Influence of Fiber Type and Fiber Orientation On Cracking and Permeability of Reinforced Concrete Under Tensile Loading 2017 Cement and Concrete Resea
Influence of Fiber Type and Fiber Orientation On Cracking and Permeability of Reinforced Concrete Under Tensile Loading 2017 Cement and Concrete Resea
a r t i c l e i n f o a b s t r a c t
Article history: During the service life of structure, cracks developed due to various internal or external stresses. These cracks
Received 20 July 2016 offer a preferential path for the ingress of water, gas and aggressive ions through the concrete. This increase of
Received in revised form 18 November 2016 concrete's permeability promotes the structure deterioration. A good understanding of structure permeability
Accepted 10 January 2017
is thus required to design durable structures. An innovative permeability device was used to measure the impact
Available online 31 January 2017
of ber orientation and ber type on the water permeability of high performance ber reinforced concrete
Keywords:
(HPFRC) tie-specimens submitted simultaneously to a uniaxial tensile loading. Specic casting techniques pro-
Durability vided tie-specimens with three ber orientations (favorable, average and unfavorable) that can be found in
Permeability ber reinforced concrete (FRC) structural elements. Experimental results showed average bers angles relative
Concrete to the loading direction of tie-specimens of about 39, 42 and 54 respectively for the favorable, average and un-
Fiber reinforcement favorable orientations. As ber orientation become less favorable (from 39 to 54), the tensile strength (ft) of
Tensile properties characterization specimens reduced up to 33%, while the increase in permeability (Kw) tie-specimens achieved
1990% at xed load and 600% at xed stress in the rebar. For HPFRC casted with the same procedure, synthetic
bers provided a lower efciency to control cracking at material and structural scales, and to limit water perme-
ability in comparison to steel bers. The impact of ber orientation is clearly more signicant on water perme-
ability of HPFRC than on their mechanical behavior. These results provides insights of the impact of ber
orientation on durability of reinforced FRC structures.
2017 Elsevier Ltd. All rights reserved.
1. Introduction During their lifetime, structures developed cracks due to various in-
ternal stresses (shrinkage, creep) and external stresses (mechanical and
Durability of reinforced concrete structures relies on the transport of environmental loads). Cracks offer an easy path for aggressive agents to
water and aggressive agents through concrete by several mechanisms penetrate into concrete by permeability. Thus, to design durable struc-
as permeability, diffusion, capillary absorption, etc. Among these mech- ture, it is necessary to identify and evaluate parameters inuencing
anisms, permeability, dened as the movement of a uid through a po- the permeability of cracked concrete. Studies performed on unloaded
rous saturated medium under a pressure gradient, can be used as a concrete specimens with one single crack showed that the water per-
durability index especially when the concrete is cracked. The increase meability of cracked concrete is mainly controlled by the crack widths
of permeability favors the ingress of water, gas and aggressive ions [48]. To be more representative of the reality of the structures, which
into the concrete, which promotes its deterioration by freeze-thaw, sul- are continually loaded and randomly cracked, some researchers studied
fate attack, alkali-aggregate reaction or reinforcement corrosion [13]. the permeability of longer concrete specimens under tensile loads [9].
These deteriorations limit the lifetime of concrete structures and raise Permeability was also measured recently on reinforced concrete tie-
their direct costs combining construction and rehabilitation works, specimens submitted to tensile loads to reproduce more efciently the
and indirect costs related to socioeconomic and environmental impacts global behavior of tensile zones found in reinforced concrete structures
resulting from these rehabilitations. In this context, durability is now [10,11].
more than ever an important concern for infrastructures owners and Research works on loaded concrete clearly showed that ber rein-
stakeholders. forced concretes (FRC) have a reduced water permeability in compari-
son to ordinary concrete. The addition of bers provides a lower
porosity and a higher tortuosity to the concrete matrix, thus reducing
Corresponding author.
permeability at uncracked state [12]. Moreover, ber's ability to bridge
E-mail addresses: thibaut.plague@polymtl.ca (T. Plagu), clelia.desmettre@polymtl.ca cracks results in thinner cracks and thus provide a lower permeability
(C. Desmettre), jean-philippe.charron@polymtl.ca (J.-P. Charron). at cracked state [10].
http://dx.doi.org/10.1016/j.cemconres.2017.01.004
0008-8846/ 2017 Elsevier Ltd. All rights reserved.
60 T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970
Table 1 Table 3
Compositions of concretes (0.75 % vol. of bers). Spreads and mechanical properties of concretes.
Logically, parameters inuencing the action of bers in the concrete with a normal workability concrete after vibration. The tie-specimens
matrix may have an impact on the permeability. They must be investi- with synthetic bers were only tested with the favorable orientation
gated to tailor FRC according to the durability expected for structural (HPFRC-SY-F) since the higher exibility of this ber type does not
applications. It was demonstrated that increasing concrete matrix allow easily ber alignment and determination of ber orientation by
strength favors anchorage of bers [13] and reduces permeability [11]. image analysis techniques.
It was also shown that increasing the dosage of steel bers [8], synthetic Four tie-specimens per conditions (HPFRC-ST-F, HPFRC-ST-A,
bers [14], or a combination of those [15], decreases water permeability. HPFRC-ST-U and HPFRC-SY-F) were submitted to permeability tests
Besides, until now no study was dedicated to the impact of ber orien- under loading. After the tests, three slices were sawn from one tie-spec-
tation on permeability of cracked concrete, although literature indicates imen of each steel ber orientation, perpendicularly to the tensile load
that bers orientation strongly inuences the tensile and bending be- and thus parallel to the cracks, to measure the ber orientation. The
haviors of FRC by modifying the ber efciency to bridge cracks [16 slices were taken adjacent to the three main cracks.
18]. Moreover, ber orientation varies within any FRC structure due to HPFRC workability and mechanical properties were measured for
formwork and rebar congurations and concrete owability, and from each condition, results are summarized in Table 3 and Fig. 2. The spread
one FRC structure to another, therefore the impact of this parameter was measured by the slump ow test according to ASTM C1611 [20].
on FRC durability must be established. The compressive strength (fc) and the Young's modulus (Ec) were deter-
The general objective of this project is to determine the inuence of mined at 28 days and 50 days (age of the permeability test) in accor-
the ber orientation and ber type on water permeability of FRC struc- dance with ASTM C39 [20]. The tensile strength (ft) was determined at
tures. A permeability device developed at Polytechnique Montreal [11, 50 days on notched core cylinders (Fig. 1a) in accordance with RILEM-
19] was used to perform water permeability and crack openings mea- TC-162 [21]. These cylinders were obtained from 95-mm diameter
surements simultaneously with the application of tensile loadings on cores extracted from larger concrete blocks that were produced with
reinforced concrete tie-specimens. the same casting techniques used for the tie-specimens production
(Section 2.2.1) in order to have a similar ber orientation. Fig. 1 show
2. Methodology pictures of typical tensile failure plans obtained for the HPFRC-ST-F,
HPFRC-ST-A and HPFRC-ST-U conditions respectively.
2.1. Experimental program The tie-specimens contain one Grade 400W 10M rebar along the
longitudinal axis. The rebar properties were measured in accordance
Tie-specimens were submitted to water permeability and cracks with ASTM A615 [22]. They had a Young's modulus of 210 GPa and
width measurements simultaneously to the application of a quasi-static yield and ultimate strengths respectively equal to 470 MPa and
uniaxial tensile loading at the age of 50 days. The tie-specimens were 580 MPa.
made of self-consolidating high-performance ber-reinforced concretes
(HPFRC) with a water to binder ratio (w/b) of 0.43 and containing 0.75%
in volume of macrobers. Two macrober types were tested: hooked- 2.2. Experimental procedures
end steel bers (HPFRC-ST) and synthetic bers (polypropylene-poly-
ethylene mix) (HPFRC-SY). The compositions of the HPFRC-ST and 2.2.1. Specimen preparation
HPFRC-SY are presented in Table 1 and the geometrical and mechanical The tie-specimens represent a rebar and the surrounding concrete
properties of both bers are summarized in Table 2. found in the tensile zone of beams, thick slabs or walls subjected to
For steel bers, three ber orientations which can be found in ber- bending loads. For these structural elements submitted to bending
reinforced concrete (FRC) structural elements were tested: favorable load, the stress gradient is small and the tensile stress is nearly uniform
(HPFRC-ST-F), average (HPFRC-ST-A) and unfavorable (HPFRC-ST-U) in the concrete cover. Applying a uniaxial tensile loading on the tie-
orientations. The favorable orientation corresponds to the situation specimen is thus representative of the loading stress in these structural
where bers are the best aligned with the tensile load direction (longi- elements. The tie-specimen has a length of 610 mm, a cross-section of
tudinal axis of the tie-specimen), on the contrary the defavorable orien- 90 90 mm2, and contains a 10 M (11.3 mm in diameter) centered
tation represent the situation where bers are the worst aligned. The rebar. These characteristics were chosen in a previous work [19] to ob-
average orientation is the intermediate condition, some tests demon- tain realistic concrete cover and cracking pattern distribution found in
strated that HPFRC-ST-A orientation is similar to the one obtained reinforced concrete structures.
Table 2
Fibers properties.
Fiber type Length (l) Diameter (d) Aspect ratio (l/d) Young's modulus Tensile strength
Fig. 1. Photos of the tensile test. (a) The specimen with its instrumentation and the bers orientation in typical failure plans (b) for HPFRC-ST-F (c) for HPFRC-ST-A (c) for HPFRC-ST-U.
The three different steel ber orientations (HPFRC-ST-F, HPFRC-ST-A a negligible effect on the specimen mechanical behavior due to its low
and HPFRC-ST-U) were obtained using different casting techniques stiffness. Two water-lled aluminum boxes were then xed on the
based on the fact that the two main parameters inuencing the ber ori- two unsealed opposite longitudinal faces by a clamping system which
entation are the casting technique (the tool used and its movement) avoids any water leakage between the boxes and the tie-specimen
and the ow direction [13]. As the HPFRC were self-consolidating con- (Fig. 5). Two high-strength threaded rods were screw to steel couplers
cretes, bers direction followed the ow direction of the concrete in connected to the ends of the rebar to ensure that rebar yielding occurs
the moulds [23]. in the tie-specimen. The opposite extremities of the threaded rods
Each tie-specimen of HPFRC-SY-F and HPFRC-ST-F conditions (fa- were xed in the jaws of the load frame.
vorable orientation) was cast in a separated mold with a hand shovel Once the tie-specimen was installed on the load frame, a pressure
whose movement followed the longitudinal axis of the tie-specimen was applied to the water on the inlet tank while the outlet tank
(Fig. 3a). To minimize wall effects on the ber orientation on tie-speci- remained at the atmospheric pressure (Fig. 4). This initiated a pressure
mens of HPFRC-ST-A condition (average orientation) and HPFRC-ST-U gradient which induced a water ow in the water-saturated hydraulic
condition (unfavorable orientation), the four specimens of each condi- circuit from the inlet tank and inlet box, going through the tie-specimen,
tion were cast together in 90 610 680 mm3 formworks. The and nally to the outlet box and outlet tank. The pressure gradient ap-
HPFRC-ST-A was placed with a tube moved at an angle of 45 with re- plied on the tie-specimen (xed at 50 kPa, equivalent to a water
spect to the formwork walls to obtain the adequate ow direction depth of 5 m and measured by the pressure sensors installed on the
(Fig. 3b). The HPFRC-ST-U was rstly placed in a vertical reservoir at clamping system) produced a unidirectional water ow through the
the extremity of the formwork and then released in the formwork two unsealed opposite faces of the tie-specimen. Differential pressure
(Fig. 3c). The HPFRC-ST-U owed by itself at an angle of 90 relatively transmitters installed at the base of the inlet and outlet tanks continu-
to the longitudinal axis of the tie-specimens. The HPFRC-ST-A and ously recorded the water head and consequently the volume changes
HPFRC-ST-U blocks were later sawn into four tie-specimens. in these tanks, which allowed assessing the incoming and outcoming
water ows through the tie-specimen during loading. Fig. 4 depicts
2.2.2. Permeability and crack opening measurement one set of tanks for sake of simplicity. Nevertheless, the permeability de-
The tie-specimens were subjected to quasi-static uniaxial tensile vice includes three sets of tanks with different diameters to obtain accu-
loading with simultaneous water permeability and crack width mea- rate measurement and a sufcient water supply at any time of the test.
surements (Fig. 4). The uniaxial tensile loading was applied by a 2.5 The water permeability coefcient Kw (m/s) was calculated with
MN servo-hydraulics load frame at a constant rate of 0.05 mm/min Darcy's law (Eq. 1). In this equation, Kw is the water permeability coef-
until reaching the rebar yielding. cient (m/s), Q the water ow (m3/s), L the ow path length (m), A the
Before installing a tie-specimen in the load frame, thirteen PI dis- area of concrete exposed to water penetration and h the differential
placements transducers with measurements lengths of 50 or 100 mm water head across the specimen (m). There are several reasons to use
were glued directly on two opposite longitudinal faces of the tie-speci- the Darcy's law in this project. First, it is often used to describe the
men. These two faces as well as the two faces at the extremity of the water ow through a homogeneous porous medium [24,25], and it
specimen were sealed with a watertight membrane (Fig. 5) which has was also used in previous works with this permeability device [10,11].
Table 4
Summary of the ber orientation results near the cracks on the HPFRC-ST tie-specimens.
Tested concrete Fiber density (bers/cm2) Fiber area in surface analyzed Mean orientation angle () Mean orientation coefcient ( = cos) Standard deviation
Table 5
Maximal and mean COD and number of cracks for different load levels (average value on 4 specimens).
Force CODmax (mm) (main crack) CODmean, secondary cracks (mm) (CODmean/CODmax) 100 (%) Average number of cracks ()
(kN) STF STA STU SYF STF STA STU SYF STF STA STU SYF STF STA STU SYF
35 0.00 0.13 0.18 0.16 0.00 0.03 0.13 0.14 25 71 86 0.7 2.3 3.0 2.0
39 0.12 0.17 0.22 0.18 0.04 0.04 0.16 0.14 37 21 71 79 3.0 3.0 4.3 2.3
40 0.13 0.26 0.23 0.19 0.03 0.07 0.16 0.15 23 27 69 82 3.3 3.7 4.8 2.3
45 0.21 0.36 0.27 0.33 0.09 0.10 0.19 0.17 41 26 71 52 4.3 6.7 4.8 4.0
Second, the water ow through the saturated system is assumed to be the permeability measurement, with PI displacement transducers.
unidirectional, laminar, incompressible and continuous, which are the- Once a macrocrack localizes in the specimen, the COD was obtained
oretical requirements for the utilization of the law. Third, despite the by removing the elastic displacement from the total displacement of
heterogeneity found in concrete microstructure, a concrete matrix and the PI transducer. In most tests, a single crack is captured by one PI
a uniformly cracked matrix can be considered as homogeneous at a larg- gage thus providing information on each crack and allowing calculation
er scale. It must be mentioned that, in cracked concrete, the water ow of individual, average or maximal crack openings in the specimen.
is controlled by the crack openings. Between cracks, the water ow is With simultaneity of loading, permeability and cracking opening mea-
much lower in the concrete matrix and negligible compared to the surements, the permeability device allows direct correlations between
bulk permeability coefcient determined with Darcy's law. the applied force, displacement, permeability, crack widths and average
stress in the rebar. The rebar stress is calculated from the specimen elon-
Q:L gation, assuming a perfect bond between the concrete and the rebar.
Kw 1
A:h
2.2.3. Characterization of steel ber orientation
The permeability device, with an accuracy of 1013 m/s on the per- Steel ber orientation was analyzed by an in-house image analysis
meability coefcient measurement [19], can evaluate the permeability program [17], which use a similar method than those developed by
of uncracked or cracked concrete. Permeability measurement of un- Wuest [26] and Mlekusch [27]. Firstly, three slices extracted from one
cracked concrete requires a complete saturation of the specimen and a specimen of each HPFRC-ST condition were polished to obtain smooth
long testing period. Nevertheless, as this research project focuses on and straight analyzing surfaces. Secondly, these surfaces were painted
the permeability of cracked concrete, the measurements were done in to highlight the steel bers in concrete and then digitalized with a
a short time period. A very high saturation of the tie-specimens was en- 2400 dpi resolution scanner. Thirdly, the image was processed by the
sured by storing them in lime-saturated water from demolding to test- program, the program began by the digitalization of the surface follow-
ing at 50 days. The ow recorded at the inlet and the outlet tanks were ed by its binearization and ber detection. Image processing is illustrat-
equal throughout the entire test conrming the steady state ow of ed in Fig. 6. As the steel bers used are cylindrical, they have elliptical
water during the entire test. Tests were carried out at 20 C. cross-section on slices whose geometrical characteristics (centroids, di-
The cracks opening displacement (COD) evolution on the tie-speci- mensions) depend on ber orientation. The program detects these geo-
men during loading was continuously measured, simultaneously to metrical characteristics for each ber in applying a morphologic lter
Table 6
Average maximal and mean COD and number of cracks for different permeability (Kw) levels.
CODmax, main crack (mm) CODmean, secondary cracks (mm) (CODmean/CODmax) 100 (%) Average number of cracks ()
Kw (m/s) ST-F ST-A ST-U SY-F ST-F ST-A ST-U SY-F ST-F ST-A ST-U SY-F ST-F ST-A ST-U SY-F
5 E-8 0.07 0.18 0.07 0.10 0.01 0.05 0.02 0.01 9 28 37 12 1.7 2.7 1.5 1
1 E-7 0.14 0.17 0.09 0.16 0.04 0.05 0.02 0.04 32 32 27 25 3.7 2.7 1.8 1.3
4 E-7 0.31 0.24 0.16 0.20 0.15 0.08 0.05 0.06 47 34 34 29 5.3 4.3 2.5 2.3
8 E-7 0.34 0.18 0.22 0.09 0.12 0.16 28 65 72 6 3 3.3
1 E-6 0.37 0.19 0.23 0.10 0.12 0.17 27 62 73 6.3 3.5 3.3
1.5 E-6 0.45 0.22 0.26 0.11 0.15 0.15 25 68 59 7.3 4 3.7
2 E-6 0.49 0.24 0.28 0.12 0.17 0.17 25 69 59 7.7 4.5 4
3 E-6 0.28 0.40 0.18 0.25 65 62 5 4
4 E-6 0.31 0.19 63 5
5 E-6 0.33 0.21 63 5
T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970 63
Fig. 2. Tensile behavior of the studied concretes. (a) Pre-peak (stress-deformation), (b) post-peak (stress-crack width).
and a global threshold to identify accurately ber outline and defects, specimens, the average ber area obtained on the surface analyzed
and using then a Hough transform [17]. This allowed the calculation of was 0.75%, with values ranging from 0.65% to 0.81%. When cuts where
the angle () between each ber and the longitudinal axis of the tie- done near main macrocracks of tie-specimens, the average ber area
specimen. It must be pointed out that the effect of the ber hooks is ig- calculated on the surface analyzed was 0.68%, with values ranging
nored in the analysis of results, they represent a small part of the ber from 0.54% to 0.81%. The slightly lower ber dosage noted close to
length (14%) and their global inuence should be minor on the average cracks comes from the fact that cracks appear and develop at weakest
ber orientation. zones of tie-specimen where local ber density is lower or local ber
orientation is unfavorable.
3. Experimental results Analyzed sections have been made as close as possible from the
crack in the tie-specimen, so they represents the orientation where
3.1. Fiber orientation of steel bers cracks formed. The average number of macrobers per unit area of con-
crete (Table 4), calculated on three slices extracted from one tie-speci-
The image analysis program detected the total number of bers on men of each HPFRC-ST condition, decreased as the ber orientation
the analyzed sections as well as the orientation angle of each ber. becomes less favorable with values of 2.3, 1.9 and 1.4 bers/cm2 for
The image analysis provides information on the cross-section of each the HPFRC-ST-F, HPFRC-ST-A and HPFRC-ST-U tie-specimens respec-
ber, thus the total ber area expressed as a percentage of the total tively. However, less ber per unit area of concrete does not mean nec-
area of concrete can be calculated for a surface. It was conrmed that essarily less ber dosage. If ber angles are individually more inclined,
this value corresponds approximately to the ber dosage of 0.75% for their cross-section on concrete surface is larger and fewer bers can
the HPFRC studied. Indeed, when random cuts were done on tie- provide an equivalent ber dosage to the one found in more favorable
Fig. 3. Casting methods and direction of concrete ow in formworks (a) HPFRC-ST-F and HPFRC-SY, (b) HPFRC-ST-A, (c) HPFRC-ST-U conditions.
64 T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970
specimen. Similar results were obtained for the two others conditions.
Fig. 7b compares distribution and average orientation angle obtained
for each HPFRC-ST condition. The average orientation angles and coef-
cients as well as the standard deviations are summarized in Table 4.
Results shows that, despite the specic casting techniques used, the
average ber orientation were 38.8, 42.3 and 54.2 for the HPFRC-ST-F,
HPFRC-ST-A and HPFRC-ST-U respectively. A difference of 3.5 between
the HPFRC-ST-F and HPFRC-ST-A, and of 15.4 between the HPFRC-ST-F
and HPFRC-ST-U were obtained. The standard deviation increased as
the ber orientation deteriorates.
Fig. 6. Image analysis evolution process. (a) Digitalization, (b) binearization, (c) ber outline detection.
bers cast in the same way (HPFRC-ST-F and HPFRC-SY-F conditions) is conditions before stabilization (rebar yielding), whereas it is a little
very signicant, synthetic bers provide a lower mechanical lower for the HPFRC-SY-F condition.
contribution.
The evolution of the cracking pattern during the tensile loading is di- 3.3. Permeability of the tie-specimens
vided into the same three phases as represented in Fig. 9a (Phases II-A,
II-B and III) in Table 5. The separation is done using different grey levels The evolution of the water permeability coefcient (Kw) is plotted in
in the table (light to darker grey levels for Phases II-A, II-B and III function of the average stress in the reinforcement (s) (Fig. 11 and Fig.
respectively). 12). The stress corresponds to the average stress measured on the entire
During Phase II-A, for a same load level, as expected CODmax, CODmean length of the rebar and include the global effect of tension stiffening. Fig.
and the average number of cracks decrease as the steel ber orientation 11a shows the observed variability in Kw within the four tie-specimens
becomes more favorable (Fig. 9b and Table 5), this is coherent with the of a same testing condition (HPFRC-ST-A), whereas Fig. 11b show the
increase of the load capacity noted in Fig. 9a. In this phase, the use of average Kw obtained for the different testing conditions. Fig. 11b is
synthetic bers with a favorable orientation (HPFRC-SY-F) provides then split into two gures to illustrate permeability against rebar stress
CODmax and average numbers of cracks quite similar to the HPFRC-ST- level and variability of results. Fig. 12a presents the average, minimum
A condition, however the CODmean are close to those of the HPFRC-ST- and maximum values of Kw for the HPFRC-ST tie-specimens, and Fig.
U condition. 12b compares the average, minimum and maximum values of Kw for
During Phase II-B, CODmax and CODmean in the HPFRC-ST-F tie-speci- the HPFRC-ST-U and HPFRC-SY-F conditions.
mens stay lower than for the other conditions at a same load level and In this study, Kw is around 5 109 m/s for the HPFRC matrix and
explain their superior mechanical behavior (Fig. 9b and Table 5). quickly achieves around 1 10 7 m/s at the formation of the rst
CODmax in the HPFRC-ST-A tie-specimens signicantly increases and be- macrocrack. Since the objective of this study is to focus on ber orienta-
comes more opened than in the HPFRC-ST-U tie-specimens (Fig. 9b). tion impact on permeability, results are shown with a linear scale for the
However, the CODmean/CODmax ratios for HPFRC-ST-A (around 25%) are crack development and propagation phase occurring mainly between
clearly lower than for HPFRC-ST-U (CODmean/CODmax ratios of around 107 m/s and 105 m/s.
80%). As the mechanical performance of HPFRC-ST-A tie-specimens Fig. 13b illustrates that Kw increases after the formation of the rst
stays superior to the one of HPFRC-ST-U (Fig. 9a), it means the mechan- macrocrack for all testing conditions (load levels of 33.6, 32.9, 27.7
ical impact of the secondary cracks is important, not only the one of the and 33.2 kN, respectively for HPFRC-ST-F, HPFRC-ST-A, HPFRC-ST-U
maximal crack. and HPFRC-SY-F) and then begins to increase. This increase is more sud-
During Phase III, the main macrocracks are formed and the slopes of den for the HPFRC-ST-U and HPFRC-SY-F than for the HPFRC-ST-F and
the load versus CODmax curves are quite similar between the HPFRC-ST HPFRC-ST-A tie-specimens for which a moderate evolution can be
Fig. 7. Distribution curves of the steel ber orientation angles. (a) For one single condition (HPFRC-ST-A) and (b) for all the HPFRC-ST conditions (average curves).
66 T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970
Fig. 8. Mechanical behavior of a single test condition (HPFRC-ST-A). (a) load-displacement curves and (b) load-CODmax curves.
noted until reaching a force of around 39 kN. After this load level, an in- For a same CODmax, Kw globally increases as the steel ber orienta-
crease of Kw is noted for the HPFRC-ST-A condition whereas the increase tion becomes defavorable (Fig. 13b). This effect is more signicative be-
of Kw in function of load stays moderate in the HPFRC-ST-F tie- tween the HPFRC-ST-A and HPFRC-ST-U conditions. This is explained by
specimens. the fact that HPFRC-ST-A tie-specimens have more cracks and lower ra-
The evolution of water permeability in function of the average stress tios CODmean/CODmax, thus the secondary cracks are numerous but thin-
in the rebar (s) shows an increase of Kw as the ber orientation be- ner and lead to a low permeability (Table 6). The HPFRC-ST-F and
comes less favorable (Fig. 11b). This effect is still signicant considering HPFRC-ST-A conditions show similar Kw versus CODmax curves until a
the minimal and maximal values plotted on Fig. 12a, even if a higher CODmax of around 0.2 mm, then permeability increases more signicant-
variability of Kw is observed as the ber orientation angle increases. ly at equivalent CODmax for the average ber orientation. Concerning the
The next comparisons for permeability results were made at stress Kw-CODmax relation with synthetic bers (HPFRC-SY-F), it behaves sim-
levels found in structures in service conditions, which means between ilarly to the HPFRC-ST-U tie-specimens until a CODmax around 0.21 mm
s = 100 MPa and s = 350 MPa. The increase of average Kw at equiv- is reached. Logically, both conditions present quite similar average
alent s, is approximately 2.1 times between the HPFRC-ST-U and number of cracks and CODmean/CODmax ratios of around 70% until this
HPFRC-ST-A and 3.5 times between the HPFRC-ST-A and HPFRC-ST-F, point (Table 6). Afterwards, Kw increases more rapidly for the HPFRC-
resulting in a total increase by 7 times between HPFRC-ST-U and ST-U tie-specimens. The permeability results were also analyzed in
HPFRC-ST-F. Moreover, use of synthetic bers with a favorable orienta- function of the total crack width (CODtotal) measured in the tie-speci-
tion (HPFRC-SY-F) provide an permeability close to the one obtained mens, however the trends mentioned previously for Kw-CODmax curves
use of steel bers with an unfavorable orientation (HPFRC-ST-U). are the same for Kw-CODtotal curves.
Fig. 13 illustrates the relation between Kw and the CODmax, for three
specimens of a same condition in Fig. 13a and for the average results ob- 4. Discussion
tained for the different conditions in Fig. 13b. Table 6 summarizes
CODmax, CODmean (for the secondary cracks) and the average number 4.1. Practical considerations on steel ber orientation
of cracks for all conditions. The empty cells in Table 6 correspond to
data after the rebar yielding and they are not plotted in Fig. 13b. Fig. Despite the given attention to obtain favorable, average and unfa-
13a shows a relatively small variability compared to those observed be- vorable steel ber orientations, it was not possible to represent the
tween the different conditions in Fig. 13b, which means that the com- whole range of orientation angles ( of 38.8 to 54.2 instead of 0 to
parison of the average curves on Fig. 13b is adequate. 90). Such result was already observed in other research projects
Fig. 9. Average mechanical behavior of the tie-specimens. (a) Load-displacement curves and (b) load-CODmax curves.
T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970 67
Fig. 10. Average load-displacement curves of tie-specimens with min-max values. (a) For HPFRC-ST and (b) for HPFRC-ST-U and HPFRC-SY-F conditions.
dealing with the impact of ber orientation on high and ultra-high per- deteriorates. The increase of the standard deviation as ber orientation
formance ber reinforced concrete [17,28,29]. The maximal variation of deteriorates can also be explained by the inuence of the adjacent -
ber orientation of around 15 observed in this project thus gives a good bers, as the space between each other increases allowing more exibil-
idea of the limitation that could be found in real FRC structures. Howev- ity for rotation of bers.
er, such a difference in the steel ber orientation goes along with an
important variation of the number of bers at crack surfaces (Table
4), this combination leads to a signicant effect on the mechanical 4.2. Effect of steel ber orientation on mechanical and permeability
and durability properties as shown in this paper. As stated in behavior
Section 3.1, less ber per unit area of concrete does not mean neces-
sarily less ber dosage. If ber angle are individually more inclined, The steel macrober orientation modify the tensile behavior of
their cross-section on concrete surface is larger and fewer bers HPFRC-ST mixes (Fig. 2), rigidity at the end of the pre-peak phase, ulti-
can provide an equivalent ber dosage to the one found in more fa- mate strength and post-peak stresses increase with more favorable ori-
vorable ber orientation with more bers and smaller orientation entation. This trend was observed in other studies [30,31]. These results
angle. This is the case in this study, where the ber dosage of 0.75% seem to indicate that steel macrobers slightly contribute to control
introduced in all the HPFRC was validated with various ber orienta- microcrack formation and led to higher ultimate strength, and then
tion measurements. strongly participate to carry load on post-peak phase [13]. It is unusual
The analysis of the average orientation angles, total number of bers to observe impact of low ber dosage on pre-peak tensile behavior of
per unit area and standard deviations between all the testing conditions HPFRC. Unfavorable ber orientation may create weaker zones in con-
shows similar trends between these parameters. In fact, as ber orienta- crete, with the alignment of less dense ber/matrix interfaces in regards
tion became less favorable (mean orientation angles increasing by 9 and to tensile stresses imposed to the specimens, and consequently reduces
39% respectively for HPFRC-ST-A and HPFRC-ST-U in comparison with matrix strength. This assumption, that remains to be validated, may ex-
the HPFRC-ST-F), the number of bers per unit area decreased (by 17 plain this observation. Since the impact of ber orientation was ob-
and 39% respectively) and the associated standard deviations increased served on pre-peak tensile behavior of both characterization
(by 9 and 22% respectively). The decrease of the total number of bers specimens and tie-specimens, this effect is similar at material and struc-
per unit area with the degradation of their orientation may be explained tural scales and should be reliable.
by the fact that bers with unfavorable orientation represent an obsta- These modications of the pre-peak tensile behavior of HPFRC-ST
cle to the positioning of adjacent bers in a same transversal section of had logically direct implication on mechanical and permeability behav-
the tie-specimens. This obstacle increases as the ber orientation iors of tie-specimens. Indeed, more favorable ber orientation led to
Fig. 11. Permeability- stress in reinforcement curves of tie-specimens. (a) For one single condition (HPFRC-ST-A) and (b) for all the conditions (average curves).
68 T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970
Fig. 12. Average permeability- stress in reinforcement curves of tie-specimens with min-max values. (a) For HPFRC-ST and (b) for HPFRC-ST-U and HPFRC-SY-F conditions.
higher rst macrocrack loads (Fig. 9a) and delayed the associated in- increases proportionally to the cube of crack width [7,11], the presence
crease of permeability (Fig. 12a). of a greater number of wider cracks with less favorable ber orientation
The impact of ber orientation on tensile post-peak behavior (stress has led to a rise of permeability.
versus crack width curves in Fig. 2b) is also found on the strain harden- The relation between the permeability (Kw) and the average CODmax
ing behavior of the HPFRC-ST tie-specimens (Fig. 9a). At the formation (Fig. 13b) indicates that CODmax did not control alone the permeability
of the macrocrack, the HPFRC-ST tensile characterization specimens and that the consideration of the secondary cracks and the number of
rstly showed a sudden stress drop which is more marked as the ber cracks is necessary to understand the permeability measured at the
orientation deteriorates (Fig. 2b). This observation is due to a lower ef- structure scale. An example is the difference of Kw-CODmax behavior be-
ciency of bers to control the crack growth with unfavorable ber tween the HPFRC-ST-U and HPFRC-ST-A tie-specimens. For a same Kw
orientation [16,17]. In parallel, higher load drops were observed at level, HPFRC-ST-U condition presents lower CODmax and a lower num-
the formation of each macrocrack in the tie-specimens with ber ori- ber of cracks (Table 6), which should provide a lower permeability for
entations more unfavorable. After this sudden stress drop, the tensile this condition. Nevertheless, as HPFRC-ST-U tie-specimens present sec-
characterization specimens presented a hardening phase, whose ondary cracks widths relatively closed to the maximal crack width
intensity was lower with more unfavorable ber orientation, follow- (CODmeam/CODmax ratios of around 6070% versus 2030% for the
ed by a gradual stress reduction as the crack opened. Therefore, tie- HPFRC-ST-A tie-specimens), each secondary crack allows more water
specimens achieved lower ultimate loads with unfavorable ber penetration, in addition to the contribution of the maximal crack. The
orientation. competition between these contrary effects resulted in a higher perme-
The permeability (Kw) versus load behavior of the HPFRC-ST tie- ability for the HPFRC-ST-U than for the HPFRC-ST-A tie-specimens for a
specimens after the formation of the rst macrocrack is also in accor- same CODmax.
dance with their tensile mechanical behavior. As the ber orientation A global view of the impact of ber orientation (mean) on the me-
become unfavorable in HPFRC-ST, at cracked state the permeability in- chanical and permeability behaviors of HPFRC is shown in Table 7,
creases in average by around 21 times for the same rebar stress levels where quantitative differences in the tensile strength of characteriza-
in service condition (from s = 100 MPa to s = 350 Mpa) (Fig. 12a). tion specimens (ft), the rst crack load (F1st crack), ultimate load (Fult)
This rise of permeability can be explained by a lower contribution of - and average permeability (Kw mean) of tie-specimens are provided in
bers due to their angle. As an example, the HPFRC-ST-U tie-specimens comparison to the HPFRC-ST-F condition. The higher differences of ft.,
presented more cracks than HPFRC-ST-F tie-specimens, with higher F1st crack, Fult and Kw mean observed for the HPFRC-ST-U (N2.5 times the
CODmax and CODmean of the secondary cracks (Table 5). As permeability differences observed for the HPFRC-ST-A) are coherent with the higher
Fig. 13. Permeability versus maximum crack width (CODmax) of tie-specimens. (a) For one single condition (HPFRC-ST-A) and (b) for all the conditions (average curves).
T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970 69
Table 7
Differences of ber orientation, strength and permeability of various conditions in comparison to specimens with steel bers cast with a favorable orientation (HPFRC-ST-F).
Tested concrete mean () ft (MPa) cylinder F1st crack (kN) tie-specimen Fult (kN) tie-specimen Kw mean (m/s) Kw mean (m/s) at equivalent stress
at equivalent load 100 s 350 MPa
35 kN F 45 kN
HPFRC-ST-F
HPFRC-ST-A +9% 7% 2% 3% +540% +250%
HPFRC-ST-U +29% 33% 18% 17% +1990% +600%
HPFRC-SY-F N/A 25% 1.2% 17% +1750% +560%
difference of ber orientation between the HPFRC-ST-U and HPFRC-ST- the secondary cracks and the number of cracks is necessary to explain
A conditions (39%) than between the HPFRC-ST-A and HPFRC-ST-F con- the results.
ditions (9%). Besides, one can conclude that the impact of ber orienta- The quantitative differences illustrated in Table 7 between HPFRC-
tion on water permeability of tie-specimens (variation of 250 to 1990%) SY-F and HPFRC-ST-F shows clearly the lower efciency of the synthetic
is clearly more signicant than on their mechanical behavior (variation bers to provide crack control (lower post-peak tensile stresses) and to
of 2 to 33%). limit water permeability in comparison to the steel bers when both
One may wonder if unfavorable ber orientation may increase per- materials are casted to favor ber orientation parallelel to tensile stress-
meability close to the one of HPC. Permeability of HPC was measured es. The lower efciency of synthetic bers to reduce water permeability
in previous works [11] with the same setup and same concrete matrix was also observed by [32].
(w/b ratio, paste volume, etc.), results are shown in Fig. 12. Despite
the signicant impact of ber orientation on water permeability, 5. Conclusion
HPFRC-ST with the worst ber orientation and HPFRC-SY are still clearly
less permeable than the HPC. At equivalent stress in service condition The general objective of this research project was to evaluate the im-
(100 s 350 MPa), HPFRC-ST-F and HPFRC-ST-U present an average pact of ber orientation on the mechanical behavior, cracking process
reduction of the permeability of 97% and 75% respectively in compari- and water permeability of reinforced concrete under uniaxial tensile
son to HPC, while HPFRC-SY has a reduction of permeability of 77%. loads. As HPFRC tie-specimens were produced with the most and the
This conrm that in any situation water and chemical agent penetration worst ber orientation, the experimental results provided a global view
by permeability will be lower in HPFRC that in HPC, therefore HPFRC of the variation of the mechanical behavior and the water permeability
will provide an extended durability. that can be found in reinforced FRC structures. The main conclusions are:
4.3. Effect of ber type (synthetic versus steel) on permeability and me- The experimental procedure provides a good repeatability for results
chanical behavior of the same testing condition and coherent relationships between
the mechanical and water permeability measurements for the differ-
In comparison to HPFRC-ST cast with the same technique, HPFRC-SY ent testing conditions.
showed a lower rigidity at the end of the pre-peak phase and a lower ul- When ber orientation become less favorable (from 39 to 54) re-
timate strength ( 25%) in the tensile behavior on characterization garding tensile solicitations in HPFRC-ST, standard deviation on orien-
specimens (Fig. 2b). Concerning the post-peak behavior, for crack tation increase and the total number of bers on crack surfaces
width from 0.1 mm to 0.5 mm, post-peak stresses were reduced ap- decreased up to 41%.
proximately by 80% with synthetic bers instead of 40% with steel - As ber orientation becomes less favorable in HPFRC-ST (from 39 to
bers. Similar trends were obtained by [32,33] on steel and synthetic 54), the tensile strength of characterization specimens (ft) reduces
bers. Some studies [12,34] indicate that synthetic bers are less ef- up to 33%, while the rst crack load (F1st crack) and ultimate load
cient than steel bers to control crack openings, this specic type of syn- (Fult) of tie-specimens decrease up to 18% and 17% respectively.
thetic bers at this volume has to be multiplied by 2 or even 3 to provide As ber orientation becomes less favorable in HPFRC-ST (from 39 to
equivalent post-peak tensile stresses [34]. This trend is most probably 54), the increase in permeability (Kw) tie-specimens achieves
related to the lowest elastic modulus of the synthetic bers (Table 2) 1990% at equivalent load and 600% at equivalent stress in the rebar
that decreases the cracking control. These differences had logically in service conditions.
direct implication on mechanical and permeability behaviors of tie- Despite the signicant impact of ber orientation on water permeabil-
specimens. ity, HPFRC-ST with the worst ber orientation and HPFRC-SY are sig-
The rst crack and ultimate loads of HPFRC-SY tie-specimens were nicantly less permeable than HPC.
respectively higher and equal to the ones of HPFRC-ST-U tie-specimens. For HPFRC casted with the same procedure, synthetic bers provide a
Moreover, macrocrack occurrences caused higher force drops in HPFRC- lower efciency to control cracking at material and structural scales,
SY tie-specimens. These observations are in accordance with the mate- and to limit water permeability in comparison to steel bers.
rial behavior noted on notched cylinders. The impact of ber orientation is clearly more signicant on water
The evolution of permeability (Kw) of HPFRC-SY-F specimens in permeability of HPFRC than on their mechanical behavior.
function of the rebar stress (s) varies between the performance of CODmax did not control alone the water permeability. The consideration
HPFRC-ST-A and HPFRC-ST-U, the trend is nearly linear after of the secondary cracks and the number of cracks is necessary to under-
s 115 MPa instead logarithmic in HPFRC-ST specimens (Fig. 11a). stand the permeability of HPFRC with various ber orientations.
This linearity is probably link to the material behavior. Indeed, the ten-
sile behavior of the HPFRC-SY-F shows a brutal stress drop at crack lo- Acknowledgments
calization, then the post-peak stresses remain nearly constant till
crack opening of 0.5 mm (crack opening found in tie-specimens), This research project has been nancially supported by the Qubec
while stress hardening is observed in HPFRC-ST in the same condition Research Fund on Nature and Technologies (FQRNT, Grant for Strategic
(Fig. 2b). Clusters) and by the Natural Sciences and Engineering Research Council
The relation between Kw and the CODmax of HPFRC-SY-A is found be- (NSERC, Grant RGPIN312508-05). The authors are grateful to Holcim,
tween the HPFRC-ST-A and HPFRC-ST-U conditions (Fig. 13b). Again Bekaert, Euclid and Sika for their material donations for achieving this
CODmax did not control alone the permeability and consideration of research project.
70 T. Plagu et al. / Cement and Concrete Research 94 (2017) 5970
References [18] L. Ferrara, N. Ozyurt, M. di Prisco, High mechanical performance of bre reinforced
cementitious composites: the role of casting-ow induced bre orientation,
[1] P. Claisse, Book review, Proceedings of the Institution of Civil Engineers - Construc- Mater. Struct. 44 (2011) 109128.
tion Materials, 161, 2008, p. 135. [19] C. Desmettre, J.-P. Charron, Novel water permeability device for reinforced concrete,
[2] P.K. Mehta, Durability of concrete - 50 years of progress, Durability of Concrete, Vols Mater. Struct. 44 (2011) 17131723.
1 and 2 1991, pp. 131. [20] ASTM-C39/C39M-15a, in: A. International (Ed.), Standard Test Method for Com-
[3] A. Neville, Chloride attack of reinforced-concrete - an overview, Mater. Struct. 28 pressive Strength of Cylindrical Concrete Specimens, ASTM International, West
(1995) 6370. Conshohocken, PA, 2015.
[4] C.M. Aldea, M. Ghandehari, S.P. Shah, A. Karr, Estimation of water ow through [21] RILEM-TC-162-TDF, Final recommendation of RILEM TC 162-TDF: Test and Design
cracked concrete under load, ACI Mater. J. 97 (2000) 567575. methods for steel bre reinforced concrete, Mater. Struct. 36 (2003) 560567.
[5] M. Hoseini, V. Bindiganavile, N. Banthia, The effect of mechanical stress on perme- [22] ASTM-A615/A615M-16, in: A. International (Ed.), Standard Specication for De-
ability of concrete: a review, Cem. Concr. Compos. 31 (2009) 213220. formed and Plain Carbon-Steel Bars for Concrete Reinforcement, ASTM Internation-
[6] V. Picandet, A. Khelidj, H. Bellegou, Crack effects on gas and water permeability of al, West Conshohocken, PA, 2016.
concretes, Cem. Concr. Res. 39 (2009) 537547. [23] L. Martinie, N. Roussel, Simple tools for ber orientation prediction in industrial
[7] K. Wang, D.C. Jansen, S.P. Shah, A.F. Karr, Permeability study of cracked concrete, practice, Cem. Concr. Res. 41 (2011) 9931000.
Cem. Concr. Res. 27 (1997) 381393. [24] C.-M. Aldea, S.P. Shah, A. Karr, Permeability of cracked concrete, Mater. Struct. 32
[8] J. Rapoport, C.M. Aldea, S.P. Shah, B. Ankenman, A. Karr, Permeability of cracked steel (1999) 370376.
ber-reinforced concrete, J. Mater. Civ. Eng. 14 (2002) 355358. [25] A. Kermani, Permeability of stressed concrete, Build. Res. Inf. 19 (1991) 360366.
[9] J.S. Lawler, D. Zampini, S.P. Shah, Microber and macrober hybrid ber-reinforced [26] J. Wuest, E. Denarie, E. Bruhwiler, L. Tamarit, M. Kocher, E. Gallucci, Tomography
concrete, J. Mater. Civ. Eng. 17 (2005) 595604. analysis of ber distribution and orientation in ultra high-performance ber-rein-
[10] C. Desmettre, J.-P. Charron, Water permeability of reinforced concrete with and forced composites with high-ber dosages, Exp. Tech. 33 (2009) 5055.
without ber subjected to static and constant tensile loading, Cem. Concr. Res. 42 [27] B. Mlekusch, Thermoelastic properties of short-bre-reinforced thermoplastics,
(2012) 945952. Compos. Sci. Technol. 59 (1999) 911923.
[11] M. Hubert, C. Desmettre, J.P. Charron, Inuence of ber content and reinforcement [28] D.-Y. Yoo, S.-T. Kang, Y.-S. Yoon, Enhancing the exural performance of ultra-high-
ratio on the water permeability of reinforced concrete, Mater. Struct. 48 (2015) performance concrete using long steel bers, Compos. Struct. 147 (2016) 220230.
27952807. [29] L. Ferrara, A. Caverzan, M. Muhaxheri, M. Di Prisco, Identication of Tensile Behav-
[12] K. Hannawi, H. Bian, W. Prince-Agbodjan, B. Raghavan, Effect of different types of - iour of SFR-SCC: Direct vs Indirect Tests, BEFIB2012 Department of Structural Engi-
bers on the microstructure and the mechanical behavior of ultra-high performance neering, Politecnico di Milano, 2012.
ber-reinforced concretes, Compos. Part B 86 (2016) 214220. [30] J. Doyon-Barbant, J. Charron, Inuence de l'orientation des bres sur le
[13] P. Rossi, Les btons de bres mtalliques, Presses de l'Ecole nationale des ponts et comportement mcanqiue des BRF, Dpartement des gnies civil, gologique et
chausses, Paris, 1998. des mines, Ecole Polytechnique de Montral, 2016.
[14] N. Banthia, A. Bhargava, Permeability of stressed concrete and role of ber reinforce- [31] L. Ferrara, Tailoring the orientation of bres in high performance bre reinforced ce-
ment, ACI Mater. J. 104 (2007) 7076. mentitious composites: part 1 - experimental evidence, monitoring and prediction,
[15] J.S. Lawler, D. Zampini, S.P. Shah, Permeability of cracked hybrid ber-reinforced Int. J. Mater. Struct. Integr. 9 (2015) 7291.
mortar under load, ACI Mater. J. 99 (2002) 379385. [32] N. Buratti, C. Mazzotti, M. Savoia, Post-cracking behaviour of steel and macro-syn-
[16] C. Oesterlee, Structural Response of Reinforced UHPFRC and RC Composite Mem- thetic bre-reinforced concretes, Constr. Build. Mater. 25 (2011) 27132722.
bers, Ecole Polytechnique Fdrale de Lausanne, Lausanne, Suisse, 2010. [33] M.N. Soutsos, T.T. Le, A.P. Lampropoulos, Flexural performance of bre reinforced con-
[17] S. Delsol, J.-P. Charron, Numerical modeling of uhpfrc mechanical behavior based on crete made with steel and synthetic bres, Constr. Build. Mater. 36 (2012) 704710.
bre orientation, RILEM-b-AFGC Int. Symposium on Ultra-High Performance Fibre- [34] J.P. Charron, C. Desmettre, C. Androut, Utilisation de bton renforc de bres (BRF)
Reinforced Concrete, UHPFRC 2013Marseille, France 2013, pp. 111. pour les glissires en bton pour chantier (GBC), Polytechnique Montral, 2016.