International Journal of Pavement Engineering

ISSN: 1029-8436 (Print) 1477-268X (Online) Journal homepage: http://www.tandfonline.com/loi/gpav20

Aggregate gradation theory, design and its impact

on asphalt pavement performance: a review

Mingjing Fang, Daewook Park, Jean Louis Singuranayo, Hao Chen &
Yueguang Li

To cite this article: Mingjing Fang, Daewook Park, Jean Louis Singuranayo, Hao Chen
& Yueguang Li (2018): Aggregate gradation theory, design and its impact on asphalt
pavement performance: a review, International Journal of Pavement Engineering, DOI:
10.1080/10298436.2018.1430365

To link to this article: https://doi.org/10.1080/10298436.2018.1430365

Published online: 05 Feb 2018.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=gpav20
International Journal of Pavement Engineering, 2018
https://doi.org/10.1080/10298436.2018.1430365

Aggregate gradation theory, design and its impact on asphalt pavement

performance: a review
Mingjing Fanga, Daewook Parkb, Jean Louis Singuranayoa, Hao Chena and Yueguang Lia
a
Department of Road and Bridge Engineering, School of Transportation, Wuhan University of Technology, Wuhan, PR China; bDepartment of Civil
Engineering, Kunsan National University, Kunsan, South Korea

ABSTRACT ARTICLE HISTORY

Gradation is one of the crucially important features of aggregate blend. The objective of this paper is to Received 15 February 2016
present a review on the theory and design method of aggregate gradation, followed by its effect on asphalt Accepted 14 January 2018
pavement performance. The results indicate that, (i) Theory of maximum density curve, particle interfering
KEYWORDS
theory and fractal theory are the main theories for aggregate gradation. (ii) The gradation design of asphalt Asphalt mixture; aggregate
mixture has two types by aggregate morphology-continuous and gap gradation, or three types by air gradation; pavement
voids of mixture-dense, open and semi-open gradation. However, the only constant principle of aggregate performance; virtual
gradation design is to strictly follow the requirements of pavement performance. (iii) Aggregate gradation reconstruction; variability
significantly influences almost all of the pavement performances. The performance-based design method control
for asphalt mixture should pay more attention on aggregate gradation. (iv) The variability of aggregate
gradation cannot be eliminated due to its objective existence from aggregate production in quarry to
mixture compaction in field. Several effective control measures are summarized to limit gradation
variability under an acceptable range. (v) The latest developments and the trend for gradation analysis are
numerical simulation based on virtual reconstruction, image-based modelling and computer-generated
modelling.

Introduction asphalt pavement. It is believed that this review can serve as a

reference for pavement engineers, especially for those who are
For any specific material type of modern building construction,
working with asphalt mixtures.
aggregates in the mixture are required to follow a certain rec-
ommended particle size distribution. This size distribution is
known as aggregate gradation. Asphalt mixture, as a well-known Gradation theories
road construction material, is widely used in asphalt pavement.
Mineral aggregate blend is the major component of asphalt mix- Asphalt mixture generally means the mixture for asphalt con-
ture and aggregate is vitally important in mixture composite. crete, defined as any composite material composed of min-
Physical and chemical characteristics, morphologies (shape, eral aggregate adhered with an asphalt binder (Roberts et al.
angularity and texture) and gradation of aggregate are the main 1996). It is a multi-phase dispersed system with a spatial net-
factors influencing asphalt mixture’s properties. Therefore, aggre- work structure formed by an aggregate skeleton and asphalt
gate gradation plays a key role on asphalt mixture’s properties binder. Essentially, asphalt mixture is a type of loose struc-
and further significantly influences the performances of asphalt ture even bonded by asphalt. The mechanical performance of
pavement. However, pavement engineers is faced with challenges an asphalt mixture depends on the cohesion (c) and internal
both in design and construction, as well as maintenance of pave- friction resistance (φ) by (Chen and Liao 2002, Mamlouk and
ments, especially in the assimilation of theories and design appli- Zaniewski 2011). The formation of asphalt concrete starts from
cations of aggregate gradation, the effect evaluation of gradation aggregates selection, and the basic requirement of its bearing
to the field performance and the quality control (QC) of the capacity comes from traffic and environmental loads. It can be
aggregate gradation variability. For a better understanding of seen that different aggregate properties produce different road
aggregate gradation issues during asphalt mixture design, it is performances of asphalt concrete pavement. Meanwhile, the
still necessary to summarise the aggregate gradation theories, strength of asphalt concrete is mainly formed by aggregates,
design methods and how gradation affects the performance of asphalt binder and their spatial distribution status in mixture.

CONTACT Mingjing Fang [email protected]

© 2018 Informa UK Limited, trading as Taylor & Francis Group
2  M. FANG ET AL.

This leads to the volumertic properties extremely important. problem had also been corroborated later by other researchers
Due to the difficulty in using voids of each aggregate to design (Walsh 1933, Besson 1935, Blanks et al. 1940). The development
asphalt mixture directly, the related mass and density are easier of more reliable grading theories was still necessary. In the early
to gain by appropriate tests. Therefore, the percentage mass is of the 1960s, the Federal Highway Administration (FHWA) pro-
adopted to represent gradation instead of percentage volume in posed a gradation curve based on the Fuller formulation, taking
an asphalt mixture. Figure 1 indicates the volumetric properties 0.45 as the value of the power n, as shown in Figure 2 (FHWA
of aggregate particles in asphalt mixture. 1962). The maximum density curve can be easily determined by
Asphalt mixture design is a process based on the volumet- this power of 0.45 in Equation (1), and which can also freely adjust
ric parameters, which is essentially different from strength- aggregate gradation. This gradation theory is still commonly used
based design to cement concrete. Thus, the aggregate gradation in current design of asphalt mixture. In addition, researchers of
with different particle composition becomes highly important the former Soviet Union proposed aggregate curve with k, rep-
(Roberts et al. 1996). Generally, three main theories related to resenting the declining coefficient of the residue on sieve, based
aggregate gradation can be summarised as, theory of maximum on the maximum density theory (Zhang 2003), which was named
density curve, particle interfering theory and fractal theory. as K-method. In the 1970s, a Chinese researcher developed a
Meanwhile, numerical simulation based on virtual reconstruc- grading calculation method with the parameter i, representing
tion is rapidly developing for gradation analysis in recent years. the percentage passing decreasing coefficient when particle sizes
As a matter of fact, the gradation theories, including the theory decreased by ½ times (Lin 1998), which was later called I-method.
of maximum density curve and the particle interfering theory, Moreover, a comprehensive analysis of aggregate composition for
are originally proposed for cement concrete. These methods are asphalt mixture was provided in the same literature. Later, the
also appropriate for asphalt mixture. equation which contained both the Superpave and Bailey design
methods was put forward (Lin 2003). The percentage passing
of different particle size, d, for aggregate composition of asphalt
Theory of maximum density curve
concrete can be calculated by the equation as below,
The research on aggregate gradation started in the beginning
Px = P0 I x (%) (2)
of 1900s. From the research by (Fuller 1907), aggregates can be
formed with maximum density in cement concrete mixture if where, P0 = percentage passing of nominal maximum aggregate
particle sizes are well graded. Furthermore, the same literature size (NMAS), D, normally controlled with range of 90 to 100%;
indicates that aggregate gradation should be controlled for its x = series expression, x = 3.32lgD/d; I = decrease coefficient of
obvious impact on the performances of cement concrete. As a passing rate, I = 0.64 ~ 0.70.
result, the simplified ideal curve of parabola maximum density
was put forward. This concept was later improved by the work
Particle interfering theory
of (Wig et al. 1916 and then became the popular Fuller’s grad-
ing curve (namely N-method), represented by the following According to a study by (Weymouth 1933), the voids between
equation, bigger particles should be filled by the smaller size, d, for obtain-
ing the maximum density of aggregate blend. Meanwhile, the
p = 100 (d∕D)n (1) size of the filling particles must be not greater than the gap, t,
where, P = percentage passing of particle size ‘d’, %; d = aggregate between the particles. Otherwise, the particles will cause interfer-
particle size, mm; D = maximum size of aggregate particle, mm; ing phenomenon. Figure 3 depicts three typical types of particle
n = grading parameter, which can control the particle distribu- distribution by interfering theory.
tion (for maximum density curve n = 0.5). In order to avoid the interference (particularly t < d), particles
However, it was difficult to control the mixture performance with different sizes should be arranged in a certain manner. The
using the Fuller’s grading curve (Talbot and Richard 1923). This space between greater particles can be determined according to

A
B
C

Figure 1. Volumetric properties of aggregate particles in asphalt mixture. A. Voids in Mineral Aggregate (VMA) filled by asphalt binder; B. Asphalt binder film; C. Solid
aggregate; D. Surface voids without absorbed asphalt; E. Inner voids; F. Surface voids with absorbed asphalt.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  3

Sieve Sizes Raised to 0.45 Power

100 100

90 90

80 80
Percentage Passing By Mass (%)

70 70

60 60

50 50

40 40

30 30

20 20

10 10

0
0.075 0.15 0.30 0.60 1.18 2.36 4.75 9.5 12.5 19.0 25.0 37.5
Sieve Size (mm)

Figure 2. Maximum density curves on FHWA 0.45 power gradation chart (FHWA 1962).

d
t D
D

(a) (b) (c)

Figure 3. Three types of particle distribution by interfering theory. (a) t > d; (b) t = d; (c) t < d.

the critical interference gap, which is calculated by the following it could make the mixture structure generated in this way more
equation, stable due to the interfering influence of particles. This theory
[( )1∕3 ] can be used to calculate not only the continuous gradation, but
𝜓0 also the gap gradation. For continuous gradation, d/D = 1/2, the
t= −1 D (3) real packing ratio can be determined by the above equation, and
𝜓a
then the required amount for different aggregates can be further
For critical interference gap, t = d, Equation (3) can be expressed obtained.
by Equation (4) as below,
( )3 Fractal theory
d
𝜓 a = 𝜓0 ∕ +1 (4) Fractal theory was established by B. B. Mandelbrot, a French
D
mathematician, in the middle 1970s. This idea was aimed to
where, t = space between greater particles, mm; d = sizes of study unordered system with self-similarity which widely existed
sub-particles, mm; D = sizes of greater particles; ψ0 = theoreti- in nature or in social activities (Mandelbrot and Wheeler 1983).
cal packing ratio = ratio of packing density and apparent density Then, the fractal geometry was introduced to material technol-
of sub-particles, %; ψa = real packing ratio of sub-particles, %. ogy in 1990s. The variability, uncertainty, fuzziness and nonlin-
The particle interference theory is constrained by the size earity of asphalt mixture’s properties, especially the distribution
and distribution of different particles. As similar as the theory of aggregate particles, air voids and asphalt binders, are very
of maximum density, the objective is still the same as that of similar with the features of fractal theory. This concept was later
the maximum density with the assumption based on the theory applied to the analysis of aggregate gradation for asphalt pave-
of maximum density curve. However, the skeleton structure by ment mixture (Meakin 1999). The detailed process to calculate
interfering theory makes aggregates mixing denser. Meanwhile, the particle size and mass of aggregate using fractal theory is
4  M. FANG ET AL.

given by Equation (5–17) (Huang 2006). Suppose r is the size of rmin

C ⋅ D𝜌kV ( 3−D )
∫rmax
3−D
particle; N(r) is the amount of particle size greater than r; N′(r) is M= dM(r) = rmax − rmin (15)
3-D
the amount of particle size smaller than r; N0 is the total amount
of aggregate particles, and then, The aggregate mass less or equal to the particles size r is,
N(r) + N � (r) = N0 (5)
rmin
[ ]( )
∫r
M(r) = dM(r) = CD𝜌kV ∕(3 − D) r 3−D − rmin
3−D
(16)
The distribution function of particle size Ф(r) can be expressed
as follows, The distribution function of aggregate mass is,
Φ(r) = N � (r)∕N0 (6) ( ) ( 3−D )
P(r) = M(r)∕M = r 3−D − rmin
3−D 3−D
∕ rmax − rmin (17)
N(r) = Cr −D
(7) The above theory was applied in the research presented
where, C = a constant, and D = the distribution parameter of in (Leonardi 2016). A correlation between asphalt mixture’s
aggregate. If the maximum size of aggregate is rmax, and the min- mechanical properties and the fractal characteristics of the
imum size is rmin, then, aggregates mixtures has been built. Hence, fractal theory can
( ) ( ) ( D ) be employed to design aggregate gradation to meet the require-
N0 = N rmin − N rmax = C rmin D
− rmax (8) ments of asphalt mixture, and even to predict the characteristics
of asphalt concrete. In addition, the research presented in (Yuan
( ) ( D ) 2004) indicated that aforementioned three methods based on the
N � (r) = N rmin − N(r) = C rmin D
− rmax (9) theory of maximum density curve, N−, K− and I-method, were
the different expressions of the fractal theory. Therefore, this
From Equations (8) and (9), the distribution function of is a valuable direction to study the unified method to analyse
aggregate particle size can be expressed as, aggregate gradation for asphalt mixture design.
� −D
N (r) rmin − r −D
𝜙(r) = = −D −D (10) Numerical simulation
N0 rmin − rmax
With the rapid development of image processing and computing
The differential of Equation (7), technology, numerical simulation based on reconstruction of aggre-
gate distribution has been developed. Yang et al. (2016) presents a
dN(r) = −C ⋅ Dr −1−D dr (11) comprehensive review on the development of heterogeneous model
The mass of aggregate is, reconstruction of stone-based composites with discrete element
modelling (DEM) and finite element modelling (FEM). From this
dM(r) = 𝜌V (r)dN(r) (12) review, there are two types of heterogeneous models for stone-based
where, ρ = density of aggregate, g/cm3; V(r) and dN(r) are the materials. One is the image-based model obtained through image
volume and particle amount of aggregate in the size region (r, processing on X-ray (Zhang et al. 2016) or optical images (You
r + dr), respectively. The volumetric equation is, 2008). Figure 4 shows the image processing work (You 2008). In
general, image-based models can capture the detailed geometrical
V (r) = kv r 3 (13) information, but it is costly and time consuming, and the model
where, kv is shape factor. Substitute Equations (11) and (13) using accuracy is highly dependent on the image processing techniques.
Equation (12), Another one is computer-generated model obtained by placing
computer-generated aggregate particles into asphalt or cement
dM(r) = −C ⋅ D𝜌kV r 2−D dr (14) matrix. Compared to image-based model, the computer-generated
The total mass of aggregate is, model is more cost effective and much easier to implement, but

Figure 4. Illustration of image processing (You 2008). (a) original optical image for asphalt concrete; (b) processed image after banalisation; (c) DE model for numerical
simulation；and three-dimensional DE model generated from two-dimensional DE models.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  5

the main concern is the accuracy of the aggregate model shapes. based on virtual reconstruction. Numerical simulation for select-
Therefore, to combine the advantages from the both is a possible ing suitable gradation can save a considerable time and effort
and effective way to make a deep analysis on aggregate gradation. before conducting the gradating tests in laboratory.
In addition, a generalised framework was developed to There are some other contributions to gradation theory.
identify the range of aggregate particle sizes using a numerical The aggregate-based performance evaluation was described in
method based on packing theory (Lira et al. 2013). The frame- (Greene et al. 2011). In which, the subsequent experimental val-
work has proven to be an effective tool for identifying those idation efforts and findings of the accelerated pavement testing
mixtures with a poor rutting performance in terms of aggregate were also summarised. The results indicated that the evaluation
gradation. Investigations on aggregate packing via discrete-ele- method can be used at the phase of mixture design to assess the
ment modelling (DEM) simulation study (Shen and Yu 2011, Yu field performance of an asphalt mixture on the basis of aggre-
and Shen 2012) were conducted to link the aggregate gradation gate gradation. In addition, (Guarin et al. 2013) focused on how
property to the VMA of asphalt mixture. A gradation weighting asphalt mixture performance was affected by changes in inter-
factor, fv, was introduced to theoretically evaluate the packing of stitial component (IC), which is the material between dominant
an aggregate structure. Models to calculate fv with different sizes aggregate size range (DASR) particles. The disruption factor (DF)
are shown in Figure 5. was introduced to evaluate the potential of IC aggregates to dis-
In another study by (Chen et al. 2015), DEM simulation rupt the DASR structure. This has been proven to be the effective
approach was applied to evaluate aggregate structure characteris- way to understand the conception of aggregate.
tics and the resistance to deformation. Three-dimensional aggre- In a word, for the theories and methods reviewed above, the
gate blend was employed to simulate penetration test of aggregate theory of maximum density curve is widely applied to support
blends numerically, which indicated that the aggregate particle mixture gradation design due to its simple and clear conception;
size played an important role on the mechanical performance of Particle interfering theory is one of mathematical expressions of
aggregate skeleton. Particles state change in graded blend before maximum density theory essentially; however, it is not as simple
and after vibrating compaction is illustrated in Figure 6. as the theory of maximum density curve. Fractal theory has
A micromechanical analysis on various asphalt mixtures more complicated expression and difficulty in understanding, but
generated from computer-simulated aggregate gradations was which can hopefully unify the theories of gradation analysis. As
performed (Singh et al. 2011). It is expected that such a process for the numerical simulation methods based on digital technol-
would save a considerable time and effort while finalising a suit- ogy, they are still in the processing and must be the developing
able gradation for the mixture design. Bruno et al. (2012) con- trend of aggregate gradation analysis, especially the combina-
ducted a study with the purpose of finalising an effective analysis tion of image-based modelling and computer-based modelling.
of asphalt section image for automatically extracting aggregate The top challenge for numerical simulation is to narrow the gap
gradation without the need of separation of the bitumen from between the calculated particle distribution and real aggregate
aggregate. A simple two-dimensional digital image processing blend. Currently, a lot of efforts need to be contributed to realise
(DIP) technique was used to obtain aggregate gradation curves pure numerical simulation for gradation design in real project.
for a set of 28 asphalt mixtures prepared with different asphalt
binders and air voids, aggregates having various NMAS and three Gradation applications
percentages of reclaimed asphalt pavement (RAP) (Moon 2015).
In another study (Liu et al. 2017), a microstructural model of Gradation design methods
asphalt mixture generated by FEM used for uniaxial compression The above-mentioned gradation theories and analysis methods
test was reconstructed based on X-ray CT scans, thus maintain- cannot directly guide the process of mixture design. There are
ing the original morphology of the aggregate. The results showed
several design methods developed on the basis of the theories
that the aggregated angularity significantly affects the mechanical
above, such as Superpave method (AI 2001), Bailey method
response of the asphalt mixture. Aggregates with different angu-
(Aschenbrener et al. 2002), Coarse Aggregate Void-filling
larities can be seen in Figure
(√ ) et al. 2017), in which, 3DTS
7 (Liu
2 (CAVF) method (Zhang et al. 2001), Aggregate Gradation
is 3D True Sphericity = 𝜋S 3 6V 𝜋
. Where, S and V represent Design and Test method for Stone Asphalt Concrete (Sha 2005).
the actual surface area and actual volume of aggregate particle, For different design conditions, the specific requirements of
respectively. These researches enriched the numerical simulation gradation are varied. As for asphalt mixture design, currently,

Figure 5. Models to calculate fv with different sizes (Shen and Yu 2011). (a) model contains aggregate larger than 4.75 mm; (b) model contains aggregate larger than
2.36 mm; (c) model contains aggregate larger than 1.18 mm.
6  M. FANG ET AL.

Figure 6. Particles in the graded blend (Chen et al. 2015). (a) before vibrating compaction; (b) after vibrating compaction.

Figure 7. Aggregates with different angularities (Liu et al. 2017). (a) 3DTS = 0.395 (real morphology); (b) 3DTS = 0.410; (c) 3DTS = 0.425; (d) 3DTS = 0.440 (near ellipsoids).

there are two main gradation types categorised by aggregate gradation are constrained in a narrow range. However, this
morphology, namely continuous gradation and gap gradation, type of gradation is generally not used for asphalt mixture.
or three types by air voids, namely dense gradation, open gra- It is often used in Portland cement concrete (PCC) mixture.
dation and semi-open gradation. Continuous gradation is the Due to the difference of aggregate source and morphology,
representation of all the standard particle sizes in a certain pro- the basic selection principle of aggregate gradation type for
portion, in which the voids between larger particles is effectively asphalt mixture design is strictly to follow the requirements
filled by smaller particles to produce a well-packed structure. of pavement performance. Figure 8 shows the schematic dia-
Gap gradation is a kind of gradation which lacks one or more gram of four gradation types (Memphis 2016).
intermediate size.
Dense-open, open and semi-open gradation in China are
usually defined by air voids with 3–6%, >18% and 6–12%,
Gradation control
respectively. In practice, continuous dense graded asphalt
mixture (AC) and asphalt treated base (ATB) are the com- There are several terminologies commonly used in gradation
monly used mixture types. They are designed to have a dense design of asphalt mixture, such as sieve screening, fine grada-
structure to prevent moisture infiltration and at the same time tion, coarse gradation, fine aggregate, coarse aggregate, mineral
maintain adequate voids to allow asphalt volume expand with filler, mineral dust, control points and so on. The first challenge
the temperature increased. Stone mastic asphalt or stone-ma- on working with asphalt mixture is to differentiate those termi-
trix asphalt (SMA) is a conventional type of gap dense graded nologies. Sieve test is the most effective way to do analysis of
asphalt with a coarse stone skeleton providing more stone- aggregate gradation. From the test, the percentage retained and
on-stone contact. Asphalt macadam (AM) for pavement base the percentage passing by mass can be calculated (see Figure 9)
belongs to gap semi-open gradation mixture. Open graded (Memphis 2016). For the fine and coarse gradation, empirically
friction coarse (OGFC) and asphalt treated permeable base and basing on visual investigation, they are conceptually sepa-
(ATPB) are the mixture with typical gap open gradation. In rated depending on whether the gradation curve is passing below
addition, it is noticed that there is another gradation type or above the maximum density 0.45 power curve (Interactive
known as uniform gradation. The particles sizes in uniform 2011),which is shown in Figure 10.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  7

Figure 8. Four typical gradations of aggregate composition. (Memphis 2016). (a) Dense gradation; (b) Open gradation; (c) Gap gradation; (d) Uniform gradation.

(a) (b) (c)

percent retained or
percent coarser, by
mass

percent passing
or percent finer,
by mass

Figure 9. Devices for sieve screening and gradation analysis (Memphis 2016). (a) Sieves; (b) Shaker; and (c) Gradation analysis.

For Superpave mixture design, the specification of American (No. 30) sieve. However, this definition was not popularly
Association of State Highway and Transportation Officials used. Meanwhile, the current definition for mineral dust is
(AASHTO) has made a division for coarse and fine gradation still 0.075 mm (No. 200) sieve, which is adopted in worldwide
using primary control sieve (PCS) (AASHTO 2007), respectively. (AI 2001).
Specifically, it is coarse gradation if percentage passing of aggre- Although hydrated lime, cement and some natural dust can
gate in PCS is less than the limit as shown in Table 1, or fine be considered as mineral filler, the most common filler is lime-
gradation if not. Meanwhile, AASHTO also defined the control stone powder. In the former method of Superpave design, VMA
point with different NMAS as shown in Table 2. ‘Trial-and-error’ was very small due to more fine aggregate by the theory of the
for gradation design is the method only based on control points. maximum density curve. Therefore, the conception of gradation
Figure 11 shows the gradation control zone and maximum in restricted zone was put forward to limit the negative effect
density curve for NMAS = 25.0, 19.0, 12.5 and 9.5 mm, respec- (ASTM 2003). However, a later report supported by National
tively (Interactive 2011). Cooperative Highway Research Program (NCHRP) was sub-
For fine aggregate, the standard definition by American mitted by National Center for Asphalt Technology (NCAT),
Society for Testing and Materials (ASTM) is the aggregate which showed that the mixture with gradation curve passing
passing through the 9.5 mm (3/8-in) sieve in 100% and almost through restricted zone had similar, even better performance
100% passing through 4.75 mm (No. 4) sieve but predomi- than the mixture gradation with considering the zone (Kandhal
nantly retained on the 0.075 mm (No. 200) sieve (ASTM 2003). and Cooley 2002). Subsequently, the report suggested AASHTO
For coarse aggregate, it is defined by ASTM as the aggregate committee to remove the concept of restricted zone. Additionally,
which almost cannot pass through the 4.75 mm (No. 4) sieve. other research reports also indicated that there was no relation-
These two definitions for fine and coarse aggregate are widely ship between the performance such as rutting and fatigue of
accepted. In 2001, Asphalt Institute (AI) defined the fine asphalt mixture and Superpave restricted zone (Prowell et al.
aggregate as the aggregate that can pass through the 0.6 mm 2005, Kim et al. 2009). Currently, the restricted zone is no longer
8  M. FANG ET AL.

No.100

3/4-inch
3/8-inch

1/2-inch
No.200

No.50

No.30

No.16

No.8

No.4
100

80 ve

Percent Passing by Mass (%)

C ur
er
ow
5P
(1) 0.4
60

40 Coarse gradation. A gradation that,

when plotted on the 0.45 power
gradation graph, falls mostly below
20 the 0.45 power maximum density line.
The term generally applies to dense
graded aggregate.
0
0.075 0.150.30 0.60 1.18 2.36 4.75 9.5 12.5 19.0
Sieve Size (mm)
No.200
No.100

3/4-inch
3/8-inch

1/2-inch
No.50

No.30

No.16

No.8

100 No.4
Fine gradation. A
gradation that, when plotted
on the 0.45 power gradation
80 graph, falls mostly above
Percent Passing By Mass (%)

rve
the 0.45 power maximum
density line. The term r Cu
generally applies to dense owe
60 5P (2)
graded aggregate. 0.4

40

20

0
0.075 0.150.30 0.60 1.18 2.36 4.75 9.5 12.5 19.0
Sieve Size (mm)

Figure 10. Diagram of coarse and fine gradation (NMAS = 19 mm) (Interactive 2011). (1) Coarse gradation; (2) Fine gradation.

Table 1. PCS of different NMAS (AASHTO M 323).

NMAS, mm 37.5 25.0 19.0 12.5 9.5
PCS, mm 9.5 4.75 4.75 2.36 2.36
Percentage passing, % 47 40 47 39 47

Table 2. Control point of different NMAS (AASHTO M323).

Maximum aggregate size- percentage passing of control points, %

37.5 mm 25.0 mm 19.0 mm 12.5 mm 9.5 mm 4.75 mm
Sieves, mm Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max.
50.0 100 – – – – – – – – – – –
37.5 90 100 100 – – – – – – – – –
25.0 – 90 90 100 100 – – – – – – –
19.0 – – – 90 90 100 100 – – – – –
12.5 – – – – – 90 90 100 100 – 100 –
9.5 – – – – – – – 90 90 100 95 –
4.75 – – – – – – – – – 90 90 100
2.36 15 41 19 45 23 49 28 58 32 67 – –
1.18 – – – – – – – – – – 30 60
0.075 0 6 1 7 2 8 2 10 2 10 6 12
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  9

3/2-inch
3/8-inch

3/4-inch
No.200
No.100

1/2-inch
No.16
No.30
No.50

1-inch

2-inch
No.8

No.4
100
(1) NMAS = 25.0 mm

Percent Passing By Mass (%)

80

60

40

Control Point
20
Maximum Density Curve
Gradation Range
0
0.075 0.15 0.30 0.60 1.18 2.36 4.75 9.5 12.5 19.0 25.0 37.5 50
Sieve Size (mm)

3/2-inch
3/4-inch
3/8-inch
No.200
No.100

1/2-inch

1-inch
No.50

No.16

2-inch
No.30

No.8

No.4
100
(2) NMAS = 19.0 mm
Percent Passing By Mass (%)

80

60

40

Control Point
20
Maximum Density Curve
Gradation Range
0
0.075 0.15 0.30 0.60 1.18 2.36 4.75 9.5 12.5 19.0 25.0 37.5 50
Sieve Size (mm)

3/2-inch
3/4-inch
3/8-inch

1/2-inch
No.100
No.200

1-inch
No.50

No.16

2-inch
No.30

No.8

No.4

100
(3) NMAS = 12.5 mm
Percent Passing By Mass (%)

80

60

40

Control Point
20
Maximum Density Curve
Gradation Range
0
0.075 0.15 0.30 0.60 1.18 2.36 4.75 9.5 12.5 19.0 25.0 37.5 50
Sieve Size (mm)

Figure 11. Gradation control zone of different NMAS (Interactive 2011). (1) NMAS = 25.0 mm; (2) NMAS = 19.0 mm; (3) NMAS = 12.5 mm; (4) NMAS = 9.5 mm.

considered as a specification for asphalt mixture design since usual (Roberts et al. 1996), the mineral aggregates properties have
2003. significant effect on asphalt mixture. A large amount of research
indicated that aggregate gradation can dramatically influence the
volumetric properties (Nukunya 2001, Shen and Yu 2011, Husain
Effect on pavement performances
et al. 2014) and pavement performances (Suresha et al. 2009,
From the theories and the design methods of aggregate grada- Sangsefidi et al. 2014, Behiry 2016). Therefore, aggregate grada-
tion, it is obviously to see the importance of aggregate gradation. tion reflects the aggregate blend composition of asphalt mixture,
In specific, making up its 93% by weight and 85% by volume in and can further affect almost all of the pavement performances
10  M. FANG ET AL.

3/2-inch
3/4-inch
3/8-inch

1/2-inch
No.200
No.100

1-inch
No.50

No.16

2-inch
No.30

No.8

No.4
100
(4) NMAS = 9.5 mm

Percent Passing By Mass (%)

80

60

40

Control Point
20
Maximum Density Curve
Gradation Range
0
0.075 0.15 0.30 0.60 1.18 2.36 4.75 9.5 12.5 19.0 25.0 37.5 50
Sieve Size (mm)
Figure 11. (Continued)

1.0
Log10 (k) = 2.183log10 (n)-3.502
R2 = 0.786
0.0 RD = 46%
n = 29
log10 (coefficient of permeability, k mm/s)

SE = 0.510
-1.0 p < 0.001

-2.0

-3.0 Log10(k) = 7.051log10(n)-8.505

R2 = 0.833
RD = 41%
-1.0 n = 34
SE = 0.390
p < 0.001
-1.0
Log10(k) = 4.208log10(n)-10.796
R2 = 0.876
-1.0 RD = 35%
n = 24 Zube (1962) (Water
SE = 0.261 permeability, field)
-1.0 p < 0.001
Vivar and Haddock (2007)
(Water permeability, laboratory)
-1.0 Goode and Lufsey (1965) (air
permeability, laboratory)
-1.0
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
Log10 (percent air voids, n%)

Figure 12. Correlation between permeability and air voids of asphalt mixture (Vivar and Haddock 2007, Goode and Lufsey 1965, Zube 1962).

of asphalt mixture (Roberts et al. 1996, Li and Ma 2007). Due to of porous asphalt mixtures. 10 aggregate gradations of OGFC,
the fact that pavement serviceability is more concerned in terms which represented typical gradations used by 20 different high-
of the objective of asphalt mixture design, this review is focused way agencies in United States, were compared by evaluating
on the gradation effects for permeability, moisture stability, low rain down, permeability, Cantabro abrasion loss, indirect tensile
and high temperature characteristic, fatigue and anti-slip. strength and rutting resistance. The test results indicated that gra-
dation significantly influences the performance of porous asphalt
mixtures. An increase in voids ratio of the aggregate structure
Effect on permeability and moisture stability
contributes to significant increase in porosity and permeability.
Although permeability and moisture stability are the two differ- Consequently, the indirect tensile strength (ITS) and durabil-
ent performances of asphalt mixture, the research by (Caro et ity (as measured by the Cantabro abrasion test) will generally
al. 2011) showed that there exists a strong relationship between decrease with the mixture porosity increased. The 10 mixtures
moisture damage and larger voids ratio. Increasing the voids showed good moisture resistance characteristics and most of the
ratio of asphalt mixture can cause the capacity of permeabil- mixtures had no significant difference in tensile strength after
ity to increase, and further makes stripping effect strengthened the moisture-conditioning procedure. Furthermore, the effects
between aggregate and binder film. Different aggregate grada- of rutting did not correlate with mixture or aggregate porosity.
tions have different voids ratio of mixture, which influences the Based on these results, it is evident that an aggregate gradation
permeability performance in usual. (Mansour and Putman 2013) for a porous asphalt mixture can be optimised, which depends
investigated the effect of aggregate gradation on the performance upon the performance requirements (permeability or strength)
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  11

based on the voids ratio of the aggregate gradation. The voids The study by (Kanitpong, Charoentham and Likitlersuang,
distribution can be linked to aggregate gradation and therefore, 2012) focused on understanding the effect of aggregate structure
a coarser asphalt concrete mixture is more likely to have inter- and its physical properties on the mechanism of moisture dam-
connected void channel. Air voids has strong correlation with age in WMA. It was found that the proper selection of aggregate
permeability, which can be proved by the data from literature type and gradation structure could reduce the moisture damage
(Vivar and Haddock 2007, Goode and Lufsey 1965, Zube 1962), problem of WMA. Husain et al. (2014) investigated the effect
as shown in Figure 12. of various aggregate gradations towards producing SFP, a com-
By conducting a comprehensive review for permeability of posite pavement. The results confirmed that different aggregate
asphalt concrete in terms of different air voids, gradations and gradations significantly affect the properties of SFP mixture. Test
binders, the findings showed that there was a strong correla- results conducted by (Sangsefidi et al. 2014) indicated that the
tion between aggregate gradation and permeability coefficient aggregate gradation had different effects on the moisture suscep-
(Vardanega 2014). Meanwhile, the maximum size of aggregate tibility of hot mixture asphalt (HMA) and warm mixture asphalt
and the percentage passing of 2.36 mm sieve were the key factors (WMA), respectively. The optimum aggregate gradation for these
to influence the permeability. Even at the same air voids, the two types of mixtures was different.
larger sizes of aggregates, the greater permeability coefficient
was. For semi-flexible pavement (SFP), similar results have also
Effect on high temperature performance
been proved in (Husain et al. 2014).
Moisture damage in asphalt mixtures refers to loss in strength High temperature performance mainly reflects the ability of
and durability due to the presence of water. Under the traffic asphalt mixture to resist the wheel tracking in high temperature
loading, the water seeped into the interface between aggregate environment or in a long time bearing, which usually uses the
and asphalt binder film can degenerate the adhesion capacity. dynamic stability as evaluation parameters. The test by (Chen and
This makes asphalt mixture deteriorated especially with the pro- Liao 2002) indicated that a good determination for the amount
cess of freezing and thawing. The results of moisture damage of fine aggregate which passed through 4.75 mm can increase
loosen aggregates, and result in some distresses like pothole. the capacity of rutting-resistance for asphalt mixture. Moreover,
The moisture damage is one of the main distresses for asphalt the function of inter-lock is mainly from coarse aggregate, which
pavement. The damage also intensifies the other distresses, such also showed that stable skeleton structure of aggregate can pro-
as bleeding, cracking, rutting and loosening. A related research duce greater internal friction for asphalt mixture. In addition,
by (Hicks 1991) showed that water damage can be influenced (Fan 2004) compared the mechanics and road performances
by numerous factors, such as aggregate properties, binder char- for three SMA mixtures and one AC mixture. The results found
acteristics, mixture types and the environmental change after that for the same type of asphalt mixture and similar aggregate
construction as well as additives, in which, the aggregate gra- gradation, there was no obvious difference for high temperature
dation directly influences moisture stability. Meanwhile, when stability when the sizes of aggregate changed. Meng et al. (2005)
the degree of aggregate segregation exceeded a certain limit, studied the high temperature performance of well-graded asphalt
the quality of asphalt mixture would deteriorate. Aggregate mixture (AC), and the result showed that the high temperature
segregation can cause the air voids distribution uneven. For stability would be decreased with increasing the amount of fine
example, the moisture stability dramatically decreases when air aggregate below 2.36 mm in AC mixture. A mixture test research
voids exceed 7%. In addition, a numerical simulation with a with 5 different gradation curves met the China’s specification
micromechanical model of moisture damage was conducted by was conducted by (Yang 2007). The results showed adding more
(Caro et al. 2011) and it showed that the initiation and evolution fine aggregate can make the high temperature performance
of moisture-related deterioration was strongly influenced by the worse. However, the high temperature performance will not be
internal air voids structural of asphalt mixture. The air voids of always increased with more coarse aggregate.
asphalt mixture are highly controlled by aggregate gradation. Wang (2008) pointed out that the high temperature perfor-
The representation of the modelling methodology is shown in mance of asphalt mixture can be represented by its shear resist-
Figure 13. ance. The bigger NMAS was, the greater the inner friction angle

Moisture Vapor F
F

4.2N

15s 30s Time

10-day moisture Fracture Test
conditioning

Figure 13. Representation of the modelling methodology (Caro et al. 2011).

12  M. FANG ET AL.

and more obvious the skeleton structure would be. Meanwhile, to evaluate the cracking-resistance in low temperature for asphalt
gradation segregation can seriously affect the resistance to shear, mixture. Hao and Zhang (2000) concluded that different grada-
especially for fine gradation, the inner friction angle is smaller tions had different effect on creep speed, critical J-integral and
compared to the coarse gradation. Thus, the shearing resistance strain energy density. For the same gradation, properly increasing
will weaken, and further causes rutting problem. The research asphalt binder would be beneficial to increase the capacity for
about rutting by the test of APA (Asphalt Pavement Analyzer) cracking-resistance in low temperature. Four parameters deter-
under low traffic showed that gradation and aggregate fineness mined by the DASR-interstitial component (IC) model were well
are the key factors affecting the rutting depth (Kim et al. 2009). related to field performance of Superpave mixtures (Chun et al.
The results by (Greene et al. 2014) confirmed that the rutting 2012), as shown in Figure 14 (Greene et al. 2014). IC character-
performance of the asphalt mixture was primarily controlled istics were more important than DASR porosity to clearly differ-
by the DASR porosity. Mixtures with marginal DASR porosity entiate field cracking performance. Therefore, the introduction
might still have had significantly better rutting performance than of DASR-IC criteria into current mixture design guidelines and
mixtures with poor DASR porosity. The gradation design method specifications will help to ensure good field rutting and cracking
of coarse aggregate skeleton was studied by (Chen et al. 2015) performance of Superpave mixtures.
based on micromechanical characteristics of aggregate particles
in asphalt mixtures. It can be seen that the aggregate gradation
Effect on fatigue performance
is very similar to that of SMA, and thus the designed asphalt
mixture has good high-temperature performance. Sousa et al. (1998) adopted four-point bending fatigue test to
Wang et al. (2016) indicated that good linear correlations were study gradation effect on fatigue performance. The results showed
observed between the fractal dimension and the high-tempera- that the fine gradation with higher content of asphalt binder
ture performance indexes, such as the dynamic stability and max- had a better fatigue performance. Hao et al. (1998) selected five
imum shear stress from the DS test and the penetration shear test. asphalt mixtures with different gradations, and conducted a serial
Different gradations cause different temperature stability, which of tests to study the effect of air voids and asphalt contents on
was also proved in (Zhang et al. 2016). The micro-structure by fatigue performance. The results showed that dense gradation
X-Ray scanning showed that the SMA-13 mixture exhibited a should be more widely considered for meeting the requirement
better compaction performance and temperature stability than of fatigue performance during the pavement structural design.
the AC-13 mixture regardless of the asphalt binder types; the Test research (Abo-Qudais and Shatnawi 2007) about fatigue
asphalt mixture of AC-13 performed a better performance on life of asphalt mixture based on accumulated damage strain
temperature stability and compaction than the AC-16 and AC-20 suggested that it was better to adopt different asphalt types and
at the same test condition. gradations to evaluate the fatigue behaviour for asphalt mixture.
Tang et al. (2009) conducted an orthogonal test about the fatigue
factors for asphalt mixture. The test results indicated that the
Effect on low temperature performance
influencing factors from the view of importance in descending
The study about thermal cracking by (Kandhal 1998) showed that order were gradation, strain level, interval time, test temperature,
there was no strong relation between cracking and aggregate, but asphalt types and bitumen aggregate ratio. Breakah et al. (2011)
highly related to asphalt characteristic. The main problem in low studied the effect of fine aggregate on the predicted pavement
temperature is cracking of asphalt pavement. There are a large life by simulating the performance differences between mixtures.
number of test methods to evaluate low temperature performance These mixtures were prepared with different sources and gra-
for asphalt mixture. However, low temperature cracking was also dations of fine aggregate following the Mechanistic-Empirical
indeed affected by the distribution of aggregate particles. Cracking Pavement Design Guide (MEPDG), and showed that fine aggre-
in low temperature can be influenced by different gradations struc- gate has important impact to pavement service life. (Saleh 2012)
ture. A concept with compressive strain energy was put forward examined the fatigue behaviour of two different types of asphalt

DASR Characterization
• DASR Porosity
-Coarse Aggregate Interlock

IC Characterization
• Disruption Factor
-Volumetric Distribution of IC
• Effective Film Thickness
-Stiffening Effect of IC on Mixture
• Fine Aggregate Ratio
-Structure of IC

Figure 14. Scheme of characterisation for DASR and IC (Greene et al. 2014).
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  13

mixtures, namely AC10-80/100 and AC14-60/70. The fatigue and the better the skid resistance of pavement would be. Suresha
models that govern the behaviour of these mixtures were derived et al. (2009) presents the study on effect of aggregate gradation
and the effect of air voids on the fatigue behaviour was investi- on the mixture design and performance properties of porous
gated. Subsequently, new fatigue models were developed. friction coarse (PFC) – mixtures. Findings of the study clearly
The correlation between dynamic response and fatigue behav- indicate that the gradations specified by various agencies will
iour of different asphalt mixture has been studied and presented have significant and different effect on the design properties of
in (Ali et al. 2015). Dynamic modulus E* can be used to compre- PFC mixtures. Meanwhile，it also helps in forming the master
hensively reflect dynamic response of asphalt mixture. Research aggregate gradation band for PFC mixtures.
for E* testing has been conducted a lot and implemented into Recently, fractal theory has also been used to study skid per-
MEPDG to link the behaviour of asphalt mixture in laboratory formance of asphalt mixture (Singh and Walia 2014, Zhang et
and in field, in which, one model, namely Witczak |E*| prediction al. 2014, Hou et al. 2016). In (Hou et al. 2016), the 2D digital
model, was typically applied. An updated version of the model image of the pavement surface is measured in terms of area frac-
(Bari and Witczak 2006) is shown as below, tal dimension and contour fractal dimension. Aggregate grada-
2 2
⎛ 6.65 − 0.032𝜌200 + 0.0027𝜌200 + 0.011𝜌4 − 0.0001𝜌4 ⎞
⎜ � �⎟
∗ �
log10 E = −0.349 + 0.754(�Gb � ∗ �−0.0052
)×⎜ Vbeff ⎟
2
⎜ +0.006𝜌38 − 0.00014𝜌38 − 0.08Va − 1.06 V + V ⎟ (18)
⎝ a beff ⎠
� V �
2.56 + 0.03Va + 0.71 V +V beff
+ 0.012𝜌38 − 0.0001𝜌238 − 0.01𝜌34
a beff
+
1 + e(−0.7814 − 0.5785 log �Gb � + 0.8834 log 𝛿b )
∗

where, E* = dynamic well modulus, psi; ρ200 = percentage of tion and British Pendulum Number (BPN) was correlated. The
aggregates (by weight of the total aggregates) passing through results indicated that the area fractal dimension of aggregate
No. 200 sieve, %; ρ4 = aggregates (by weight) retained on No. 4 provided a simple way to acquire the continuous gradation of
sieve, %; ρ38 = aggregates (by weight) retained on 3/8 inch sieve, asphalt concrete sample and the contour fractal dimension was
%; ρ34 = aggregates (by weight) retained on 3/4 inch sieve, %; an available parameter to characterise roughness and friction
Va = air voids (by volume of the mix), %; Vbeff = effective binder of pavement surface texture microscopically. Figure 15 presents
content (by volume of the mix), %; |Gb*| = dynamic shear mod- the surface textures of five different types of pavements: AC,
ulus of binder, psi; δb = phase angle of binder associated with micro-surfacing (MS), OGFC, Rubber AC (Interactive) and
|Gb*|, degree. SMA. From the texture in microscopy, it is easy to show that
From the model above, it is obviously to see that different aggregate gradation can significantly influence the skid resist-
gradation can influence the dynamic modulus E*, and then fur- ance of asphalt pavement with different types. In another word,
ther cause different mechanical performance of asphalt concrete International Ride Index (IRI) and BPN can well estimate the
mixture, including fatigue behaviour of asphalt pavement. In skid resistance of asphalt pavement in macroscopic and micro-
addition, fatigue parameter can be derived from E* of asphalt cosmic view, respectively. Both of them are related to aggregate
mixtures to evaluate the resistance to fatigue (Ye et al. 2009). |E*| gradation of asphalt mixture.
× sinδ can be used to estimate fatigue parameter.

Gradation control
Effect on skid resistance of asphalt pavement
Factors of gradation variability
Wang and Zhang (2004) conducted a comparative test analysis
for the three antiskid surface mixtures. As far as safety is highly From the above review, aggregate gradation can almost influ-
concerned, the interface between wheel and pavement should ence all the performances for asphalt pavement. Therefore, it is
have sufficient friction. There exists obvious relation between quite necessary to know the importance to keep the gradation
the roughness of pavement and aggregate gradation. Zhao et al. stability during the asphalt mixture’s production, transportation,
(2005) conducted a test in laboratory using five gradations with paving and compacting. However, each step of asphalt mixture
different design curves and different NMASs to get the attenu- construction easily leads to gradation variability, which will cause
ation law of skid resistance. The results showed that the mass mixture uniformity and premature damage of pavement. Peng
content of the coarse aggregate above 4.75 mm and the maximum and Sun (2006) conducted a quantitative study on the relation
aggregate size had significant influence on skid resistance of between mixture uniformity and air voids, dynamic stability and
pavement. At the same time, the determination of the maximum mechanical strength. The results showed that the aggregate gra-
size of aggregate should be integrated by taking into considera- dation has great influence on the uniformity of asphalt mixture,
tion the layer thickness and temperature and so on. Huang et al. with the significance of the aggregate on the sieve of 4.75 mm
(2008) used discrete exposure size function to study the random greater than that on the size of 2.36 mm. Yu et al. (2007) stud-
distribution characteristic with surface micro-texture of asphalt ied the variability of gradation and bitumen aggregate ratio and
mixture. The findings indicated that the fractal dimension of the related influence on pavement performance. The findings
micro-texture was different for different aggregation. The greater revealed that gradation segregation was an objective existence
the fractal dimension, the greater the exponent of size function and gradation variability is much related with the strength and
14  M. FANG ET AL.

Figure 15. Surface textures of five different pavement types (Hou et al. 2016).

segregation degree. The results showed that the mixture with

Table 3. Allowable range of gradation fluctuating.
gradation curve of S shape and above the maximum density
Allowable fluctuating range/% curve can keep the gradation stability during the production and
Highway and first-class Other classified transportation of mixture. In addition, due to different lithology,
Terms highway highway processing equipment and technologies from different quarries,
Sieves /mm 0.075 ±2 ±2 the specifications of processed aggregates are varied and their
≤2.36 ±4 ±4
≥4.75 ±6 ±7 qualities are not always stable. Mixing with different aggregate
source also causes gradation variability during asphalt mixture
design. Therefore, aggregate segregation caused by production,
durability of asphalt mixture. Huang et al. (2009) conducted a transportation, paving and compaction of asphalt mixture can
research about the relation between six different gradations and lead to gradation variability to some extent.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  15

Control measures for gradation variability control, especially in the processing of production,
mixing, transportation and paving the mixture should
Ji and Wu (2008) summarised the gradation fluctuating from the
be monitored strictly to make the mixture uniformity
tests in laboratory, and the related range of gradation fluctuation
and to decrease gradation variability.
was recommended as indicated in the following Table 3.
(7) Due to the cost and time consuming, numerous related
In addition, the control of gradation variability should be
works can be done with numerical modelling before
linked to the whole processing from the aggregate selection,
conducting laboratory tests, such as DEM and FEM.
mixture production to paving and then compacting. Basing on
The challenge for numerical simulation is to narrow the
material selection, storage, adding additives, strictly controlling
gap between the calculated particle distributions and
for maximum aggregate size and gradation variability, optimising
real aggregate blend, and the combination of image-
construction devices of materials (Yuan et al. 2009), the main
based modelling and computer-generated modelling is
control measures were summarised.
the developing trend for gradation analysis.

Conclusions
Acknowledgements
This paper provided a review on several common issues of aggre- The authors highly appreciate unknown reviewers of this manuscript in
gate gradation for the asphalt mixture design, referred to the several times of revision, as well as thanks to Prof. Imad Al-Qadi and other
theories and applications of aggregate gradations, the effect of editors for their great patience and suggestions during revising.
gradation on pavement performances, the gradation variability
and the related control measures. From the review, several con-
clusions can be drawn as follows.
Disclosure statement
No potential conflict of interest was reported by the authors.
(1) Aggregate gradation is a certain recommended particle
size distribution. Using percentage passing rate by mass
(%) presents aggregate gradation instead by volume Funding
due to the difficulty to gain its volume via tests directly. This work was supported by National Natural Science Foundation of China
The required performances of asphalt mixture signifi- [grant number 51308429].
cantly depend on appropriate aggregate gradation.
(2) The gradation design of asphalt mixture has two References
types by aggregate morphology, continuous grada-
tion and gap gradation, or three types by air voids, AASHTO, 2007. Standard specification for superpave volumetric mix design.
American Association of State Highway and Transportation Officials.
dense gradation, open gradation and semi-open gra- Abo-Qudais, S. and Shatnawi, I., 2007. Prediction of bituminous mixture
dation for asphalt mixture. No matter methods, the fatigue life based on accumulated strain. Construction and Building
essential requirement for gradation selection should Materials, 21, 1370–1376.
be highly consistent with site performance of asphalt AI, 2001. Superpave mix design. Lexington: Asphalt Institute.
pavement. Ali, Y., et al., 2015. Sensitivity analysis of dynamic response and fatigue
behaviour of various asphalt concrete mixtures. Fatigue & Fracture of
(3) Different aggregate gradations form different types Engineering Materials & Structures, 38, 1181–1193.
of asphalt mixture. AC, ATB, SMA, AM, OGFC and Aschenbrener, T., et al., 2002. Bailey method for gradation selection in
ATPB, popularly used in asphalt pavement, are the typ- hot-mix mixture design. Transportation Research Board. Transportation
ical in practice. Asphalt mixture with dense gradation Research E-Circular, E-C044.
(AC) is the commonly applied for a dense structure, ASTM, 2003. Standard terminology relating to concrete and concrete
aggregates. West Conshohocken, PA: American Society for Testing and
preventing moisture infiltration and also maintaining Materials Standard.
adequate voids to allow asphalt expansion. Bari, J. and Witczak, M., 2006. Development of a new revised version of
(4) Coarse gradation and fine gradation define the related the Witczak E* prediciton model hot mix asphalt mixtures. Journal of
gradation curve below and above, respectively, on the Association of Asphalt Paving Technologists, 75, 381–423.
maximum density curve −0.45 power gradation curve. Behiry, A., 2016. Optimisation of hot mix asphalt performance based on
aggregate selection. International Journal of Pavement Engineering, 17,
AASHTO divides the coarse and fine gradation by 924–940.
defining the PCS, prime control sieve. Especially for Besson, F., 1935. Case against surface area and fineness modulus.
fine aggregate, the current definition is 0.075 mm (No. Engineering News Record, 114.
200) sieve, as the same as the mineral dust sieve defini- Blanks, R., et al. 1940. Properties of Concrete Mixes. American Concrete
tion. Importantly, there is no restricted zone in the cur- Institute Journal, 11.
Breakah, T., et al., 2011. The impact of fine aggregate characteristics on
rent specifications of asphalt mixture design since 2003. asphalt concrete pavement design life. International Journal of Pavement
(5) The aggregate gradation can influence almost all the Engineering, 12, 101–109.
performances for asphalt mixture. Therefore, the Bruno, L., Parla, G., and Celauro, C., 2012. Image analysis for detecting
source of pavement quality control (QC) should be aggregate gradation in asphalt mixture from planar images. Construction
tracked to aggregate, including aggregate selection, and Building Materials, 28, 21–30.
Caro, S., et al., 2011. Stochastic micromechanical model of the deterioration
mixing, transportation, paving and compaction. of asphalt mixtures subject to moisture diffusion processes. International
(6) Aggregate gradation is not only with regard to its influ- Journal for Numerical and Analytical Methods in Geomechanics, 35,
ence on asphalt mixture properties, but also on strict 1079–1097.
16  M. FANG ET AL.

Chen, J., et al., 2015. Micromechanical characteristics of aggregate particles Leonardi, G., 2016. Fractal dimension for the characterization of
in asphalt mixtures. Construction and Building Materials, 91, 80–85. asphalt mixture properties. American Journal of Applied Sciences, 13,
Chen, J.S. and Liao, M.C., 2002. Evaluation of internal resistance in hot- 877–882.
mix asphalt (HMA) concrete. Construction and Building Materials, 16, Li, L. and Ma, X., 2007. Influence research of gradation segregation on
313–319. performance of asphalt mixture. Journal of Tongji University (Natural
Chun, S., Roque, R., and Zou, J., 2012. Effect of gradation characteristics on Science), 35, 1622–1626.
performance of superpave mixtures in the field. Transportation Research Lin, X., 1998. Structural design method for flexible pavement. Beijing:
Record, 43–52. People Communication Press.
Fan, T., 2004. Influence of different gradation of aggregate on the Lin, X., 2003. The calculating formula of aggregate composition of rational
mechanical properties and performance of asphalt mixture. Journal of bituminous and concrete. East China Highway, 82–84.
Chongqing Jiaotong University, 23, 22–25. Lira, B., Jelagin, D., and Birgisson, B., 2013. Gradation-based framework
FHWA, 1962. Aggregate gradation for highways: aggregate gradation, for asphalt mixture. Materials and Structures, 46, 1401–1414.
simplification, standardization and uniform application; and a new Liu, P., et al., 2017. Modelling and evaluation of aggregate morphology
graphical chart for evaluating aggregate gradation. Washington, DC. on asphalt compression behavior. Construction and Building Materials,
Fuller, W.B., 1907. The laws of proportioning concrete. Journal of 133, 196–208.
Transportation Division, American Society of Civil Engineering, 59, 67–143. Mamlouk, M. and Zaniewski, J., 2011. Materials for civil and construction
Goode, J. and Lufsey, L., 1965. Voids, permeability, film thickness vs. engineers. Upper Saddle River: Prentice Hall.
asphalt hardening. Journal of Association of Asphalt Paving Technologists, Mandelbrot, B. and Wheeler, J., 1983. The fractal geometry of nature.
34, 430–463. Journal of the Royal Statistical Society, 147, 468.
Greene, J., Chun, S., and Choubane, B., 2014. Enhanced gradation Mansour, T. and Putman, B., 2013. Influence of aggregate gradation on the
guidelines to improve asphalt mixture performance. Transportation performance properties of porous asphalt mixtures. Journal of Materials
Research Record: Journal of the Transportation Research Board, 2456, in Civil Engineering, 25, 281–288.
3–10. Meakin, P., 1999. A historical introduction to computer models for
Greene, J., Kim, S., and Choubane, B., 2011. Accelerated pavement testing fractal aggregates. Journal of Sol-Gel Science and Technology, 15,
and gradation-based performance evaluation method. Transportation 97–117.
Research Record, 119–127. Memphis, T.U.O., 2016. Aggregate gradation and sampling [Online].
Guarin, A., et al., 2013. Disruption factor of asphalt mixtures. International Available from: http://www.ce.memphis.edu/3137/Powerpoint%20
Journal of Pavement Engineering, 14, 472–485. Handouts/2%20-%20Aggregate%20Gradation%20and%20Sampling.
Hao, P. and Zhang, D., 2000. Research on evaluation target for the cracking pdf [Accessed 20 July 2017].
resistance property of bituminous mixture in low temperature. Highway, Meng, Y., Li, Y., and Zhang, Y., 2005. Influence analysis of gradation
63–67. composition on high temperature stability of bituminous mixtures.
Hao, P., Zhang, J., and Zhang, D., 1998. Study on fatigue characteristic of Journal of Highway and Transportation Research, 10, 1–6.
asphalt mixture with different gradations. Petroleum Asphalt, 12, 20–24. Moon, K., et al., 2015. Analyzing aggregate size distribution of asphalt
Hicks, R., 1991. Moisture damage in asphalt concrete. NCHRP Synthesis mixtures using simple 2D digital image processing techniques. Arabian
of Highway Practice. Washington, DC: Transportation Research Board. Journal for Science and Engineering, 40, 1309–1326.
Hou, Y., et al., 2016. Fractal analysis on asphalt mixture using a two- Nukunya, B., 2001. Evaluation of aggregate type, gradation, and other
dimensional imaging technique. Advances in Materials Science and volumetric properties for design and acceptance of durable superpave
Engineering, 2016, 1–7. mixtures. Dissertation/Thesis, ProQuest Dissertations Publishing.
Huang, B., et al., 2008. Fractal method based on quantitative evaluation of Peng, Y. and Sun, L., 2006. Study on influence factors of homogeneity
asphalt pavement anti-slide performance. China Journal of Highway and of asphalt mixture. Tongji Daxue Xuebao / Journal of Tongji
Transport, 21, 12–16. University(Natural Science), 34, 59–63.
Huang, J. and Huang, P., 2006. Relativity study on aggregates fractal and Prowell, B., Zhang, J., and Brown, E., 2005. Aggregate properties and the
mechanics indexes of asphalt mixture. Journal of Tongji University performance of superpave-designed hot mix asphalt. Washington, DC:
(Natural Science), 34, 1632. National Cooperative Highway Research Program (NCHRP) Report 464.
Huang, Q., et al., 2009. Relationship between gradation type and segregation Roberts, F., et al., 1996. Hot mix asphalt materials, mixture design and
degree for asphalt mixture. Journal of Traffic and Transportation construction, Construction Equipment.
Engineering, 9, 1–6. Saleh, M., 2012. Effect of aggregate gradation and percentage of air voids on
Husain, N., et al., 2014. Effects of aggregate gradation on the physical the fatigue behaviour of hot mix asphalts. Road & Transport Research,
properties of semiflexible pavement. Advances in Materials Science and 21, 53–63.
Engineering, 2014, 1–8. Sangsefidi, E., Ziari, H., and Mansourkhaki, A., 2014. The effect of
Interactive, P. 2011. “Gradation Test” 1 July 2011. http://www. aggregate gradation on creep and moisture susceptibility performance
pavementinteractive.org, http://www.pavementinteractive.org/article/ of warm mix asphalt. International Journal of Pavement Engineering, 15,
gradation-test/ [Online]. [Accessed 20 January 2017]. 133–141.
Ji, Y. and Wu, H., 2008. Study on effect of gradation fluctuation and asphalt Sha, Q., 2005. Aggregate gradation design methods for SAC and other
content variety in construction to performance of asphalt mixture. coarse aggregate gap gradation. Highway, 143–150.
Journal of China and Foreign Highway, 28, 240–243. Shen, S. and Yu, H., 2011. Analysis of aggregate gradation and packing for
Kandhal, P.A.P. and Parker Jr, F., 1998. Aggregate tests related to asphalt easy estimation of hot-mix-asphalt voids in mineral aggregate. Journal
concrete performance in pavements. Washington, DC: Transportation of Materials in Civil Engineering, 23, 664–672.
Research Board. Singh, P. and Walia, G., 2014. Application of fractal analysis in pavement
Kandhal, P.S. and Cooley Jr, L.A., 2002. Investigation of the restricted zone materials. Journal of Basic and Applied Engineering Research, 1, 38–42.
in the superpave aggregate gradation specification (with discussion). Singh, A., Das, A., and Basu, S., 2011. A numerical study on the effect
Journal of the Association of Asphalt Paving Technologists, 71. of aggregate gradation on mechanical response of asphalt mix. KSCE
Kanitpong, K., Charoentham, N., and Likitlersuang, S., 2012. Investigation Journal of Civil Engineering, 16, 594–600.
on the effects of gradation and aggregate type to moisture damage Sousa, J., et al., 1998. Effect of aggregate gradation on fatigue life of
of warm mix asphalt modified with Sasobit. International Journal of asphalt concrete mixes. Transportation Research Record: Journal of the
Pavement Engineering, 13, 451–458. Transportation Research Board, 62–68.
Kim, Y., et al., 2009. Effects of aggregate structure on hot-mix asphalt Suresha, S., George, V., and Shankar, A., 2009. Effect of aggregate gradations
rutting performance in low traffic volume local pavements. Construction on properties of porous friction coarse mixes. Materials and Structures,
and Building Materials, 23, 2177–2182. 43, 789–801.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING  17

Talbot, A. and Richard, F., 1923. The strength of concrete and its relation to Ye, Q., Wu, S., and Li, N., 2009. Investigation of the dynamic and fatigue
the cement, aggregate, and water. Bulletin. properties of fiber-modified asphalt mixtures. International Journal of
Tang, J., Gong, F., and Yuan, H., 2009. Analysis of orthogonal test on Fatigue, 31, 1598–1602.
influence factors of fatigue performance of asphalt mixture. Construction You, Z., Adhikari, S., and Dai, Q., 2008. Three-dimensional discrete
Machinery and Construction Technology, 51–53. element models for asphalt mixtures. Journal of Engineering Mechanics,
Vardanega, P., 2014. State of the art: permeability of asphalt concrete. 134, 1053–1063.
Journal of Materials in Civil Engineering, 26, 54–64. Yu, H. and Shen, S., 2012. Impact of aggregate packing on dynamic
Vivar, E. and Haddock, J.E., 2007. Hot-mix asphalt permeability and modulus of hot mix asphalt mixtures using three-dimensional discrete
porosity. Journal of Association of Asphalt Paving Technologists, 76, element method. Construction and Building Materials, 26, 302–309.
953–979. Yu, H., Li, X., and Jiang, T., 2007. Analysis on expressway asphalt mixture
Walsh, H. 1933. Simplified concrete mix design. American Concrete gradation & asphalt content variability. China Municipal Engineering,
Institute Journal, 5. 12–16.
Wang, H., 2008. Study on the structure and material of asphalt pavement on Yuan, Y. 2004. Research on Rheological Characteristics of Asphalt Mortar
the zones with the feature of heavy loads and high temperature (Ph.D). and Gradation Optimization Based on Superpave (Ph.D).
Central South University. Yuan, Y., E, Y., and Xing, J., 2009. Bituminous pavement segregation and
Wang, H. and Zhang, Z., 2004. Design and comparative analysis of anti- control measures. Journal of Jilin Architectural and Civil Engineering, 26,
skid surface of HMA. Journal of Highway and Transportation Research 31–34.
and Development, 21, 15–17. Zhang, C., et al., 2016. Compaction characteristics of asphalt mixture with
Wang, H., et al., 2016. The effect of morphological characteristic of coarse different gradation type through superpave gyratory compaction and
aggregates measured with fractal dimension on asphalt mixture’s X-ray CT scanning. Construction and Building Materials, 129, 243–255.
high-temperature performance. Advances in Materials Science and Zhang, D., 2003. Engineering guide for asphalt pavement. Beijing: China
Engineering, 2016, 1–9. Communication Press.
Weymouth, C., 1933. Effects of particle interference in mortars and Zhang, X., Wang, S., and Wu, K., 2001. CAVF method for asphalt mix
concretes. Rock Products, 36, 26–30. gradation design. Highway, 17–21.
Wig, R., Williams, G., and Gates, E., 1916. Strength and other properties Zhang, X., et al., 2014. Research on skid resistance of asphalt pavement
of concrete as affected by materials and methods of preparation. based on three-dimensional laser-scanning technology and pressure-
Technical Papers of the Bureau of Standards. No. 58. Washington, DC: sensitive film. Construction and Building Materials, 69, 49–59.
US Government Printing Office. Zhao, Z., Zhang, Z., and Hu, C., 2005. Influence of gradation on anti-
Yang, X., 2007. Influence of grading curve trend on high temperature skidding performance of asphalt pavement. Journal of Chang’an
performance of asphalt mixture. Technology & Economy in Areas of University (Natural Science Edition), 25, 6–9.
Communications, 6–10. Zube, E., 1962. Compaction of asphalt concrete pavement as related to the
Yang, X., et al., 2016. Review on heterogeneous model reconstruction of water permeability test. Prceedings of 41st Annual Meeting Highway
stone-based composition in numerical simulation. Construction and Research Board, Transportation Research Board, 1962 Washington DC,
Building Materials, 117, 229–243. 12–37.