Construction and Building Materials 242 (2020) 118186

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Parametric study of functionally graded concretes incorporating steel

fibres and recycled aggregates
Ricardo Chan ⇑, Xingzi Liu, Isaac Galobardes
Department of Civil Engineering, Xi’an Jiaotong-Liverpool University (XJTLU), Ren’ai Road 111, 215123 Suzhou, China

h i g h l i g h t s

FGC mechanical performance is inferior to FRC but superior to FRRAC.

FGC could be used in lower loading applications, as car parks or cycling lanes.

In terms of CO2 and costs, fibre content affects more than recycled aggregates.

Design parameters can be correlated with reinforced thickness and fibre content.

CO2 and costs can be correlated with reinforced thickness and fibre content.

a r t i c l e i n f o a b s t r a c t

Article history: Even presenting several drawbacks, especially regarding environmental impact, concrete is the most con-
Received 12 March 2019 sumed human-made material worldwide. In this sense, incorporating steel fibres and recycled aggre-
Received in revised form 5 November 2019 gates, and applying the functionally graded material concept may help in the sustainability of
Accepted 14 January 2020
concrete. In this paper, relationships between mechanical properties, embodied CO2, cost, among other
parameters of functionally graded concretes (FGC) incorporating steel fibres and recycled aggregates
were studied. The test results show that the FGC mechanical performance is inferior to fibre reinforced
Keywords:
concrete but superior to fibre reinforced recycled aggregate concrete. Therefore, the FGC studied could
Functionally graded concrete
Fibre reinforced concrete
be used in lower loading capacity applications, such as car parks and cycling lanes. In terms of embodied
Recycled aggregate concrete CO2 and cost, the FGC studied are more affected by the content of fibre than the content of recycled aggre-
Fibre reinforced recycled aggregate concrete gates. Also, equations correlating design and sustainability parameters with reinforced thickness and
Parametric study content of fibre were proposed.
Correlation Ó 2020 Elsevier Ltd. All rights reserved.
Sustainability

1. Introduction can be taken into account, such as improvement of concrete prop-

erties; and innovation in concrete constituents [3].
With an estimated production of 25 billion tonnes every year Steel fibre-reinforced concrete (SFRC) was created as one alter-
[1], concrete is the most consumed human-made material in the native to address this sustainability demand in concrete produc-
world and, by 2050, its demand is expected to increase by 12– tion. SFRC is a composite material characterized by improved
23%, in comparison with 2014 [2]. The plasticity in the fresh state, energy absorption capacity, fatigue strength, impact, and abrasion
the strength, and durability in the hardened state, the affordability resistance. These enhanced properties are due to the presence of
and abundance of raw materials are some reasons that make con- steel fibres, which control the cracking of the matrix, altering its
crete so massively used as a construction material. However, con- post-crack behaviour [4]. As a result, the use of SFRC has increased
crete production has several negatives environmental impacts, in a wide variety of structures, such as rigid pavements; shotcrete
such as energy consumption; emission of greenhouse gases; and and precast structures; dams; seismic structures; and bridges [4].
use of natural resources, for example, sand and rock. In order to In the case of rigid pavements, the application of SFRC reduces
achieve a more sustainable concrete production, some aspects the slab thicknesses, increases the lifecycle, and requires fewer
maintenance costs [5–8]. These factors contribute to the reduction
of raw material and energy consumption, greenhouse gases emis-
sions, and also the total cost during the whole lifecycle. Despite
⇑ Corresponding author. the advantages aforementioned, the high cost of steel fibres still
E-mail address: [email protected] (R. Chan).

https://doi.org/10.1016/j.conbuildmat.2020.118186
0950-0618/Ó 2020 Elsevier Ltd. All rights reserved.
2 R. Chan et al. / Construction and Building Materials 242 (2020) 118186

represents an obstacle for the use of this material in many coun- embodied CO2 and costs among the concrete mixes produced in
tries [6–8]. this study; and (4) a statistical analysis was performed to study
To optimize the use of steel fibres in SFRC and reduce their the correlation between design and sustainability parameters with
impact in cost, the concept of functionally graded material (FGM) other variables, such as the type of aggregate, layer configuration
can be used in the production of SFRC, obtaining a functionally and content of fibre.
graded concrete (FGC). FGM is a class of material which presents
composition and structure that gradually changes over the volume, 2. Experimental program
resulting in a material with properties corresponding to this grada-
tion [9]. In this sense, the FGC can be reinforced with steel fibres 2.1. Materials
only in the lower layers of concrete structures under bending,
where they are more effective in terms of resisting the flexural 2.1.1. Cement, water and aggregates
stresses introduced by loads [10–13]. A Portland cement of ordinary early strength class and standard
Besides, to increase the sustainability of concrete, the use of strength between 42.5 and 62.5 MPa (CEM I-42.5N), which fulfilled
recycled aggregates (RA) produced from construction and demoli- the specifications defined in BS EN 197-1:2011 [35], was adopted
tion waste (CDW), as a replacement for natural aggregates (NA), in this study. Tap water at room temperature (20 °C) was used.
has been studied since the end of the Second World War [14,15]. Natural and recycled aggregates were used for both coarse and fine
Since aggregates represent roughly 75% of the volume of typical aggregates. The natural aggregates were composed by river sand
concrete mixes [3], this alternative has a great potential reducing and limestone gravel (see Fig. 1a and b, respectively), while the
the impact in landfills and preserving virgin aggregates. However, recycled aggregates were produced from demolition waste (see
studies developed in past decades suggest that concretes produced Fig. 1c and d).
with RA, also called recycled aggregate concrete (RCA), present The sieving analysis of coarse and fine aggregates was done
lower strength and durability, in general [16–20]. In the attempt according to BS 812-103.1:1985 [36], and the results are shown
to overcome the drawbacks of RCA, the addition of steel fibres, pro- in Fig. 2. Then, the resulting size particle distributions were com-
ducing fibre-reinforced recycled aggregate concrete (FRRAC), has pared to the grading limits presented in BS 882:1992 [37]. The lim-
been studied and showed positive responses [12,21,30–32,22–29]. its are referred to the 10 mm single sized aggregate (see Fig. 2a)
Therefore, this paper aims to carry out a parametric study and the overall limits (see Fig. 2b) for coarse and fine aggregates,
between design and sustainability parameters with other vari- respectively. The fine aggregates particle size distributions are
ables, focusing on the flexural behaviour of FGC incorporating recy- within limits, but the coarse aggregates are slightly out of them.
cled aggregates and steel fibres. In addition, the relation between However, since the concretes assessed were produced using the
compressive and flexural strengths of FGC was assessed to verify same aggregates, the fact that the course aggregates are out of lim-
the agreement with the same relation for conventional concretes its does not affect in the test results and analysis.
[33,34]. To achieve these goals, this study was divided in four Table 1 presents the main properties of the aggregates consid-
parts: (1) an experimental program was designed to produce FGC ered for the production of the concrete samples. These are the oven
with natural and recycled aggregates, considering different layer dry density (qrd), the water absorption (WA), and the coefficients of
configurations and contents of fibre; (2) the results obtained from uniformity (Cu) and curvature (Cc). qrd and WA were obtained
the experimental program were presented and discussed; (3) a according to BS EN 1097-6:2013 [38], and Cu and Cc were calcu-
simplified sustainability analysis was done to compare the lated by Eqs. (1) and (2), respectively [39].

Fig. 1. Samples of (a) natural fine and (b) coarse aggregates, and (c) recycled fine and (d) coarse aggregates.
 R. Chan et al. / Construction and Building Materials 242 (2020) 118186 3

Fig. 2. Particle size distribution and grading limits for (a) coarse and (b) fine aggregates.

Table 1
Main characteristics of the aggregates.

Type of aggregate qrd (mg/m3) WA (%) Cu Cc

Natural Coarse 2.65 1.09 2.43 1.28
Fine 2.70 2.62 4.94 0.90
Recycled Coarse 2.25 7.54 5.53 2.13
Fine 1.89 15.07 10.33 1.02

D60
Cu ¼ ð1Þ
D10

ðD30 Þ2
Cc ¼ ð2Þ
D10  D60 Fig. 3. Concrete families considered: (a) FRC, (b) FRRAC, (c) PCC + FRC and (d)
PCC + FRRAC.
The oven-dry density of natural aggregates is higher than the
recycled aggregates. On the other hand, the water absorption
results are much lower for natural aggregates in comparison with was only considered in the bottom layer, where steel fibres can
recycled aggregates. This is probably due to the old mortar provide an essential increase in concrete flexural and flexural
attached to the recycled aggregate, which increases the porosity residual strength [4,41,42]. Also, to guarantee an adequate com-
and, therefore, decreases the density and increases the water pressive strength in the top layer, recycled aggregates were not
absorption [33,34]. Furthermore, the coefficients of uniformity used due to their negative effect on the concrete compressive
and curvature indicate that the recycled aggregates are better strength [19,43,44].
graded than the natural aggregates.
2.3. Mix designs and sample production
2.1.2. Steel fibre
A hooked-end steel fibre that fulfilled the specifications Since this study involves the comparison between concretes
required by BS EN 14889-1:2006 [40] was used in this study. produced with two types of aggregates, two reference mixes (M1
According to the supplier, these fibres have 60 mm of length (L), and M2) were designed and are presented in Table 2, being speci-
0.75 mm of diameter (D), aspect ratio of 80 (L/D), and tensile fied in terms of the mass of different materials required to produce
strength of 1150 MPa. For the production of each concrete mix, one cubic metre of concrete. M1 was designed for PCC targeting a
three contents of fibre (cf) were adopted: 20, 40 and 60 kg/m3, characteristic compressive strength of 30 MPa, using a method
which are equivalent to approximately 0.25, 0.50 and 0.75% in vol- based on the mass of materials in a unit volume of fully compacted
ume, respectively. concrete assuming that the aggregates were at a saturated surface-
dry condition [45]. Then, M2 was designed using the same propor-
2.2. Types of concrete tions of M1 but substituting the natural aggregates for recycled
aggregates and adjusting the quantities according to the difference
In this study, four concrete families are considered, as shown in in saturated surface-dry density.
Fig. 3. Two families were produced with one concrete mix through A cement content of 475 kg/m3 and free water/cement ratio
their entire volume: FRC and FRRAC (see Fig. 3a and b), while the (w/c) of 0.45 were adopted for both mixes. Also, notice that the
FGM concept was applied in the other two families. In that sense, aggregates were used in the air-dried condition. Therefore, the
two-layered systems were adopted being the top layer produced aggregates were pre-wetted according to their moisture content
with plain cement concrete (PCC) and the bottom layer with FRC
(see Fig. 3c) and FRRAC (see Fig. 3d), respectively.
Table 2
The FGC configuration was chosen in order to improve the effi- Reference mix designs (kg/m3).
ciency of concrete in structures subjected to bending, such as
Reference Cement Water Type of Fine Coarse
beams and slabs. In this kind of structure, the top zone works in
mix aggregate aggregate aggregate
compression, while the bottom zone in tension. Therefore, since
the addition of steel fibres does not influence the compressive M1 475 215 Natural 790 890
M2 Recycled 805 715
strength of concrete [32,41,42] significantly, the reinforcement
4 R. Chan et al. / Construction and Building Materials 242 (2020) 118186

presented before the production of concrete and their respective Four prismatic samples were produced for each FGC mix, total-
water absorption [45] shown in Table 1. izing 96 samples. All samples were demoulded 24 h after casting
M1 was used to produce PCC and FRC, whereas M2 was used to and cured in water at a temperature of approximately 20 °C for
produce FRRAC. Besides, as presented in Fig. 4, the FGC were stud- 28 days following the requirements of the standard BS EN
ied considering different relationships between reinforced layer 12390-2:2009 [47].
height and total height (h/H): 0.25, 0.50, and 0.75.
Considering h/H (0.25, 0.50 and 0.75) and cf (0.25, 0.50 and 2.4. Test methods
0.75%), a total of 24 FGC were produced. Note that the cf is referred
only to the reinforced layer. To identify the mixes, the following 2.4.1. Bending test
codification was used: First number + Letter + Second number. The flexural behaviour of the FGC was studied using the four 4-
The first number stands for the relation h/H, the letter is N for nat- point bending test described in the ASTM standard C 1609/C
ural aggregates and R for recycled aggregates, and the second num- 1609M-06 [48]. Four prismatic samples per mix had their loading
ber stands for the content of fibre in the reinforced layer (cf). For surface prepared following the procedure as mentioned earlier and
example, 0.50N0.75 is the concrete with reinforced layer repre- were tested at the age of 28 days. In this test, the load (F) is con-
senting half of the total height, which is produced with natural trolled, and the net deflection (d) is measured using a pair of LVDTs
aggregates and a cf equal to 0.75% in volume of the reinforced sustained by a steel frame, obtaining the F–d curve (see Fig. 6a).
layer. For homogeneous concretes, the relation h/H is equal to These results are used to determine the flexural strength at the first
1.00 since their whole volume is reinforced with steel fibres. Also, crack (f1), and the residual flexural strengths are corresponding to a
for simplification, all concrete mixes produced with only natural net deflection equal to L/600 (0.75 mm) and L/150 (3.0 mm) (f0.75
aggregates will be referred to as N-concretes, and concretes con- and f3.0, respectively). These parameters were calculated using
taining recycled aggregates, R-concretes. Eq. (3), which depends on the load (F), the span length (L), and
Prismatic samples with 150  150 mm cross-section and the width and the depth of the sample (B and D, respectively.
550 mm span length were produced. The casting procedure
described in BS EN 14651:2005 + A1:2007 [46] was followed for FL
f ¼ ð3Þ
the production of homogeneous concretes. Regarding the function- B  D2
ally graded concretes, they were produced according to the follow-
ing method: (1) the concrete of the bottom layer was mixed and 2.4.2. Compressive test
placed in the mould (see Fig. 5a); (2) then it was compacted for Compressive strength of concrete is the main parameter to be
20 s by mechanical vibration, and its height was verified after- considered for design and quality control of structures. Besides, it
wards (see Fig. 5b), and (3) the upper layer was cast about can be related to f1 values obtained from the bending test
30 min after finishing the previous layer, being vibrated for half [33,34]. Therefore, the compressive strength of the FGC was stud-
the time spent for the bottom layer (10 s) in order to avoid mixing ied. In that sense, two cubic samples (150  150  150 mm) were
the layers to each other (see Fig. 5c). This production method was extracted from each prism after assessing their flexural and resid-
used in other previous research and assured that the samples are ual flexural strengths. The extraction was done as follows: (1) two
well compacted, and the bonding between layers is enough to marks 50 mm distant from each end were made, then additional
guarantee that they work together [12]. marks were drawn every 150 mm (see Fig. 7a); (2) the prismatic

Fig. 4. FGC mixes regarding the bottom layer thickness/total thickness relation (h/H).

Fig. 5. FGC sample production: (a) casting the bottom layer; (b) verification of bottom layer thickness, and (c) casting the top layer.
 R. Chan et al. / Construction and Building Materials 242 (2020) 118186 5

Fig. 6. Four-point bending test (a) and compressive strength test (b).

Fig. 7. Extraction of cubes from prismatic samples: (a) marking; (b) cutting, and (c) selection.

samples were cut in the marked lines (see Fig. 7b), and (3) both behaviour. Considering that the content of fibre has no significant
ends and the central part were discarded, avoiding the influence influence in the modulus of elasticity [22,23,51,52], the higher
of the wall effect in the fibre orientation [49], and the influence modulus of elasticity of the non-cracked plain cement concrete,
of the crack produced in the bending test, respectively. In the in comparison to the recycled aggregate concrete [19,44,53],
end, one of the two remaining parts was selected for testing (see reduces the stresses transferred to the bottom layers of the sample
Fig. 7c). [54]. Thus, in the case of homogeneous FRRAC, which presents a
The procedure presented in BS EN 12390-3:2009 [50] was fol- lower modulus of elasticity, the concrete cracks under lower loads.
lowed for the surface preparation and the assessment of the com- Furthermore, the hardening post-cracking behaviour is more pre-
pressive strength of the specimens. For this test, four extracted sent in N-concretes, suggesting that this phenomenon is linked
cubic specimens (150  150  150 mm) per type of concrete were not only to the content of fibre but also to the compressive
tested at the age of 28 days. The samples were loaded according to strength of concrete [55]. In this sense, the lower compressive
the casting direction, as shown in Fig. 6b. For each tested sample, strength of R-concretes may decrease the pull-out strength, dimin-
the compressive strength (fcm) was calculated by Eq. (4), which ishing the post-cracking response [56–59], as evidenced in Fig. 9.
depends on the maximum compressive load at failure (F) and the Fig. 9a shows that FGC with FRC present straightened fibres
specimen’s cross-section area on which the compressive load acts whereas FGC with FRRAC present fibres not-straightened after
(Ac) [50]. the test (see Fig. 9b). This suggests that there is a different fibre-
matrix interaction in between FGC with FRC and FRRAC, and is
F based on the strength of the matrix. The stronger is the matrix,
f cm ¼ ð4Þ
Ac the more straightened is the fibre, as expected [56,57,59].
Table 3 presents the results of flexural strength at the first crack
(f1) and residual flexural strengths corresponding to a net deflec-
3. Results and discussion tion equal to 0.75 mm and 3.0 mm (f0.75 and f3.0, respectively).
The average results are expressed in MPa, and the coefficient of
3.1. Bending test variation is added in parenthesis.
Flexural strength (f1) results tend to present a lower coefficient
The average F–d curves obtained from the bending test for each of variation than the results for residual flexural strengths (f0.75 and
type of concrete are presented in Fig. 8. In general, it is observed f3.0), with average values of 6.93, 21.07 and 16.75%, respectively.
that an increase in h/H or cf results in higher post-cracking This tendency was expected [4,41]. Regarding the type of aggre-
response, being translated in softening post-cracking behaviour gate, in general, the results for R-concretes are systematically
in concretes with lower values of h/H and contents of fibre lower than those for N-concretes, when the same values of h/H
(h/H  0.50, cf  0.50%), and hardening behaviour for higher values and cf are considered. Therefore, the FGC studied could be used
of h/H and contents of fibre (h/H  0.75, cf  0.75%). This may hap- in structural applications requiring lower loading capacity, such
pen because the post-cracking response depends on the size of the as low traffic roads, car parks, cycling lines, or pedestrian
cracking area containing fibres, which increases with h/H and cf pavements.
[4,12,13,41,42]. The f1 results for the N-concretes present no significant change
An exception occurs when comparing the curves between dif- with h/H. However, R-concretes suggest decreasing f1 values with
ferent values of h/H for R0.50 concretes. In this case, when h/H is the increase of h/H. This may be explained by the dependence of
equal to 1.00, which refers to homogeneous concrete, even con- flexural strength on the properties of the concrete matrix rather
taining fibres in its whole volume, the post-cracking response falls than the content of fibre [4,41]. In this sense, increasing h/H in R-
back to a similar level when h/H is equal to 0.50. This fact high- concretes imply in reducing the elastic modulus, since the com-
lights the influence of the type of aggregate in the post-cracking pressive strength is also diminished, as discussed in Section 3.2,
6 R. Chan et al. / Construction and Building Materials 242 (2020) 118186

Fig. 8. F–d curves of concretes with different cf for h/H equal to (a) 0.25, (b) 0.50, (c) 0.75 and (d) 1.00.

Fig. 9. Pull-out behaviour of steel fibres in FGC with (a) FRC and (b) FRRAC.

and the elastic modulus and compressive strength are positively cracking response [4,12,13,41,42], especially for larger net deflec-
related [33,34]. Thus, an increase of h/H results in a reduction in tions, such as 3.0 mm, where the cracked area advances towards
f1 results of R-concretes. the top layer of the specimen. For R-concretes, two different trends
In the case of f0.75, for all concretes except for the ones with can be observed: a growing tendency with a maximum point
0.75% of cf, the parameter rises until h/H reaches the value of within the h/H range of 0.75–1.00 for mixes with 0.25 and 0.50
0.75, after this value, the results drop. For the R0.75 concrete, the of cf; and a growing tendency similar to the one presented by N-
residual flexural strength at a net deflection of 0.75 mm starts to concretes for the mix with 0.75 of cf.
fall after h/H hits 0.50, while the f0.75 results for the N0.75 concrete In the case showing a growing tendency with a maximum point,
do not decrease with increasing h/H. The increase of f0.75 with h/H the response may be due to the influence of the modulus of elastic-
is expected, since the area containing fibres is increased with h/H, ity. As explained previously, homogeneous FRRAC cracks under
enhancing the post-cracking response [4,12,13,41,42]. On the other lower loads than FGC with FRRAC because it presents a lower mod-
hand, the decrease of f0.75 after reaching a particular value of h/H ulus of elasticity than FRC [19,44,53]. Furthermore, an optimal h/H
may be explained by the random distribution of fibres throughout ratio is expected within the range of 0.75–1.00, in which a maxi-
the reinforced layer. In that sense, since the cracking area is con- mum f3.0 would be reached, showing the benefits of using FGC with
centrated in the bottom of the sample when the net deflection is FRRAC instead of homogeneous FRRAC. This optimized value of h/H
0.75 mm, the parameter f0,75 is conditioned by the number of fibres has to be assessed in future studies. In the case of the R0.75 con-
in the cracked area, which can fluctuate randomly [4,41]. crete, the trend presented by N-concretes is followed possibly
Regarding the values of f3.0, N-concretes show a growing ten- due to the higher cf, which may compensate for the lower modulus
dency for this parameter with increasing reinforced layer thick- of elasticity.
ness. This is expected since, as presented previously, the area Finally, regarding the effect of the content of fibre (cf) in the
containing fibres is augmented with h/H and enhances the post- flexural behaviour of concretes, its impact can be assessed consid-
 R. Chan et al. / Construction and Building Materials 242 (2020) 118186 7

Table 3 The average values presented for these concretes are 33.5, 32.1,
Results obtained using the bending test. 26.6 and 25.9 MPa for h/H equal to 0.25, 0.50, 0.75 and 1.00,
Concrete f1 (MPa) f0.75 (MPa) f3.0 (MPa) respectively. In that sense, for the relation equal to 1.00, the fcm
0.25N0.25 4.92 (11.40%) 2.52 (61.54%) 1.35 (45.80%) of R-concretes is equivalent to 53% of the fcm of N-concretes. Nev-
0.25N0.50 5.36 (6.86%) 3.35 (11.48%) 1.66 (16.73%) ertheless, for other cases, where h/H is equal to 0.75, 0.50 or 0.25,
0.25N0.75 5.14 (7.68%) 3.42 (38.22%) 2.11 (38.40%) the difference is less than 30%. The reduction of compressive
0.50N0.25 5.31 (4.80%) 2.65 (4.97%) 1.68 (2.35%) strength of R-concretes with higher values of h/H is expected, since
0.50N0.50 5.44 (10.63%) 5.21 (7.88%) 3.03 (21.94%)
0.50N0.75 5.81 (8.48%) 5.61 (7.75%) 4.97 (16.15%)
the content of recycled aggregate increases with h/H [19,43,44].
0.75N0.25 5.67 (9.94%) 4.65 (28.52%) 2.63 (22.54%) Furthermore, regarding the effect of the content of fibre (cf), its
0.75N0.50 5.60 (6.77%) 5.51 (6.76%) 4.57 (10.56%) influence in the compressive strength of the concretes tested in
0.75N0.75 5.81 (6.63%) 6.23 (8.46%) 5.55 (14.87%) this study is not significant, as expected [29,52,60].
1.00N0.25 3.64 (4.29%) 3.03 (10.16%) 2.86 (15.95%)
Finally, since a compressive test is more convenient to perform
1.00N0.50 4.73 (3.02%) 4.74 (7.41%) 5.75 (21.79%)
1.00N0.75 6.18 (8.20%) 6.89 (4.66%) 7.44 (14.37%) than a bending test, the relation between f1 and fcm is commonly
0.25R0.25 5.10 (7.34%) 1.28 (32.39%) 0.74 (28.50%) obtained for conventional concretes to facilitate the quality control
0.25R0.50 5.01 (9.15%) 1.47 (8.54%) 0.91 (10.92%) of concrete structures under bending [33,34]. Thus, the relation
0.25R0.75 5.06 (14.83%) 4.05 (35.55%) 1.79 (9.17%) between the values of f1 and fcm obtained in this study is presented
0.50R0.25 3.13 (2.35%) 2.22 (35.46%) 1.46 (16.37%)
in Table 5. An average value of 13.1% and 13.7% is found for N-
0.50R0.50 4.34 (2.04%) 3.05 (9.41%) 2.79 (11.00%)
0.50R0.75 4.86 (3.19%) 4.59 (9.01%) 3.21 (16.50%) concretes and R-concretes, respectively. These values are similar
0.75R0.25 4.11 (6.33%) 2.72 (8.24%) 1.91 (19.13%) to the relation found for other concretes [33]. Thus, a relation
0.75R0.50 3.60 (15.46%) 3.74 (14.97%) 4.10 (12.15%) between f1 and fcm can also be experimentally stablished and used
0.75R0.75 3.84 (13.69%) 4.03 (22.49%) 4.38 (11.76%)
for the quality control of FGC.
1.00R0.25 2.41 (17.20%) 1.61 (10.86%) 1.95 (8.40%)
1.00R0.50 3.02 (8.80%) 2.54 (8.36%) 3.15 (11.87%)
1.00R0.75 4.13 (12.95%) 3.89 (4.13%) 4.70 (7.71%)
4. Simplified sustainability analysis

ering two stages: before and after the first crack. Before the first A sustainability analysis was performed to compare the embod-
crack takes place, the flexural strength of concretes is not strongly ied CO2 and costs between the concrete mixes produced and
affected by cf. However, after the first crack, it is observed that all assessed in this study. In this comparative analysis, the different
the results are enhanced with the increment of cf, as expected concrete mixes are considered to be used for the same applications
[4,12,41,42]. and building processes. Therefore, the simplified methodology
used by Cáceres et al. [61] was adopted, which focus on the impact
of CO2 emissions and costs of the production of materials, rather
3.2. Compressive test their application. In that sense, several processes are not taken into
account, such as construction, maintenance, and demolishing. Also,
The results obtained in the compressive test are presented in once the quantity of cement used for each concrete mix is the
Table 4. The mean compressive strength (fcm) and the coefficient same, only the quantities of natural aggregates, recycled aggre-
of variation (presented in parenthesis) are calculated for each type gates and steel fibres were considered.
of concrete. Table 6 presents the embodied CO2 (EC) and the cost, per ton of
N-concretes present higher compressive strength than R- material, for natural fine and coarse aggregates (NFA and NCA,
concretes, as expected [19,43,44], independently of cf and h/H val- respectively), recycled fine and coarse aggregates (RFA and RCA,
ues. This difference is probably caused by the lower density and respectively) and steel fibres (SF). The cost values are the same val-
higher water absorption of recycled aggregates [33,34]. On the ues used in the analysis presented by Liu et al. [12], which were
other hand, the use of recycled aggregates did not have a visible given by the suppliers. Meanwhile, the embodied CO2 values for
impact in the variation of results, since the average coefficient of
variation (CV) for N-concretes is 5.22% and for R-concretes is 5.50%.
The compressive strength results of N-concretes are not Table 5
affected by the variation h/H since both layers are produced with Relation between f1 and fcm.

natural aggregates. The average values of fcm for these concretes Concrete f1/fcm Concrete f1/fcm
and values of h/H of 0.25, 0.50, 0.75 and 1.00 are 38.5, 42.5, 37.9 0.25N0.25 14.0% 0.25R0.25 16.1%
and 48.4 MPa, respectively. On the other hand, fcm values for 0.25N0.50 11.3% 0.25R0.50 12.4%
R-concretes tend to decrease with higher values of h/H. 0.25N0.75 15.6% 0.25R0.75 17.1%
0.50N0.25 12.1% 0.50R0.25 10.2%
0.50N0.50 14.7% 0.50R0.50 13.2%
Table 4 0.50N0.75 12.5% 0.50R0.75 14.8%
Compressive strength test results. 0.75N0.25 17.7% 0.75R0.25 15.5%
0.75N0.50 11.2% 0.75R0.50 11.9%
Concrete fcm (MPa) Concrete fcm (MPa) 0.75N0.75 18.3% 0.75R0.75 16.6%
1.00N0.25 8.7% 1.00R0.25 9.8%
0.25N0.25 35.2 (3.99%) 0.25R0.25 31.6 (3.80%)
1.00N0.50 9.7% 1.00R0.50 10.7%
0.25N0.50 47.3 (5.79%) 0.25R0.50 39.4 (8.86%)
1.00N0.75 11.3% 1.00R0.75 16.0%
0.25N0.75 32.9 (8.88%) 0.25R0.75 29.6 (2.83%)
0.50N0.25 44.1 (5.77%) 0.50R0.25 30.6 (7.54%)
0.50N0.50 36.9 (9.23%) 0.50R0.50 32.8 (4.79%)
0.50N0.75 46.4 (2.90%) 0.50R0.75 32.9 (3.41%)
0.75N0.25 32.0 (4.82%) 0.75R0.25 26.6 (11.47%) Table 6
0.75N0.50 50.1 (3.24%) 0.75R0.50 30.1 (2.37%) Unit embodied CO2 and cost of materials.
0.75N0.75 31.7 (5.51%) 0.75R0.75 23.1 (5.57%)
Parameter NFA NCA RFA RCA SF
1.00N0.25 41.7 (6.77%) 1.00R0.25 24.7 (5.10%)
1.00N0.50 48.7 (4.35%) 1.00R0.50 27.2 (5.80%) EC (kgCO2/ton) 23.00 32.00 12.00 11.00 2680.00
1.00N0.75 54.9 (1.44%) 1.00R0.75 25.8 (4.49%) Cost ($/ton) 28.57 38.10 6.15 7.99 1714.25
8 R. Chan et al. / Construction and Building Materials 242 (2020) 118186

aggregates and steel fibres were obtained from studies published only variable changed is the type of aggregate and, as presented
by Hossain et al. [62] and Stengel and Schießl [63], respectively. in Table 6, recycled aggregates are less polluting and also cheaper
The total EC and costs (in parenthesis) values to produce one than natural aggregates. Furthermore, the difference in both total
cubic meter of each concrete mix, along with the values for each EC and costs values between R-concretes and N-concretes
material, are presented in Table 7. Besides, Fig. 10a shows the con- increases with increasing h/H, while it decreases with increasing
tribution of each material in the embodied CO2 of the concrete cf. This behaviour indicates that the influence of the total amount
mixes studied, whereas Fig. 10b shows the contribution of each of steel fibres, in terms of embodied CO2 and costs, is more signif-
material regarding the cost of one cubic meter of concrete. icant than replacement ratio of natural aggregates for recycled
From Table 7, it can be observed that, for the same h/H and cf, aggregates, which is illustrated in Fig. 10.
the total EC and costs values for R-concretes mixes are always When homogenous fibre-reinforced concrete mixes are consid-
lower than the values for N-concretes. This is expected since the ered, the effect of the steel fibres in the total EC can be as high as

Table 7
Embodied CO2 and costs (in parenthesis) to produce one m3 of concrete.

Concrete NFA NCA RFA RCA SF Total

0.25N0.25 18.17 (22.57) 28.48 (33.91) – (–) – (–) 13.07 (8.36) 59.72 (64.84)
0.25N0.50 18.17 (22.57) 28.48 (33.91) – (–) – (–) 26.13 (16.71) 72.78 (73.19)
0.25N0.75 18.17 (22.57) 28.48 (33.91) – (–) – (–) 39.20 (25.07) 85.85 (81.55)
0.50N0.25 18.17 (22.57) 28.48 (33.91) – (–) – (–) 26.13 (16.71) 72.78 (73.19)
0.50N0.50 18.17 (22.57) 28.48 (33.91) – (–) – (–) 52.26 (33.43) 98.91 (89.91)
0.50N0.75 18.17 (22.57) 28.48 (33.91) – (–) – (–) 78.39 (50.14) 125.04 (106.62)
0.75N0.25 18.17 (22.57) 28.48 (33.91) – (–) – (–) 39.20 (25.07) 85.85 (81.55)
0.75N0.50 18.17 (22.57) 28.48 (33.91) – (–) – (–) 78.39 (50.14) 125.04 (106.62)
0.75N0.75 18.17 (22.57) 28.48 (33.91) – (–) – (–) 117.59 (75.21) 164.24 (131.69)
1.00N0.25 18.17 (22.57) 28.48 (33.91) – (–) – (–) 52.26 (33.43) 98.91 (89.91)
1.00N0.50 18.17 (22.57) 28.48 (33.91) – (–) – (–) 104.52 (66.86) 151.17 (123.34)
1.00N0.75 18.17 (22.57) 28.48 (33.91) – (–) – (–) 156.78 (100.28) 203.43 (156.76)
0.25R0.25 13.63 (16.93) 21.36 (25.43) 2.42 (1.24) 1.97 (1.43) 13.07 (8.36) 52.43 (53.38)
0.25R0.50 13.63 (16.93) 21.36 (25.43) 2.42 (1.24) 1.97 (1.43) 26.13 (16.71) 65.50 (61.74)
0.25R0.75 13.63 (16.93) 21.36 (25.43) 2.42 (1.24) 1.97 (1.43) 39.20 (25.07) 78.56 (70.10)
0.50R0.25 9.09 (11.28) 14.24 (16.96) 4.83 (2.47) 3.93 (2.85) 26.13 (16.71) 58.22 (50.28)
0.50R0.50 9.09 (11.28) 14.24 (16.96) 4.83 (2.47) 3.93 (2.85) 52.26 (33.43) 84.35 (67.00)
0.50R0.75 9.09 (11.28) 14.24 (16.96) 4.83 (2.47) 3.93 (2.85) 78.39 (50.14) 110.48 (83.71)
0.75R0.25 4.54 (5.64) 7.12 (8.48) 7.25 (3.71) 5.90 (4.28) 39.20 (25.07) 64.00 (47.19)
0.75R0.50 4.54 (5.64) 7.12 (8.48) 7.25 (3.71) 5.90 (4.28) 78.39 (50.14) 103.20 (72.26)
0.75R0.75 4.54 (5.64) 7.12 (8.48) 7.25 (3.71) 5.90 (4.28) 117.59 (75.21) 142.39 (97.33)
1.00R0.25 – (–) – (–) 9.66 (4.95) 7.87 (5.71) 52.26 (33.43) 69.79 (44.09)
1.00R0.50 – (–) – (–) 9.66 (4.95) 7.87 (5.71) 104.52 (66.86) 122.05 (77.52)
1.00R0.75 – (–) – (–) 9.66 (4.95) 7.87 (5.71) 156.78 (100.28) 174.31 (110.94)

(a)

(b)
Fig. 10. Contribution of each material in (a) embodied CO2 and (b) costs.
 R. Chan et al. / Construction and Building Materials 242 (2020) 118186 9

77% and 90%, for N-concretes and R-concretes, respectively. Regard- of fibre (cf). In that sense, in order to identify possible correlations
ing the total costs, steel fibres can contribute as much as 64% and between these parameters with other variables, a linear correlation
90%, for N-concretes and R-concretes, respectively. Therefore, a analysis among experimental results for N-concretes and R-
cut in the total volume of steel fibres, due to a reduction of h/H or concretes was run, and the results are presented in Table 8. These
cf, is substantially significant to minimize the total EC and costs of are the linear correlation coefficients (r) and the p-values between
concrete. Although the replacement ratio of natural aggregates for each pair of parameters. Note that, for this study, it is considered
recycled aggregates does not affect so much as the content of steel that two variables are correlated when r  0.50 and that the corre-
fibres, the role played by the recycled aggregates cannot be lation is significant when p-value  0.05.
neglected, specifically for higher values of h/H, when the reduction Regarding N-concretes results, the parameters f0.75 has linear
in EC and cost can be as high as 29% and 51%, respectively. correlations with h/H, cf and f1, but the correlation is only significant
with cf and f1. This may be an indication that, in the initial moments
5. Correlations following the first crack, content of fibre and flexural strength play a
major role in the residual flexural strength at small net deflections,
The design of fibre-reinforced concrete elements is based on the for example 0.75 mm, which is expected, since the cracked area is
post-cracking parameters: f0.75 and f3.0, since the concretes were still small and concentrated in the bottom layer of the sample. On
characterized using the standard ASTM C 1609/C 1609M-06 [48]. the other hand, when considering larger net deflections, such as
According to ACI 544.4R [64], the parameter f0.75 is commonly used 3.0 mm, the variable f3.0 is significantly linearly correlated with
for designs based on service limit state (SLS), where the crack width h/H and cf. This behaviour can be explained by the higher depen-
and deflection control are essential requirements. For designs based dence of flexural residual strength to the size of cracked area rein-
on the ultimate limit state (ULS), where the failure strength is forced with fibres than to the strength of concrete. Considering the
focused, the parameter f3.0 is used. Apart from that, the considera- parameters EC and Cost, they present a good and significant correla-
tion of sustainability parameters, such as EC and Cost, in the selec- tion with h/H and cf, which is logical considering that consumption
tion of the material for a specific application is essential. of steel fibres is the only variable affecting them.
In that sense, the determination of equations correlating design Regarding R-concretes, the correlation analysis brings results
and sustainability parameters with other variables used in this that differ substantially from those obtained for N-concretes, par-
study, such as type of aggregate, reinforced layer thickness, and ticularly for the variable f0.75, which only presents a significant lin-
content of fibre, has great importance. Therefore, the following sta- ear correlation with the content of fibre. This response is
tistical analysis processes were used: first, a linear correlation predictable, considering that f0.75 results increase until a particular
analysis; second, a multiple linear regression analysis; and third, value of h/H, after which they decrease. In that sense, a non-linear
a nonlinear regression analysis. The linear correlation analysis relationship can be more adequate between f0.75 and h/H. Regard-
was initially carried out in order to sort out the main parameters ing f3.0, this parameter is linearly correlated to h/H, cf and f1, though
that are most likely to be correlated to the post-cracking parame- having a significant correlation specifically with h/H and cf. Simi-
ters. Next, using the parameters selected from the first analysis, a larly to the case for N-concretes, the correlations between f3.0,
multiple linear regression analysis was performed to identify the h/H and cf for R-concretes can also be explained by the reliance
compatibility of using a linear equation to predict results for the of residual flexural strength to the size of cracked area reinforced
parameters f0.75, f3.0, EC and Cost. Finally, a nonlinear regression with fibres. As for the sustainability parameters, on the one hand,
analysis was done to verify the accuracy of nonlinear equations a good linear correlation is obtained between EC, h/H, and cf, but
in comparison to linear equations. on the other hand, Cost is only linearly correlated to cf. Note that
this behaviour can be modified by the unit embodied CO2 and cost
adopted for each material.
5.1. Linear correlation analysis

From the experimental results presented and discussed in sec- 5.2. Multiple linear regression
tions 3.1 and 4, it was verified that the parameters f0.75, f3.0, EC
and Cost are strongly affected by the relation between reinforced From the previous section, it was observed that the variables
layer thickness and total thickness (h/H), and also by the content f0.75 and f3.0 presented a good and significant linear correlation

Table 8
Correlation coefficients and correlation significance among the experimental results.

N-concretes
Parameters h/H cf fcm f1
f0.75 r 0.507 0.678 0.285 0.672
p-value 0.093 0.015 0.369 0.017
f3.0 r 0.710 0.626 0.498 0.412
p-value 0.010 0.029 0.099 0.184
EC r 0.707 0.646 0.453 0.404
p-value 0.010 0.023 0.139 0.192
Cost r 0.707 0.646 0.453 0.404
p-value 0.010 0.023 0.139 0.192
R-concretes
Parameters h/H cf fcm f1
f0.75 r 0.152 0.823 0.269 0.187
p-value 0.638 0.001 0.398 0.560
f3.0 r 0.631 0.641 0.512 0.223
p-value 0.028 0.025 0.089 0.485
EC r 0.585 0.740 0.461 0.114
p-value 0.046 0.006 0.132 0.724
Cost r 0.300 0.871 0.253 0.164
p-value 0.343 0.000 0.428 0.610
10 R. Chan et al. / Construction and Building Materials 242 (2020) 118186

with cf for both FGC, with natural and recycled aggregates. How- correlation between h/H and f0.75 is not significant. Hence, these
ever, the parameter h/H is only significantly related to f3.0 indepen- results suggest that Eq. (5) in conjunction with the coefficients pre-
dently of the type of concrete. Apart from that, the variable f1 can sented in Table 9 can be a useful tool for the estimation of embod-
be correlated exclusively with f0.75 for N-concretes. Taking into ied CO2 and cost of the concretes produced in this study.
account that the parameters h/H and cf have the best correlation Nevertheless, the equation and coefficients shown in this section
with f0.75 and f3.0, a multiple linear regression (MLR) analysis was are only valid for the experimental dataset obtained in this study,
performed considering the groups of parameters for N-concretes thus further adjustments may be required when using results from
and R-concretes, separately, as follows: (1) h/H, cf and f0.75; (2) other studies.
h/H, cf and f3.0; (3) h/H, cf and EC; and (4) h/H, cf and Cost. The Finally, the relationship between experimental results (f0.75, f3.0,
MLR of the groups as mentioned earlier is represented by Eq. (5), EC and Cost) and estimated results (f0.750 , f3.00 , EC0 and Cost’) calcu-
where: Y is f0.75, f3.0, EC or Cost; and A, B are the coefficients related lated using Eq. (5) are presented in Figs. 11 and 12, for N-
to h/H and cf, respectively. concretes and R-concretes, respectively. The graphs illustrate the
goodness of the equation through R-square values from linear cor-
Y ¼ A  h=H þ B  cf ð5Þ relations between experimental and estimated results for the
parameters as mentioned earlier.
The results obtained from the multiple linear regression analy-
sis are the R-square (R2), the coefficients A and B, and their respec-
5.3. Nonlinear regression
tive p-value (in parenthesis). These results are presented in Table 9.
From Table 9, it is observed that the multiple linear regression
As pointed out in Section 5.2, a linear correlation between the
for f0.75, f3.0, EC and Cost, present an R2 above 0.95, indicating a good
parameters h/H and cf, with f0.75 and f3.0 is not suitable. Therefore,
linear relationship between them and the parameters h/H and cf.
a nonlinear regression (NLR) analysis was carried out using the
However, the coefficient A, related to the parameter h/H, for f0.75
curve fitting software based on the Levenberg-Marquardt algo-
of R-concretes has a p-value of 0.139, indicating that the linear
rithm, named LAB Fit [65], considering the same three groups
assessed in the multiple linear regression analysis. The best fitting
Table 9 obtained is represented by Eq. (6), which R2, and coefficients A, B,
Multiple linear regression analysis results correlating h/H and cf, with f0.75, f3.0, EC, and and C, with the respective p-value in parenthesis, are shown in
Cost.
Table 10. The NLR analysis presents better results than the MRL
Concrete Y R2 A B analysis considering the values of R2 and p-value, indicating that
N f0.75 0.97 2.92 (1.28E3) 5.22 (9.53E5) h/H and cf are nonlinearly correlated to f0.75 and f3.0.
f3.0 0.95 3.28 (1.60E3) 3.53 (4.59E3) 1
EC 0.99 92.07 (1.01E5) 111.98 (1.45E5) Y ¼ A  Bh=H  C cf ð6Þ
Cost 0.99 76.65 (1.37E6) 98.26 (1.24E6)
Finally, Figs. 13 and 14 present the relationship between exper-
R f0.75 0.96 0.84 (1.39E1) 4.75 (2.97E5)
f3.0 0.95 2.06 (4.88E3) 2.79 (3.29E3) imental and estimated results for the parameters f0.75 and f3.0 for
EC 0.98 62.95 (2.39E4) 111.98 (1.45E5) N-concretes and R-concretes, respectively. The scatter between
Cost 0.99 30.83 (2.19E3) 98.26 (1.24E6) experimental data and the results calculated using Eq. (6) and
the coefficients presented in Table 10, shown in Figs. 13 and 14,

Fig. 11. Relationship between experimental and estimated results (abscissa and ordinate axis, respectively) from the multiple linear regression for N-concretes, correlating h/
H and cf, with: (a) f0.75; (b) f3.0; (c) EC; and (d) Cost.
 R. Chan et al. / Construction and Building Materials 242 (2020) 118186 11

Fig. 12. Relationship between experimental and estimated results (abscissa and ordinate axis, respectively) from the multiple linear regression for R-concretes, correlating h/
H and cf, with: (a) f0.75; (b) f3.0; (c) EC; and (d) Cost.

Table 10
Nonlinear regression analysis results correlating h/H and cf, with f0.75 and f3.0.

Concrete Y R2 A B C
N f0.75 0.99 9.21 (3.59E6) 0.84 (1.83E9) 0.51 (6.10E4)
f3.0 0.99 13.62 (7.38E7) 0.66 (4.16E9) 0.79 (5.26E6)
R f0.75 0.97 5.68 (1.38E4) 0.93 (3.32E8) 0.72 (2.60E3)
f3.0 0.99 8.43 (9.84E6) 0.71 (3.35E8) 0.75 (1.16E4)

Fig. 13. Relationship between experimental and estimated results (abscissa and ordinate axis, respectively) from the nonlinear linear regression for N-concretes, correlating
h/H and cf, with (a) f0.75 and (b) f3.0.

is smaller than the scatter shown in Figs. 11 and 12, as observed by 6. Conclusions
the higher values of R2 obtained from the linear correlations for
N-concretes and R-concretes, respectively. Therefore, the empirical The parametric study between design and sustainability param-
correlations have a good potential to optimize, in terms of steel eters, focusing on the flexural behaviour of FGC incorporating recy-
fibre and natural aggregates consumption, the design of structural cled aggregates and steel fibres, was assessed through: an
elements produced with functionally graded concretes incorporat- experimental program, where the mechanical properties were
ing steel fibres and recycled aggregates. However, notice that the evaluated; a simplified sustainability analysis; and a parametric
equation and coefficients presented in this section are based on study correlating the results obtained from the experimental pro-
experimental data obtained in this study. Thus, proper adjustments gram and sustainability analysis with variables such as type of
in the equation and coefficients may be required when using aggregate, reinforced layer thickness and content of fibre. In sum-
different datasets. mary, the experimental results indicate that the FGC studied could
12 R. Chan et al. / Construction and Building Materials 242 (2020) 118186

Fig. 14. Relationship between experimental and estimated results (abscissa and ordinate axis, respectively) from the nonlinear linear regression for R-concretes, correlating
h/H and cf, with (a) f0.75 and (b) f3.0.

be used in structural applications requiring lower loading capacity
Although the replacement ratio of natural aggregates for recy-
such as low traffic roads, car parks, cycling lines, or pedestrian cled aggregates does not affect so much as the content of steel
pavements. Furthermore, the estimated results suggest that the fibres, in terms of EC and cost, the role played by recycled aggre-
equations obtained from the parametric study can be useful tools gates cannot be neglected, specifically for higher values of h/H,
for the optimization of the design of structural elements produced when the reduction in EC and cost can be as great as 29% and
with FGC using FRRAC, considering sustainability factors such as 51%, respectively;
embodied CO2 (EC) and costs, and design parameters such as h/H
From the parametric study was concluded that design parame-
and cf. Apart from that, the following conclusions are extracted ters (f0.75 and f3.0) and sustainability parameters (embodied CO2
from this study: and cost) present a good correlation with the variables h/H and
cf. A multiple linear regression approach showed to be adequate

Under a bending load, the effect of h/H is different on flexural for correlations considering the sustainability parameters. On
strength (f1), according to the type of aggregate used in the rein- the other hand, better correlations were obtained for design
forced layer. For N-concretes, h/H has little influence in f1, while parameters using a nonlinear regression approach, and
increasing h/H results in lower f1 in R-concretes. As for the influ-
Empirical equations correlating design parameters (f0.75 and f3.0)
ence of cf, no significant change was detected in f1 results when and sustainability parameters (EC and cost) with the variables
cf was increased from 0.25 to 0.75%, for both N-concretes and R- h/H and cf were proposed. These may be useful tools for the
concretes; optimization of the design process of structural elements pro-

The post-cracking behaviour of FGC reinforced with steel fibres duced with FGC.
under bending is dependent on thickness and content of fibre of
the reinforced layer. A softening post-cracking behaviour is pre-
sented in concretes with lower values of h/H and contents of
Declaration of Competing Interest
fibre (h/H  0.50, cf  0.50%), and a hardening behaviour can
be identified for higher values of h/H and contents of fibre
The authors declare that they have no known competing finan-
(h/H  0.75, cf  0.75%);
cial interests or personal relationships that could have appeared

All concretes tested showed enhanced residual flexural strength
to influence the work reported in this paper.
when increasing cf. However, the increase of h/H only enhances
these parameters for N-concretes. For R-concretes, residual flex-
ural strength increases for h/H up to 0.75, then drops when h/H Acknowledgments
is equal to 1.00, reaching similar values for h/H equal to 0.50. In
that sense, further studies are needed to understand the beha- The authors would like to thank the Xi’an Jiaotong-Liverpool
viour of residual flexural strength for R-concretes with h/H University (XJTLU) Research Development Fund for the financial
between 0.75 and 1.00, in order to assess an optimum h/H support received from the project with reference RDF-16-02-42.
value;

Compressive strength of FGC is highly dependent on the type of References
aggregate, while the content of fibre (cf) indicated to not affect
this property. For values of h/H under 0.75, R-concretes pre- [1] CSI, Recycling Concrete, 2009. www.wbcsdcement.org.
sented values of compressive strength (fcm) up to 30% lower [2] IEA, CSI, Technology Roadmap: Low-Carbon Transition in the Cement Industry,
2018. www.wbcsdcement.org.
than the corresponding N-concretes. On the other hand, for [3] J. de Brito, N. Saikia, Recycled Aggregate in Concrete, Springer, London, 2013.
h/H = 1.00, the reduction in fcm is increased to 53%, between doi: 10.1007/978-1-4471-4540-0.
N-concretes and R-concretes; [4] A. Bentur, S. Mindess, Fibre Reinforced Cementitious Composites, 2nd ed.,
Taylor & Francis, London and New York, 2007.

The values of f1 are, in average, equivalent to 13% of the values [5] M.A. Kamel, Quantification of benefits of steel fiber reinforcement for rigid
fcm in the FGC studied. Thus, a relation between f1 and fcm can be pavement, Am. J. Civ. Eng. Archit. 4 (2016) 189–198, https://doi.org/10.12691/
experimentally stablished and used for the quality control of AJCEA-4-6-1.
[6] C. Achilleos, D. Hadjimitsis, K. Neocleous, K. Pilakoutas, P.O. Neophytou, S.
FGC;
Kallis, Proportioning of steel fibre reinforced concrete mixes for pavement

The sustainability analysis indicates that a cut in the total vol- construction and their impact on environment and cost, Sustainability 3
ume of steel fibres, due to a reduction of h/H or cf, is substan- (2011) 965–983, https://doi.org/10.3390/su3070965.
tially significant to minimize the total EC and cost of concrete, [7] A.G. Graeff, K. Pilakoutas, K. Neocleous, M.V.N.N. Peres, Fatigue resistance and
cracking mechanism of concrete pavements reinforced with recycled steel
since steel fibres can represent up to 90% of the total EC and fibres recovered from post-consumer tyres, Eng. Struct. 45 (2012) 385–395,
cost; https://doi.org/10.1016/j.engstruct.2012.06.030.
 R. Chan et al. / Construction and Building Materials 242 (2020) 118186 13

[8] W.A.M.H. Elsaigh, E.P. Kearsley, J.M. Robberts, Steel fibre reinforced concrete [34] A.M. Neville, Properties of Concrete, 5th ed., Pearson Education Limited,
for road pavement applications, in: 24th Annu. South. African Transp. Conf. Harlow, England, 2011.
(SATC 2005) Transp. Challenges 2010, University of Pretoria, 2005, pp. 191– [35] BSI, BS EN 197-1:2011, Cement – Composition, specifications and conformity
201. criteria for common cements, BSI Standards Limited, London, England, 2011.
[9] Y. Miyamoto, W.A. Kaysser, B.H. Rabin, A. Kawasaki, R.G. Ford, Functionally doi: 10.3403/30205527.
Graded Materials, Springer, US, Boston MA, 1999. doi: 10.1007/978-1-4615- [36] BSI, BS 812-103.1:1985, Testing aggregates - Method for determination of
5301-4. particle size distribution - Sieve tests, BSI Standards Limited, London, England,
[10] B. Shen, M. Hubler, G.H. Paulino, L.J. Struble, Functionally-graded fiber- 1998. doi:10.3403/00139627.
reinforced cement composite: Processing, microstructure, and properties, [37] BSI, BS 882:1992, Specification for aggregates from natural sources for
Cem. Concr. Compos. 30 (2008) 663–673, https://doi.org/10.1016/j. concrete, third ed., BSI Standards Limited, London, England, 2002. doi:10.3403/
cemconcomp.2008.02.002. 02522741.
[11] J. Roesler, G. Paulino, C. Gaedicke, A. Bordelon, K. Park, Fracture behavior of [38] BSI, BS EN 1097-6:2013, Tests for Mechanical and Physical Properties of
functionally graded concrete materials for rigid pavements, Transp. Res. Rec. J. Aggregates – Determination of Particle Density and Water Absorption, BSI
Transp. Res. Board. 2037 (2007) 40–49, https://doi.org/10.3141/2037-04. Standards LimitedLondon, England, 2013. doi:10.3403/30218643.
[12] X. Liu, M. Yan, I. Galobardes, K. Sikora, Assessing the potential of functionally [39] R.D. Holtz, W.D. Kovacs, An Introduction to Geotechnical Engineering,
graded concrete using fibre reinforced and recycled aggregate concrete, Prentice-Hall, New Jersey, United States, 1981.
Constr. Build. Mater. 171 (2018) 793–801, https://doi.org/10.1016/ [40] BSI, BS EN 14889-1:2006, Fibres for Concrete – Steel Fibres – Definitions,
j.conbuildmat.2018.03.202. Specifications and Conformity, BSI Standards Limited, London, England, 2006.
[13] R.S. Ravindrarajah, C.T. Tam, Flexural strength of steel fibre reinforced concrete doi:10.3403/30110332.
beams, Int. J. Cem. Compos. Light. Concr. 6 (1984) 273–278, https://doi.org/ [41] A.D. de Figueiredo, N.L. Nunes, J. Tanesi, Mix design analysis on steel fiber
10.1016/0262-5075(84)90022-8. reinforced concrete, in: P. Rossi, G. Chanvillard (Eds.), 5th Int. RILEM Symp.
[14] A.D. Buck, Recycled concrete as a source of aggregate, ACI J. Proc. 74 (1977) Fibre-Reinforced Concr., RILEM Publications SARL, 2000: pp. 103–118.
212–219, https://doi.org/10.14359/11004. [42] R.V. Balendran, F.P. Zhou, A. Nadeem, A.Y.T. Leung, Influence of steel fibres on
[15] P.J. Nixon, The use of materials from demolition in construction, Resour. Policy strength and ductility of normal and lightweight high strength concrete, Build.
2 (1976) 276–283, https://doi.org/10.1016/0301-4207(76)90082-9. Environ. 37 (2002) 1361–1367, https://doi.org/10.1016/S0360-1323(01)
[16] P.J. Nixon, Recycled concrete as an aggregate for concrete - a review, Matériaux 00109-3.
Constr. 11 (1978) 371–378, https://doi.org/10.1007/BF02473878. [43] R.V. Silva, J. de Brito, R.K. Dhir, The influence of the use of recycled aggregates
[17] T.C. Hansen, Recycled aggregates and recycled aggregate concrete second on the compressive strength of concrete: a review, Eur. J. Environ. Civ. Eng. 19
state-of-the-art report developments 1945–1985, Mater. Struct. 19 (1986) (2015) 825–849, https://doi.org/10.1080/19648189.2014.974831.
201–246, https://doi.org/10.1007/BF02472036. [44] J. Xiao, J. Li, C. Zhang, Mechanical properties of recycled aggregate concrete
[18] R.K. Dhir, K.A. Paine, J.M.C.L. de Brito, M.L. Etxeverria, N.Y. Ho, C.S. Poon, V.W.Y. under uniaxial loading, Cem. Concr. Res. 35 (2005) 1187–1194, https://doi.org/
Tam, Use of recycled and secondary aggregates in concrete: an overview, in: 10.1016/j.cemconres.2004.09.020.
1st UKIERI Concr. Congr. Concr. 21st Century Constr., Shroff Pub & Dist. Pvt. [45] D.C. Teychenné, R.E. Franklin, H.C. Erntroy, Design of Normal Concrete Mixes,
Ltd, 2011: pp. 157–182. doi: 10.13140/RG.2.1.4378.2166. second ed., Construction Research Communications Ltd, Watford, England,
[19] K. McNeil, T.H.-K. Kang, Recycled concrete aggregates: a review, Int. J. Concr. 1997.
Struct. Mater. 7 (2013) 61–69, https://doi.org/10.1007/s40069-013-0032-5. [46] BSI, BS EN 14651:2005+A1:2007, Test Method for Metallic Fibred Concrete –
[20] B. González-Fonteboa, S. Seara-Paz, J. De Brito, I. González-Taboada, F. Measuring the Flexural Tensile Strength (Limit of Proportionality (LOP),
Martínez-Abella, R. Vasco-Silva, Recycled concrete with coarse recycled Residual), BSI Standards Limited, London, England, 2008. doi:10.3403/
aggregate. An overview and analysis, Mater. Construcción 68 (2018) 151, 30092475.
https://doi.org/10.3989/mc.2018.13317. [47] BSI, BS EN 12390-2:2009, Testing Hardened Concrete – Making and Curing
[21] V. Vytlačilová, Fibre Concrete With Recycled Aggregate - Masonry and Specimens for Strength Tests, BSI Standards Limited, London, England, 2009.
Concrete, in: 35th Conf. Our World Concr. Struct., CI-Premier PTE LTD, 2010. doi:10.3403/30164903.
http://cipremier.com/100035058. [48] ASTM C 1609/C 1609M–06, Standard test method for flexural toughness and
[22] V. Vytlačilová, The fibre reinforced concrete with using recycled aggregates, first-crack strength of fibre-reinforced concrete (Using beam with third-point
Int. J. Syst. Appl. Eng. Dev. 5 (2011) 359–366. http://www.universitypress.org. loading), ASTM International, West Conshohocken, United States, 2006. doi:
uk/journals/saed/20-229.pdf. 10.1520/C1609_C1609M-06.
[23] J.A. Carneiro, P.R.L. Lima, M.B. Leite, R.D. Toledo Filho, Compressive stress– [49] E. Galeote, A. Blanco, S.H.P. Cavalaro, A. de la Fuente, Correlation between the
strain behavior of steel fiber reinforced-recycled aggregate concrete, Cem. Barcelona test and the bending test in fibre reinforced concrete, Constr. Build.
Concr. Compos. 46 (2014) 65–72, https://doi.org/10.1016/j. Mater. 152 (2017) 529–538, https://doi.org/10.1016/j.conbuildmat.2017.07.028.
cemconcomp.2013.11.006. [50] BSI, BS EN 12390-3:2009, Testing hardened concrete - Compressive strength of
[24] V.W.Y. Tam, O. Mirza, S. Senaratne, W.H. Kang, Sustainable structural material test specimens, BSI Standards Limited, London, England, 2011. doi: 10.3403/
combining recycled aggegate and steel fibres, in: Proc. 3rd World Constr. 30164906.
Symp. 2014 Sustain. Dev. Built Environ. W. Forw, Ceylon Institute of Builders, [51] V. Ramakrishnan, W. Coyle, V. Kulandaisamy, E. Schrader, Performance
Colombo, Sri Lanka, 2014, pp. 567–574. http://2014.ciobwcs.com/. characteristics of fiber reinforced concretes with low fiber contents, ACI J.
[25] P.S. Reddy, R.M. Rao, Study on properties of recycled aggregate concrete with Proc. 78 (1981) 388–394, https://doi.org/10.14359/10524.
hybrid fibres at low volume fractions, Int. J. Technol. Res. Eng. 2 (2015) 1304– [52] W. Tameemi, R. Lequesne, Correlations between Compressive, Flexural, and
1308. Tensile Behavior of Self-Consolidating Fiber Reinforced Concrete, Kansas,
[26] S. Senaratne, D. Gerace, O. Mirza, V.W.Y. Tam, W.H. Kang, The costs and United States, 2015. doi:10.13140/RG.2.1.1808.8084.
benefits of combining recycled aggregate with steel fibres as a sustainable, [53] R.V. Silva, J. de Brito, R.K. Dhir, Establishing a relationship between modulus of
structural material, J. Clean. Prod. 112 (2016) 2318–2327, https://doi.org/ elasticity and compressive strength of recycled aggregate concrete, J. Clean.
10.1016/j.jclepro.2015.10.041. Prod. 112 (2016) 2171–2186, https://doi.org/10.1016/j.jclepro.2015.10.064.
[27] S. Hesami, I. Salehi Hikouei, S.A.A. Emadi, Mechanical behavior of self- [54] R.C. Hibbeler, Mechanics of Materials, tenth ed., Pearson Education Limited,
compacting concrete pavements incorporating recycled tire rubber crumb and Boston, United States, 2017.
reinforced with polypropylene fiber, J. Clean. Prod. 133 (2016) 228–234, [55] J.-H. Lee, B. Cho, E. Choi, Flexural capacity of fiber reinforced concrete with a
https://doi.org/10.1016/j.jclepro.2016.04.079. consideration of concrete strength and fiber content, Constr. Build. Mater. 138
[28] J.A. Ortiz, A. de la Fuente, F. Mena Sebastia, I. Segura, A. Aguado, Steel-fibre- (2017) 222–231, https://doi.org/10.1016/j.conbuildmat.2017.01.096.
reinforced self-compacting concrete with 100% recycled mixed aggregates [56] F. Isla, G. Ruano, B. Luccioni, Analysis of steel fibers pull-out. Experimental
suitable for structural applications, Constr. Build. Mater. 156 (2017) 230–241, study, Constr. Build. Mater. 100 (2015) 183–193, https://doi.org/10.1016/
https://doi.org/10.1016/j.conbuildmat.2017.08.188. j.conbuildmat.2015.09.034.
[29] E. Mohseni, R. Saadati, N. Kordbacheh, Z.S. Parpinchi, W. Tang, Engineering and [57] T. Simões, C. Octávio, J. Valença, H. Costa, D. Dias-da-Costa, E. Júlio, Influence of
microstructural assessment of fibre-reinforced self-compacting concrete concrete strength and steel fibre geometry on the fibre/matrix interface,
containing recycled coarse aggregate, J. Clean. Prod. 168 (2017) 605–613, Compos. Part B Eng. 122 (2017) 156–164, https://doi.org/10.1016/
https://doi.org/10.1016/j.jclepro.2017.09.070. j.compositesb.2017.04.010.
[30] D. Gao, L. Zhang, M. Nokken, Mechanical behavior of recycled coarse aggregate [58] D.-Y. Yoo, Y.-S. Yoon, N. Banthia, Flexural response of steel-fiber-reinforced
concrete reinforced with steel fibers under direct shear, Cem. Concr. Compos. concrete beams: effects of strength, fiber content, and strain-rate, Cem. Concr.
79 (2017) 1–8, https://doi.org/10.1016/j.cemconcomp.2017.01.006. Compos. 64 (2015) 84–92, https://doi.org/10.1016/j.
[31] D. Gao, L. Zhang, M. Nokken, Compressive behavior of steel fiber reinforced cemconcomp.2015.10.001.
recycled coarse aggregate concrete designed with equivalent cubic [59] D.Y. Yoo, J.J. Park, S.W. Kim, Fiber pullout behavior of HPFRCC: Effects of matrix
compressive strength, Constr. Build. Mater. 141 (2017) 235–244, https://doi. strength and fiber type, Compos. Struct. 174 (2017) 263–276, https://doi.org/
org/10.1016/j.conbuildmat.2017.02.136. 10.1016/j.compstruct.2017.04.064.
[32] D. Gao, L. Zhang, Flexural performance and evaluation method of steel fiber [60] P. Balaguru, M.G. Dipsia, Properties of fiber reinforced high-strength semi-
reinforced recycled coarse aggregate concrete, Constr. Build. Mater. 159 (2018) lightweight concrete, ACI Mater. J. 90 (1993) 399–405, https://doi.org/
126–136, https://doi.org/10.1016/j.conbuildmat.2017.10.073. 10.14359/3865.
[33] P.K. Mehta, P.J.M. Monteiro, Concrete: Microstructure, Properties, and [61] A.R.E. Cáceres, V.M. John, A.D. de Figueiredo, Comparison between
Materials, third ed., McGraw-Hill Education, 2006. doi: 10.1036/0071462899. pavements of reinforced concrete and fiber reinforced concrete with
14 R. Chan et al. / Construction and Building Materials 242 (2020) 118186

respect to global CO2 emissions, in: An. Do 56o Congr. Bras. Do Concreto, Schmidt, S. Stürwald (Eds.), Proc Second Int. Symp. Ultra High Perform. Concr.,
IBRACON, 2014. Kassel University Press, 2008, pp. 461–468.
[62] M.U. Hossain, C.S. Poon, I.M.C. Lo, J.C.P. Cheng, Comparative environmental [64] ACI Committee 544, 544.4R-18, Guide to Design with Fiber-Reinforced
evaluation of aggregate production from recycled waste materials and virgin Concrete, American Concrete Institute (ACI), 2018.
sources by LCA, Resour. Conserv. Recycl. 109 (2016) 67–77, https://doi.org/ [65] W.P. Silva, C.M.D.P.S. Silva, LAB Fit Curve Fitting Software (Nonlinear
10.1016/j.resconrec.2016.02.009. Regression and Treatment of Data Program), (2016). http://www.labfit.net.
[63] T. Stengel, P. Schießl, Sustainable construction with UHPC–from life cycle
inventory data collection to environmental impact assessment, in: E. Fehling,