Construction and Building Materials 248 (2020) 118710

Contents lists available at ScienceDirect

Construction and Building Materials

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

Modeling and performance evaluation of asphalt mixtures and aggregate

bases containing steel slag
Mohamed Shiha a,b, Sherif El-Badawy a, Alaa Gabr a,⇑
a
Public Works Engineering Department, Mansoura University, Mansoura 35516, Egypt
b
Misr Higher Institute for Engineering and Technology, Mansoura, Egypt

h i g h l i g h t s

Steel slag (EAFS)/limestone blends were investigated for asphalt mixtures and bases.

EAFS up to 60% of the coarse aggregates met the Egyptian standards for binder layer.

Increasing EAFS content in asphalt, gave superior stability and loss of stability.

As the EAFS increases, the shear strength and resilient modulus generally increase.

Increasing EAFS content in bases, showed better performance than limestone blends.

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

Article history: The demand on using industrial by-products and waste materials in the construction industry has
Received 4 November 2019 increased in recent years. This study focused on the feasibility of using Electric Arc Furnace Steel slag
Received in revised form 8 March 2020 (EAFS) as a replacement to the natural aggregates (Limestone, LS) in both Hot Mix Asphalt (HMA) and
Accepted 9 March 2020
granular base layers. Four HMAs were investigated, two mixes were prepared using proportions of the
EAFS/LS of 0/100 and 100/0% of the total aggregate weight, while for the other two mixes, the coarse lime-
stone aggregate fractions were replaced by 60 and 80% coarse EAFS aggregates. The granular base EAFS/LS
Keywords:
blends were (0/100, 20/80, 40/60, 60/40, 80/20, and 100/0%) of the total aggregate weight. EAFS, LS aggre-
EAFS
Marshall
gates and asphalt binder were first characterized in the laboratory, and then the performance of the
Resilient modulus investigated asphalt mixtures in terms of Marshall stability, loss of stability, indirect tensile strength
Dynamic modulus and tensile strength ratio was evaluated. On the other hand, the granular base blends were characterized
FEM through routine tests and static and Repeated Load Triaxial (RLT) tests to evaluate the behaviour of these
Performance blends under traffic loading. RLTT results showed that as the EAFS increases, the stiffness generally
AC rutting increases. A new revised universal model is proposed to incorporate the EAFS effect on the resilient mod-
Fatigue cracking ulus (Mr) prediction having an excellent accuracy. The (E*) values for the investigated HMA were pre-
dicted using Witczak NCHRP 1-37A model. While, the Mr values for the investigated granular base
blends were predicted by the universal model at the anticipated field stress, which were obtained by
‘‘ANSYS” Finite Element Analysis software. Moreover, the field pavement performance in terms of
Asphalt Concrete (AC) rutting and fatigue cracking was predicted for different typical sections using
the Quality-Related Specifications Software (QRSS). Results showed better performance for HMA mixes
and unbound aggregate bases containing EAFS compared with the control section containing only natural
aggregates.
Ó 2020 Elsevier Ltd. All rights reserved.

1. Introduction construction. Steel slag is considered as one of the most wide-

spread by-product materials, since it is a cost-effective and envi-
Recently, utilizing by-product materials has played a vital role ronmentally acceptable alternative to the conventional natural
in the construction industry especially in the field of pavement aggregates [1]. Previous studies have confirmed that slag is a suit-
able material for concrete construction, where durability, compres-
⇑ Corresponding author. sive strength and tensile strength of concrete were improved
E-mail addresses: shi7a89@gmail.com (M. Shiha), sbadawy@mans.edu.eg compared to the limestone (LS) aggregate [1–3]. In addition, slag
(S. El-Badawy), eng-alaa1400@mans.edu (A. Gabr). has good potential to be used in pavements, since slag aggregates

https://doi.org/10.1016/j.conbuildmat.2020.118710
0950-0618/Ó 2020 Elsevier Ltd. All rights reserved.
2 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710

improved the resistance to stripping, rutting, and stability of the by steel slag aggregate, improved the mechanical properties of the
hot mix asphalts (HMAs) [3]. Worldwide, more than 400 million AC mixes. Wen et al. [15] investigated the use of four AC mixes
tons of iron and steel slag are produced annually [4]. Every ton containing steel slag with percentages of (0, 20, 40, and 60%). They
of the iron or steel production generates about 200 to 400 kg of found that as the steel slag increased, the tire wear resistance,
slag [4,5]. The United States (US) Geological Survey (USGS) data dynamic modulus values, thermal cracking resistance, and rutting
reported that the total production of slag in 2019 was 17 million resistance increased. They concluded that up to 60% steel slag
tons, in which Blast Furnace Slag (BFS) represented 50% of this replacement, the bitumen content was reduced from 6.1 to 4.8%
amount and the remainder was for the production of the electric compared to the virgin mixes. Sebaaly and Khnaizer, [16] evalu-
arc furnace slag (EAFS) [6]. In the US, slag is primarily used as ated the Marshall mix design criteria for asphalt base course mix-
aggregate in road construction [1]. In Europe, as of 2016, BFS pro- tures with different steel slag/gabbro aggregate contents of (30/70,
duction was about 25 million tons. The cement industry utilized 40/60, 50/50, 60/40, 70/30, 80/20%). They found that only the mix-
about 81.4% of this amount, while the road construction consumed ture of 50/50 steel slag/gabbro aggregate met the Qatar National
about 20.9% of the total slag production [7]. Furthermore, the steel Construction Standards (QCS 2010) requirements and can be used
slag production in the same year was about 18.4 million tons, of in asphalt base course. On the other hand, five different blends of
which about 46% was utilized in pavements, and only about 4.4% steel slag/gabbro aggregates as base material with percentages of
was used in cement production [7]. In Egypt, the annual produc- (0/100, 20/80, 40/60, 60/40 and 80/20%) were investigated through
tion of BFS is approximately 600,000 metric tons. Only 400,000 routine tests. These routine testes included gradation, density-
metric tons per year of BFS are utilized in cement production and moisture relationship, soundness, Los Angeles Abrasion (LAA), Cal-
the rest are wasted, while the produced annual amount of EAFS ifornia Bearing Ratio (CBR), flakiness and elongation indices, sand
is 300,000 metric tons [8]. equivalent, Atterberg limits, aggregate crushing value, and clay
lumps and friable particles. They found that 80% of steel slag aggre-
gate replacement achieved the QCS 2010 requirements, however,
2. Literature review the Qatar ministry of environment recommended the use of 60%
of steel slag instead.
Numerous studies were conducted on the laboratory character- Pasetto and Baldo [17,18] compared the use of two different
ization of slag for the use in the construction of Asphalt Concrete types of EAFS as a replacement to the natural aggregates in base
(AC), granular base, and subbase layers. Relevant studies for the course and road base asphalt concrete pavements using both Mar-
use of EAFS in pavement construction are as follows: shall and gyratory compaction methods. They conducted perfor-
Shatnawi et al. [9] compared the use of 100% steel slag in wear- mance testing through fatigue test, indirect tensile strength test,
ing course mix with the conventional wearing course mix using permanent deformation test, and stiffness modulus at various
dense graded LS at the same gradation and bitumen contents. temperatures. They reported that the slag improved Marshall sta-
The steel slag mix yielded higher stability and higher resistance bility, Marshall quotient, fatigue resistance, water damage resis-
to rutting compared to the conventional aggregate mix. Fistrić tant, and the resistance to permanent deformations. Behiry [19]
et al. [10] used 75% of steel slag aggregate with 25% LS rock aggre- found that the best percentage of a blended slag/LS was
gate in AC mixes. Results showed that the produced steel slag from (70/30%) to achieve the highest density and strength of a subbase
water-cooled process yielded good resistance to permanent defor- layer. This evaluation was based on theoretical analysis to predict
mation, high stability with good flow properties and high stiffness the vertical deformations, vertical and radial stresses and vertical
of bituminous materials. Zumrawi and Khalill, [11] investigated strains of subbase under heavy truck loads. He reported that as
four different percentages (0, 50, 75, and 100%) of steel slag aggre- the slag content increased, the resistance to deformation also
gate as a replacement to natural aggregates in asphalt mixtures increased. Maghool et al. [20] evaluated the use of Ladle Furnace
using Marshall mix design. They observed that the addition of pure Slag (LFS) as a base/subbase material. They found that the LFS had
steel slag to HMA increased both density and stability and higher maximum dry density (MDD) and CBR compared with typ-
decreased both flow and air voids at the optimum asphalt content ical quarry materials. They also reported higher Mr for the sam-
of each mix in comparison with HMA prepared with 100% natural ples prepared at 7, and 28 days curing time than uncured
aggregates. However, they did not study the effect of moisture sus- samples.
ceptibility. Louzi, [12] used steel slag in AC mixtures with different In summary, several studies confirmed the applicability of steel
percentages (0, 15, 30, and 45%) of the total weight of coarse aggre- slag in pavement construction due to its high density and hardness
gates. Fatigue, indirect tensile strength, loss of indirect tensile [21]. This is expected to help in sustaining the heavy traffic loads
strength and the resilient modulus (Mr) laboratory tests were per- on Egyptian roads. The Egyptian road network is subjected to sev-
formed on all investigated HMAs. He concluded that the steel slag ere damage, especially rutting and fatigue cracking, due to the
improved the properties of asphalt mixtures in terms of fatigue life, extremely heavy traffic loads and the hot climate. Trucks/trailers
indirect tensile strength, Mr, and stability. Hainin et al. [13] com- are usually loaded beyond the permissible load limits with weak
pared the use of steel slag with two different nominal maximum traffic laws enforcement for the legal load limits. In fact, more than
sizes in AC mixes with a conventional mix. Marshall mix design 120 tons (the allowable load limit is 70 tons) trucks are wandering
in accordance with the Malaysian specifications was used. Samples on the Egyptian road network. This causes severe distresses and
of AC mixes were subjected to the Mr and creep tests. They con- reduces the pavement service life. Large amounts of EAFS are pro-
cluded that steel slag mixes showed lower permanent deformation duced and wasted annually as previously mentioned, which can be
than the conventional asphalt mixes. used as aggregate in both HMA, and rigid pavements as well as in
Asi et al. [14] replaced the LS aggregates in asphalt mixes by the unbound granular base/subbase layers. In Egypt, few studies inves-
steel slag with percentages of 0, 25, 50, 75, and 100% of the total tigated the effect of using EAFS as aggregate on reducing the dam-
weight of aggregates. The suitability of the steel slag aggregate age of the AC and granular base layers, while no firm conclusions
was characterized by indirect tensile strength test, Mr test, wheel can be derived from these limited studies due to inconsistency in
track test, creep test, and stripping resistance test. They concluded results. Thus, there is an important need to assess the benefits
that steel slag aggregate can be used in AC mixes, since its proper- and risks of utilizing the EAFS as aggregate in pavements (asphalt
ties met both Superpave properties and the Jordanian standards. mixes and unbound aggregate base layers). Furthermore, to date,
They also found that replacing up to 75% of the LS coarse aggregate the steel slag was used only in few trial pavement sections
 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710 3

conducted by the General Authority for Roads, Bridges, and Land compounds exist in the investigated steel slag from the electric
Transport (GARBLT), Egypt. arc furnace as provided by [28]. It consists primarily of CaO,
Fe2O3, SiO2, Al2O3, MgO, and MnO, of more importance is the min-
3. Objectives eralogical form of the slag, which is highly dependent on the rate of
slag cooling in the steel-making process [28]. EAFS contains free
The main objective of this study is to investigate the suitability lime and magnesium oxides with percentages of 37.9 and 7.53%,
of using EAFS as a replacement to the LS aggregate in pavement respectively, which are not completely consumed in the steel slag.
layers (unbound base and HMA) for improving the mechanical Routine tests were conducted on the investigated materials (LS
properties of these layers. In addition, preliminary recommenda- and EAFS) to evaluate their engineering properties for the prepara-
tions for evaluating the effect of EAFS on pavement performance tion of asphalt mixtures and granular base layers. Table 3 summa-
are suggested. The secondary objectives of this research work are: rizes the general engineering properties of the investigated LS and
EAFS aggregates. The EAFS showed superior properties compared

Determine the optimum blending ratio of EAFS/LS aggregates in to the LS aggregate. This is evident by the higher sand equivalent
the HMA mixtures according to the Egyptian specifications. and bulk specific gravity, lower water absorption, lower LAA, and

Investigate the Mr of EAFS/LS blends for use as unbound granu- lower soundness in comparison with those obtained for the LS. In
lar base material. addition, all physical and mechanical properties of both aggregates

Study the applicability of the general universal model for Mr were within the range of the ECP specification limits as shown in
prediction for EAFS/LS blends as unbound granular base layer. Table 3 [22].

Predict the pavement performance in terms of AC rutting and Gradation of LS and EAFS aggregates as HMA aggregate, was
fatigue cracking for a typical pavement section based on the fractionated to fall between the specification limits of the binder
Mechanistic-Empirical Pavement Design Guide (MEPDG) course (3-D) in accordance with the ECP [22], as shown in
principles. Fig. 3a. For the granular base layer, one target gradation (D)
was used for the preparation of EAFS/LS blends as shown in
Fig. 3b. Each fraction of both materials (EAFS and LS) was
4. Materials and experimental work
blended with a proportion in order to control the gradation of
the materials.
The materials investigated in this research are bitumen, LS,
EAFS aggregate, and designed Marshall HMA mixtures. In this
4.3. Marshal mix design results
research, a series of laboratory tests including routine and
advanced tests, were conducted to characterize the investigated
Conventional Marshall mix design was conducted to determine
materials and determine their general engineering properties and
the optimum bitumen content (OBC) for the investigated mixtures
performance. Fig. 1 outlines the experimental work program.
with the different LS/EAFS proportions for a 3-D binder course in
The properties of the investigated materials are explained in the
accordance with the Egyptian specifications [22]. Table 4 summa-
following subsections.
rizes the Marshall properties of the investigated asphalt mixtures.
The 100% EAFS mix had the highest bulk density, stability and
4.1. Bitumen rigidity compared with those of other mixes. It is believed that
the reason for this may be due to the higher specific gravity and
In this study, a typical 60/70 penetration grade asphalt binder lower LAA values of the EAFS compared to LS aggregates as pre-
from the Amreya Petroleum Refining Company (APRC), Alexandria, sented previously in Table 3. The OBC of the investigated mixes
Egypt was used. Penetration, specific gravity, softening point, flash as well as the percentages of the total air voids in the mix (Va),
point, and kinematic viscosity tests were conducted on the selected and voids in mineral aggregate (VMA) at a given OBC were found
bitumen. Table 1 summarizes the results of the routine tests per- to slightly decrease with the increase in EAFS percentage. The rea-
formed on the bitumen. It is noted from the table that all bitumen son for this is probably due to the presence of free hydrated lime in
properties comply with the Egyptian standards [22]. EAFS, which contributed to greater densification during com-
paction compared with LS [3,11,14,17]. Moreover, the aggregate-
4.2. Limestone and EAFS aggregate to-aggregate contact along with the rough texture of EAFS surface
(Fig. 2-a) as compared with LS, attributed to better adhesion with
Limestone was collected from Allam Elmrqab Quarry, Alexan- asphalt as agreed with [11,17]. Consequently, the absorbed asphalt
dria, Egypt, while EAFS was sourced from the Contra Steel Com- slightly increased, but with relatively decreased effective asphalt
pany, Alexandria, Egypt. As shown in Fig. 2, EAFS was crushed content. Conversely, the voids filled with asphalt (VFA) increased
manually in the laboratory into suitable sizes of (1.500 –100 ), with the increase in EAFS.
(100 –3/400 ), (3/400 –3/800 ), (3/800 –#4), (#4–#200), and (passing #200) All investigated asphalt mixes prepared at OBC achieved the
for the use in both asphalt mixtures and granular base. The per- ECP [22], limits of stability of more than 700 kg for the 3-D binder
centage of EAFS in the investigated asphalt mixes was 0, and course asphalt mixes for heavy traffic. Slight decrease in the flow
100% of the total weight of aggregates for two out of the four mixes values with the increase of EAFS percentage in the mix was
and 60, and 80% of the coarse limestone aggregate weight for the noticed. Moreover, all specimens at OBC were within the limits
other two mixes. The reason for employing only the coarse EAFS of flow of 2–4 mm for heavily trafficked roads as recommended
fractions is that the fine EAFS was found to agglomerate while mix- by the ECP [22]. VMA values at OBC for the 100 and 80% EAFS
ing with LS and asphalt. This caused poor compaction and segrega- mixes were lower than the minimum requirement by the ECP
tion to the asphalt mix during compaction, which is also reported [22], of 15%. However, the Asphalt Institute recommended that
by other researchers [3,14,17]. Blends of EAFS/LS with percentages the minimum VMA for aggregates with a nominal maximum size
of 0/100, 20/80, 40/60, 60/40, 80/20 and 100/0% of the total aggre- of 3/400 should be 14%.
gate weight were utilized as a granular base. The loss of stability decreased with the increase in EAFS per-
The chemical composition of the EAFS is usually expressed in centage in the asphalt mixtures. Moreover, all investigated asphalt
terms of simple oxides calculated from elemental analysis deter- mixtures showed lower values of loss of stability than the ECP
mined by X-ray fluorescence. Table 2 presents the percentages of specification limit of 25% [22]. The tensile strength ratio (TSR)
4 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710

Fig. 1. Experimental Work Program.

Table 1 In summary, based on Marshall, loss of stability, and indirect

Physical properties of the asphalt cement used in this study. tensile strength (IDT) testing results, up to 60% EAFS coarse aggre-
Properties Standard (ECP-2008) Measured gate is recommended to be used in asphalt mixtures wherein this
Limits values content achieved the requirements of the ECP.
Specific gravity [23] – 1.023
Penetration [24] 60–70 62
Softening point (°C) [25] 45–55 °C 52 4.4. Static and repeated load triaxial testing on aggregate
Kinematic viscosity (cst/s) [26] minimum 320 374
Flash point (°C) [27] minimum 250 270
Three replicate samples of each blend with 300 mm height and
150 mm diameter, were compacted inside a split mould according
to the modified Proctor at 100% MDD and OMC. According to the
increased with the increase in EAFS percentage in asphalt mix- modified Proctor effort [36], each specimen was divided into six
tures, meaning that asphalt mixtures containing EAFS are less layers to achieve the compactive effort of 2700 N/m3 requiring
prone to stripping. All the HMA mixtures achieved the TSR value 114 blows per layer. Each specimen was left in the split mould
of 80% minimum as specified by ECP [22]. As well, the stripping for two days to cure and assure that the specimen would not fail
test results confirmed that EAFS showed good behavior in bonding during extrusion. After extrusion, a rubber membrane was
with bitumen with stripping value less than 5% as specified by the wrapped around the specimen and two O-rings were used to seal
ECP [22]. the sample at the bottom and top platens. Fig. 4 shows the
 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710 5

(a) EAFS before Crushing.

(b) EAFS after Crushing.

Fig. 2. EAFS Before and After Crushing.

Table 2 The results of the MDD, OMC, and CBR for the different blends
Chemical properties of EAFS [28]. are summarized in Table 5. MDD values showed an obvious
Component Percentage % increase with the increase in EAFS content in the investigated base
Free Lime (CaO) 37.90
blends and their values varied from 2.233 to 2.847 t/m3. Con-
Silicon dioxide (SiO2) 21.99 versely, the OMC decreased with the increase in EAFS percentage
Manganese oxide (MnO) 2.24 in the blend varying from 7.7 to 3.5%. CBR values after 4-days soak-
Magnesium oxide (MgO) 7.53 ing were found to range between 89 and 340% for the different
Aluminum oxide (Al2O3) 7.90
investigated blends. A significant increase in the CBR with an aver-
Iron Oxide (Fe2O3) 22.78
age of about 69% was observed with the increase in EAFS content
up to 60%. Replacing LS by more than 60% EAFS, exhibited a notice-
able increase in CBR value with an average of 22%. All blends had
preparation steps of the static and Repeated Load Triaxial Testing CBR values greater than that recommended by the ECP [22], for
(RLTT) specimens starting from compaction to the installation base layers of 80%.
and placing the triaxial cell inside the Universal Testing Machine Table 6 summarizes the values of c and u for the investigated
(UTM-25). The RLTT and static triaxial shear tests were conducted EAFS/LS blends. It is noticed that the six blends of EAFS/LS have rel-
on the investigated EAFS/LS blends in accordance with the AASHTO atively high angles of internal friction, u, with values ranged from
T-307 [37] in order to determine the Mr and shear strength param- 36 to 53°. The values of the apparent cohesion varied between 49
eters (cohesion, c, and internal angle of friction, u), respectively. and 100 kPa.

Table 3
Engineering properties of LS and EAFS used in this study.

Property Standard ECP, [22] Limits LS EAFS

Flat/elongated (1:5 ratio), % [29] 10% max 0 0
Sand equivalent, % [30] 45% min 58 88
Coarse aggregate specific gravity [31] N/A 2.535 3.300
Coarse aggregate water absorption, % 5% max 2.27 1.38
Fine aggregate specific gravity [32] N/A 2.611 3.356
Fine aggregate water absorption % 5% max 3.45 2.16
Clay lumps and friable particles, % [33] 1% max 0.14 0
LAA (500 rev), % [34] 40% max 25.9 21.8
Soundness by (Mg SO4), % [35] 18% max 3.50 2.10
Soundness by (Na2 SO4), % 12% max 2.40 1.82

N/A: Not Applicable.

6 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710

Upper Gradation Limit Upper Gradation Limit

Used Gradation Used gradation
Lower Gradation Limit Lower Gradation Limit
100 100
90 90
80 80
70

% Passing
70
% Passing

60 60
50 50
40 40
30 30
20 20
10 10
0 0
0.01 0.1 1 10 100 0.01 0.1 1 10 100
Sieve size (mm) Sieve size (mm)
a) HMA Gradation b) Granular Base Gradation
Fig. 3. Particle Size Distribution for the Investigated EAFS/LS for HMA and Base.

Table 4
Summary of Marshall Results for the Investigated Asphalt Mixtures.

Properties Mixture Type

(0% EAFS/100 LS) of total (60% EAFS/40% LS) of coarse (80% EAFS/20% LS) of coarse (100% EAFS/0% LS) of total ECP, [22]
aggregate by weight aggregate by weight aggregate by weight aggregate by weight Limits
O.B.C (%) 4.85 4.73 4.62 4.40 N/A
Stability (kg) 1100 1440 1570 1930 Minimum
700
Gmb 2.321 2.598 2.675 3.038 N/A
Gmm 2.390 2.701 2.808 3.162 N/A
Absorbed asphalt 0.14 0.51 0.70 0.90 N/A
by weight (%)
Effective asphalt by 4.71 4.22 3.92 3.50 N/A
weight (%)
Air voids (%) 4.9 4.4 4.2 3.90 3–8
Flow (mm) 3.65 3.5 3.3 2.90 2–4
Rigidity (kg/mm) 301 415 476 666 NA
VMA (%) 15.4 15.1 14.6 14.3 Minimum
15
VFA (%) 68.0 70.0 72.0 72.7 NA
Loss of stability (%) 23.3 18.8 16.4 10.2 Maximum
25
TSR (%) 83.2 92.5 95.2 99.5 Minimum
80
Stripping value (%) 4.5 N/A N/A 3.0 Maximum
5
Gmb Bulk Specific Gravity of the Mix. VMA Voids in Mineral Aggregate.
VFA Voids Filled with Asphalt. O.B.C Optimum Bitumen Content.
TSR Tensile strength ratio N/A Not Applicable
Gmm Maximum theoretical specific gravity

Fig. 5 shows the relationship between the Mr and deviator where, K1, K2, and K3 are material regression coefficients;
stress (rd) at different confining stress levels for all investigated soct is the octahedral shear stress¼ 13 
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
EAFS/LS blends. As seen in the figure, the Mr increased with the
ðr1  r2 Þ2 þ ðr1  r3 Þ2 þ ðr2  r3 Þ2 ;
increase in both deviator and confining stresses. The 100% EAFS
has the highest resilient moduli at the same stress levels compared Pa is the atmospheric pressure (Pa = 101.3 kPa).
with other blends. Table 7 summarizes the values of the regression coefficients of
Eq. (1) as well as the goodness of fit statistics for each blend. The K2
4.5. Resilient modulus modelling values of the model are all positive while all K3 values are very
small (close to zero) meaning a stress hardening behavior of the
The universal Mr model as presented in Eq. (1) [38], was applied investigated materials which is typical for granular materials.
to the different EAFS/LS blends. The solver function in Microsoft The prediction accuracy of the universal model for each individual
Excel was used to estimate the regression coefficients, K1, K2, and blend is excellent based on the goodness of fit statistics shown in
K3 based on the least square method. the table (R2 of 0.96 to 0.987).

K2
K3 The universal model was also applied to all the RLTT data for all
h soct blends (90 points for all six blends neglecting the preconditioning
Mr ¼ K1 Pa þ1 ð1Þ
Pa Pa sequence) in order to obtain a one set of constants for the model.
 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710 7

b) Removing sample from the

a) Samples of each blend after split mold. c) Sample between the end
2 days curing. plates.

d) Membrane expander. e) Sample sealed with f) Sample inside UTM

membrane and O-rings. ready for testing.
Fig. 4. RLTT Test Samples Preparation.

Table 5 EAFS is the content of Electric Arc Furnace Steel slag, %; and the
MDD & OMC, and CBR Results for the Tested Granular Base Blends. other parameters are as defined previously. This model yielded
EAFS/LS (%) OMC (%) MDD (t/m3) CBR (%) excellent accuracy with R2 of 0.98 as indicated in Fig. 6.
0/100 7.7 2.233 89
20/40 7.0 2.320 153.6 5. Pavement performance
40/60 6.8 2.463 235.0
60/40 6.4 2.664 296.8
80/20 5.6 2.840 318.1 A typical pavement section consists of 7.6 in. (19 cm) asphalt
100/0 3.5 2.847 340.8 layer over a 16 in. (40 cm) granular base layer resting on a sub-
grade as shown in Fig. 7-a was chosen for the pavement perfor-
mance prediction. Table 8 shows the four different sections (A, B,
C and D) with different percentages of EAFS/LS in both the HMA
and granular base layer to compare pavement performance. Only
The prediction accuracy in terms of R2 was found to be poor with a
two contents of EAFS were chosen as base materials in all investi-
value of 0.32 and need further modification. In order to improve
gated sections, 0% EAFS (lowest modulus) as control and 100% EAFS
the accuracy of the universal prediction model for all blends, the
that yielded the highest resilient modulus. For HMA, the mix with
model form was modified as shown in Eq. (2) by adding the per-
60% EAFS was the mix that surpassed the ECP requirements, which
centage of EAFS in the blend with one more regression coefficient
was fixed in sections C, and D against the two contents of 0 and
(K4).
100% EAFS in base sections. Section, B with 100% EAFS in both lay-

K2
 ers was chosen for comparison, although the 100% EAFS in HMA
h soct K3 was not optimal and did not achieve the mix requirements.
Mr ¼ K1  Pa    ðEAFS þ 1ÞK4 ð2Þ
Pa Pa Two traffic loading conditions were applied to these sections;
the first is for medium traffic with about 3 million 18 kips (80
where, K1, K2, K3 and K4 are material regression coefficients; KN) Equivalent Single Axle Load (ESAL) and denoted as A-1, B-1,

Table 6
Shear strength parameters for EAFS/LS blends.

EAFS/LS (%) 0/100 20/80 40/60 60/40 80/20 100/0

Apparent Cohesion, c, (kPa) 48.68 58.13 90.58 100.2 92.26 98.97
Friction Angle, u, ° 36.39 46.30 48.68 47.39 51.54 53.23
8 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710

400 400
Mr Measured (MPa) 350 350

Mr Measured (MPa)
300 300
250 250
200 200
150 150
100 σ3=20.7kPa σ3=34.5kPa 100 σ3=20.7kPa σ3=34.5kPa
σ3=68.9kPa σ3=103.4kPa σ3=68.9kPa σ3=103.4kPa
50 50
σ3=137.9kPa σ3=137.9kPa
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Deviator stress (kPa) Deviator stree (kPa)
(a) (b)
500 600

Mr Measured (MPa)
Mr Measred (MPa)

400 500

400
300
300
200
σ3=20.7kPa σ3=34.5kPa 200
100 σ3=68.9kPa σ3=103.4kPa σ3=20.7kPa σ3=34.5kPa
100 σ3=68.9kPa σ3=103.4kPa
σ3=137.9kPa σ3=137.9kPa
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Deviator stress (kPa) Deviator stress (kPa)
(c) (d)
800 1000
700
Mr Measured (MPa)

Mr Measured (MPa)

800
600
500 600
400
300 400
200 σ3=20.7kPa σ3=34.5kPa σ3=20.7kPa σ3=34.5kPa
σ3=68.9kPa σ3=103.4kPa 200
100 σ3=68.9kPa σ3=103.4kPa
σ3=137.9kPa σ3=137.9kPa
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Deviator stress (kPa) Deviator stress (kPa)
(e) (f)
Fig. 5. Effect of Stress Level on Resilient Modulus of EAFS/LS Base Blends: (a) 0/100, (b) 20/80, (c) 40/60, (d) 60/40, (e) 80/20, (f) 100/0.

Table 7
Universal model coefficients and the goodness of fit statistics parameters.

EAFS/LS (%) 0/100 20/80 40/60 60/40 80/20 100/0

K1 1.783 1.994 2.463 2.865 3.290 3.819
K2 0.302 0.313 0.351 0.363 0.393 0.425
K3 0.015 0.056 0.091 0.006 0.048 0.046
SES 767 1492 1621 3038 5032 4537
Se 7.995 11.150 11.624 15.912 20.478 19.445
Sy 51.400 57.771 79.886 98.398 129.787 167.644
Se/Sy 0.156 0.193 0.146 0.162 0.158 0.116
R2 0.976 0.963 0.979 0.974 0.975 0.987
R2Adj 0.972 0.957 0.975 0.969 0.971 0.984

SES = Sum of the squares of the errors (SES) between the measured and the predicted resilient modulus.
Se = Standard error of estimate; Sy = Standard deviation of the measured resilient modulus values.
R2 = Coefficient of determination; R2adj = Adjusted coefficient of determination.
 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710 9

900

800

700

600
Predicted Mr (MPa)

500

400

300

200

100

0
0 200 400 600 800
Measured Mr (MPa)

Fig. 6. Measured versus Predicted Resilient Modulus Values for EAFS/LS Blends using (Eq. (2)).

C-1 and D-1. Whereas A-2, B-2, C-2 and D-2 are denoted for heavy layer. A two-dimensional axis-symmetrical Finite Element Model
traffic with about 30 million ESALs. (FEM) with 8-nodded elements was built using the ANSYS software
Two design speeds were selected; low speed of 10 mph [43] to predict the anticipated deviator and confining stresses at
(16 km/hr) and high speed of 50 mph (80 km/hr) to simulate the the mid-depth of the granular base for predicting its anticipated
loading frequency impact. Furthermore, two climatic data sets for resilient modulus value. Linear elastic finite element analyses were
different climatic regions in Egypt; Alexandria (representing a performed for the typical pavement structure presented in Fig. 7.
moderate costal climate) and Aswan (representing a hot climate) Both predicted HMA E* values by QRSS and initial base Mr values
were utilized. For the pavement performance prediction, the by the Mr-CBR relationship as given in Table 9, were incorporated
HMA dynamic modulus, E* was firstly predicted using Witczak into the FEM.
NCHRP 1-37A model [39]. The Mr of the base layer was predicted A dual tires axle with equivalent load of 80-kN (18-kip) was
based on the anticipated field stresses computed by the ANSYS modeled representing a uniform pressure of 0.828 MPa over a cir-
finite element analysis model. Finally, the AC rut depths and cular area with a radius of 124 mm. Fine mesh was used under the
fatigue cracking were computed by the Quality Related Specifica- wheel with applied boundary conditions; Ux = 0 for the sidewalls of
tions Software (QRSS) software. The QRSS is developed by Fugro the model and was fixed at the bottom in the two directions,
Consultants, Inc. and Arizona State University (ASU) under the Uy = 0 and Ux = 0. The effective depth at the fixed bottom was taken
NCHRP 9–22 projects as a simplification of the MEPDG [40,41]. about 50-times (6200 mm) the contact radius (a) of loading area,
A total of 32 QRSS simulation runs were conducted on the while the effective horizontal distance to the roller boundary
investigated typical pavement sections shown in Fig. 7 in order was about 12-times (1488 mm) the contact radius from loading
to predict the E* values for the different sections by changing the center [45].
traffic conditions (medium and high), design speeds (10 and 50 The anticipated field stresses were predicted by the ANSYS FEM
mph) and the two climate region data (Alex and Aswan). The cli- at the mid-depth of the granular base layer in order to estimate the
matic effect on the dynamic modulus is evaluated based on the anticipated resilient modulus. Fig. 8 exemplifies the contour plots
critical temperature (effective temperature) concept [40,42]. Initial of the vertical and radial stresses for pavement section, A under
values for the base resilient modulus were inputted in the QRSS Alexandria climate condition. Fig. 9 summarizes the deviator and
using the Mr-CBR relationship presented in Eq. (3) [38], until it fur- confining stress values, respectively for all pavement sections.
ther computed based on both the anticipated field stresses by the The values of deviator stresses (rd) varied from 44.15 to
ANSYS [43] model and the universal model coefficients (K1, K2, and 62.21 kPa (6.4 to 9 psi), while the confining stress (r3) ranged
K3). Poisson’s ratio was assumed for all layers based on literature between 9.12 and 18.11 kPa (1.3 to 2.6 psi). Jones and Witczak
[44]. [46] recommended that the design Mr value can be estimated at
a deviator stress of 41.4 kPa (6 psi) and confining stress of
Mr ¼ 2555ðCBRÞ0:64 ð3Þ 13.8 kPa (2 psi). Ji et al. [47] reported a range of deviator stresses
where Mr is the resilient modulus, psi; and CBR is the California of 89.6 to 110.3 kPa (13 to 16 psi) and 27.5 to 55.2 kPa (4 to 8
Bearing Ratio, %. psi) for confining stresses. This contradiction between studies is
due to the differences in materials properties, applied pressures
and contact loading radii.
5.1. Pavement response by ANSYS software
Based on the values of the anticipated field stresses at the mid-
depth of the base layer, Eq. (1) was used by applying the regression
Since the resilient modulus of the unbound layer is stress
coefficients K1, K2 and K3 (Table 7) for granular base blends to
dependent, it was important to determine the stresses in the base
10 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710

a) Pavement Cross Section Configuration

b) Layers Properties
Fig. 7. Typical Pavement Section using EAFS/LS Blends in HMA/Granular Base Layers.

Table 8 determine the anticipated resilient modulus. Table 10 presents the

Summary of EAFS/LS Percentages in the Investigated Sections. values of the field resilient moduli for the different blends, which
Section HMA Base correspond to the minimum and maximum anticipated field stres-
A (control) 0% EAFS 0% EAFS
ses (rd and r3).
B 100% EAFS 100% EAFS Fig. 10 shows a comparison between the anticipated field resi-
C 60% EAFS 0% EAFS lient moduli given in Table 10 and the predicted resilient moduli
D 60% EAFS 100% EAFS from Eq. (3) for the different blends. It can be observed from the
figure that the resilient moduli from both predictions always
increased with the increase in EAFS content. Eq. (3) accuracy is rea-
Table 9
sonable with R2 of 0.83. However, the predicted resilient moduli
Material Properties of the Investigated Sections in Alex and Aswan.
from Eq. (3) were higher compared to those obtained from the uni-
Section Location E*Max (psi) Initial Mr (psi) versal model (Eq. (1)) at the minimum and maximum anticipated
A Alex 522,800 45,186 stresses. This means that the literature models tend to overesti-
Aswan 409,400 mate the resilient modulus of the granular materials. Conse-
B Alex 532,400 64,074
quently, Eq. (3) multiplier was modified as described in Eq. (4) to
Aswan 414,600
C Alex 524,700 45,186 a range of 1045 to 1253 psi with an average of 1140 psi instead
Aswan 414,600 of 2555 psi.
D Alex 524,700 64,074
Aswan 414,600
Mr ¼ ð1045  1253ÞCBR0:64 ð4Þ
 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710 11

a) Vertical stresses b) Vertical stresses versus depth at radius=0

c) Radial stresses d) Radial stresses versus depth at radius=0

Fig. 8. Example of the Contour Plots of Vertical and Radial Stresses for section (A), Alexandria Climate.

Horizontal stress Deviator stress

Anticipated field stress (kPa)

80 62.21 62.21
49.96 50.64 49.15 51
44.21 44.15
60

40
17.7 18.11 17.74 18.11
20 9.14 10.11 9.12 10.01

Fig. 9. Anticipated Mid Depth Field Stresses for the Investigated Pavement Sections.

It should be noted that the multiplier and exponent in Eq. (4) ranged between 4920 and 8540 and the exponent was 0.48 to
are valid only for the investigated blends, but they can be cali- 0.38, respectively. Other studies by [49,50], also found that the
brated for other materials. For example, another research by same model form but with different regression constants, is valid
[48], developed a comprehensive model based on 10 unbound for different blends of construction & demolition (C&D) waste as
granular base/subbase materials collected from different on- well as blends of reclaimed asphalt pavement (RAP) and virgin
going road construction projects in Egypt. The multiplier in psi aggregates in Egypt.
12 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710

Table 10 increase in the rut depth for sections (A-2, B-2, C-2 and D-2) than
Summary of resilient moduli values at the anticipated field stresses. (A-1, B-1, C-1 and D-1) sections is due to the increase in ESALs. On
EAFS/LS Mr Max. (MPa) at (rd = 62.21, Mr Min. (MPa) at (rd = 44.15, the other hand, there is no obvious variation in AC rut depth values
Blend, % r3 = 18.11) kPa r3 = 9.12) kPa at the same traffic level due to the little variation in the predicted
0/100 186.1 160.7 E* value for the four sections. It should be also noted that these AC
20/80 206.2 177.6 rut depth values are valid only for comparison purposes, since the
40/60 253.8 215.2 QRSS prediction models need calibration for Egyptian conditions.
60/40 302.4 253.4
80/20 345.2 285.2
100/0 402.8 328.3
5.3. Predicted fatigue cracking

The QRSS predicted fatigue cracking for all investigated sections

5.2. Predicted AC rutting is presented in Fig. 12. Sections (B and D) with 100% EAFS as gran-
ular base had lower fatigue cracking values compared with other
Fig. 11 compares the predicted AC rut depth for all investigated sections (A and C with 0% slag in base) due to the higher resilient
sections at the two climatic conditions and speeds. As excepted, moduli of the granular base, and hence lower tensile strain at the
the AC rut depth of the slow speeds (10 mph) was higher than bottom of asphalt layer at both high and low design speeds for
those at the higher speeds (50 mph) for all the investigated sec- the two climatic conditions. EAFS seems to work as a stiff aggre-
tions due to the rate of loading effect and the viscous nature of gate, causing the HMA mixtures and granular bases to be stiffer,
the asphalt materials. Moreover, the rutting of HMA was higher which gives the asphalt concrete a higher resistance to fatigue
in Aswan because of the hotter temperature. As well, the obvious cracks. On the other hand, the fatigue cracking of sections (A and

120000
Mr (psi) at min
anticipated field stress
100000
Mr (psi) at max
80000 anticipated field stress
Mr (psi)

MR = 2555 CBR^0.64
60000

40000 MR = 1045 CBR^0.64

20000
MR = 1253 CBR^0.64

0
0 20 40 60 80 100 EAFS (%)
Fig. 10. Predicted resilient moduli from Eqs. (3) and (4), and anticipated field stresses versus EAFS content.

2
Alex (10mph) Alex (50mph) Aswan (10mph) Aswan (50mph)
1.8

1.6

1.4
Rut depth (in.)

1.2

0.8

0.6

0.4

0.2

0
A-1 A-2 B-1 B-2 C-1 C-2 D-1 D-2
Fig. 11. Predicted AC Rut Depth from QRSS simulation at Alex and Aswan for different sections at 10 and 50 mph Design Speeds.
 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710 13

35
Alex (10mph) Alex (50mph) Aswan (10mph) Aswan (50mph)
30

25
Fatigue cracks (%)
20

15

10

0
A-1 A-2 B-1 B-2 C-1 C-2 D-1 D-2
Fig. 12. Predicted Fatigue Cracking Percentage from QRSS simulation at Alex and Aswan for different sections at 10 and 50 mph Design Speeds.

C) at lower design speed (10 mph) was found to be higher than owing to that LS was used for base material for both sections. Fur-
those obtained for higher design speed (50 mph) at the same cli- thermore, even though sections B-2 and D-2 have the highest total
matic data. Conversely, sections (B and D) with 100% EAFS in the initial construction cost of 544 and 524 EGP/m2, they have the low-
granular base had lower fatigue cracking for the low design speed est damage of about 5–6% fatigue cracking yielding the longest
(10 mph) compared to those for the high design speed (50 mph) at pavement life of almost 35 years due to the stiffer base layer,
the same climatic conditions. Finally, the figures show that the which helps in reducing the tensile strain at the bottom of the
variation of climatic conditions in Egypt had no significance effect AC layer. It can be concluded that EAFS is a feasible by-product
on the predicted fatigue cracking. This is because of the warm nat- material that can be used in pavement applications and can benefit
ure of the climate and fatigue is more pronounced in colder the environment, nonetheless the wear of equipment due to slag
climates. should be studied.

6. Cost analysis 7. Summary and conclusions

Table 11 exemplifies the conducted cost-benefit analysis for the EAFS/LS blends were investigated for the construction of both
investigated pavement sections. The table provides the operation asphalt mixtures and granular bases. A comprehensive experimen-
cost for the plant mix and the cost for a hauling distance around tal program was conducted to determine the basic engineering
100 km in Egyptian Pounds (EGP) per cubic meter for base and properties of bitumen and aggregate in addition to the Marshall
HMA materials. The table also shows the total construction cost mix design, loss of stability, and IDT tests. Moreover, advanced
in EGP per square meter including the cost of prime coat, tack coat, tests including the static and RLTT tests for base materials, were
administrative fees, and taxes as well as the layers compaction and conducted to evaluate the behavior of EAFS/LS blends under traffic
operation for the selected typical pavement section (19 cm HMA loading and to estimate the resilient modulus and shear strength
layer + 40 cm base layer). In addition, the table gives the allowable parameters. The RLTT data of the granular base blends was used
number of repetitions (pavement life) predicted by the MEPDG to estimate the regression constants of the universal Mr prediction
fatigue model [38] for the selected sections, A-2, B-2, C-2, and D- model. Then, the universal model was also modified to incorporate
2 under conditions of heavy traffic, 50 mph design speed, and mod- the EAFS percentage in the model for better accuracy. Pavement
erate climate (Alex). It is evident from the data that the control sec- performance in terms of AC rutting and fatigue cracking was pre-
tion (A-2) has the lowest total initial construction cost of 479 EGP/ dicted for typical pavement sections using the QRSS for two differ-
m2 with the shortest pavement life of 19 years due to the predicted ent climatic conditions (Alexandria and Aswan) and two speeds of
fatigue cracking of almost 25% for section A-2 as shown in Fig. 12. (10 and 50 mph). Finally, a cost-benefit analysis was performed to
Although section C-2 has the second rank regarding the total initial study the economic impact and feasibility of using the investigated
construction cost of 499 EGP/m2, it has almost the same pavement EAFS materials as a by-product alternative in pavement construc-
life and same damage as predicted by QRSS and shown in Fig. 12 tion. The main findings of this study are summarized as follows:

Table 11
Summary of the Cost-Benefit Analysis for the Investigated Pavement Sections.

Item Layer/Section HMA Base A-2 B-2 C-2 D-2

Plant Mix Operation and Hauling Costs (EGP/m3) 0% EAFS 1100 150 1250 — 1340 —
60% EAFS 1190 — — — 1390
100% EAFS* 1277 200 — 1477 —
Total Construction Cost (EGP/m2) for 19 cm HMA and 40 cm Base 0% EAFS 371 108 479 — 499 —
60% EAFS 391 — — — 524
100% EAFS 411 133 — 544 —
Allowable number of repetitions, pavement life (Years) 19.0 35.0 19.3 35.0
*
The hauling cost is higher due to the higher specific gravity of the slag
14 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710

7.1. EAFS in HMA layers CRediT authorship contribution statement

EAFS showed superior characteristics compared to the natu- Mohamed Shiha: Data curation, Writing - original draft. Sherif
ral limestone aggregate. EAFS has high bulk specific gravity, El-Badawy: Conceptualization, Writing - review & editing. Alaa
low LAA value, low water absorption and high adhesion Gabr: Methodology, Investigation, Supervision, Writing - review
with bitumen, which improved the performance of HMA & editing.
mixes.

Increasing the EAFS up to 60% of the coarse aggregates was
Declaration of Competing Interest
found to meet the requirements of the current Egyptian specifi-
cations for the binder course layer, 3-D.
The authors declare that they have no known competing finan-

Increasing the EAFS percentage yielded superior Marshall sta-
cial interests or personal relationships that could have appeared
bility, loss of stability, and TSR compared to conventional lime-
to influence the work reported in this paper.
stone aggregate mixtures.

References
7.2. EAFS in unbound granular base layers
[1] S. El-Badawy, A. Gabr, R. Abd El-Hakim, Recycled materials and by-products for

The EAFS content had significant effect on the CBR, shear pavement construction, in: Martínez et al. (Eds., Chapter in Handbook of
Ecomaterials, Springer International Publishing AG, 2018, pp. 1–22, ISBN: 978-
strength, and resilient modulus of the granular material. Gener-
3-319-48281-1 (Print & Online), DOI: 10.1007/978-3-319-48281-1_168-1.
ally, an increase in CBR, shear strength, and Mr occurred with [2] S. Wu, Y. Xue, Q. Ye, Y. Chen, Utilization of steel slag as aggregates for stone
the increase in the EAFS content due to the rough surface of mastic asphalt (SMA) mixtures, Build. Environ. 42 (7) (2007) 2580–2585,
the EAFS compared to the LS aggregate. https://doi.org/10.1016/j.buildenv.2006.06.008.
[3] R. Collins, S. Ciesielski, Recycling and use of waste materials and by-products

A new revised universal model was developed incorporating the in highway construction, National Cooperative Highway Research Program
EAFS content with excellent prediction accuracy (R2 of 0.98) for (NCHRP Synthesis of Highway Practice 199), Transportation Research Board,
all blends. Washington, D.C., 1994.
[4] World Steel Association (WSA). Fact Sheet - Steel industry co-products. World

ANSYS FEM showed that the anticipated field deviator stresses Steel Association, Belgium, https://www.worldsteel.org/en/dam/jcr:
(rd) at the mid-depth of the granular base layer were 44.15 1b916a6d-06fd-4e84-b35d-c1d911d18df4/Fact_By-products_2018.pdf, 2018
to 62.21 kPa (6.4 to 9 psi) and confining stresses (r3) between (accessed on 23rd February 2020).
[5] D.M. Proctor, K.A. Fehling, E.C. Shay, J.L. Wittenborn, J.J. Green, C. Avent, R.D.
9.12 and 18.11 kPa (1.3 to 2.6 psi). Bigham, M. Connolly, B. Lee, T.O. Shepker, M.A. Zak, Physical and chemical

The traditional Mr-CBR relationship was found to over predict characteristics of blast furnace, basic oxygen furnace, and electric arc furnace
the resilient moduli of the investigated blends. The multiplier steel industry slags, Environ. Sci. Tech. 34 (8) (2000) 1576–1582.
[6] U.S. Geological Survey (USGS), Mineral Commodity Summaries 2020, p. 200,
constant of this relationship was modified to range between https://doi.org/10.3133/mcs2020, Reston, Virginia, https://www.usgs.gov,
1045 and 1253 psi instead of the 2555 psi. 2020 (accessed on 23rd February 2020).
[7] The European Slag Association, Statistics 2016, http://www.euroslag.org, 2016
(accessed on 23rd February 2020).
7.3. Effect of EAFS on pavement performance [8] N. Nemerow, F. Agardy, J. Salvato, ‘‘Environmental Health and safety for
municipal infrastructure”, land use and planning and industry ISBN: 978-0-
470-08305-5 for Municipal Infrastructure, 6th edition., John Wiley & Sons, Inc.,

Little variation in AC rut depth values at the same traffic level Hoboken, NJ, 2009.
was found due to the slight variation in the predicted E* values [9] A.S. Shatnawi, M.S. Abdel-Jaber, K.Z. Ramadan, Effect of Jordanian steel blast
of the four investigated sections. furnace slag on asphalt concrete hot mixes, Jordan J. Civ. Eng. 2 (3) (2008).
[10] M. Fistrić, A. Strineka, R. Roskovic, Properties of steel slag aggregate and steel

Increasing the percentage of EAFS/LS in base blends showed
slag asphalt concrete, in: Proceedings of CETRA 2010 Conference, Institute IGH,
better performance than natural aggregate blends in terms of Zagreb, Croatia, https://www.scribd.com/document/266216118/Properties-
the fatigue cracking resistance of the AC layer. As the stiffness of-steel-slag, 2010 (accessed on 7th March 2020).
of base layer increases, the tensile strain at the bottom of the [11] M. Zumrawi, F. Khalill, Experimental study of steel slag used as aggregate in
asphalt mixture, Int. J. Civ. Environ. Struct. Constr. Arch. Eng. 9 (6) (2015).
AC layer decreases leading to better resistance to fatigue cracks [12] N. Louzi, Modification of hot asphalt mixtures in jordan and syria by using
and pavement life. steel slag, Jordan J. Civ. Eng. 6 (3) (2012).
[13] M.R. Hainin, N.I. Yusoff, M.F. Sabri, M.A. Abdul Aziz, M.A. Hameed, M.F. Reshi,
Steel slag as an aggregate replacement in malaysian hot mix asphalt, ISRN Civ.
8. Further research Eng., Malaysia (2012).
[14] I.M. Asi, H.Y. Qasrawi, F.I. Shalabi, Use of steel slag aggregate in asphalt
concrete mixes, Canadian J. Civ. Eng. 34 (8) (2007) 902–911, https://doi.org/

Dynamic modulus (E*), flow number, and beam fatigue should 10.1139/L07-025.
be measured in the laboratory for mixes containing different [15] H. Wen, S. Wu, S. Bhusal, Performance evaluation of asphalt mixes containing
percentages of EAFS for better understanding of the perfor- steel slag aggregate as a measure to resist studded tire wear, J. Mater. Civ. Eng.
28 (5) (2016), https://doi.org/10.1061/(ASCE)MT.1943-5533.0001475.
mance and to develop new specifications for the use of EAFS [16] H. Sebaaly, J. Khnaizer, Usage of steel slag aggregate in HMA mixtures and
in different applications. aggregate base, in; 3rd Middle East Society of Asphalt Technologists (MESAT)

A large-scale field performance is needed to investigate the Conference, American University in Dubai, UAE, April 6–8, 2015.
[17] M. Pasetto, N. Blado, Experimental evaluation of high performance base course
HMA mixtures and granular base blended with EAFS aggregate and road base asphalt concrete with electric arc furnace steel slags, J. Hazard.
to be compared with the predicted QRSS pavement perfor- Mater. 181 (1–3) (2010) 938–948.
mance and calibrate it, if warranted. [18] M. Pasetto, N. Blado, Mix design and performance analysis of asphalt concerts
with electric arc furnace slag, Constr. Build. Mater. 25 (2011) (2011) 3458–

A comprehensive chemical testing is required to acquire the 3468.
surface texture, chemical composition of compounds, and func- [19] A.I.A. Behiry, Evaluation of steel slag and crushed limestone mixtures, as
tional groups of the investigated materials and fully interpret subbase material in flexible pavement, Ain Shams Eng. J. 4 (1) (2013) 43–53,
https://doi.org/10.1016/j.asej.2012.07.006.
some of the testing results i.e., HMA moisture damage.
[20] F. Maghool, A. Arulajah, S. Horpibulsuk, Y. Jun Du, Laboratory evaluation of

Utilization of steel slag aggregate can benefit the environment ladle furnace slag in unbound pavement-base/subbase applications, J. Mater.
and at the same time reduce the amount of limestone and dolo- Civ. Eng. 29 (2) (2016), https://doi.org/10.1061/(ASCE)MT.1943-
mite applications in highway construction. However, some bar- 5533.0001724.
[21] A. Behnood, M. Ameri, Experimental investigation of asphalt mixtures
riers i.e., equipment wear owing to dealing with the slag containing steel slag, Scientia Iranica, Trans. A: Civ. Eng. 19 (5) (2012) 1214–
materials should be studied. 1219, https://doi.org/10.1016/j.scient.2012.07.007.
 M. Shiha et al. / Construction and Building Materials 248 (2020) 118710 15

[22] Egyptian Code of Practice for urban and rural roads (ECP), Road materials and [38] ARA, Inc., ERES Consultants Division, ‘‘Guide for Mechanistic-Empirical Design
their tests (ECP-2008 part four), The Ministry of Housing, Utilities and Urban of New and Rehabilitated Pavement Structures NCHRP 1-37 A Final Report,
Communities, Egypt, 2008. Transportation Research Board, National Research Council, Washington, D.C,
[23] American Society for Testing and Materials (ASTM), Standard test method for 2004.
density of semi-solid bituminous materials (Pycnometer Method) ASTM D 70, [39] M.W. Witczak, D. Andrei, W. Mirza, Development of Revised Predictive Model
ASTM, West Conshohocken, PA, 2018. for the Dynamic (Complex) Modulus of Asphalt Mixtures, Interim technical
[24] American Association of State Highway and Transportation Officials report, National Cooperative Highway Research Program (NCHRP) Project 1-
(AASHTO), Standard method of test for penetration of bituminous materials 37A, University of Maryland, 1999.
Test Procedure T49-15, AASHTO, Washington, D.C, 2019. [40] NCHRP, A Performance-related Specification for Hot-mixed Asphalt NCHRP
[25] AASHTO, Standard method of test for softening point of bitumen (ring-and- Report No. 704, Transportation Research Board, Washington, D.C., 2011.
ball apparatus) Test Procedure T53-,, AASHTO, Washington, D.C, 2018. [41] H. Ezzat, S.M. El-Badawy, A. Gabr, S. Zaki, T. Breakah, Predicted performance of
[26] AASHTO, Standard method of test for kinematic viscosity of asphalt (Bitumen) hot mix asphalt modified with nano-montmorillonite and nanosilicon dioxide
Test Procedure T201-15, AASHTO, Washington, D.C, 2019. based on egyptian conditions, Int. J. Pavement Eng., Taylor & Francis (2018),
[27] AASHTO, Standard method of test for flash and fire points by Cleveland open https://doi.org/10.1080/10298436.2018.1502437.
cup Test Procedure T48-17, AASHTO, Washington, D.C, 2018. [42] S. El-Badawy, M. Jeong, M. El-Basyouny, Methodology to Predict Alligator
[28] Contra steel company, The chemical composition of electric arc furnace steel Fatigue Cracking Distress based on AC Dynamic Modulus, Transportation
slag, http://www.yellowpages.com.eg/en/profile/contra-steel/68383, 2014 Research Record, J. Transport. Res. Board, No. 2095, Transportation Research
(Viewed on 14th August 2014). Board of the National Academies, Washington, DC, 2009, pp. 115–124.
[29] ASTM, Standard Test Method for Sand Flat Particles, Elongated Particles, or [43] ANSYS Users’ Manual, ANSYS Inc., Canonsburg, PA, 2002.
Flat and Elongated Particles in Coarse Aggregate ASTM D 4791, ASTM, West [44] R. Mousa, S. El-Badawy, A. Azam, A. Gabr, M. Arab, Resilient modulus
Conshohocken, PA, USA, 2010. characterization for granular base material in Egypt, 8th International
[30] ASTM, Standard. Test Method for Sand Equivalent Value of Soils and Fine Engineering Conference, Sharm Al-Shiekh, Egypt, 2015.
Aggregate ASTM D 2419, ASTM, West Conshohocken, PA, USA, 2014. [45] J.M. Duncan, C.L. Monismith, E.L. Wilson, Finite Element Analyses of
[31] AASHTO, Standard method of test for specific gravity and absorption of coarse Pavements, in: Highway Research Record No. 228, TRB, National Research
Aggregate Test Procedure T85-14, AASHTO, Washington, D.C, 2018. Council, Washington, D.C., 1968, pp. 18–33.
[32] AASHTO, Standard method of test for specific gravity and absorption of fine [46] M. Jones, M. Witczak, Subgrade modulus on the San Diego test road, Transp.
Aggregate Test Procedure T84-13, AASHTO, Washington, D.C, 2017. Res. Rec. 641 (1977) 1–6.
[33] AASHTO, Standard method of test for clay lumps and friable particles in [47] R. Ji, N. Siddiki, T. Nantung, D. Kim, Evaluation of resilient Modulus of subgrade
aggregate Test Procedure T112-00, AASHTO, Washington, D.C, 2017. and base materials in Indiana and its implementation in MEPDG, Sci. World J.
[34] AASHTO, Standard method of test for resistance to degradation for Small Size (2014) 14.
Coarse aggregate by abrasion and impact in the Loss Angeles machine Test [48] A.S. El-Ashwah, E. Mousa, S.M. El-Badawy, M.A. Abo-Hashema, Advanced
Procedure T96-15, AASHTO, Washington, D.C, 2019. Characterization of Unbound Granular Materials for Pavement Structural
[35] AASHTO, Standard method of test for soundness of aggregate by sodium Design in Egypt, Transportation Research Board 99th Annual Meeting, Paper#
sulfate or magnesium sulfate Test Procedure T104-03, AASHTO, Washington, 20-02764, Transportation Research Board, Washington D.C., USA, 2020.
D.C, 2016. [49] A. Arisha, A. Gabr, S. El-Badawy, S. Shwally, Performance evaluation of
[36] AASHTO, Standard method of test for moisture-density relations of soils using construction and demolition waste materials for pavement construction in
a 4.54-kg (10-Ib) Rammer and a 475-mm (18-in) Drop Test Procedure T180- Egypt, ASCE’s J. Mater. Civ. Eng. 30 (2) (2018) 1–14, https://doi.org/10.1061/
04, AASHTO, Washington, D.C, 2019. (ASCE)MT.1943-5533.0002127.
[37] AASHTO, Standard method of test for resilient modulus of subgrade soils and [50] E. Mousa, A.M. Azam, M. El-Shabrawy, S.M. El-Badawy, Laboratory
untreated base/subbase materials Test Procedure T307-12, AASHTO, characterization of reclaimed asphalt pavement for road construction in Egypt,
Washington, D.C, 2017. Can. J. Civ. Eng. 44 (6) (2017) 417–425, https://doi.org/10.1139/cjce-2016-0435.