Kuwait J. Sci. Eng. 30(2) pp.

169-182, 2003

Using Marshall test to assess asphalt-concrete modulus

for mixes used in Kuwait

AHMAD H. AL-JASSAR1*, ABDULAZIZ M. AL-SHEHAB2,

AREF AL-ZAABI1
1
Civil Eng. Dept., Kuwait University,
P.O.Box 5969, Safat 13060, Kuwait, Fax: 4817524
2
Experts Department, Ministry of Justice, P.O.Box 23937, Safat 13100, Kuwait

ABSTRACT
The structural (thickness) design of asphalt pavement layers is a function of many
factors, one of the most important of which is the sti€ness of the asphalt mix. In Kuwait,
the structural design of pavements is performed using assumed values for sti€ness that
may not represent the actual sti€ness of the used material. Sti€ness modulus testing is
time and e€ort consuming and requires skilled labour and special equipment. Therefore,
it is not performed on a routine basis to check the sti€ness of the asphalt-concrete mixes
in Kuwait. This study was performed to establish a relationship between the sti€ness
(represented by the resilient modulus) and the Marshall test parameters obtained from
the Marshall mix design procedure which is the standard used in Kuwait to design
asphalt-concrete paving mixes. The Marshall parameters considered in this study are
stability/¯ow ration and the Marshall Index, which represents the slope of the straight
portion of the load-deformation curve obtained during Marshall testing.

Regression analysis demonstrated good correlation between sti€ness and stability/

¯ow ratio at 608C at various loading frequencies. The resulting model showed a linear
relationship between the two parameters that are independent of the mix type.

Keywords: Asphalt concrete; Kuwait; Marshall test; MATTA; sti€ness modulus.

INTRODUCTION

Background
Asphalt concrete durability is an important criterion to ensure that a pavement
layer maintains its desired properties. With time, asphalt pavement becomes
dense and failure may occur due to changes in the aggregate, permanent
deformation, cracking, or bleeding of asphalt, among other changes. One way to

* Corresponding author.
170 Ahmad H. Al-Jassar, Abdulaziz M. Al-Shehab, Aref Al-Zaabi

minimize such failure is to select the best mix properties. Asphalt mix design is
performed to select the best blend of aggregate and optimum binder content.
The standard method of Marshall mix design is used in Kuwait (MPW 1987a).
In Kuwait, asphalt mixtures are divided into three main types depending on
their speci®c use. A base course mise of a large aggregate size up to 37.5 mm is
placed directly on the subgrade. It is characterized by low asphalt content due to
large size aggregates. The second mix type is the binder course mix, which serves
as an intermediate layer between the surface and the underlying layers. It
consists of large aggregate sizes up to 25 mm with low bitumen content. The
third type of asphalt mixes is the surface course mix, which is a surface layer
designed to have sucient stability and durability to carry the expected trac
loads, sucient skid resistance, and resistance to changes in weather conditions.
This type of mix contains the highest bitumen content and smaller sizes of
aggregate that provide a smooth and dense surface. A fourth type of surface
mix, called open-graded friction course (OGFC) may be used to increase surface
skid resistance and to provide a free-draining layer for the surface water to seep
to the edges of the pavement. It also contains a relatively high bitumen content
mixed with small size aggregates (MPW 1987a).
Resilient modulus is an important design parameter in ¯exible pavement
design. The 1986 AASHTO guide for design of pavement structures
incorporated the resilient modulus into design procedures (Little et al. 1992).
Apart from design, the sti€ness modulus is also a key parameter in the
evaluation of pavements (Thom et al. 1977).
The indirect tensile test (ITT) has been identi®ed as an economical and
practical means of measuring sti€ness modulus and has been standardized
(Nunn 1996). To minimize human errors, a fully automated closed loop testing
and data acquisition system was recommended for ITT (Banksdale et al. 1997).
The Nottingham Asphalt Tester (NAT) and Material Testing Apparatus
(MATTA) are two recently developed experimental systems to measure sti€ness
modulus of asphalt concrete (Leech & Sexton 1996). MATTA testing system
was utilized in this research.
A review of the literature indicates that a number of researchers have studied the
relationship between the resilient modulus and other characteristics of asphalt
concrete. Relationships between the resilient modulus and other bearing
characteristics of the pavement layer were developed. These include LBR
(Limerock Bearing Ratio) and FWD (Falling Weight De¯ectometer)
backcalculated modulus (Ping et al. 2001). An equation was developed for the
elastic modulus of asphalt concrete and the results were compared with the Hashin
and Shtrikman theoretical bounds and Henkelom and Klomp equation (Li et al.
1999). Using two identical asphalt mixtures, a research programme for comparing
 Using Marshall test to assess asphalt-concrete modulus for mixes used in Kuwait 171

the Marshall test with the splitting tension test was conducted in Germany
(Adedimila 1986). The repeated-load indirect tensile test was utilized to develop
relationships between various engineering properties and optimum asphalt contents
(Gonzales et al. 1975). In a recent study in Saudi Arabia, temperature correction
factors and resilient modulus estimation equations from basic material physical
properties were developed using statistical procedures to a high degree of reliability
(Al-Abdul Wahab et al. 2001). The e€ect of temperature on sti€ness modulus and
load carrying capacity was studied in Kuwait in 1972 (Bissada, 1972).

Problem statement
The Ministry of Public Works (MPW) in Kuwait uses a locally-developed
manual for pavement structural (thickness) design (MPW 1987b). The manual
covers a range of values for the sti€ness modulus (E) of the local asphalt layers.
Currently, designers use the manual with assumed (typical) E-Values that may
not represent the actual sti€ness of the used material.
Use of these assumed values may result in thickness over-design, which means
an extra cost for constructing the pavement, or under-design that may cause
premature failure of the pavement layers. Therefore, actual sti€ness modulus
values should be used to avoid such problems. This is important since E-values,
according to the Kuwait Design Manual, signi®cantly a€ect the design thickness
of a given layer (MPW 1987b).
Therefore, E-values of the used asphalt mix should be measured by
conducting a standard E-test. Alternatively, they may be estimated from other
mix properties if a correlation exists between E and such properties of the mix.

Objective
A study was conducted to establish a relationship between the sti€ness modulus
(E) and several Marshall parameters (Al-Shehab 1999). The Marshall
parameters considered are stability/¯ow ratio (STF) and slope of the load-
deformation curve plotted during the loading process in the Marshal machine,
which is called the Marshall Index (MI), i.e.;

E ˆ f…STF† …1†

or

E ˆ f…MI† …2†
172 Ahmad H. Al-Jassar, Abdulaziz M. Al-Shehab, Aref Al-Zaabi

From such a relationship, the predicted modulus is used in the structural

design of a pavement to determine its required layer thickness. The advantage of
this is that the mix design results are used to estimate the sti€ness of the mix
without conducting an E-test.

ASPHALT PAVEMENTS IN KUWAIT

Asphalt Concrete (AC) is the most common type of paving material used in
Kuwait. A typical pavement section includes a prepared subgrade, base course,
binder course, and wearing course. The four AC mix types used in Kuwait are
Types I, II, III, and IV. The mix names are local, and their gradations are shown
in Table 1. The AC mixes used for base and binder courses are Type I and Type
II, respectively. type III is used for the wearing course. Type IV is of limited use
as a wearing course (MPW 1987a).

Table 2. Marshall design criteria for AC mixes used in Kuwait (MPW 1987a)

Type I Type II Type III Type IV

(Base Course) (Binder Course) (Wearing Course) (Wearing Course)

Marshall Criteria Min. Max. Min. Max. Min. Max. Min. Max.
No. of blows 75 - 75 - 75 - 75 -
Stability, kg 1660 - 1700 - 1800 - 1800 -
Flow, 0.01 in. 8 16 8 16 8 16 8 16
V.M.A. 13 - 14 - 15 - 15 -
Air voids, % 3 8 3 8 3 6 3 6
Filler/bitumen ration - 1.6 - 1.5 - 1.4 - 1.4

The mix design of the AC used in Kuwait is based on the standard Marshall
method of mix design (Asphalt Institute 1991) using an approved aggregate
blend and asphalt cement. The Marshall design criteria used in Kuwait are
shown in Table 2 (MPW 1987a). The structural (thickness) design of asphalt
pavements in Kuwait is performed using the "Kuwait Design Manual for
Asphalt Pavements in Arid and Hot Climates" (MPW 1987b). The Design
Manual requires the following input:
- Subgrade soil load-bearing capacity (CBR)
- Design axle loads (in ESALs)
- Material sti€ness (E)
Each of the above input parameters is important in the structural design of
 Using Marshall test to assess asphalt-concrete modulus for mixes used in Kuwait 173

AC pavements as it a€ects the selected layer thicknesses. Because of the time,

e€ort, and skilled-labour requirements, the sti€ness of the AC mixes is not
determined experimentally, but rather assumed values are used from experience.
Such assumed values may no longer represent the mixes being used because of
many factors including change of mix proportioning, change of aggregate
source, and change of asphalt cement type.
This study was performed to develop a model relating the sti€ness modulus to
some parameters obtained by conducting the Marshall test on samples of AC
mixes to be used for pavements.

METHODOLOGY
The study started with preparing representative AC mixes satisfying the
following requirements:
- Aggregates are cement coated, typical of what is used locally.
- Aggregate blends satisfy the local requirements as shown in Table 1.
- 60/70 penetration grade asphalt cement was used, as per local requirements
(MPW 1987a).
The aggregate size gradations of the tested mixes are shown in Table 1. The
mixes were prepared according to the standard method of Marshall mix design
ASTM D-1559 (Asphalt Institute 1991), and the optimum asphalt cement
content (OAC) was determined for the four considered mix type as shown in
Table 2.
At each optimum value of asphalt cement content, three new identical
specimens were prepared for each mix type. These specimens were tested for
resilient modulus as per ASTM D-4123 (ASTM 1982) using the MATTA testing
system at three temperatures and three loading frequencies. The considered
temperatures were chosen to cover the range from room temperature (238C),
408C, and 608C.
The resilient moduli of the specimens were determined using the following
loading frequencies of 1.0 Hz, 0.5 Hz, and 0.33 Hz. Therefore, 108 values of
resilient modulus were obtained in the testing program.
The specimens were then tested in the Marshall testing machine. The same
specimens were used so that related values of Marshall parameters and resilient
modulus would correspond to identical material. The following results were
obtained for each of the tested specimens:
- Marshall stability
 174
Table 1. Aggregate gradation and OAC for the tested AC mixes

Cumulative % Passing Speci®cation Limits

Sieve Size Type I Type II Type III Type IV Type I Type II Type III Type IV
37.5 mm (1.5") 100 - - - 100 - - -

25.0 mm (1") 86 100 - - 72 - 100 100 - -

19.0 mm (3/4") 75 91 100 - 60 - 89 82 - 100 100 -

12.5 mm (1/2") 61 72 81 100 46 - 76 60 - 84 66 - 95 100

6.25 mm (3/8") 54 62 71 90 40 - 67 49 - 74 54 - 88 80 - 100

4.75 mm (No. 4) 42 45 54 65 30 - 54 32 - 58 37 - 70 55 - 75

2.36 mm (No. 8) 33 34 39 43 22 - 43 23 - 45 26 - 52 35 - 50

1.18 mm (No. 16) 26 25 29 - 15 - 36 16 - 34 18 - 40 -

Ahmad H. Al-Jassar, Abdulaziz M. Al-Shehab, Aref Al-Zaabi

600 mm (No. 30) 19 19 22 24 10 - 28 12 - 25 13 - 30 18 - 29

300 mm (No. 50) 14 14 16 19 6 - 22 8 - 20 8 - 23 12 - 25

150 mm (No. 100) 9 9 11 13 4 - 14 5 - 13 6 - 16 8 - 18

75 mm (No. 200) 5 6 7 7 2-8 4-7 4 - 10 4 - 10

Asphalt cement content 5.0 5.0 5.5 6.5 3.5 - 5.0 4.0 - 6.5 4.5 - 7.5 4.5 - 7.5
(% by total weight of aggregate)
 Using Marshall test to assess asphalt-concrete modulus for mixes used in Kuwait 175

- Marshall ¯ow
- A graph relating the applied load to deformation - representing the Marshall
loading conditions - using a plotter connected to the Marshall testing
machine. A typical graph is shown in Figure 1.

Fig.1. Marshall load-deformation curve for sample A1

of AC mix type I at 5.0% asphalt cement content
176 Ahmad H. Al-Jassar, Abdulaziz M. Al-Shehab, Aref Al-Zaabi

The ratio of Marshall stability to ¯ow (STF) was calulated for each specimen.
From the load-deformation graph, the slope of the straight portion (assumed to
cover the range of actual trac loading) was calculated. This slope was called
the "Marshall Index" (MI).

Finally, results from resilient and Marshall tests were analyzed to establish a
model correlating E to Marshall parameters.

RESULTS AND ANALYSIS

The results of resilient modulus testing are shown in Table 3. It can be seen that,
in general, the resilient modulus decreases with decreases in loading frequency.
Also, variation in testing temperatures show that the material sti€ness decreases
signi®cantly with increases in temperature, as expected.

Table 3. Resilient modulus test results (resilient modulus average values, MPa)

Temperature Loading Frequency (Hz)

(8C) Mix Type 1.0 0.5 0.33
I 795.7 829.2 836.1

23 II 3814.6 3732.1 2839.0

III 5141.4 4790.5 5059.0

IV 3330.1 3247.7 3075.3

I 621.7 584.7 549.4

40 II 582.2 540.3 523.5

III 720.1 678.9 664.1

IV 634.8 565.0 534.5

I 181.6 166.1 163.1

60 II 243.8 231.8 221.2

III 229.8 212.5 208.0

IV 216.1 205.1 203.7

Testing the specimens in the Marshall machine resulted in stability and ¯ow
values, and load-deformation curves similar to the one shown in Figure 1. The
stability/¯ow ratio (STF) was calculated for each specimen, and from the load-
deformation curve, the Marshall Index (MI) was calculated.
 Using Marshall test to assess asphalt-concrete modulus for mixes used in Kuwait 177

Correlation analysis was performed to evaluate the relationship between the

resilient modulus at di€erent temperatures and loading frequencies, and the
obtained STF and MI values. The best correlations were found to be between
the resilient modulus at 608C and STF. This may be due to the fact that STF is
obtained from the Marshall test at the same temperature. It may also be due to
the sensitivity of AC specimens to deformations at low temperatures (high E).
Such correlations are shown in Figures 2 to 5, which resulted in the following
relationships:
E1:0 Hz ˆ 1:155 STF R2 ˆ 0:76 …3†

E0:5 Hz ˆ 1:075 STF R2 ˆ 0:81 …4†

E0:33 Hz ˆ 1:044 STF R2 ˆ 0:82 …5†

where,
Ei Hz = Resilient modulus at 608C and i loading frequency, MPa;
STF = stability/¯ow ratio, kg/0.01 in.;
2
R = correlation factor for the developed relationship.

Fig.2. E versus STF at 608C and 1.0 Hz

178 Ahmad H. Al-Jassar, Abdulaziz M. Al-Shehab, Aref Al-Zaabi

Fig.3. E versus STF at 608C and 0.5 Hz

Fig.4. E versus STF @ 608C and 0.33 Hz

 Using Marshall test to assess asphalt-concrete modulus for mixes used in Kuwait 179

Fig.5. E versus STF @ 608C and three loading conditions

CONCLUSIONS
The resilient modulus of asphalt concrete (AC) mixes is an important input in
the structural desing process of AC pavements. In Kuwait, assumed (typical)
values for this modulus are used to determine the required pavement layer
thicknesses. This study shows that a relationship exists between certain
parameters from the Marshall test and the resilient modulus. Three correlations
were obtained relating the resilient modulus to the ratio of Marshall stability-to-
¯ow at three di€erent loading frequencies: 1.0, 0.5, and 0.33 Hz. During the
course of testing and analysis, some conclusions were obtained, including:
- The sti€ness of the AC mixes decreases with increases in temperature.
- The measured sti€ness of an AC mix decreases with decreases in loading
frequency, as shown in Figure 5. Therefore, slow trac is expected to cause
more deformation in pavements than faster trac.
- The best relationships for predicting the AC resilient modulus are obtained
by correlating it to the stability-to-¯ow ratio.
- The resilient modulus is linearly proportional to stability/¯ow ratio.
- The developed relationships shown in Figures 2 to 5 passed through the
origin. This indicates that when stability equals zero, the material sti€ness
equals zero. This logical relationship strengthens the validity of the developed
models.
180 Ahmad H. Al-Jassar, Abdulaziz M. Al-Shehab, Aref Al-Zaabi

- The developed models of the form E = f (STF) may be used to predict the
resilient modulus of AC mixes using Marshall test results (stability and ¯ow).
- The developed relationships may only be used within the range of moduli
shown in Figures 2 to 5. Prediction of E-values outside this range needs
further investigation. However, because the relationships pass through the
origin of Figures 2 to 4, there is more con®dence in using them in the lower.

ACKNOWLEDGEMENTS
The authors would like to thank Dr. Mohammed Othman (formerly in the
Maintenance Division, Ministry of Public Works, Kuwait) for providing test
materials, support, and constructive comments. The authors would also like to
acknowledge the help provided by the Transportation Lab (Civil Engineering
Department, Kuwait University) in conducting the experimental work for this
study, with special appreciation to Dr. M. Ahmed Ali and Mr. Quddusur
Rehman. Appreciation is also due to Mini Kora and Daisy Mathew for their
e€ort in typing this manuscript.

REFERENCES
Adedimila A.S., 1986. A comparison of the Marshall and indirect tensile tests in relation to asphalt
mixture design. Proceedings - Institution of Civil Engineers Pt. II. Research and Theory,
London.
Al-Abdul Wahab, H.I., Asi, I.M., Rawadhan R.H., 2001. Modeling resilient modulus and temperature
correction for Saudi roads. Journal of Materials in Civil Engineering, 13(4): 298-306.
Al-Shehab, Abdulaziz M. 1999. Predicting Sti€ness Modulus of Asphalt Mixes Using Marshall
Parameters. M.Sc. Project, Department of Civil Engineering, Kuwait University, Kuwait.
American Society for Testing and Materials, ASTM 1982. Indirect Tension Test for Resilient
Modulus of Bituminous Mixes. ASTM D4123, USA.
Asphalt Institute 1991. Mix Design Methods for Asphalt Concrete and Other Hot-Mix Types.
Manual Series No. 2 (MS-2), USA.
Banksdale, R.D., Alba, J., Khosla N.P., Kim R., Lambe P.C., Rahman, M.S., 1997. Laboratory
determination of resilient modulus for ¯exible pavement design. NCHRP Project 1-28 Final
Report, NCHRP, Washington DC.
Bissada, A.F., 1972. Asphalt Pavement Temperatures Related to Kuwait Climate. Highway
Research Record, TRB, Washington, D.C.
Gonzales, G., Kennedy T.W. Anagnos, J.N., 1975. Evaluation of the resilient elastic characteristics
of asphalt mixtures using the indirect tensile test. Research Report 183, Texas University,
Austin, USA.
Leech, D. & Sexton, B., 1996. Standardization trail of performance tests for bituminous materials.
TRL Report 158, 0968-4107, Transport Research Laboratory, UK.
Li G., Li Y., Metcalf J.B., Pang S.S., 1999. Elastic modulus of prediction of asphalt concrete.
Journal of Materials in Civil Engineering. 11(3): 236-241.
Little D.N., Crockford, W.W., Gaddam, V.K.R. 1992. Resilient Modulus of Asphalt Concrete,
Final Report. FHWA, Washington DC.
 Using Marshall test to assess asphalt-concrete modulus for mixes used in Kuwait 181

Ministry of Public Works, MPW, 1987a. General Speci®cations for Kuwait Motorway/Expressway
System. Kuwait.
Ministry of Public Works, MPW, 1987b. Kuwait Design Manual for Asphalt Pavements in Arid
and Hot Climates, Kuwait.
Nunn, M.E.L., 1996. The characterization of bituminous macadams by indirect tensile sti€ness
modulus. TRL Report 160; 0968-4107, Transport Research Laboratory, UK.
Ping W.U., Liu C., Yang, Z., 2001. Implementation of resilient modulus in the Florida ¯exible
pavement design procedure. Final Report FHWA, Washington D.C.
Thom N.H., Elliot R.C., Cheung, L.W., 1997. Comparisons between laboratory and in-situ
determined asphalt concrete moduli. Proceedings of Eighth International Conference on
Asphalt Pavements, Vol. II. University of Washington, Seattle, USA.

Submitted : 11/6/2001
Revised : 10/8/2002
Accepted : 2/2/2003
182 Ahmad H. Al-Jassar, Abdulaziz M. Al-Shehab, Aref Al-Zaabi

?!9A}yG {|9g| ~++tAy d9VQ9| Q9=ALG eGOMASG

B*(wyG ,p ?|OMAT}yG ?+AzqS G J9azMzy

1
,<9fRyG aQ9f h 2H9%WyG R*RgyGO=f h 1Q9TGyG r+azyGO=f O(}I O}IC
B*(wyG ?g|9F - dhQA=yGh ?SO"%yG ?+zv - ?+!O}yG ?SO"%yG ~Ts1
13060 B*(wyG - I9qZyG 5969 .H.U
B*(wyG ,dOgyG IQGRh ,AGQ=MyG IQGOE2

? Y ;L

9%}$C #| {|G(f IOf )zf ?+AzqS G bQayG rYQ J9t=ay ,F9W! G ~+}ZAyG O}Ag*
rvYQvyG J9vtv=vavy ,vF9vWv! G ~v+v}vZvAvyG {v}vf ~vAv*h .?v+vAvzvqvS G J9vtv=vavyG J9vavzvL ?v!9vAv|
~+tyG {D}@ : Os gh ,?!9A}yG {|9g}y ?+\GQApE ~+s eGOMAS9< IO9f B*(wyG ,p ?+AzqS G
k
9Ash ?!9A}yG {|9g| Q9=ALG AGQFE >zaA*h .?|OMAT}yG BzqS G J9t=_ ?!9A}y ?+t+tJyG
k
{}f IO9f ~A* : xyh ,?Y9L ?*Q=AM| JGOg| )yE ?p9\ 9< ,?<QO| ?y9}fh GO%Fh
.B*(wyG ,p ?+AzqS G J9azMyG ?!9A| #| utJAzy ,"+@hQ AGQF{v Q9=AL:G xyP
#| ?+AzqS G J9azMzy ?!9A}yG {|9g| #+< ?s;f O9G* ?SGQOyG gP$ B*QFC Osh
k
GP$ ~A*h .J9azMyG xz@ ~+}ZAy ;YC eOMAT}yG d9VQ9| Q9=ALG HF9A"< 9%a<Q d;L
QY9"gyGh .d9VQ9| Q9=ALG QY9"fh ?!9A}yG {|9g| #+< ?s;f O9G*E d;L #| Q*OtAyG
)QV||( {+yO ($ QL6G QZ"gyGh ,upOAyG/J9=DyG ?=T! ,$ ?SGQOyG gP$ ,p ?|OMAT}yG
H@9"yG yW9j]!:G - {}JyGz )"J"| #| ~+tAT}yG ARGyG {+| #f IQ9=f ($h ,d9VQ9|
.,S9+tyG d9VQ9| Q9=ALG #|
/J9=DyG ?=T!h ?!9A}yG {|9g| ~+s #+< ?s;f O(Fh ,gFGQAyG {+zJAyG #| #+=@ Osh
#| hO=*h .?qzAM| {+}J@ JGOOQ@ O"f xyPh ?*(8| 60 IQGQI ?FQO O"f upOAyG
upOAyG/J9=DyG ?=T!h ?!9A}yG {|9g| #+< ?+aL ?s;f O(Fh {+zJAyG #| H@9"yG LP(}"yG
.IQ=AM}yG ?+AzqS G ?azMyG ?+f(! #f Qd"yG ^j<