mixture, the overall mean values of calculated theoretical and
measured deflections for all six mixtures should correspond, and the variability
about the means should be reasonably similar. Instead the following
discrepancies can be identified:
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 71
1. The UBC MTS, NU and HBT systems measure deflections that are on
average significantly lower than the theoretical values by respectively
26%, 14% and 7%. Agreement is good, within 2%, for the remaining
three systems. This reflects the accuracy of the systems independent
of how first crack has been identified. The mean of the differences
between theoretical and measured values in the last line of Table A l ,
reflects any tendency in the measurement system to consistently
underestimate or overestimate deflection. For a reliable system it
should be close to zero as it is for UC, UBC INS, and UL systems.
For single mixtures differences of 10 to 16 (30% to 46% of theoretical)
are worst cases and occur in the UBC MTS and NU data for Mix 5 and
Mix 6.
Given that the six concrete matrices are very similar and that the types and
percentages of fibers used are unlikely to greatly change the first-crack load
and the associated theoretical first-crack deflection, and that any changes in
mixture characteristics would be the same for all participating labs, the means
for first-class deflection (and load) should agree if all labs were using the same
approach to identify first crack. The coefficients of variation about the mean
should be similar if each participant was consistent in identifying first crack
from one mixture to another. Instead, the following discrepancies can be
identified:
1. On average for ail six mixtures, NU clearly identifies first crack at the
lowest theoretical deflection and therefore load, with UBC (both)
somewhat higher, and the remaining three in close agreement and
higher still.
--``,`,-`-`,,`,,`,`,,`---
72 Chen et al
Nowhere are the discrepancies in the defining first crack more obvious,
and surprisingly so,than for the plain concrete (Mix 1 in Table 4 of the paper)
where the first-crack strength derived solely from load varies from a low of
3.93 MPa for Nu to 4.64 MPa for UBC MTS, and 5.23 to 6.1 MPa for the
remaining four participants (mean 5.7 MPa with C of V of 6%). Given the
fairly widely held view that first crack and ultimate loads are essentially equal
for plain concrete, it is surprising, at least to this writer, that Nu and UBC
MTS identify the first-crack load 31 % and 19% lower respectively than the
mean from the other four participants.
In the paper, Table 8 is used to support the view and the associated
conclusion (item 2 in Conclusions) that the Japanese standard parameters TjscE
and FjscE appear to better distinguish performance differences between
mixtures. Yet the increase from Mix 5 to Mix 2 for R30,50 is 67%, higher
than the corresponding 49% for TJSCE and FjScE. Since the two Japanese
parameters are in fact directly proportional any ranking of performance is the
same for both, so only FjscE will be considered in the following analysis.
To compare ail parameters on the same basis, the value for the lowest
level of performance (Mix 5 with 0.2% polypropylene) is expressed as a
percentage of the value for the highest level of performance (Mix 4 with
--``,`,-`-`,,`,,`,`,,`---
1.27% steel) in the last line of Table A2. This percentage is about the same
for FJscE and R30,50, so the argument that the former is better than the latter
is not supported. Moreover, since the performance difference indicated by this
percentage is greater for R3?,50 than for R20,30 or it is probable that
R50,100 which has an end-point deflection comparable to the 2 mm end point
for FJscE would distinguish the performance difference better still. Given that
the end-point deflection for FjscE corresponds to a serviceability condition
with crack width of the order of 2 mrn, that FJscE does not address
serviceability conditions with lower end-point deflections and smaller crack
widths more appropriate for many applications, and that R30,50, R,0,30 and
even R10,20 do address serviceability conditions with lower end-point
deflections while distinguishing performance levels reasonably effectively,
Testing of FRC 73
Summary
--``,`,-`-`,,`,,`,`,,`---
The writer agrees with much of the content of the paper and with the
essence of the conclusions other than item 2. Despite the criticai analysis
offered in this discussion, he believes that much can be learned from the
project. Obviously, achieving accurate and consistent deflection measurement
is more difficult than was thought at the outset. Refining the existing
provisions in ASTM Cl018 for identifying first crack and better
communicating their intent remain a priority. The fact that the effects of any
instability in the load-deflection curve immediately following first crack can be
important and can often be eliminated by the use of residual strength factors
in preference to toughness indices needs to be more widely recognized.
Finally, the reaiity is that engineers involved in design understand the
significance of residual strength better than toughness or toughness indices for
most applications, so specifications and test methods should move towards
more emphasis on residual strength at an end-point deflection serviceability
condition appropriate to the application. In this regard, a residual strength
factor based on the average load over a portion of the load-deflection curve
immediately preceding the specified end-point deflection (e.g. R30,50or
R,,,,,) is more meaningful than a factor based on the average load over the
whole curve including the portion prior to first crack (e.g. FJscE>.
-
1 042 032 036 038 O40 027
II
038 (4) 024 (8) 029 (7) O41 (-3) 032 (8) 022 (5)
2 038 037 036 039 040 027
I,
039 (-1) 032 (5) 036 (O) 036 (3) 038 (2) 032 (-5)
3 040 034 036 039 O40 034
II
4
II
-
Max.
Min.
Range
Mean
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 75
I30
18.1í3)
25.8í2)
3 1.4(1)
15.9(5)
18.0(4)
--``,`,-`-`,,`,,`,`,,`---
SP 155-4
--``,`,-`-`,,`,,`,`,,`---
Kevwords: Cracking (fracturing); fiber reinforced concretes; fibers; load-
deflection curve; toughness a;
Introduction
Testing of FRC 79
--``,`,-`-`,,`,,`,`,,`---
0.24 mm (0.0094in.), while all the other references describe tests with
beam sizes of 100 x 100 x 350 mm (4 x 4 x 14 in.), span 300 mm (12 in.),
or 150 x 150 x 500 mm (6 x 6 x 20 in.), span 450 mm (18 in.) leading to a
real first crack deflection of approximately 0.04 mm (0.0016 in.) or 0.06
mm (0.0024 in.) respectively. The fact that a larger pari of the load-
deflection curve is considered explains the better sensitivity of the I,,
value.
The other references also indicate that using a larger pari of the
load deflection curve I>,, I,, or more, or even the Japanese definition of
toughness (9),going to a deflection of 1/150 of the span length, is much
more capable of distinguishing the differences in toughness due to fiber
type and/or fiber volumes. Several authors also mentioned the difficulty
of measuring accurately the first-crack deflection and the influence of
including extraneous deformations in the determination of toughness
indices (2 - 4,8,10,11). The consequence is that it is very difficult to
determine I, and I,, correctly. Although these conclusions were known at
the time of the latest revision of ASTM 1018, the requirement to determine
I, and I,, was maintained. Indeed, a proposal to omit I, and R,.,,,make I,,
mandatory and to make I, and R,,,3, optional was voted down.
Commercial aspects may have played a more decisive role than scientific
reasons for defeat of this proposal. It may be desirable to modify the
standard so that it is only valid for the case in which a minimum toughness
is provided (no need for substantially brittle materials to look for
toughness values).
But why are the indices chosen as they are? In the original
C 1018 edition published in 1984, the same values, I, and I,, ,were
specified as obligatory, but here I, was also specified as optional
(deflection of 15.5 times 6).
This means also that the 1984 proposed I, toughness index was
indeed calculated using a pari of the L- D curve after first crack equal to 2
times the measured first-crack deflection and thus at least 6 times the real
first-crack deflection. For I,, , it was approximately 15 times the real first-
crack deflection, and for I,, , the proposed end of the test, it was more
than 40 times the real first-crack deflection. This explains why in these
1984 tests it was possible to distinguish the influence of fiber type and
fiber amount with the I, and I,, indices.
I. The prescribed toughness values I, and I,, for the small beams
100 x 100 x 300 mm or 150 x 150 x 450 mm (4x4~12or 6 x 6 ~ 1 8
in.) only consider a very small part of the L - D curve and are not
at all capable of distinguishing the influence of fiber type and
fiber amount on toughness. This is especially true since the
correct determination of the L - D curve immediately after first
crack is often very difficult.
Testing of FRC 81
New Proposal
Reauirements:
“x” may be taken as the deflection at first crack calculated from the
formula:
Testing of FRC 83
--``,`,-`-`,,`,,`,`,,`---
Fig. 8 illustrates how first crack load is determined using the above
definition. A large scale is used to illustrate the method. It can be
seen that the possible error, if any, is relatively small and will have
only a limited influence when used to determine relative toughness
(the deflection at first crack is no longer used as a parameter).
a) R,,,,,5 for the deflection interval of 0.5 mm (0.02 in.) to 1.5 mm (0.6
in.), and
b) R,,,,, for the deflection interval of 1.5 mm (0.6 in.) to 3.0 mm (1.2
in.).
the considered deflection interval, and the value of the first-crack load
(or equivalent first-crack stress). These values are comparable with
the previous Residual Strength Values, Rx,y,as defined in ASTM
C1018. They also can be used in relation to the serviceability
requirements of the structure
5. The argument that the test requires too much time if large deflections
must be measured can easily be solved by increasing the rate of
increase of deflection. Increasing this rate to 0.5 mmlmin. (0.02
in./min.) has almost no influence on test results (9).
References
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 85
8. Mindess, S., Chen, L., and Morgan, D.R., ?Determinationof the First-
Crack Strength and Flexural Toughness of Steel Fiber Reinforced
Concrete,? To be published in ACBM Journal, 5th issue.
1I.
El-Shakra, Zeyad M., and Gopalaratnam, Vellore S., ?Deflection
Measurements and Toughness Evaluations for FRC,? Cernent and
Concrete Research, Vol. 23, 1993, pp. 1455-1466.
:
<II
8
2
3
O
n z
--``,`,-`-`,,`,,`,`,,`---
Testing of FRC 87
.x -
o
c
-
o
L
o -
o
--``,`,-`-`,,`,,`,`,,`---
r
-I
a
/
A
Load F(W
3.0 -
2.0 -
1.0 -
O
'6' 9 6 . 1 5.56.2 0.3
Defiection (rnm)
Load f F(kN)
1 2
--``,`,-`-`,,`,,`,`,,`--- Deflection (mrn)
Copyright American Concrete Institute
Provided by IHS under license with ACI Fig. 4-1-D turves recorded with two sytems
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-355 95 Ohh29Y9 0523378 437 =
Testing of FRC 89
First crack
21.2 kN
20
10
I I
I
I I I
*
I I I
O I , I
02 0.3
-s,+ w Deflection (m)
Fig. !i-L-D curve for apparently very tough concrete (large scale)
0 1
1.o 2.0
--``,`,-`-`,,`,,`,`,,`---
DekUOn (mm)
Copyright American Concrete Institute Fig. &Same L-0 curve as Fig. 5 at normal scale
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 Ob62949 0 5 2 3 3 7 9 373
7
70
-i
50
45
.^ I
Beam
i50 X 150 X 500 (450) mm3
,..-.
'- .'
Daflarttvi immi
y
16i1
o
0X-J o! 1!0 115 2!0 2!5 3!0
Load
fc A
fc B
fc c
fc D
. .
Testing of FRC 91
4
--``,`,-`-`,,`,,`,`,,`---
SP 155-5
--``,`,-`-`,,`,,`,`,,`---
Antonio Arino is a research assistant at the Department of Civil
Engineering, Arizona State University. His research interest is in the
area of high strength fiber reinforced concrete.
INTRODUCTION
Testing of FRC 95
G(a,) =R(a,) - - --
aa ¿%I
a a= a, (2)
Testing of FRC 97
+
By measuring (P, a =a, Aal values at successive intervals of crack
growth, the compliance-crack length relationship is constructed.
The stress intensity factor at the tip of the effective crack may be
obtained using Equations 5 and 6, and reported as the R-Curve, KR.
This definition of the R-Curve is referred to as a modified LEFM
approach:
--``,`,-`-`,,`,,`,`,,`---
1.99-0((1-a)(2.15-3.93a +2.7a2)
F(a) = (6)
Jx(i+2a)(1- ~ r ) ~ n
G ( u ) = 1 -
- a c p 2
2 t au
terms as:
1 8%
G * ( u )= G ( u )+ - -
2 t aa
Equation 8 includes the effect of residual displacements in the
strain energy release term. This parameter is used t o define an
elastically equivalent effective toughness, KR, referred t o as the total
toughness:
where E', = E, / (1- v:) for plane strain and E, for plane stress. E,
and v, represent the elastic modulus and the poisson's ratio of the
composite.
EXPERIMENTAL PROGRAM
Testing of FRC 99
--``,`,-`-`,,`,,`,`,,`---
the ductility by bridging the matrix cracks. Load deflection response
of a specimen reinforced with 4% PP fibers that are 12 mm long is
shown in figure 3. The nonlinear effects due to inelastic deformation.
and the toughness increase are significant. Work of fracture was
defined as the area under the entire load-deflection curve. i n
comparison to the plain matrix (shown in Fig. 21, it is observed that
the work of fracture increases as much as fifteen times. The ductile
response beyond the maximum load is due t o the closing pressure
exerted by the fiber pullout. Note that the response of the beam
under the load may be compared to an elastic-plastic solid.
DISCUSSION OF RESULTS
methods are used to derive the KRat various crack lengths for four
replicate FRC specimens with 1% Alumina fibers. One approach is
based on the use of equation 5 (Modified LEFM approach), and the
other is based on the use of equation 7 followed by equation 9
(compliance method). These methods are compared in Fig. 5. The R-
Curve increases considerably beyond the maximum load and
asymptotically approaches a constant toughness level. This may be
viewed as a steady state crack growth condition. The KRcurve at
this level is as much as 70% higher that its value at the maximum
load. The differences between the t w o methods are well within the
scatter of the test results.
R-Curves were defined using the strain energy release rate and
an elastically equivalent LEFM model. The energy release rate is a
fraction of the total potential energy of the system that is not
consumed due to the irreversible processes. In the present approach,
the contribution of the inelastic deformation to the total toughness
was measured. The work of fracture method however, includes the
energy dissipation due t o the crack growth and other dissipating
mechanisms in the process zone.
CONCLUSIONS
Acknowledgements
References
1. Mindess, S.,"The Fracture Process Zone in Concrete,"
Toughening Mechanisms in Quasi-Brittle Materials, S.P. Shah
(ed.), 1991, Kluwer Academic Publishers, pp. 271-286.
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 9 5 I0662949 0523191 9'70 I
2. Lenain J. C., and Bunsell A. R., J. of Mat. Sci., Vol. 14, pp.
31 2-332 (1979).
11. Mai, Y.W., and Hakeem, Slow Crack Growth in Cellulose Fibre
Cements," Journal of Materials Science, 19, ( 1984) 501-508.
íMPa) (GPa)
alumina 0.762 2.5 1725 105 2.70
carbon 1 .o 25 1800 230 1.90
PP 12.0 35x250 340-500 8.5-1 2.5 0.91
Crack Extension
7
%= ao+Aai =u a
Fig. la-R-curves representing resistance to crack growth for brittle, quasi-brittle, and fiber
reinforced materials. The instability condition is the point of tangency of R-curve with the strain
energy release rate, G
CMOD H H
S = 4b
--``,`,-`-`,,`,,`,`,,`---
900 I
--``,`,-`-`,,`,,`,`,,`---
Deflection, mm
Fig. 2-Cyclic load-deflection response of plain mortar and specimen reinforced with 1 percent
alumina fibers
1000 I
4% PP
800
--``,`,-`-`,,`,,`,`,,`---
600
400
200
-
O 1 1 1 1 1 1 1 1 1 ( 1 ' 1 1 1 1 ' 1 1
Fig. 3-Cyclic load-deflection response of specimen reinforced with 4 percent polypropylene fibers
Fig. &Loading and unloading compliance of mortar and specimen reinforced with 4 percent
polypropylene fibers as function of crack mouth opening displacement
40 - * *
-
O O *
I
O 1 1 1 1 ~ 1 ~ ' ~ ~ ~ '
O 10 20 30 40
Crack Extension, mm
Fig. E-R-curve responses based on two methods of modified LEFM approach and compliance
approach
--``,`,-`-`,,`,,`,`,,`---
I.5
E
E 1.0
ai
O
2
o
o
.-
ci
(B
-a
(II
0.5
-t
--``,`,-`-`,,`,,`,`,,`---
o.oq-p-l , , ~
I I I I I l I I I l I I
O 10 20 30 40
Crack Extension, mm
Fig. b l n e l a s t i c deformation as function of track extension obtained using compliance calibration
technique
300
-- e- Mortar, Modified LEFM
Mortar, Total
-
* -
-*
8% Carbon, Modified LEFM
--
8% Carbon, Total
- 4% PP, Total
loo/
--``,`,-`-`,,`,,`,`,,`---
O 10 20 30 40
Crack Extension, mm
Fig. 7-R-Aa curves for mortar and composites with 8 percent carbon and 4 percent PP fibers.
The energy term due to inelastic deformations is a significant portion of the R-curve
Deflection
Fig. &Definition of cyclic energy and total energy dissipation during crack growth experiments.
The work of fracture is obtained by summation of incremental values
0.025 3,
Mortar
0.020
+ Total
,* cyclic
E 0.015
E
t
c
e 0.010
0.005
0.60 1
I /
4% PP
* Total
Cyclic
E 0.40
I /
??
L
O
0.20-
1 #
-
0.00
0.00 0.25 0.50 0.75 1.00 1.25
CMOD, mm
--``,`,-`-`,,`,,`,`,,`---
(b)
Fig, 9-Work
Copyright American Concrete Institute of fracture as function of crack mouth opening for (a) mortar, and (b) 4 percent
PP fiber
Provided by IHS under license with ACI
No reproduction or networking reinforced
permitted composite
without license from IHS Not for Resale
A C 1 SP-155 95 Ob62949 0523202 5 ô b W
E
???1
350
--``,`,-`-`,,`,,`,`,,`---
z?
Mortar \
Fig. 10-Cornparison of total work of fracture and its cyclic component for various fiber
composites
SP 155-6
Svnousis: Two test methods are presented that can be used for evaluating
durability of polymeric fibers subjected to: alkaline environment present in
concrete, and UV light exposure. The test methods were used to evaluate three
polymeric fibers namely: nylon, polypropylene, and polyester. Durability of the
fibers in an alkaline environment was ascertained by measuring the flexural
toughness of fiber reinforced concrete specimens that had been aged in lime
saturated water maintained at 50°C. The UV light exposure test was conducted at
a temperature of 65°C with intermittent water spray. The wet spray was used to
simulate conditions in the field. Durability of the fibers was determined by
measuring the retained tensile strength of the fibers after light exposure and by
observing the surface characteristics of fibers under a microscope.
The test results indicate that nylon and polypropylene fibers are durable in
alkaline environment present in concrete. The nylon fibers, which were light
stabilized, were determined to be stable under UV light exposure. Polypropylene
fibers deteriorated under W light, and the deterioration of the polypropylene
single filament fibers was more rapid than for the fibrillated fibers. Hence these
fibers should not be used in applications where the fiber contribution is needed at
cracked-exposed sections.
INTRODUCTION
Polymeric fibers have been used in concrete for the past 15 years. The
fibers include: nylon, polyester, polyethylene, polypropylene, and polyvinylalcohol
[1,2]. In most cases, these fibers are used at low volume fractions to reduce the
shrinkage cracking during the initiai and final setting period. For this type of
application, the fiber content is normally less than 1 kg/m3. Recently polymeric
fibers are being used at higher volume fractions for repairs, tunnels, and canal
linings. At these higher volume fractions, fiber contribution to ductility is utilized.
Even under low volume fractions, fiber contribution to ductility is considered as a
factor for choosing the fiber type.
The fibers exposed to the elements should be durable in order to sustain the
composite action. For example, the fibers in a cracked overlay will be exposed to
light and rain. If these fibers degrade due to the exposure to light, then the
composite action will be lost. It could be argued that most fibers are protected by
concrete. But the need for fiber contribution is critical at the cracked section to
reduce crack widths and reduce long-term crack opening. In fact, the fibers are not
needed in hardened concrete if it can be assured that concrete will not crack.
Hence, the fibers exposed at the cracked sections should be durable so that they
can transfer forces across the crack. “Although UV radiation makes up
approximately 5% of the total sunlight, it causes almost all of the damage to
durable materials” [3]. Some of the most common attacking elements are UV light
and moisture. UV light component is strong both in sunlight and artificial light.
This paper presents the test methods and results for popular polymeric
fibers for both durability tests involving alkaline environment and W light. The
alkaline durability test was adopted from the method used for glass fibers, and the
UV test method was adopted from an existing ASTM test used for evaluating the
deterioration of geotextiles.
--``,`,-`-`,,`,,`,`,,`---
DURABILITY TEST FOR ALKALINE ENVIRONMENT
The popular test for evaluating the fibers in concrete is the accelerated
aging test in which fiber reinforced concrete samples are stored in lime-saturated
water maintained at elevated temperature levels. The lime-saturated water prevents
the leaching away of naturally occurring lime. The elevated temperature levels
varying from 50 to 80°C accelerate the aging process. For example, 1 day
immersion of the sample in lime-saturated water maintained at 50°C has been
shown to be equivalent to 101 days of natural weather exposure in the United
Kingdom with a mean annual temperature of 10.4OC [1,4,5]. Typically, test
specimens stored in water baths are tested either in tension or flexure at various
time intervals to determine strength and ductility. Normally the accelerated tests
are run up to 52 weeks.
accelerated aging 50, 60 and 80°C. When the accelerated aging curves at various
temperatures are horizontally displaced to the right, they tend to coincide with
natural weathering with good accuracy, Fig 2. Hence, it can be concluded that the
results obtained at various accelerated aging temperatures correlate well with
natural weathering conditions.
Table 1 presents some of the xenon arc methods that could be used for testing
fibers. The American Association of Textile Chemists and Colorists (AATCC) test
number 16E is intended for determining the colorfastness (fade-resistance) of
textile fibers to light. AATCC 16E is not a good test for concrete reinforcing fibers
because it lacks the shorter W part of the spectrum that is present in sunlight. The
Society of Automotive Engineers (SAE) JI885 is an accelerated exposure test for
evaluating automotive interior trim components which are often polymeric in
composition. While nylon, polyester, and polypropylene are often used in
automotive interiors, the SAE JI 885 test is too severe for concrete reinforcing
fibers as the test temperature is 89°C.
In the ASTM D 4355 test method, the fibers are exposed to UV light for
specified time intervals then tested for percent breaking strength retained. The
fibers were also examined using Scanning Electron Microscopes (SEM) to
determine the extent of deterioration.
--``,`,-`-`,,`,,`,`,,`---
EXPERIMENTAL PROGRAM
For the first phase dealing with the durability of fibers in concrete, fiber type
was the only independent variable. The concrete was proportioned to obtain a 28
day compressive strength of about 20 MPa. The durability was studied using
--``,`,-`-`,,`,,`,`,,`---
flexural test specimens. Both flexural strength and flexural toughness were
measured at various stages of accelerated aging.
For the second phase, nylon, fibrillated polypropylene, and single filament
polypropylene fibers were tested using ASTM D 4355 test method. Fiber strengths
and surface characteristics of fibers after various exposure periods were used to
evaluate the durability.
Materials
The constituent materials for the beam specimens consisted of ASTM
Type I cement, natural sand, crushed stone, tap water, and air entraining
admixtures. The aggregates met the ASTM gradation requirements. Ail the three
types of fibers were 19 mm long. Nylon and polyester fibers were made of single
filaments whereas polypropylene fibers were fibrillated.
Mix Proportions
The matrix composition for flexural specimens consisted of 307, 813,
1068, and 177 kg/m3 of cement, sand, coarse aggregate, and potable water
respectively. In addition, 500 and 709 mum3 of high range water reducing and air
entraining admixtures were used. A high fiber content of 4.75 kg/m3 was used in
order to obtain well-defined and repeatable load-deflection responses. The well-
defined load-deflection curves were needed in order to compute toughness indices
explained in a later section of this paper.
Specimen Preparation
The primary test specimens for phase I study were 100 x 100 x 360 mm
prisms. These prisms were cast using Plexiglas molds and compacted using a table
vibrator. After casting, the specimens were left in their molds for 24 hours and
covered with polyethylene sheets. After 24 hours, they were demolded and cured
for 27 days in a moist room maintained at 100 percent relative humidity.
Companion 150 x 300 mm cylinders were made to obtain the 28 day compressive
strength. These cylinders were made using plastic molds and were compacted
using a vibrating table. The curing scheme was the same as the one used for
flexural specimens.
For phase II study, the fiber specimens were obtained from commercial
suppliers. Fibers of the same cross-sectional area were selected for physical testing
to insure valid physical test results. The faces of the clamps that grip the fibers
were resurfaced with very fine crocus cloth to insure fibers would not slip when
tested. Strain related properties of the fibers were not considered. Inaccuracies in
the measurement of short gauge lengths and the penetration of the forces into the
clamping region make determination of effective gauge length difficult. Therefore
only the breaking strengths of the fibers were determined. At least 10 specimens of
each type of fiber were tested for breaking strength. The remaining fibers were
evaluated by SEM. The samples were either 50 mm or 19 mm long.
TEST METHODS
The fresh concrete was tested for workability and air content. Workability
was measured using both the standard and inverted slump cone test, ASTM C143
and C995 respectively. The air content was measured using the pressure method,
ASTM C23 1. The 150 x 300 mm cylinders were tested in compression at 28 days
to obtain the compressive strength.
The accelerated aging test was conducted by storing the specimens in lime-
saturated water maintained at 50°C. Flexural strength and flexural toughness were
used to measure the contribution fibers provide to the concrete matrix. For each
group, twenty-four 100 x 100 x 360 mm prisms were made and cured for 28 days
before placing them in hot water bath.
Three samples were tested in flexure after O, 4, 8, 16, 32, and 52 weeks of
accelerated aging, The tests were conducted using ASTM C1018 procedure, in
which the beams were loaded at mid third points over a simply supported span of
300 mm. The deflection was measured at mid span. The load-deflection curves
were used to obtain the modulus of rupture and toughness index. Maximum load
and uncracked section properties were used for the computation of modulus of
rupture. Toughness index values Is, Ilo and 40were computed using ASTM C1018
I
definitions [8]. The ratios I& and 13&, were used to estimate the ability of the
fiber reinforced concrete specimens to sustain loads at large deflection.
ASTM D4355 method was used to test the fibers. This method subjects the
fibers to a light spectrum with an intensity that matches winter sunlight. The
exposure cycle consisted of 90 min. light only and 30 min. light plus water spray.
The intensity of light was controlled by monitoring the irradiance level at a
specified wavelength, Table 1. Approximately 20 specimens of each fiber type
were exposed in one xenon arc weatherometer at the same time to insure equal
weathering conditions. Tensile strength tests were conducted after O, 150, 300,
and 500 hours of exposure and the percent tensile strength was determined.
Scanning Electron Microscopic (SEM) photos were also taken at magnificationsof
400X and 1200X. For smaller diameter short fibers, only photos were taken
because tension tests could not be conducted.
The load deflection curves obtained at various stages of aging are shown in
Fig. 3, 4, and 5 for nylon, polypropylene, and polyester fibers respectively. The
toughness indices are presented in Table 2. The modulus of rupture and the ratios
of 1 3 0 / I ~ ,are presented in Fig. 6 and 7 respectively. A careful study of Table 2 and
the figures lead to the following observations:
'The post-peak resistance decreases with aging for polyester fibers. For the
nylon and polypropylene fibers, the post-peak performance improves in certain
instances.
'The toughness index IS is about the same for all three fiber types and
different aging periods. This is in agreement with earlier results that indicate that Is
is not sensitive enough to distinguish fiber contributions [i].
--``,`,-`-`,,`,,`,`,,`---
'Nylon and polypropylene fibers had higher 110and 130 values than polyester
fibers. Decreases in the 130 values for polyester fibers were observed with aging.
This difference can be observed by noting the variation in i3O/I10 values, Fig. 7.
The literature points out that "solar irradiation of PP and PE leads to the
embrittlement of the polymers and the formation of oxidation products" [IO]. For
polyamides photooxidation is initiated by impurities in the polymer.
SEM photos for the various fibers are shown in Fig. 9 to 15. For the
polypropylene fibrillated fibers both 400X and 1200X magnification are shown
because these fibers are more coarse and it is difficult to see the entire fiber under
1200X magnification. Fig. 9 to 15 confirm the disintegration process for the
polypropylene fibers and the results shown in Fig. 8. All polypropylene fibers
become brittle after exposure to W light. The fibers with smaller diameters
disintegrate more rapidly due to increased exposed surface area as shown in Fig.
-
15.
CONCLUSIONS
Based on the results presented in this paper and observations made during
the experimental investigation, the following conclusions can be drawn.
'At a fiber content of 4.75 kg/m3, all three fiber types (nylon, polypropylene
and polyester) provide post-peak resistance.
'If the structure is totally protected from UV light, then both nylon and
polypropylene fibers are expected to be durable. But if the fibers are exposed to
REFERENCES
1. Balaguru, P., and Shah, S. P., Fiber Reinforced Cement Composites, McGraw-
Hill, 1992, 535 pages.
4. Litherland, K. L., Oakley, D. R., and Proctor, B. A., “The Use of Accelerated
--``,`,-`-`,,`,,`,`,,`---
Aging Procedures to Predict Long-term Strength of GFRC Composites”,
Journal of Cement and Concrete Research, Vol. 11, 1981, pp. 455-466.
10. Encyclopedia of Polymer Science & Engineering, 2nd edition, Volume 4, 1986.
--``,`,-`-`,,`,,`,`,,`---
Age Weeks
OPP - polypropylene
+PY - polyester
4c - U.K.weathering
ni e Accelerated nging at
i O
2:
Wl
- 30- % *
O
A
50°C
60°C
a ** 80°C
3r $*a
D
$ 20-
0 4
h
O
0& B WO$*O,o
g
CI
10-
s
--``,`,-`-`,,`,,`,`,,`---
12
z
2¿
2
O 4
I4
I I I I
O 1 2 3 4
DEFLECTION, m m
Fig. &Load deflection behavior of beams subiected to accelerated aging: nylon fibers
Y /
- - - - - -b
I
o 1 2 3 4
DEFLECTION, m m
Fig. +Load deflection behavior of beams subieaed to accelerated aging: polypropylene fibers
--``,`,-`-`,,`,,`,`,,`---
I'
4
Fig. 5-load deflection behavior of beams subjected to accelerated aging: polyester fibers
6
r 52
ACCELERATED AGE, UEEKS:52
32
o
H 4
w
e:
E
0.0
2 3
cri
O
VI
P
c
l
302
O
E
1
NYLON
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
500 hrs exposure (400X Magnification)
taken
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
Restrained Shrinkage Tests on Fiber
Reinforced Cementitious Composites
by N. Banthia, M. Azzabi, and M. Pigeon
--``,`,-`-`,,`,,`,`,,`---
INTRODUCTION
It is generally recognized that uniaxial tensile tests are desirable fiom the
point of view of obtaining fimdamental material information and also in order to
understand the mechanisms of fiber reinforcement when under drying shrinkage
conditions. Based on this recognition, it was undertaken to develop a rational test
technique capable of subjecting cement-based fiber reinforced composites
EXPERIMENTAL
In an actual test, the freshly mixed cement mixture was poured in the mould
and the whole assembly was transferred to the drying chamber. Later, when
possible, the specimen was demoulded and further drying occurred in the chamber.
In this state, the only restraint in the specimen came fiom the end-anchors in the
longitudinal direction. AU crack observations and measurements were canied out in
the drying chamber itself on the most severely cracked surface for up to twenty-four
hours.
Test Procram
The equipment described above was used for studying cracking in fiber
reinforced cement paste and mortar under one-dimensionai restraint when subjected
to a severe dtying environment. M e r several unsuccessful attempts at inducing
cracking in milder environments, a particularly severe environment with a
temperature of 50°C and a relative humidity of less than 50% was chosen. Seven
fiber types belonging to two major categories macro(1arge) and micro(fine) were
investigated. The details of these fibers are given in Table 1 where .the extremely fine
s i e of micro-fibers as compared to the macro-fibers can be noted. At equal volume
fractions, one could expect two to three orders of magnitude more micro-fibers in
--``,`,-`-`,,`,,`,`,,`---
the composite than macro-fibers. Consequently, it was anticipated that the micro-
fibers would provide reinforcing mechanisms distinctly different from those of the
macro-fibers.
Two matrices (cement paste and cement mortar) were investigated, although
only the results obtained with the mortar matrix are presented in this paper. Results
with the paste matrix may be found elsewhere (16). The mortar matrix proportions
were (cement:water:silica-fÙme:sand) 1.0:0.4:0.2:2.0,and appropriate quantities of
superplasticizer were added to obtain a satisfactory workability. The use of silica
fume was considered essential for an effèctive dispersion of high specific surface area
micro-fibers (17) and for a better fiber-matrix bond. A routine mortar mixer was
used. The following fiber volume fractions were investigated (see Table 1 for
notation):
Data Analysis
--``,`,-`-`,,`,,`,`,,`---
transfer length and with sufficient buiid-up of stresses in the matrix, the matrix will
develop another crack parallel to the first one (18). This multiple cracking can go on
until the entire element develops cracks which are approximately equidistant and run
perpendicular the direction of tensile stresses.Multiple cracking can occur, however,
only if the fibers at a section can cany the stresses higher than that carried by the
matrix at cracking i.e., if the critical fiber volume fiaction is exceeded, if the fibers
are long enough and also if fiber breakage across the crack does not occur.
Based on the above discussion, it is clear that fibers influence the restrained
shrinkage behavior in two ways: first, they distribute cracking more evenly over the
entire length resulting in closely spaced cracks, and second, they reduce crack widths
through effective crack bridging. It appears logical, therefore, that any attempt at
quantifj4ng fiber effectiveness must consider both these mechanisms. In this study, a
non-dimensional parameter called thefiber eficiency factor was adopted in order to
quantimng the effectiveness of a given fiber. It was defined as,
where, as defined before L, is the cumulative crack length and w, is the cumulative
crack width over a given surface of observation. Clearly, an unreidorced matrix that
usuaiiy failed with a single wide crack, had a low eficiency factor.
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C I sp-155 95 0 6 6 2 9 4 9 0 5 2 3 2 3 0 440
RESULTS
General Observations
Specimens without the fibers, as expected, cracked with one wide crack
spanning the entire width of the specimen. In the case of the macro-fiber reinforced
composites, however, the specimens developed a number of closely spaced cracks
and each crack stayed substantially narrower than the sigle wide crack observed in
the unreinforced specimens. In the micro-fiber reinforced composites, on the other
hand, the specimens usually cracked with one crack in the middle of the specimen,
and this crack stayed very narrow. In Figures 2 and 3, the number of cracks, n, as a
function of fiber volume fraction for the macro and micro-fiber composites,
respectively, are plotted.
Effectiveness of Macro-Fibers
Detailed results are given in Table 2. The maximum observed crack widths,
,w are plotted as a íùnction of fiber volume fraction in Figure 4 for the macro-fiber
composites. Note a consistent decrease in w,, with an increase in the fiber volume
fiaction. W e specimens without fibers always developed one wide crack in the
middle, fiber reinforced specimens developed several well distributed cracks much
narrower in widths; at large fiber volume fiactions, as many as 15 cracks were
observed. Even with those many cracks, the cumulative crack widths, w,,for the
fiber reinforced specimens stayed considerably smaller than the width of a Sigle
--``,`,-`-`,,`,,`,`,,`---
,I
crack in the unreinforced specimen. Even more interestingly, as the number of cracks
I
grew larger with the fiber volume fiaction, the cumulative crack width, w,,in fact,
decreased.
I In Figure 5, thejìber eflciency factors (Eqn. 1) are plotted as a function of
fiber volume fraction. It may be noticed that a fiber deformed along the entire length
(Fiber F3) is more effective than those deformed oniy at the ends (Fibers FI and F2).
Also, if the geometry is the same, a longer fiber (Fiber F2) is more effective than a
shorter fiber (Fiber F1) in reducing the crack widths, in causing more multiple
cracking and in the overall reinforcing efficiency.
Effectiveness of Micro-Fibers
CONCLUDING REMARKS
ACKNOWLEDGEMENTS
The authors wish to thank the Natural Sciences and Engineering Research
Council of Canada (NSERC) for continued financial support.
REFERENCES
3. Paillere, A.M., Buil, M. and Serrano, J.J., Effect of Fiber Addition on the
Autogeneous Shrinkage of Silica Fume Concrete, AC1 Materials Journal,
86(2), March-April 1989, pp. 139-144.
--``,`,-`-`,,`,,`,`,,`---
10. Krenchel, H. and Shah, S.P., Restrained Shrinkage Tests with PP-Fiber
Reinforced Concrete, Fiber Reinforced Concrete Properties and
Applications, SP-105, American Concrete Institute, Detroit, 1987, pp. 141-
158.
13. Kovler, K., Sikular, J. and Bentur, A, Restrained Shrinkage Tests of Fiber
Reinforced Concrete Ring Specimens: Effect of Core Thermal Expansion,
Materials and Structures,RILEM, 26, 1993, pp. 23 1-237.
15. Kraai, P.P.,A Proposed Test to Determine the Cracking Potential due to
Drying Shrinkage of Concrete, Concrete Construction, 1985, pp. 775, 778.
16. Banthia, N., Maha, A. and Pigeon, M., Restrained Shrinkage Cracking in
Fiber Reinforced Cementitious Composites, Materials and Structures
WEM), 26, 1993, pp. 405-413.
18. Avenston, A., Cooper, G.A. and Kelly, A., Single and Multiple Fracture, in
The Properties of Fiber Composites, Proc. Conference of National Physical
Laboratories, IPC, Science and Technology Press, LJK, 1971, pp, 15-24.
8
e4
m
L
o
n
.-
L
--``,`,-`-`,,`,,`,`,,`---
I Mortar I
~
1
Maximum Observed Crack Width in mm
Total Cumulative Crack Width in mm
3
Total Cumulative Crack Length in mm
* Number of Cracks
LJw,: Fiber EfficiencyFactor
I Mortar I
Fibre Vf, % W1- d L3c LJWC n*
0.00 2.85 2.85 40.0 14.03 1
--``,`,-`-`,,`,,`,`,,`---
40x40 x500
i
In
i
ímm)
Mortar (Macro-Fibers) 0 F1
' F2
+ F3
--``,`,-`-`,,`,,`,`,,`---
Maar (Micro-Fibers) 0 c1
--``,`,-`-`,,`,,`,`,,`---
c2
+ s
" P
T T +
O I I I
O 1 2 3 4
Fiber Volume Fraction, V€,%
Fig. &Number of shrinkage induced cracks, n, in micro-fiber reinforced mortars. Notice that
composites usually cracked with a single crack without any multiple cracking
g 3.0
Mortar (M acro-Fibers)
s 0 F1
8z F2
24
2.0; + F3
j2g3 1.0-
O
ti I
0.0 I I I
Mortar(MaaeFibers) o
cl I
F2
+ F3
Q
iz 0.0 I I l I
3.0
Mortar (Micro-Fibers) o c1
c2
2.0
i'I \\ + s
1.o
0.0 I I I
Fig. ó-Maximum crack widths (w ) observed in cement composites reinforced with micro-
fibers under restrained shrinkage comndxitions
--``,`,-`-`,,`,,`,`,,`---
Mortar (Micro-Fibers) 0 c1
3s
I=-
+ s
c2
L 2oo.ol M P
Fig. 7-Fiber efficiency factors (LJwJ plotted as function of fiber volume fraction for various
micro-fibers
SP 155-8
INTRODUCTION
~
Enhanced tensile strength of concrete is the most important
physical property that fibers contnbute to the brittle concrete, and yet
tensile testing has not become a part of FRC acceptance criteria.
Many previous tensile tests resulted in relatively high standard
deviations within batches with corresponding coefficient of variations
load, wliicii is sliortiy after the mortar or concrete has artaineci initial
set. The molds also are lined with an expanded metal reiiiforciiig that
is bolted to the side wall for assuring a good mechanical as well as a
friction bond to the concrete to be tested. Eye-bolts are threaded into
a welded nui on top of a threaded rod that runs tlirough the side wall
sections that make up the top and bottom portions of the specimen
inold to form a siinple link at both ends of a cylindrical specimen (see
Fig. 2). I11 this direct tension test, the iiiteiit is to keep the weak link in
the center of the 6.7 inch ( 1 7 cm) length at the duct-tape joint of the
3.2 inch (8. i cm) diameter cylinder. Simplicity of mold construction
to help assure repeatability was a primary objective.
Test Set-up
I n these direct tension tests only the mortar fiaction of concrete
mixtures was employed. Future tests will expand the specimen sizes
larger than 4 inch ( I O cin) diameters so that concrete mixtures with
coarse aggregates can be tested. Figures 3 through 5 illustrate the
--``,`,-`-`,,`,,`,`,,`---
specimens, apparatus, test set-up arid procedures used. It is important
to note that the bottom is capped and the top of the specimen is left
open in order to more closely siinulaie iypicai siab sur%aceexpusure to
drying conditions. After casting, the specimens are placed in a
controlled chamber for the first 6 Iiours before testing. The chamber
is controlled at 80 degrees F. and air is circulated by internal fan.
Fiber Cement (IFCj. Mixtiires ii 4-6 are basically the saine bui
adjusted to equal yield for a more precise volume comparison.
Intergrwnd Fiber Cernent M ateritils
Research on the interground fiber cement process in a laboratory
griJldiJig inill indicated that fibers become iinpregiated with grains of
cement (6). Figures 6 and 7 show some of the roughened
polypropylene fiber surfaces and cement impregnation. In this series
of lab tests, milling was controlled to simulate the actual cement
milling process where fibers more closely resemble the roughened
fiber shown in Fig. 7 microgaph. The same cement was used in all
1FC mixtures. The intergrinding of the 3/4 inch (19 mm) long
polypropylene fibers and cement was controlled at 15 minutes in a lab
ball inill. A sieved IFC sample of 131 fibers indicated an average
fiber length of 1 /2 inch ( 1 3.4 inm) after grinding.
--``,`,-`-`,,`,,`,`,,`---
until tlie cylinder was completely filled. The rop was levekd aiid
troweled once. Wet inortar is placed in a separate container to check
the rate of hardening. The test specimens are placed in the
temperature controlled chamber with continuous air circulation. T h e
is noted and 6 hours added as tim-to-test. Setting characteristics are
monitored with a Vicat needle as shown in Fig. 3. The cement paste
inust set sufficiently for testing. In each test series tlie identical
iiiixing and handling procedure is repeated for tlie three separate
batches.
Direct 1,oadioE Procedure
--``,`,-`-`,,`,,`,`,,`---
steel pipe with annular riiigs weided at the center fiactiire seairi
provided a more rigid mold. The steel pipe inold is shown in Figres
9 and 10. Not eiiough later-age experiments using the steel molds
were coiidiicted to provide statistically reliable data for this report.
A summary of the five series of direct tensile tests results is
shown and expressed in terms of stren@h improvement percentages in
Table 3.1, which illustrates the substantial tensile strength increases
attributed to the Intergroiuid Fiber Cement process.
--``,`,-`-`,,`,,`,`,,`---
CONCLUSIONS
These 6-hour tensile test results relate to and help study early plastic
shrinkage cracking of slabs exposed to evaporative conditions.
Plastic shrinkage has already been established as acceptance criteria
for fiber reinforced concrete by one code ageiicy (7). Their criteria
requires tests nui side by side of plain and fiber reinforced slabs, and
comparing the square area of cracks, requiring a ininimuin 40%
reduction with FRC. Further study of the relationship of tensile strain
capacity at 6 hours with early cracking inay provide answers on how
fibers contribute quantifiably to the crack-resistance performance of
hydraulic cement composite inaterials. Future research will definitely
be needed in this area. Obviously more sophisticated improvements
will be made based on use and experience. This brief discussion will
lay the groundwork for stimulating renewed interest in new and more
effective tensile tests. Determining the necessary tensile strength to
ensure a crack-free concrete in the field was Paul Kraai's goal. With
his passing that goal lives on iii the form of a new FRC research tool,
the Kraai tensile test.
I
References
1. Kral, S. & Gebauer, J. "Shrinkage and Cracking of Concrete at Early Ages";
"Advances in Concrete Slab Technology" (Eds. E. K. Dihr and J. G. Munday),
Pergarnoti Press, Oxford & New York, 1980, pp. 412-420.
2. Neville, A. M. "Hardened Concrete: Physical and Mechanical Aspects", AC'I
Momgraph No. 6, ACI, Detroit, 19ï1, pp. 37-56.
3. Kraai, P. P. "Concrete Drying Shrinkage Facts and Fallacies", A C ï SZ'-76-3
"Designing for Crack & Shrinkage in Concrete Structures", AC1 Detroit,
1982, pp. 25-51, Ref 5, p. 36.
4. Mather, Bryant, Faxed Letter to Vondran, U.S. Department of the Army,
Vicksburg, April 20, 1994, pp. 1-3.
1 5. Vondran, G. L. "Interground Fiber Cement", United States Patent Number
5,298,071 issued March 29, 1994, pp. 1-6.
I 6. Vondran, G. L. "Interground Fiber Cement, A New Process", Proceedings
for the National Science Foundation-Universityof Sheffield Workshop on
"Fibre Reinforced Cement and Concrete" (Eds. R.N. Swamy & V.
Ramakrkhnan), July 28-30, 1994, pp 18-37.
7. ICBO AC32, "Acceptance Criteria for Concrete with Synthetic Fibers",
International Council of Building Officials, Whittier, CA, July 1993, pp. 1-13.
--``,`,-`-`,,`,,`,`,,`---
YC
O 1 2 3 4 6 6 7 8
TIME AFTER WATER ADDITION, hr
Fig. 1-Vertical and horizontal shrinkage tests by Kral and Gebauer (3) show high-volume
change in second phase where most cracking occurs. After 4 hn shrinkage percent and cracking
level off
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
Fig. 2-Direct tensile specimen molds
--``,`,-`-`,,`,,`,`,,`---
Fig. ó-Micrograph of cernent grains imbedded in fibers after grinding in lab ball mill. Note
plain fiber cylinder shape
--``,`,-`-`,,`,,`,`,,`---
Fig. 7-Electron micrographs showing fibers before and after intergrinding in cement plant
production run
4.3
2
1
U
PLAIN Mixed-In Fiber Interground
Fig. U o m p a r a t i v e results of mixed-in and interground fibers tensile strengths (average for 5
tests of 2 cylinders each)
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
SP 155-9
Svno~is: This paper describes the construction of two simple impact machines--
one small with a capacity of 100 Joules and the other large with a capacity of loo0
Joules--designed to conduct impact tests on fiber reinforced mortars and concretes
in the uni-axial tensile mode. During a test, the applied load, accelerations and
velocities are measured such that with a proper analysis scheme, the raw data can
be analyzed to obtain fundamental material properties under impact loading.
Carbon, steel and polypropylene micro-fiber reinforced mortars and steel fiber
reinforced concrete were tested and it was demonstrated that the proposed
technique is a simple and rational method of obtaining meaningful material
properties. In general, fiber reinforced composites were found to be more impact
--``,`,-`-`,,`,,`,`,,`---
resistant than their unreinforced counterparts and the improvements were
proportional to the fiber volume fraction. In addition, both the unreinforced matrix
as well as fiber reinforced composites were found to b e stress-rate sensitive, but the
extent of sensitivity observed was smaller than usually reported in the literature for
cement-based materials under uni-axial tensile loading.
Kouki Chokri was a graduate student at Laval University, Quebec, Canada where
this work was carried out. He investigated strain-rate effects in micro-fiber
reinforced cement composites.
--``,`,-`-`,,`,,`,`,,`---
done extensive investigation of static and impact behaviour of steel fiber reinforced
concrete.
INTRODUCIïON
Several technique have been developed in the past to test concrete under
impact. The most commonly used is the Drop Weight Test (2-5) in which a mass is
raised to a predetermined height, and then allowed to drop directly on a concrete
specimen. Similar in principle are the SwingingPendulum Machines (6-8)--as in the
conventional Charpy or Izod impact machines used by the metallurgists--where a
swinging pendulum is allowed to strike a specimen in its path thereby transferring
momentum and causing high stress-rates. Other significant impact tests include the
Split Hopkinson Pressure Bar Test (9,lO) in which the specimen is sandwiched
between two elastic bars and high stress-rates are generated by propagating a pulse
through one of the elastic bars using a drop weight or a similar device. In most
modern impact test systems, sufficient instrumentation is provided such that along
with loads and deformations, additional specimen responses such as accelerations,
velocities, etc., are also measured; these are needed for a proper analysis of the
data later. Using the various test techniques, plain as well as fiber reinforced
concrete have been shown to be stress-rate sensitive under all modes of loading
(11-15).
RESEARCH SIGNIFICANCE
EXPERIMENTAL
Two separate impact machines were designed. One with a smaller capacity
(u 100 Joules) was suitable for testing smaller specimens made of fiber reinforced
mortars, and the other with a larger capacity (u loo0 Joules) was designed for
testing fiber reinforced concrete specimens.
The Concept
The concept behind the two impact machines may be described with respect
to the schematic drawn for the 1000-Joule machine shown in Figure 1. A close-up
of the specimen gripping and loading arrangement is also shown in Figure 1 (inset).
In principie, the specimen bridges two supports, A and B, with support B mounted
o n rollers (called the trolley) and support A being fuced. Support B is struck by the
swinging pendulum on impact points located on either side of the specimen and in
the same plane as the specimen. Support A being f i e d , this causes tensile loading
in the specimen.
The impact hammer carries two dynamic load cells mounted on either side
that record the contact load vs. time pulse during the impact. Under an impact, the
specimen fractures and the trolley travels toward the shock absorbers. O n its way,
the trolley passes through two photocell assemblies where its post-event velocity is
--``,`,-`-`,,`,,`,`,,`---
The objective of the impact tests was to obtain strength and fracture energy
values for various fiber reinforced cementitious composites as a function o€ the
applied stress-rate. Considering the free body diagram of the troZZey (part B) itself,
the horizontal force equilibrium (ignoring damping) may be written as:
The hammer load Pt(t) in the above equation is the sum of the loads
recorded by the two load cells mounted on either side o€ the hammer and Ps(t) is
the specimen load. Pj(t) is the inertial load (2) given rise to by the accelerations
in the system. If at(t) is the trolley acceleration and mt and m, the masses of the
trolley and the specimen, respectively, Eqn. 1 can be written as:
a,(t) and 4 in the above Eqn. are the specimen stress and the cross-sectional area,
respectively. Note that it has been tacitly assumed that the loading is perfectly
aligned along the specimen centroidal axis without any eccentricity. This was
independently verified by the data from the two load cells as described later. With
trolley accelerations at(t) recorded by the accelerometer, Eqn. 2 can be solved for
a,(t), the peak value o€which could then be taken as the tensile strength of the
composite under impact.
The fracture energy consumed by the specimen was determined using the
principle of conservation o€energy. Using the impulse-momentum principle (6,7,8),
the energy lost by the pendulum (E,,) during its contact with the trolley can b e
written as:
where,
=i/2gh
g =earth’s gravitational acceleration
h =height of hammer drop
If one can ignore the frictional and other losses of energy, the energy lost by the
pendulum can b e regarded as the sum of the energies consumed by the specimen
during fracture (Es) and that gained by the trolley as post-fracture kinetic energy
(EJ. In other words,
Eh - Et +Es (4)
(5)
With all quantities on the right hand side known, the fracture energy consumed by
--``,`,-`-`,,`,,`,`,,`---
the true specimen loading. After the first peak, the accelerations died down and the
inertial forces could be ignored.
It was also found that because of hard steel-to-steel impact, the inertial
loads were very high and strong vibrations persisted during the test. Consequently,
in order to reduce the trolley accelerations and machine vibrations, the rate of
transfer of momentum between the pendulum and trolley was reduced by
introducing a rubber pad (stiffness = 2.83 MN/m) at the pendulum-trolley contact
points. The use of a rubber pad reduced the inertial load, delayed the occurrence
of both the peaks and reduced the rate of transfer of momentum (16). Indeed, the
rubber pad would absorb some energy from the hammer at first contact. However,
since the rubber pad is elastically unloaded at the end of the test, the energy
absorbed by the pad should not appear in the final equation of energy balance
(Eqn. 6).
RESULTS
--``,`,-`-`,,`,,`,`,,`---
ho, - -N+1
1
h
1
BU + - In (Oy-, - (7)
N+1
--``,`,-`-`,,`,,`,`,,`---
where, N is a material constant (slope of the stress intensity factor, KI, vs. crack
velocity, V, plot on a logarithmic scale), subscripts i and f refer to initial and final
conditions, respectively, and B is given by (13,lS):
2K;iN(N+1)
B -
A (N-2)
whereA and Y are constants. In Table 2, the values of constant N obtained on the
basis of Eqn. (7) are given. Notice that N varies between wide limits and is, in
general, higher than normally reported, particularly for the plain unreinforced
matrices. In other words, the sensitivity of concrete to stress-rate as observed in
these tests is less pronounced than that observed by others (15, 19, 20). This is
surprising given that concrete is expected to be far more sensitive t o stress-rate in
uni-axial tension than in any other mode. The following are the probable causes:
i) Estimated Stress Rate: The imposed stress-rates under impact were calculated
from the slopes of the contact load vs. time pulses in rising part of the "second
hump" (see Figures 3, 4, 5 and 6) in the load-time plots and then averaged. This
gives only an approximate stress-rate.
iil Eccentricity in Loading: The load eccentricity in rapidly applied impact loads
(where the specimen did not get much time to align itself in the direction of the
load) was greater than that in slow static tests. This may have led t o decreased
apparent strengths under impact loading and hence lower impact/static strength
ratios.
iii) Lack of a Linear ResDonse: As pointed out by Mindess (13), the assumption of
a linear elastic fracture response assumed in Eqn. (7) is not entirely valid. Concrete
is not ideally brittle and the o-Eresponse for both concrete and its fiber reinforced
composites is far from linear. The values of constant N therefore may not be
expected to capture the true nature of stress-rate sensitivity in these materials.
iv) Stress-Rate Vs. Strain Rate: Assuming the material is linearly elastic, the
imposed stress-rate ( u ) in a test can be related t o the imposed strain-rate (i)
through a simple Equation:
ir-E& (9)
where E is the elastic modulus. Eqn. 9 implies that for a given applied stress-rate
a stiffer materials would be subjected to lower strain-rates. Which means that if the
failure is governed by a limiting strain rather than a limiting stress value, the data
must be normalized and different materials must be compared only on an equal
strain-rate basis. This is, however, not attempted here given the lack of appropriate
--``,`,-`-`,,`,,`,`,,`---
Higher fracture energy values recorded for plain concrete with a higher
compressive strength (Figure 5 and Table 2) is, in general, against the prevalent
belief that high strength concretes are more brittle than normal strength concretes.
O n the other hand, the observed increases in the fracture energy absorption due
to fiber reinforcement (Figures 4 and 6; Table 2) are well anticipated and in tune
with most published data. When fracture energy values are normalized by dividing
them by the area of the specimen, the 1000-Joule machine appears t o give much
higher specific fracture energy values. This indicates that the normalized fracture
energy values are specimen size dependent and also dependent o n the machine
capacity. There is also doubt over this way of normalizing the fracture energy
values.
CONCLUSIONS
1. The paper proposes a simple technique of testing concrete and its fiber
reinforced composites under impact. It is demonstrated that with the
proposed technique, meaningful material properties under impact can
be obtained, which can then be useful in designing structures subjected
to impact loads.
--``,`,-`-`,,`,,`,`,,`---
are both stronger as well as tougher under impact loading as compared
to static loading.
ACKNOWLEDGMENTS
REFERENCES
10. Malvern, L.E., Tang, T., Jenkins, D.A and Gong, J.C., Dynamic
Compressive Strength of Cernentitious Materials, In Mindess, S. and Shah,
S.P. (eds.), Cement-Based Composites: Strain Rate Effects on Fracture,
Materials Research Society Symposia Proceedings, Vol. 64, Materials
Research Society, Pittsburgh, pp. 119-138.
12. Reinhardt, H.W., Strain Rate Effects on the Tensile Strength of Concrete
as Predicted by Thermodynamic and Fracture Mechanics Models. In
Mindess, S. and Shah, S.P. (eds.), Cement-Based Composites: Strain Rate
Effects on Fracture, Materials Research Society Symposia Proceedings, Vol.
64, Materials Research Society, Pittsburgh, 1986, pp. 1-14.
17. Banthia, N. and Trottier, J.-F., Impact Resistance of Concrete and Fiber
Reinforced Concrete under Uni-axial Tension, ACI, SP on Impact and
Impulsive Loading, submitted.
18. Nadeau, J.S., Bennet, R. and Fuller, E.R. Jr., An Ekplanation of the Rate-
of-Loading and Duration-of-Load Effects in Wood in Terms of Fracture
Mechanics, J. of Mat. Science, 17, 1982, pp. 2831-2840.
20. Zielinski, A., and Reinhardt, H.W., Stress-Strain Behavior of Concrete and
Mortar at High Rates of Tensile Loading, Cement and Concrete Research
12, 1982, pp. 309-319.
Il
Y
c>
O 0
52
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 75 0 6 6 2 9 4 9 0 5 2 3 2 b î 305
e
a
U
E
--``,`,-`-`,,`,,`,`,,`---
Fig. 2-Load-time plots with and without specimen for small 100-Joule impact machine
ao I I I 1 I I I I I
70
60
50
5
n- 40
Id
O
30
20
10
n
O 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Time, ms
Copyright American Concrete Institute Fig. 3-Load-time plots with and without specimen for large 1000-Joule impact machine
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 9 5 I0662949 0 5 2 3 2 7 3 T 6 3
Time, ms
Fig. 40-Impact load-time plots for mortars reinforced with 1, 2, and 3 percent by volume of
carbon fiber
Fig. 4b-Impact load-time plots for mortars reinforced with 1, 2, and 3 percent by volume of
steel fiber
--``,`,-`-`,,`,,`,`,,`---
Time, ms
Fig. 4-Impact load-time plots for mortars reinforced with 1, 2, and 3 percent by volume of
polypropylene fiber
1 O0 I I l 4 I I I
I I
O 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Time, ms
Fig. %Impact load-time plots for concretes with various strengths (high strength: 85 MPa;
midstrength: 52 MPa; normal strength: 40 MPa)
--``,`,-`-`,,`,,`,`,,`---
1 O0 I
90 -
ao -
70 -
-
5 60 _ - - -Plain
U- 50 - Plain + Fiber FI
m
9 40-
30 -
20 .
10 -
O 1 .o 2.0 3.0 4.0 5.0 6.0
O
Time, ms
Fig. 60-Impact load-time plots for steel fiber reinforced concrete with 0.51 percent by volume of
hooked-end fiber F1 (the plain matrix had static compressive strength of 52 MPa, which increased
to 56 MPa with fiber addition)
100 r I I I
Time, ms
Fig. ób-Impact load-time plots for steel fiber reinforced concrete with 051 by volume of
crimped fiber F2 with circular section (the plain matrix had static compressive strength of 52 MPa,
which remained unchanged with fiber addition)
--``,`,-`-`,,`,,`,`,,`---
1O0
90 -
80 -
70 -
- - - - - Piain
5 60
Plain + Fiber F3
x- 50 -
40-
30 -
20
--``,`,-`-`,,`,,`,`,,`---
10 -
O
O 1 .o 2.0 3.0 4.0 5.0 6.0
Time, rns
Fig. óc-Impact load-time plots for steel fiber reinforced concrete with 0.51 percent by volume of
crimped fiber F3 with crescent section (the plain matrix had static compressive strength of 52
MPa, increased to 53 MPa with fiber addition)
100 I I I I 1
---- Plain
Plain + Fiber F4
u-
m
O
-I
I 1 4
Fig. ód-Impact load-time plots for steel fiber reinforced concrete with 0.51 percent by volume of
twin-cone fiber F4 (the plain matrix had static compressive strength of 52 MPa, which increased
to 53 MPa with fiber addition)
SP 155-10
Kewords: Fibers; impact tests; loading rate; pullout tests; slippage; steels
--``,`,-`-`,,`,,`,`,,`---
Gold Medal Award. He has been awarded the Alexander von Humboldt
Senior Visiting Scientist Award t o Germany as well as the NATO
Visiting Senior Scientist t o France.
INTRODUCTION
inclination, a fiber can pass plate 1 in hole A, then meet plate 2 in hole
B'. In order t o keep a symmetric geometry, the second row of fibers
were inclined with the same angle but in the opposite direction. During
testing this guide plate was used t o separate the pullout section of the
specimen from the anchored end and provide for the transfer of the
load. Both ends of fibers are embedded in the matrix. However, part of
the fiber within the long embedment side was periodically deformed t o
assure that the fiber was pulled-out from the end with the short
embedment length.
extensometers, with a range of 0.075 in. (1.9 mm) and a gage length
of 0.5 in. (12.7 mm) were mounted on both sides of the specimen t o
measure slip displacement. The average reading from the t w o
extensometers was used as the feedback signal t o control loading rate.
Once the debonding process was completed the rest of the fiber was
pulled out from the matrix a t a constant rate under stroke control with
the same speed as that observed during the first part of the test. A
scheme of this configuration can be found in Fig. 3.
CHARPY MODIFICATION
and C in Fig. 4.) and the accelerometer (sensor 4).The amplifiers had
a sampling rate of 500 ns per point. A Tektronix 2232 oscilloscope
was used t o store the slip signals measured by the t w o MULTI-VIT’S
(sensors R and 13.Load cell A was used to trigger the whole system
as shown in Fig. 4.
where FA and FE are the force measured from load cells A and B,
respectively.
where F, represents the pull-out load measured using load cell C.It is
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 0 6 6 2 9 4 9 0523280 T7b W
EXPERIMENTAL PROGRAM
Initially, the pull-out load increases with the slip almost linearly.
Nonlinearity in the pull-out load and slip curves, which is often
regarded as an indication of propagation of an interfacial crack
(debonding), is usually observed before the peak load. After the pull-
out load reaches the maximum value it decreases slowly with the
increasing slip. Softening type load-slip curves are observed from the
pull-out test. One can observe similar slopes prior t o the peak load for
both series of specimens. However, both the peak pull-out load and the
pull-out load after the peak increase with fiber inclination. The effect
of fiber inclination on the peak pull-out load was theoretically predicted
by Ouyang e t al. [ I 31 using energy considerations.
Since the slip value prior to peak load is relatively small, the
effect of frictional stress on pull-out of inclined metal fibers may not be
dominant before the peak load. The influence of frictional stress
becomes ever more important after the peak' load because more
slipping occurs. This is particularly evident through the pull-out load
and slip curves of series B specimens shown in Fig. 8.
Spalling of the matrix at exit point of the fiber was not observed.
However, some radial matrix cracking at the exit of the fiber from the
matrix were detected.
Fig. 11 shows a plot of the average peak pull-out force per fiber
versus slip rate. Values from aligned and 37 degrees inclined
specimens were selected. Scatter for series B specimens (8 fibers) is
larger than for series A, but still does not mask the trend. 1141.
Fig. 12 shows the effect of the rate of slip on the ratio s/ipd/!/ip"
where slipdis the slip a t the point of p , for the dynamic rates and slip"
is the slip at the point of p,, for the reference rate of 0.000005 in./s.
Results from specimens with aligned and 27 degrees inclined fibers are
represented.
Effect of Inclination
--``,`,-`-`,,`,,`,`,,`---
Although the effect of number of fibers on the debonding force
is not fully understood, a certain trend can be observed from the
experimental results.
It should be noted that the unit values of p,, are higher for
series B specimens with aligned fibers. The rest of the inclinations
present higher values for series A specimens, showing a different
effect for impact rates than quasi-static rates.
CONCLUSIONS
ACKNOWLEDGEMENTS
--``,`,-`-`,,`,,`,`,,`---
REFERENCES
(8) Brand, A. M., "On the optimal direction of short metal fibers
in brittle matrix composites," Journal of Materials Science,Vol. 20,
1985, 3835-3841, 1985.
(13) Ouyang, C., and Pacios, A., and Shah, S . P., "Pull-Out of
Inclined Fibers from Cement-Based Matrices," accepted for publication,
Journal of Engineering Mechanics, ASCE, 1994.
(14) Pacios, A., Ouyang, C., and Shah, S'. P., "Rate Effect on
Interfacial Response between Fibers and Matrix," accepted for
publication in the Materials and Structures, 1994.
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
t
2a
-
3
(u
m
-
t
>
hl
a
-
t
>
r
m
-
3,
a
-
t
>
O
m
--``,`,-`-`,,`,,`,`,,`---
Debonding
and slipping
Yielding
of fiber
...
51
--``,`,-`-`,,`,,`,`,,`---
f
I
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
1400
1200
1O00
Ki
o 600
400
200
O
O 3 5 8 10 13 15 18
Time (ms)
400
300
200
1 O0
O
O 5 10 15 20 25 30 35 40
Time (ms)
600
500
400
300
200
IO0
O
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0
Time (ms)
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
800
700
A
600
0Y
500
400
300
200
1O0
O
O 0.01 0.02 0.03 0.04 0.05 0.06
Slip (in)
--``,`,-`-`,,`,,`,`,,`---
275 t
250 $
225 &
U 200 m
-2 40 175
O
7
ti
? 150 43
7 30
125 y
Q
--``,`,-`-`,,`,,`,`,,`---
Y
1 20 100
75 5
.-E
.i IO 50 .E
i7
H 25 z
O O
1E-6 1E-5 1E-4 1E-3 1E-2 1E-I IE+O 1E+1 1E+2
Slip rate (ink)
E. 200 -
t
175 2
ic
150
TI
c
a
125 0"
c
-
-P 100
3
2
B 20 B
75 $j
a#
..-E 10 50
Q
E
X
25
5 5
O o-
1E-6 IE-5 1E-4 1E-3 1E-2 IE-I IE+O IE+l 1E+2
Slip rate (inls)
Fig. 11-Effect of loading rate on maximum peak pull-out load per fiber
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-1.55 9 5 I0662949 0523303 5 2 0
4.0 1 111111
- 1~1111111
16 fibers
I 1111111 I 111111 I 1,11111 I 1111111 I
3.5 - t
- + Odegrees
3.0
1 27degrees' ' . I
2.0
I.o
0.0
1E-4 1E-3 IE-2 1E-I 1 E+O IE + l 1E+2
Slip rate (ink)
--``,`,-`-`,,`,,`,`,,`---
90.0 400
h
80.0 350 z
à5
LI
70.0
300
60.0 Q
250 :
-
O
50.O
40.0
30.0
20.0
10.0
0.0 O
O 5 10 15 20 25 30 35 40
8 (grades)
70.0
300
h
60.0 tL
250 4
IC
50.O L
aJ
Q
200 u
m
40.0 -
O
30.0
Y
100
20.0 Q
10.0 50
6
.-E
X
2
0.0 O
O 5 IO 15 20 25 30 35 40
--``,`,-`-`,,`,,`,`,,`---
8 (grades)
SP 155-11
--``,`,-`-`,,`,,`,`,,`---
SvnoDsis: Recent natural disasters involving high wind
events have demonstrated the fact that building
envelopes, including structural walls and roofs, can lose
structural integrity as a result of penetration by
missile objects. As a result, there is heightened
interest in the testing of components and cladding that
are used as a part of building envelopes of habitable
structures.
A large missile impact test has been designed and is
being evaluated in laboratories around the country. The
test, discussed in this paper, is suitable for laboratory
or field applications and is currently undergoing
scrutiny by the ASTM Task Force of Committee E6,
Performance of Buildings. Adoption of the test by the
South Florida Building Code came in the wake of Hurricane
Andrew in 1992.
The test has been applied to numerous types of wall
systems and building products including a fiber
reinforced cellular concrete panel which is designed to
be used as an alternate to masonry infill construction,
architectural precast, demising walls and security
fencing. Additional tests of the missile impact
resistance of fiber reinforced cellular concrete
involving the use of large caliber ballistics are also
discussed. The high energy impact resistance of fiber
reinforced systems is demonstrated and discussed.
INTRODUCTION
--``,`,-`-`,,`,,`,`,,`---
impact velocity conventional concrete block construction
could not resist penetration of the missile prompting the
eventual reduction of the missile impact velocity
specification. The photograph of Figure 7 depicts a
concrete block wall which was subjected to missile impact
at the 80 f p s (24.38 m/s) speed.
FRCC wall panels were tested prior to the reduction of
the 80 fps (24.38 m/s) missile impact speed. The panels
exhibited little damage and complete resistance to
penetration. Full scale panel prototypes were 7-112
inches (19.05 cm) thick, 8 feet (2.44 m) in length, 4
feet (1.22 m) in width and were simply supported over the
long span against a rigid wall. The photograph of Figure
8 shows the indentation made by the standard missile with
area dimension at approximately the same size as that of
the dimensional lumber which is the missile. Two missile
impacts within approximately 12 inches (30.48 cm) of each
other are shown. Penetrations extended to a depth of
approximately 314 inches (1.91 cm) A closer view of one .
of the indentations is depicted in the photograph of
Figure 9.
--``,`,-`-`,,`,,`,`,,`---
but one which has gained interest and acceptance based on
relative simplicity and economics as well as its
suitability to the task.
Fiber reinforced concrete has long been known for its
impact and cyclic fatigue loading toughness
characteristics. The large missile impact test well
documents and is aptly applied to the testing of FRC
building wall systems. The resistance to bullet
penetration by FRCC wall panels has also been
demonstrated.
REFERENCES
--``,`,-`-`,,`,,`,`,,`---
2 x 4 DIMENSIONAL LUMBER
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
Fig. 8-Impact test of FRCC panel
--``,`,-`-`,,`,,`,`,,`---
230 Hays and Zoll0
--``,`,-`-`,,`,,`,`,,`---
SP 155-12
SvnoDsis: One of the possible factors inhibiting the wider application of fiber
concretes is the fact that to characterize the engineeringproperties of fiber concrete,
both cubedcylindm and prism have to be cast and tested, in addition to the
determination of workability. This paper shows that with the use of superplasticizer,
the slump test can be used to give guidance on the flowability characteristics of the
fresh fiber concrete. The advantage of the slump test is that it is easy to carry it out
in the field, apart fim its simplicity and convenience. It is further shown that the
--``,`,-`-`,,`,,`,`,,`---
equivalent cube strength obtained from the broken pieces of a flexural test can
adequateiyrepresent the compressivestrength of fiber concretes. It is thus shown that
it is possible to characterize the engineering properties of fiber concretes from oniy
onesetofprisms, say, of lOOxlOOx5OOmmsize. Aparîfrombt crackload, moduius
of rupture and fracture toughness properties, the prisms can be used to give additional
information as appropriate such as shrinkage and expansion, and through puise
velocity, internai microcracking. It is suggested that by rationalizing our approach to
testing, it is possible to reduce not oniy the cost and inconvenience associated with
Werent sizes of test specimens but also to enhance the relevance of some of the
information obtained from such testing.
234 Spadea et al
Prof. Giuseppe Spadea received his degree at the University of Pira, Italy.
Presently he is Associate Professor of Civii Engine- wiîh inîerest in design and
analysis of structures at the Department of Structural Engine- University of
Calabria, Italy. His research interests are experimental characterization of structural
materials and modehg of the behavior of reinforced concrete and FRP structures.
Rosaiba Cava is a PhD candidate at the University of Calabria, Italy, where she
earned her degree in structural engineering. Her research interest is in the area of
fiber reinforced concrete.
INTRODUCTION
There are also other quations related to the relevance of some of the recammended
test methodologies to characterize fiber concrete. Two examples are quoted here.
One relates to the so calied workability of fiber concrete. Whilst it is universaiiy
recognized that the conventionalsiump test gives a useful measure of the consistency
of concrete (without fiber), which forms oniy one aspect of the complex property of
workabiiity, there is considerable evidence to show that the m-ed versions of
tests for Cesh ñber concreteworkability &o do not give a clear idea of their flowability
properties; indeed they show h@iy variable resuiîs and exhibit no consistentrelations
between the various measured values (7-12). Indeed, the most imporîant Equirement
of fiber reinforced concrete, namely, the uniform distribution of fibers w i t h the
body of concrete, is not met or measured by any of the current test methods on fresh
fiber concrete. The second relates to toughness indices. Here again there is widespread
scepticism on the relevance and usefulness of the various toughness indices for
practical applications, although this aspect is not discussed in this paper.
The aim of this paper is to tackle some of the problems related to testing for FRC
discussed above. In particular, the following aspects are considered.
steel and polyacryloniûiie (non-metallic fibers) when used in concrete without and
with SF?
236 Spadea et al
4. Can prisms (for example, 10Ox1OOx500 m m size) be used as the sole test
specimens cast to obtain the engineering properties of FRC? To illustrate this,
shrinkage, expansion, pulse velocity and i6rst crack load in flexure are reported.
Since flexural test specimens are normally used in any case to obtain moduius of
rupture and fracture toughness parameters, these characteristics are not reported
here.
This paper thus shows that apart from the test for the flowability properties of
FRC,10Ox10Ox500mm prisms alone need to be cast to characterize the engineering
properties of FRC for practical appiications.
RESEARCH SIGNIFICANCE
The aim of this paper is to show that it is possible to characterize the engineering
properties of fiber concretes by using oniy one set of test specimens i.e. prism, say,
of 1OOx1OOx500 mm size. Further, the slump test can be used to give practical
information on the flowability properties of fiber concrete, particularly when
superpiasticizers am incorporated, although one recognizes that by the very nature
of the slump test, it cannot give any directiy measurable quality of workabiliíy.
EXPERIMENTAL PROGRAM
--``,`,-`-`,,`,,`,`,,`---
Materiais
Rapid hardening portland cement was used for all mixtures. The minimumvahies
of flexural and compressive strength required by Italian Standards for this type of
cement are respectively 6 MPaand 32.5 MPa at 7 days and 7 MPaand 42.5 MPa at
28 days. Saca fume (SF) was used as partiai repiacement of the portland cement on
an equal-weight basis.
The fine aggregate used was river sand with a maximum size of 1.5 mm and a
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
A C 1 SP-155 95 0662949 0523323 191
water absorption ot 1.58 per cent. i he coarse aggregate was washed and dried
crushed gravel mostly of rounded shape, with a maximum size of 10 mm and a
water absorption of 0.54 per cent.
Hooked-end steel fibers glued together in bundles were used. These fibers
had a nominal length of 30 mm and a diameter of O. 5 mm (aspect ratio 60), with
a specific gravity of 7.1 g/cm3and an elastic modulus of 210 GPa. The
polyacrylonitrile fibers used in these tests were a high modulus synthetic fiber
with a kidney-shaped cross-section. These fibers are bound together in sheaves
by a special water-soluble sizing agent which then softens during mixing releasing
the singie fibers. Polyacrylonitrilefibers were 24 mm long with a nominal diameter
of 16 pm, a specific gravity of 1.18 g/cm3and an elastic modulus of 13.5 GPa.
Mixtures
--``,`,-`-`,,`,,`,`,,`---
For the second group the concrete matrix was modified with 5.0 per cent of
the cement replaced by s i k a fume. The two types of fibers, steel and
polyacrylonitrile, were used with a volume percentage of 1.0 and 0.25,
respectively.
From each mixture six 100x100x500 mm prisms and at least three 150
m m cubes were cast in steel and polystyrene molds respectively.
238 Spadea et al
capacity of 0.35 m3. The mixing procedure was as follows: dry coarse and fine
aggregate were homogenized together and mixed with their absorbed water to
reach a saturated surface-dry condition; after that cement, silica fume if present,
and miXing water were added. Steel fibers, superplasticizer and polyacrylonitrile
fibers were haüy introduced in the mixture. The fibers were gradually scattered
by hand in the matrix. in ail fiber concrete mixtures, care was taken to ensure
that the fibers suffered as iittle damage as possible consistent with thorough
mixing and uniform distribution of fibers. Based on the experience acquired
during the fabrication of trial mixtures, the polyacrylonitrile fibers were added
right at the end of the mixing process when the matrix had been thoroughly
mixed into a homogeneous and flowable mixture because otherwise balling of
fibers could be observed. The specimens were cast in the molds and compacted
on a vibrating table. For each mixture of the first group two batches were
successively prepared.
Specimens were cured by covering them with a plastic sheet while they were
sídi inside the molds. The specimens were demoulded after 24 hours and then
exposed to two curing regimes:
-
Continuous water curing in uncontrolled internal environment.
-
Seven days water curing followed by exposure to uncontrolled internal
environment.
During the curing regimes the temperature and relative humidity variations
were not recorded.
Testine Details
Flowability - whilst fully recognizing that the conventional slump test has
little direct scientific significance to the property of workability, with a
superpiasticizerthe test can give useful practical information to the mixdesigner.
This test was therefore adopted in this study according to B.S. 1881:Parî 2.
-
Compressive S t r e d Compressive strength was determined both from
cubes and the equivalent cubes of the broken parts of the flexural test prisms.
Three cubes and six equivalent cubes were tested for each mixture.
two discs were as far as possible in the same plane. The first readings were taken
two hours after the fWng of the discs. Regular measurements were then
-
subsequentlytaken for the duration of the tests every day up to 28 days, then
twice a week for 8 weeks and then once every week up to about 200 days. The
shrinkagelexpansion value of each specimen was an average of the readings on
the four faces. The development of shrinkage and expansion was calculated
treating the fist reading as the initial reading (zero reading). For each mixture
six specimens were tested.
Flexural Test -- AAer the long term exposure, the specimenswere tested under four-
pointbending overaspanof4ûûmm Dunngthetesiaconstantcross-hadrateofO.05
d m i n was kept. Deflections were measured at the midspan section by means of two
LWTs witha m;urimUmtrawl of 20 mm. The specimensm d in water were prepared
for bending test mthe fobwingway: after takangthem out of waterthey were died with
a paper and wex left mthe laboratmy for about three hours before testing. The resulis of
this patt of the study are not reported here except for the fmícrack load
Equiwient Cube Strennth -This property was obtainedby canying out an equivalent
cube test on the broken p& of the prism specimens afìer they had been tested m
bending.This was done by locating square steel plates, equident m section to the cross-
sectionoffheprism,beîweenMe~andtestingnaachinebtransnatthecompressniP:
load@& 1).
The resutoS of the tesis on fie& concrete flowabiiitypr@ are reportedin Table
1. It shouid be mentioned here that duringthe siump tests, the behavior of themixes was
carefülly monitored in terms of ability to flow,segregation of constituents, ability
240 Spadea et al
to remain cohesive and homogeneous, and ability of the fibers to separate and
remain embedded in the cement matrix. The discussion below is based on these
close observations. The results in Table 1on slump measuremenis have many practical
Unpiications. They show that whilst a good flow properties can be obtained with a
relatively high volume of steel fibers, the quantity of superplasticizer needs to be
increased as the fiber volume is increased. From the close test observation, a slump
of about 100 mm is considered to be acceptable to ensure good flowabiiity in the
field to ensure proper compaction.
A non-metallic fiber like the polyacrylonitde fiber used here required tugher
amounts of superplasticizer even for low fiber volumes: even then, the slump values
--``,`,-`-`,,`,,`,`,,`---
obtained were less than those of comparable steel fiber concrete mixtures. Further,
even with high superplasticizer contents, oniy slumps of 30-50mm were obtained
with the poiyacrylonitriie fiber. It was considered that about 3 per cent by weght of
cementitious content is about the maximum acceptable superplasticizer content both
ftom cost point of Mew, and the avoidance of secondary effects such as increased
setting times and segregation.
The use of silica h e , without and with fibers, also aáverseiy affected the
workability properties of the mixtures. The presence of silica fume produces an
increase in the fines content, and hence an increase in water demand. Even with
increased amounts of superplasticizer, slumps of oniy about 80 mm and 25 mm
were obtainable for 1.0 per cent and 0.25 per cent steel and poiyacrylonitriiefibers
respeckly. The silica fume mixtures, on the other hand,were found to be cohesive,
and indicated that slump values do not always provide an accurate indication of the
flow properties of concrete mixture, whether it contained fibers or not.
The results in Table 1 emphasize the need of technical compromises in the use
and appiication of flowable and workable fiber concrete in practice - between cost-
effectiveness, workability required and/or obtainable, acceptable amounts of
superplasticizer without side effects on other properties such as setting times and
segregation, types and amounts of fibers, amount of silica fume and the engineering
properties required for a specific use. There is the further point that these factors
need to be considered even at the mix proporîioning and fabIication stage, as these
would d u e n c e the subsequent períormance of the material in situ (13).
One of the aims of this stuây was to find out if the conventionai slump test can be
accepted as an adequateindicator of flowabiiityof fiber concretemixes when different
types and amounts of fibers and silica fume are used as additionai constituents. The
resuits show that provided a reasonable dump is designed for (say about 50 to 120
mm), and a superpiasticizer is incorporated, the slump test could be a good field
gui& ofworkabilitypropertiesfmpracíicai appiications. The results ofthe dump tests
obtained in this study confirm the conclusions arrived by Popovics on his analysis of
slump test data reported recentiy (14). The authom of this paper umcur with these
tindings and suggest that the slump test, although not scientifically based, can give a
lot of useful information to the practical concrete technologist, and that it is a simple,
convenient and cost-effective method of quality control of fiber concrete in the
field.
Silica fume is nowadays widely used for enhancement of both strength and
durability. However, when fibers and SF are both incorporated, serious problems
of lack of flow properties could arise without a superplasticizer. Even when a
SP is used, the amounts required could be quite high: in the tests reported here
2 to 3 per cent by weight of cementitious content was required to obtain slumps
of 75 to 85 mm for steel fibers and 20 to 30 mm for polyacrylonitrile fibers for
moderate volumes of fibers. The need to proportion fiber concrete mixtures
bearing these factors in mind cannot be overemphasized, and perhaps these
problems highlight why fiber concrete mixtures are still not as widely accepted
--``,`,-`-`,,`,,`,`,,`---
ComDressive Strenpth
Table 2 shows the 28 day cube compressive strength (Table 2a) and the
equivalent cube strength results obtained from testing the broken parts after
flexural strength testing (Table 2b). The latter were obtained at the ages of 170
and 220 days as indicated in the Table. One of the aims of this study was to fmd
out if the equivalent cube test can replace conventional cube or cylinder tests
for compressive strength. The greater emphasis of the discussion below is
therefore on this aspect.
The results given in Table 2 show that the addition of steel or polyacrylonitrile
fibers or the incorporation of 5 per cent of SF as a cement replacementmaterial
has no significant effect on compressive strength. This result is as it should be
considering the amounts and roles of fibers and SF used in these tests. The
numericalvalues in Table 2 are ail within acceptable experimental errors associated
with the variability of concrete mixtures.
More importantly, the equivalent cube strengths tested at later ages, are
representative of the 28 day cube strengths, and further, reflect the effects of
curing and ageing on the 28 day result. Both the cube strengths and the equivalent
cube strengths are ail of the same order: the data, although limited, give confidence
that equivalent cube strengths can be relied upon to give realistic values of
compressive strength. There are other published data on piain concrete which
conñrm that the equivalent cube test is a good test for compressivestrength: the
test is also included in the British Standards. The study presented here shows
that this test can be confidently extended to fiber concretes without and with
silica fume.
Shrinkage (including both plastic and drying shnnkage) and expansion data are
often not quoted for fiber concrete, but in practice, the effects of plastic shnnkage
cracking and of the incorporation of fibers in creating internal voids, for example,
are important facton to be considered by the engineer. This is ilhistrated by some of the
re& obtainedinbiisstudy.
The advantage of the puise wlocity measurementsis that they can give some idea of
the development of the intemal structm~,Particuiariy with time. Fig. 5 and 6 show that
air drying does create m t e d micro crac^ and botfi types of fiber concretes show a
reduction in puise velocity with time, the greater reduction occurring with the
poiyacryimiûiie ñbers.
The results ofthe dika fume umcrete in Fig. 7,on the ohm hand, confirm lhe positive
roie of îhk mineraiadtrMrhirem umcrete. The differences between the mixturesexposed
towetanddrymgcniditionsaremuchlesscanparedtoportlandcementconcreteniixtirres
shownin Fig. 5 and 6. indeed ifone considered oniy compre* skngth and the
effects of wetting or drynig, the daîa presented here show that the mcorporatonof 5 O h
silica b e as a cement repiacement materiaí can be just as effecthe as the addiibn of
steel or p o i y a u y l d e fibers.
--``,`,-`-`,,`,,`,`,,`---
Table 3 shows the fnrst crack loads determined h m the load-transit time c m and
from the load-displacement curves. The puise velocity technique gives better and
more reliable first crack load values compared to those obtained fimload-deflection
curves.
The use of the puise velocity measurements in concrete has thus two majorbenefits
- firsts, it can give some idea of the development of internal microcraclung with the
progress of dryrng, secondiy, it can give more reliable data of first crack load values
compared to those obtained from load-deflection curves.
CONCLUSIONS
1. Because of its simplicity and convenience, the slump test is easy to carny out in
the field. It is shown that provided a reasonable slump is designed for (about 50 to
120 mm), and the mix incorporates a superplasticizer,the slump test can be a good
guide of flow properties for field applications of concretes containing steel and non-
metallic fibers without or witb silica fume.
2. Sincefibers and silica fume both affect adverselythe flow properties of concrete,
the need for carehi mix proportioning of concretes containing these constituents to
produce concretes of adequate workability can be a key to wider application of fiber
concretes.
3. The equivalent cube strength test carried out on the broken pieces &er a
flexural test can represent the cube compressive strength of fiber concretes.
244 Spadea et al
Apart Bom the slump test tor Iresh concrete, the oniy other specimens that
i.
need to be cast to characterize the engineering properties of fiber concrete are
lOOx100x500 mm prisms. 'Tests on these specimens can give inîormation on
shrinkage, expansion, puise velocity, first crack load, fracture toughness, modulus
of rupture and equivalent cube strength.
REFERENCES
1. RILEM, Testing and Test Metho& ofFiBer Cement Comparites, RILKM hi.
Symp. Sheffield, April 1978, Editor R. N. Swamy,The Conshniction Press, pp. 545.
6. Fibre Reinforced Cement and Concrete, Proc. of the Fourth RILEM Int.
Sympasium,Edited by R. N. Swamy, E & FN Spon, London, 1992, pp.1347.
7. El-Refsu, F. E., and Morsy, E. I€, "Some Properiies of Fibre Reinford Concrete
Wiai SuperpIastìcim,''mZC86,DevetOpments in FibreRei$omd Cement andconcret,
Ed. R N. Swamy, R L WagstaíTe and D. R Oakiey, V. 1, Paper 3.2, 1986.
8. Rossi, P., Jhrrouche, N., and de Lanard, E, "Method for Opthking the
Compition of M e i a b F h R e i r r f d Conc~&~,"
Fibre Reidowed Cementr and
Concretes :Recent Developments, Edited by R. N. Swamy and B. Barr,1989, pp.
1-10.
Y. uyan, M.. Yiidirim, H., and bryaman, A. H., "Workability and Durability of
Steel Fiber Reinforced Concrete Cast with Nomial Plasticizers," Fibre Raznforced
Cement andc'oncrete, Edited by R. N. swam^, lW2, pp. 70-81.
10. Swamy, R. N., and Jojagha, A. H., "Workability of Steel Fibre Reinforced
Lightweight Aggregate Concrete," Int. Journal of Cement Composites and
Lightweight Concrete, V. 4, No. 2, May 1982, pp. 103-109.
12. Swamy, R. N., and Stravides, H., "Some Properties of High Workability
Steel Fibre Concrete," Proc. RILEM Symp. 1975, Constr.Press Ltd.
-
14. Popovics, S., "The Slump Test Is Useless or Is it?," Concrete International,
V. 16, No. 9, September 1994, pp. 30-33.
15. Swamy, R. N., "Steel Fibre Concrete for Bridge Deck and Builduig Floor
Applications," Steel Fibre Concrete, Swedish Cemenl and Concrete Research
Institute, 1985, pp. 443-478.
--``,`,-`-`,,`,,`,`,,`---
246 Spadea et al
3 ".
3
Y
O
vi
9
3 w
m
I æ
c
L3 G
i-
æ
Y
Di E
m
I-
V>
c
V>
Y
r Y
m
k2
Y
æ
U
m
L I- Y 09
Y d
3
Y z
J d
m a
æ
U Y
I I
n 3
N
O
--I
YI
N
2
Y 3
Y
2
d
m m
s 3
3 2
d
--``,`,-`-`,,`,,`,`,,`---
E
E
O
52
--``,`,-`-`,,`,,`,`,,`---
--``,`,-`-`,,`,,`,`,,`---
248 Spadea et al
y -1000 T
2 I
m -800
....-.DR1 -air
d; -(jo0
-400
PC -water
-200
1O00
' g o I
.-8
' t
400
ii
a
600 i
y -1000 -
d -.-_PC -air
. _ . _PAN1
_ . - air
2
-800 -PAN2 - air
PC -water
3 -600 __ ........... PANI - water
3
;n -400
-200
1O00
b
W
O I
8
s
.e
400
600 i
Fig. L S h r i n k a g e and expansion of mixes without SF (polyacrylonitrile fibers)
2
PCSF -air
-800 --
E/
rn
Y
- - - - - . D N S F -air
'a
b, 200 --
Age, days
.-8 --
400
ila
38 4.3
- _ - .PC -air
3'gl
PC -water
- - - - - . D R 1-air ........... DR1 -water
DR2 - air DR2 - water
i 3.5 I I I , I
O 50 1O0 150 200
Age, days
Fig. %Relationship between pulse velocity and age for mixes without SF (steel fibers)
250 Spadea et al
4.7
o
jj8 4.3
!
I
2 3.9 I
1 PC
..-...PAN1 -air
-air PC -water
3 I ........... PAN1 - water
PAN2 - air PAN2 - water
3.5
O 50 1O0 150 200
Age, daYs
Fig. &Relationship between pulse velocity and age for mixes without SF (polyacrylonitrile
fibers)
4.7
T
-.-__
O ...............................
8 7:.7:-7 : z y. --y-.'
jj 4.3
2 3.9 _.
- _ -PCSF
_ -air PCSF -water
& DR2SF -air DR2SF -water
...... PAN2SF - air
~
...........
PAN2SF - water
3.5
O 50 100 150 200
Age, daYs
Fig. 7-Relationship between pulse velocity and age for mixes with SF
--``,`,-`-`,,`,,`,`,,`---
Length
Area
Volume (capacity)
Force
Mass
Temperatureg
' 'íük selected list gives practical conversion faciors ofunits found in wncrete technoIoa. The reference sources for information
on SI units and more exact conversion factors are ASIU E 380 and E 621. Symbois of metric units are given in parenthesis.
t E Indicates that the factor given is exact.
t One liter (cubic decimeter) equals 0.001 m' or 1000 cm'.
i These equations wmert one temperature reading to another and inciude the necessary scale wrrections. To wnvert a
difference in temperature from Fahrenheit degrees to Celsius degrees, divide by 1.8 only,ie., a change from 70 to 88 F represents
a change of 18 F or 18/1.8 = 10 C deg.
--``,`,-`-`,,`,,`,`,,`---
Index 253
INDEX
254 Index