Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279610970

Test methods for flexural toughness characterization of fiber reinforced


concrete: Some concerns and proposition

Article  in  Aci Materials Journal · January 1995

CITATIONS READS

93 1,794

2 authors, including:

Nemkumar Banthia
University of British Columbia - Vancouver
317 PUBLICATIONS   5,049 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Shrinkage of concrete View project

High performance and sustainable composites based on alkali-activated materials (AAM) View project

All content following this page was uploaded by Nemkumar Banthia on 29 February 2016.

The user has requested enhancement of the downloaded file.


ACI MATERIALS JOURNAL TECHNICAL PAPER
Title no. 92-M06

Test Methods for Flexural Toughness Characterization


of Fiber Reinforced Concrete:
Some Concerns and a Proposition

by Nemkumar Banthia and Jean-François Trottier

The major advantage of fiber reinforced concrete over its unreinforced ASTM C 1018, STANDARD TEST METHOD OF FRC
counterpart is in the improved energy-absorption capability, or toughness. TOUGHNESS CHARACTERIZATION
The current methods of characterizing the toughness of fiber reinforced The ASTM C 10184 standard method is based on deter-
concrete, however, have proven to be largely inadequate and have caused a
great deal of dissent and confusion. This paper discusses some of the major
mining the amount of energy required first to deflect and
difficulties with these standard methods and demonstrates their susceptibil- crack an FRC beam loaded at its third points and then to
ity to human judgment errors. The paper also proposes an alternate tech- selected multiples of the first-crack deflection (Fig. 1).
nique that addresses some of these concerns and is capable of Toughness indexes I5, I10, I20, I30, etc., are then calculated by
characterizing fiber reinforced concrete toughness in an objective manner.
taking the ratios of the energy absorbed to a certain multiple
of first-crack deflection and the energy consumed up to the
Keywords: cracking (fracturing); fiber reinforced concretes; flexural
strength; flexural tests. occurrence of first crack. Expressed in general terms

The usefulness of fiber reinforced concrete (FRC) in Energy absorbed up to a certainmultiple of first crack deflection-
I N = ---------------------------------------------------------------------------------------------------------------------------------------------------------
various civil engineering applications is indisputable. Fiber Energy absorbed up to the first crack
reinforced concrete has so far been successfully used in
slabs-on-grade, shotcrete, architectural concrete, precast The subscripts N in these indexes are based on the elasto-
products, offshore structures, structures in seismic regions, plastic analogy such that, for a perfectly elasto-plastic mate-
thin and thick repairs, crash barriers, footings, hydraulic rial, the index IN would have a value equal to N. The scheme,
structures, and many other applications. thus, compares a given FRC with a conceptual material that
The enhanced performance of fiber reinforced concrete behaves in an ideally elasto-plastic manner. Implicitly, the
over its unreinforced counterpart comes from its improved scheme also assumes that plain concrete is ideally brittle and,
capacity to absorb energy during fracture. While a plain hence, the various toughness indexes in its case assume a
unreinforced matrix fails in a brittle manner at the occur- constant value of 1. The strength remaining in the material is
rence of cracking stresses, the ductile fibers in fiber rein- characterized by the residual strength factors (R) derived
forced concrete continue to carry stresses beyond matrix from the toughness indexes (Fig. 1). Expressed in general
cracking, which helps maintain structural integrity and cohe- terms RM,N, the residual strength factor between Indexes IM
siveness in the material. Further, if properly designed, fibers and IN (N > M) is expressed as
undergo pullout processes, and the fractional work needed
for pullout leads to a significantly improved energy-absorp- RM,N = C{IN – IM}
tion capability. Thus, fiber reinforced concrete exhibits
better performance not only under static and quasi-statically
applied loads but also under fatigue, impact, and impulsive where constant C = 100/(N – M) chosen such that for an
loadings.1 This energy-absorption attribute of FRC is often ideally elasto-plastic material the residual strength factors
termed “toughness.” assume a value equal to the stress at which the elastic-to-
There is still considerable debate on how the toughness of plastic transition takes place. Plain concrete, with its ideally
fiber reinforced concrete should be measured.2,3 The two
widely used standard test methods are based on determining ACI Materials Journal, V. 92, No. 1, January-February 1995.
the energy required to deflect and fracture beam specimens Received May 5, 1993, and reviewed under Institute publication policies. Copyright
© 1995, American Concrete Institute. All rights reserved, including the making of
under four-point loading as discussed in the following copies unless permission is obtained from the copyright proprietors. Pertinent discus-
sion will be published in the November-December 1995 ACI Materials Journal if
section. received by Aug. 1, 1995.

48 ACI Materials Journal/January-February 1995


From this measure of flexural toughness, a flexural tough-
ACI member Nemkumar Banthia is an associate professor of civil engineering at the
University of British Columbia, Vancouver, Canada. He is a member of ACI Committees ness factor (FT) is calculated as shown in Fig. 1. Note that
544, Fiber Reinforced Concrete; and 549, Ferrocement and Other Thin Reinforced the flexural toughness factor (FT) has the units of stress such
Products. His main research interests include instrumented impact testing, fracture that its value indicates, in a way, the post-matrix cracking
studies, low-temperature effects, and fiber reinforced concrete. He has authored or
coauthored more than 60 papers on concrete and fiber reinforced concrete. residual strength of the material when loaded to an arbitrary
deflection of span/150. Clearly, the flexural toughness factor is
ACI member Jean-François Trottier is an assistant professor at the Technical dependent on specimen geometry and other testing variables.
University of Nova Scotia in Halifax, Canada. He received his PhD in civil
engineering from Laval University in Québec City. He is investigating static and The chosen deflection of span/150 for its calculation is purely
impact behavior of fiber reinforced cement-based materials. arbitrary and not based on serviceability considerations.
A large number of concerns have been raised about these
two standardized test methods. The purpose of this paper is
brittle response, therefore, has residual strength factors equal
to highlight some of their limitations, to describe some of the
to zero.
difficulties commonly encountered, and to caution against
Both toughness indexes and the residual strength factors
some of the pitfalls. Commonly made errors in conducting
provide information on the shape of the load-deflection plot
these tests and in analyzing the test results are also identified.
and are presumably independent of the specimen size and
Finally, a simplified technique of analyzing the data from
other testing variables. Notice that an accurate assessment of
these tests is proposed, which addresses some of these
the energy at first crack is of critical importance, since its use
concerns.
is made later in the determination of all performance parame-
ters. Equally important is an exact determination of the beam
deflections both before and after the first crack. RESEARCH SIGNIFICANCE
Fiber reinforced concrete (FRC), on account of its superior
energy-absorption capability or “toughness,” is a material
JAPAN SOCIETY OF CIVIL ENGINEERS (JSCE)
STANDARD SF-4 METHOD OF FRC TOUGHNESS suitable for numerous applications. There is, however, a
CHARACTERIZATION great deal of debate over how the toughness of FRC should
In this technique, the area under the load-versus-deflection be characterized. The inadequacy of the currently available
plot up to a load point deflection of span/150 is obtained.5 test methods is widely recognized and need is felt to devise

Fig. 1—ASTM C 1018 and JSCE SF-4 techniques of fiber reinforced toughness characterization.

ACI Materials Journal/January-February 1995 49


a better characterization tool. This paper describes the defi- both sides of the specimen and averaging the measured
ciencies of the current techniques and the difficulties values. In a flexural specimen under a transverse load,
encountered during their use. The paper also proposes a however, the bigger source of error is the settlement of the
better characterization scheme. specimen supports themselves, such that the measured
displacement of the load points comprises not only the true
CONCERNS WITH THE ASTM C 1018 TEST METHOD displacement due to the response of the beam material to the
Measuring true specimen deflections applied stress but also those arising from seating and the
Given that toughness characterization is based on energy
computations, an accurate measurement of deflections is downward movement of the beam as a rigid body. This is
very important. In the language of the code, “...exercise care particularly true at the instant of first crack, where deflec-
to ensure that the measured deflections are the net values tions are small and the error can be relatively large. If not
exclusive of any extraneous effects due to seating or twisting properly considered, the settlement in the supports can lead
of the specimen on its supports or deformation of the support to a gross overestimation of the first-crack energy and,
system... .” Given this rigorous requirement, a great deal of hence, to erroneous indexes. This is illustrated in Fig. 2,
the data available in the literature based on inaccurate deflec- where load-versus-deflection plots obtained on a steel fiber
tion measurements is worthless. reinforced concrete beam with deflections measured in three
The error in the measured deflections due to twisting of the different ways are shown. Deflections were measured as the
specimen can be minimized by measuring displacements on movement of the loading cross-arm itself, by an LVDT in

Fig. 2—Load-deflection curves obtained using three different deflection measurement


techniques (1 ksi = 6.895 MPa; 1 in. = 25.4 mm; 1 kip = 4.448 kN).

50 ACI Materials Journal/January-February 1995


contact with the top of the specimen and, finally, by using a Locating the first crack
device called the “Yoke” shown in Fig. 3. While the LVDT The calculations of toughness indexes require an accurate
and the cross-arm displacements include the settlement of assessment of the first-crack energy which constitutes the
supports, in the case of the “Yoke,” the rigid body displace- denominator in the definition of the various indexes. Any
ments are automatically subtracted from the measured error in the first-crack energy will lead to a significant error
deflections such that only the true deflections are recorded. in the values of the various indexes. As demonstrated in the
Note in Fig. 2 that the first-crack displacements measured following paragraphs, locating the point on the load-deflec-
using the “Yoke” alone match the theoretical value6 of tion curve where the first crack occurs is not so simple.
0.038, and the cross-arm and top LVDT displacements are, Fig. 4 compares the load-versus-deflection curves for
respectively, 60 and 28 times greater. Such an error in first- identical specimens of plain and steel fiber reinforced
crack displacement clearly leads to large errors in the calcu- concrete. The deflections in both cases were measured accu-
lations of the first-crack energy and, hence, in the toughness rately using a “Yoke,” and both load and deflection data were
index calculations, as shown in Fig. 2. acquired digitally at a frequency of 10 Hz. In Fig. 5, the initial
ascending parts of theses curves are magnified to locate the first
crack. Notice that both these curves have substantial nonlin-
earity even prior to attaining the peak load, making it almost
impossible to accurately locate the first crack.
According to ASTM C 1018, “... first crack is the point on
the load-deflection curve at which the form of the curve ‘first’
becomes nonlinear (approximately the onset of cracking in the
concrete matrix)... .” Based on this, the first crack may be
placed at Point A where the deflection is 0.0175 mm. On
comparing this deflection value with the range of deflections
suggested in the standard (for an identical specimen) of
between 0.038 to 0.064 mm, it is apparent that the chosen
deflection is 2.17 to 3.65 times smaller than the code-
suggested deflection. If one places the first crack at the point
of “significant” nonlinearity with a decrease in concrete stiff-
ness, as suggested recently by Johnston,7 one may place the
first crack at Point B corresponding to a deflection of 0.03
Fig. 3—Use of a Yoke to measure true specimen deflections. mm, which is still outside the suggested range.

Fig. 4—Load-deflection curves for plain and steel fiber reinforced concrete beams (1 in. = 25.4 mm; 1 kip = 4.448 kN; 1 lb =
0.4536 kg; 1 ft3 = 0.02832 m3).

ACI Materials Journal/January-February 1995 51


There are compelling arguments to locate the first crack at of the curve decreases, and the curves for a very high-
Point C at a deflection of 0.04 mm in Fig. 5. First of all, if the strength matrix become almost linear. Quantitatively, while
objective of the test method is to quantify what FRC can do the nonlinear strains/deformations in the case of normal
over and above plain concrete, Point C needs to be chosen as strength concrete may be as high as 30 percent of the total,
the first crack. Besides, for low-fiber fraction FRC, the peak those for the high-strength matrix may be as low as 2
load is logically the end of the matrix contribution and essen- percent. This is well expected since, in the case of a high-
tially the commencement of fiber reinforcing action. Point C strength matrix, less microcracking before the peak load
is, therefore, justifiably the “real” first crack. This point also occurs, which reduces the apparent nonlinearity. This also
lies in the suggested range of deflections. affirms that the prepeak nonlinearity observed with normal
In Table 1, the indexes are calculated based on the three strength matrixes (Fig. 5) is real and not merely an artifact of
possible locations of the first crack at Points A, B, and C the test apparatus.
corresponding to first-crack deflections of 0.0175, 0.03, and
0.04 mm, respectively. Notice that, based on where the first Instability after peak load
crack is assumed to occur, extreme variations in the index Fig. 6 shows that the prepeak response for both plain and
values may occur. Even if standardized equipment with high FRC is nearly identical, and it is only after the occurrence of
precision is used as in this case, the placement of first crack the peak load that any real influence of fibers is observed.
is highly subjective, which is bound to lead to large varia- The point of peak load occurrence, however, is also the point
tions between operators and laboratories. Included in Table of instability for the loading machine which, if not stiff
1 are also the indexes for plain concrete. Notice that the I5 for enough, will undergo sudden unloading and release large
plain concrete is 1 only when the first crack is placed at Point amounts of energy. This sudden release of energy has major
C; otherwise, it assumes values significantly greater than effects on the load-deflection curve immediately following
unity and even as high as that for FRC! the peak load, particularly for FRCs with low fiber contents
The ascending parts of the load-deflection curves for and with very high-strength matrixes.
FRCs based on matrixes with different compressive The instability problem is illustrated in Fig. 7 where an
strengths (40, 55, and 85 MPa) are shown in Fig. 6. Curves unstable curve for FRC-A with a lower fiber content of a less
for specimens without fibers are also shown. These are identical efficient fiber is compared with a stable curve for FRC-B
to those with fibers. Notice that, as the compressive strength with a higher fiber content of a relatively more efficient
of the matrix is increased, the nonlinearity in the prepeak part fiber. In both cases, the machine releases the same amount of

Fig. 5—Initial ascending parts of curves in Fig. 4 magnified (1 in. = 25.4 mm; 1 kip = 4.448 kN)

52 ACI Materials Journal/January-February 1995


energy at Point A. In the case of FRC-B, the material, due to The problem of instability is exemplified in the case of a
its high fiber content, absorbs this energy and averts insta- high-strength matrix, where in some cases even I30 can be
bility. However, in the case of FRC-A with a low fiber measured in the unstable zone. A high-strength matrix can
content, the incoming energy causes instability, and the support a higher peak load, which inadvertently leads to a
curve jumps suddenly from Point A (deflection 0.037 mm) bigger energy release at the occurrence of peak load. The
to Point B (deflection 0.33 mm) in a matter of 1/10 sec (the failure can, therefore, transform from stable to unstable by
data-logger, even when running at a high-acquisition merely increasing the matrix strength, as shown in Fig. 8.
frequency, usually records no points in between). With Point
B being the true level of post-peak load-carrying capacity of Table 1—Toughness indexes as affected by first-
crack placement
the specimen, one may postulate that an energy equal to that
Toughness indexes
contained in the triangle ABC is the amount by which the
First crack at First crack at First crack at
performance has been overestimated. With the portion of the 0.0175 mm, 0.03 mm, 0.04 mm,
Material Point A Point B Point C
curve immediately following the peak load being of primal
I5 = 5.73 I5 = 3.82 I5 = 3.33
importance in index calculations, the indexes falling in this
I10 = 10.94 I10 = 7.08 I10 = 5.96
region will be incorrect. This is readily apparent in the values
of the toughness indexes, as shown in Fig. 7. FRC-A, even FRC 40 kg/m3 I20 = 21.27 I20 = 13.09 I20 = 11.15

with a lower fiber content, registered higher I5 and I10 I30 = 31.62 I30 = 19.24 I30 = 16.50
indexes (measured in the unstable zone) than FRC-B with a I60 = 61.42 I60 = 38.02 I60 = 32.85
higher fiber content! Plain concrete I5 = 4.57 I5 = 1.65 I5 = 1.00

Fig. 6—Prepeak load-deflection curves for plain and fiber reinforced concretes with low-,
medium, and high-strength matrixes (1 in. = 25.4 mm; 1 kip = 4.448 kN; 1 lb = 0.4536 kg;
1 ft3 = 0.02832 m3).

ACI Materials Journal/January-February 1995 53


Fig. 7—Stable and unstable failures in fiber reinforced concrete (1 in. = 25.4 mm; 1 kip = 4.448 kN).

Fig. 8—Change in nature of failure from stable to unstable with increase in compressive strength (1 in. = 25.4 mm; 1 kip =
4.448 kN)

54 ACI Materials Journal/January-February 1995


The problem associated with instability can be remedied PROPOSED TECHNIQUE
by using a closed-loop servo-controlled test system. This Rationale
kind of sophisticated equipment, unfortunately, is rarely 1. The ascending part of the load-deflection curve up to the
used by the commercial laboratories. occurrence of the peak load will always be nonlinear, and
locating the point where the matrix cracking first occurs will
CONCERNS WITH THE JSCE SF4 TEST METHOD always be subjected to human error. The technique, there-
This method also relies on the measurement of energy fore, must not be sensitive to errors made in locating the first
absorbed by the specimen and, as such, an accurate measure- crack like the current ASTM method.
ment of specimen deflections is of importance. Note in Fig. 2. The instability observed in the load-deflection curve at
2 that the FT values when displacements are measured in matrix cracking in the case of FRC with either a low-fiber
three different ways are also affected, but not as dramatically volume or with a high-strength matrix will occur in all load-
as the toughness indexes. Identifying the correct location of controlled tests. This should be properly dealt with.
the first crack, which is crucial and one of the main problems 3. With the current design practice, the use of energy (or
with the ASTM method (Fig. 5), is not a concern with the toughness) as a material property in designs is somewhat
JSCE method. intangible. Until advanced design tools are developed to
Unlike the ASTM method, the instability in the load- design for toughness, the characterization of FRC with
deflection plot soon after the first crack is not of major residual strength values after matrix cracking, as done in the
concern in the JSCE method, as shown in Fig. 7; the end- JSCE method, appears to be more appropriate.
point deflection of span/150 is too far out in the curve to be Based on these guidelines, analysis of the load-deflection
affected by the instability in the initial portion. curves (obtained by testing flexural specimens under third-
The JSCE technique, however, is also not without limita- point loading as in the other standard techniques) is proposed
tions and concerns. First of all, the flexural toughness factors as follows (see Fig. 9):
are specimen geometry-dependent, which makes an exact 1. Obtain the load-versus-deflection curve (as in the case
correlation with the field performance of FRC rather diffi- of other methods) with accurate deflection measurements
cult. Also, the end-point chosen on the curve at a deflection using a “Yoke” or a similar device.
of span/150 (Fig. 1) is often criticized for being much greater 2. Locate the peak load and divide the curve into two
than the acceptable deflection/serviceability limits. The regions: the prepeak region before the occurrence of the peak
behavior immediately following the first crack, which may load and the post-peak region after the peak load. Note the
be of importance in many applications, is not indicated in the value of load at the peak and measure the area under the
flexural toughness factor in any way. Finally, the technique curve up to the peak load. This measure of energy is termed
may be criticized for failing to distinguish between the prepeak energy and denoted as Epre in Fig. 9. For some
prepeak and the post-peak behaviors by adopting a smeared composites with high fiber volume fractions, two load peaks
approach of using the combined area under the curve to may occur: one at the end of the matrix contribution and the
calculate the flexural toughness factors. other when the fibers reach their ultimate capacity. In such

Fig. 9—Proposed technique.

ACI Materials Journal/January-February 1995 55


Fig. 10—Characterizing FRC using proposed technique. Corresponding toughness indexes and Japanese flexural toughness
factors are also shown (1 ksi = 6.895 MPa; 1 in. = 25.4 mm; 1 kip = 4.448 kN).

cases, clearly, the first load peak, corresponding to the end of can be chosen depending on the application, and it may even
the matrix/composite contribution, should be considered. coincide with the Japanese value of 150.
3. Locate points on the curve in the post-peak region with As expected, the instability problem will exist even in the
specimen deflections equal to various fractions of the span L/m1, calculation of the proposed PCSm values. It is proposed that
L/m2, etc. The suggested fractions are between L/3000 the triangle ABC in Fig. 7 should not be considered in the
and L/150. Measure the areas under the curve up to these calculation of the PCSm, and the curve should be considered
deflections denoted as Etotal,m (measured at a deflection of L/m). to follow the path A-C-B. However, the test report must
4. Subtract the prepeak energy Epre from the various identify the PCSm values that fall in the unstable zone CB.
values of Etotal,m to obtain the post-peak energy values to a
deflection of L/m, Epost,m. EXAMPLE
5. Calculate post-crack strengths (PCSm) in the post-peak Consider the curves for two different fiber reinforced
region at the various deflections. The post-crack strength at concretes (Beams A and B) as shown in Fig. 10. Note that
a deflection of L/m, PCSm, is defined as Beam A, in general, may perform superior to Beam B, with
the exception that in the beginning Beam B outperforms
( E post,m )L Beam A. Notice from the table in Fig. 10 that, in spite of
PCS m = -------------------------------------
L
⎛ ----- – δ peak⎞ bh
2 using the best possible equipment, the ASTM C 1018 anal-
⎝M ⎠ ysis technique is not able to distinguish between these two
very different composites. The JSCE technique identifies the
Notice that the PCSm values are calculated from the post- general superior performance of Beam A but fails to indicate
peak energy (Epost,m) and not the total energy (Etotal,m), such the initial superior performance of Beam B. This technique
that the prepeak energy (Epre) has been omitted from the also incorrectly combines the elastic and the post-elastic
calculations. The advantage over the ASTM C 1018 tech- parts of the curve to produce a coupled flexural toughness
nique is that Epost,m is obtained by subtracting Epre from the factor at a very high deflection.
total energy Etotal,m (unlike the ASTM technique, where divi- Figure 10 also reports the proposed PCSm values calcu-
sion by the first-crack energy is involved) and as such the lated at beam deflections equal to various fractions of the
resulting PCSm values are not sensitive to small errors made span L/m. Notice that PCSm values readily recognize that
in the calculation of the prepeak energy Epre. The value of m Beam B is better performing initially up to a deflection of

56 ACI Materials Journal/January-February 1995


span/1500, after which Beam A takes over and outperforms niques and the prevalent concerns. The paper also proposes
Beam B, as seen visually. an alternate method of analyzing the curves by calculating
The important benefit of the proposed technique is that post-crack strength values at various deflections. The
there is no ambiguity in either identifying the peak load on proposed technique leads to FRC attributes that are not
the curve or in calculating the prepeak energy absorbed. The susceptible to human judgmental errors.
proposed PCSm values, therefore, are not susceptible to any
human judgmental error. ACKNOWLEDGMENTS
The continued financial support of the Natural Sciences and Engineering
One apparent criticism of the proposed analysis technique, Research Council of Canada is greatly appreciated.
as also with the Japanese technique, is the inability of the
PCSm values to distinguish clearly between low fiber volume REFERENCES
composites where only toughening occurs and high fiber 1. Banthia, N.; Mindess, S.; and Bentur, A., “Impact Resistance of
volume composites where both strengthening and tough- Concrete, Materials and Structures,” RILEM 20 (119), pp. 293-302.
2. Gopalaratnam, V. S., et al., “Fracture Toughness of Fiber Reinforced
ening take place. However, if strengthening takes place in Concrete,” ACI Materials Journal, V. 88, No. 4, July-Aug. 1991, pp. 339-353.
the composite after matrix cracking, the peak load will be 3. Johnston, C. D., Discussion of “Fracture Toughness of Fiber Reinforced
higher and located at a greater deflection, such that the Concrete,” by Gopalaratnam, V. S., et al., ACI Materials Journal, V. 89,
prepeak energy Epre will be higher. Any strengthening in the No. 3, May-June 1992, pp. 304-309.
4. ASTM C 1018, “Standard Test Method for Flexural Toughness and
system, if it occurs, will result in the higher values of Ppeak First Crack Strength of Fiber Reinforced Concrete (using Beam with
and Epre. Third-Point Loading (4.02),” American Society of Testing and Materials,
The use of the proposed technique to analyze various fiber Philadelphia, 1990, pp. 637-644.
5. Japan Society of Civil Engineers, “Method of Test for Flexural
reinforced concretes has indicated that the PCSm values are Strength and Flexural Toughness of Fiber Reinforced Concrete,” Standard,
sensitive not only to the characteristics and volume fraction SF-4, 1984, pp. 58-66.
of the fiber but also to the strength and composition of the 6. Banthia, N., and Trottier, J.-F., Discussion of “Fiber Type Effects on
concrete matrix.8 the Performance of Steel Fiber Reinforced Concrete,” by Soroushian, P.,
and Bayasi, Z., ACI Materials Journal, V. 89, No. 1, Jan.-Feb. 1992, pp.
106-107.
CONCLUSIONS 7. Johnston, C. D., Methods of Evaluating the Performance of Fiber
The use of the ASTM C 1018 and JSCE SF-4 techniques Reinforced Concrete, Materials Research Society Symposium Proceedings,
V. 211, S. Mindess and J. Skalny, eds., Boston, 1990, pp. 15-24.
of fiber reinforced concrete toughness characterization have
8. Banthia, N., and Trottier, J.-F., “Concrete Reinforced with Deformed
caused a great deal of confusion. This paper describes some Steel Fibers, Part II: Toughness Characterization,” ACI Materials Journal,
of the difficulties commonly experienced with these tech- V. 92, No. 2, Mar.-Apr. 1995, pp. 146-154.

ACI Materials Journal/January-February 1995 57


View publication stats

You might also like