You are on page 1of 12

Effects of Supersingle Tire Loadings on Pavements

Daehyeon Kim, M.ASCE1; Rodrigo Salgado, M.ASCE2; and Adolph G. Altschaeffl3

Abstract: Due to their efficiency and economy, supersingle tires have gradually been replacing conventional dual tires in the trucking
industry. According to recent studies, supersingle tires generate much higher vertical contact stresses than do conventional dual tires,
resulting in larger deformations and more severe damage to the subgrade. In order to better assess the larger stresses generated by
supersingle tires and their effect on the subgrade, analyses are done taking into account soil plasticity. In this paper, the effects of
Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

supersingle tires on subgrades for typical road cross sections are investigated using plane-strain two-dimensional and three-dimensional
static and dynamic finite element analyses. The analyses focus on the sand and clay subgrades rather than on asphalt and base layers. The
subgrades are modeled as saturated in order to investigate the effects of porewater pressures under the most severe conditions. The
analyses suggest that current flexible pavement design methods are unconservative for the increased loads imposed by supersingle tires on
the pavement system.
DOI: 10.1061/共ASCE兲0733-947X共2005兲131:10共732兲
CE Database subject headings: Tires; Subgrades; Finite element method; Plasticity; Pavements; Trucks.

Background the asphalt and base layers rather than on the subgrade layer, with
the implied assumption that the asphalt layer and base layer
would take most of the tire loadings, and the subgrade would be
Research Motivation
in an elastic state. These studies also assume that the contact
The pursuit of increased efficiency and economy by the trucking stresses induced by tire loadings are the same as the inflation
industry has led to increasing use of wide-base 共“supersingle”兲 pressures. The results of these studies are probably acceptable for
tires and heavier truck loadings. Studies by De Beer et al. 共1997兲, lower inflation pressures and conventional dual tires, but they
Myers et al. 共1999兲, and Siddharthan and Sebaaly 共1999兲 have would be unconservative for increased inflation pressures and
indicated that these tires create higher vertical and transverse con- contact stresses 共De Beer et al. 1997; Siddharthan and Sebaaly
tact stresses across larger contact areas with the pavement. These 1999兲. In order to ascertain the severity of the potential effects of
studies indicated that vertical contact stresses as large as 2.5 times the use of wide-base tires on the subgrades, more advanced analy-
the tire inflation pressures are created when these tires are used, ses taking into account increased contact stresses and wider con-
especially at high loads and inflation pressures. The increased tact areas are required. These analyses and the lessons they offer
trend toward larger truck loads, higher tire inflation pressures, and are the main focus of the present paper.
the use of wide-base tires has caused concern about pavement
integrity. As the area of contact between the tires and the pave-
Contact Area
ment and the contact stresses increase, the subgrade may also be
affected. Although the contact area of a truck tire with the pavement is
Excessive deformation of the subgrade has always been rec- assumed to be circular in multilayered elastic theory, the contact
ognized as a major problem. The larger loads, higher tire contact area of a truck tire is in reality closer to a rectangular than a
pressures, and larger tire contact areas send significant stress circular shape. Weissman 共1999兲 and Tielking 共1994兲 also con-
pulses deeper into the pavement system. The larger stresses may firmed that the contact areas of truck tires are rectangular regard-
lead to larger permanent deformations. However, in modeling a less of the type of tire. As identified in the study by Tielking
road structure, most of the studies found in the literature focus on 共1994兲 and evidenced by data collected from the tire industry
共Michelin 1997兲, the contact area of a supersingle tire is larger in
1
Geotechnical Research Engineer, Research Division, INDOT, the transverse than in the longitudinal direction, contrary to what
West Lafayette, IN 47906 共corresponding author兲. E-mail: is typically observed for a conventional dual tire. The contact area
dkim@indot.state.in.us of a truck tire is primarily dependent upon axle load, inflation
2
Professor, School of Civil Engineering, Purdue Univ., West pressure, and tire type. Most researchers have either assumed that
Lafayette, IN 47907-1284. E-mail: rodrigo@ecn.purdue.edu the contact area is circular or used the equivalent contact area
3
Professor Emeritus, School of Civil Engineering, Purdue Univ., West shown in Fig. 1共a兲 when doing analysis of pavement loading. The
Lafayette, IN 47907-1284. E-mail: altsch@ecn.purdue.edu equivalent contact area is commonly used in finite element analy-
Note. Discussion open until March 1, 2006. Separate discussions must
sis for dual conventional tires. The equivalent contact area ap-
be submitted for individual papers. To extend the closing date by one
month, a written request must be filed with the ASCE Managing Editor.
proximates the shape of the actual contact area of dual tires 关Fig.
The manuscript for this paper was submitted for review and possible 1共b兲兴 共Huang 1993兲 using the following equation:
publication on February 2, 2004; approved on January 21, 2005. This
paper is part of the Journal of Transportation Engineering, Vol. 131, Ac = ␲共0.3L兲2 + 共0.4L兲共0.6L兲 = 0.5227L2 共1兲
No. 10, October 1, 2005. ©ASCE, ISSN 0733-947X/2005/10-732–743/
$25.00. where Ac = equivalent contact area.

732 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005

J. Transp. Eng. 2005.131:732-743.


contact stress = K1 + K2 ⫻ 共inflation pressure兲 + K3 ⫻ 共load兲
共2兲

where K1, K2, and K3 = regression coefficients that are always


positive numbers. They take different values for the normal stress,
longitudinal, and transverse shear stresses.
In our FE analysis, the maximum contact stresses of the three
components for the Michelin 425/65R22.5 wide base tire were
obtained for an inflation pressure of 862 kPa 共125 psi兲 and for a
Fig. 1. Contact area for dual tires 共after Huang 1993兲 tire load of 51 kN 共11,400 lb兲, which are the maximum inflation
pressure and tire load recommended by the tire company for this
tire. Based on the above regression equation, an increase in either
the load or the inflation pressure increases the maximum vertical
and transverse contact stresses. It is intuitive that vertical contact
Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

In the finite element 共FE兲 analysis discussed later in the stresses are more sensitive to changes either of the inflation pres-
present paper, Eq. 共1兲 was used to calculate the value of the con- sure or of the tire load than are the other two components. More-
tact area for conventional dual tires. An equivalent rectangular over, the ratio of the maximum vertical contact stress to the lateral
contact area was used, shown in Fig. 1共b兲. As the 425/65R22.5 and longitudinal stress is 10:1.6:0.8 approximately 共Siddharthan
supersingle tire is most typically used 共Akram et al. 1993; et al. 1998兲. This means that the vertical contact stress is the most
Bonaquist 1992; Tielking 1994兲, the ratio of width to length ap- significant component. Siddharthan et al. 共1998兲 reported that
propriate for the typical supersingle tire 关1:0.85 for 425/65R22.5 the impact of contact shear stresses on tensile strains at the
共Michelin 共1997兲兲兴 was used for supersingle tires in order to bottom of the asphalt layer were insignificant. Accordingly, in our
model them more realistically. FE analysis, we neglect transverse and longitudinal shear stress
components.
Contact Stresses
Empirical Studies
When a tire loading is applied to the pavement surface, three
contact stress components are generated: vertical, transverse, and Rutting and fatigue cracking are two major concerns associated
longitudinal 共Yap 1988兲. Multilayered elastic theory assumes a with pavement structures. The design of a pavement structure
concentrates on limiting strains in order to achieve satisfactory
uniform contact stress, equal to the inflation pressure. Most analy-
pavement performance within a certain design period.
ses using the finite element method 共FEM兲 are based on the same
Fatigue cracking and rutting predictive models have been used
assumption. Recent studies 共Marshek et al. 1986; Yap 1988; Ford to calculate the number of load repetitions leading to fatigue fail-
and Yap 1990; Tielking 1994; De Beer et al. 1997; Myers et al. ure and rutting. The prediction of the failure of a pavement sys-
1999兲 have shown that the vertical contact stresses are not uni- tem has been empirically developed by the correlation between
form and have a peak value around 2.5 times the inflation pres- the multilayered elastic theory and field tests such as the AASHO
sure. De Beer et al. 共1997兲 indicated that, for South African roads, road test 共AASHO 1962兲. For pavement rutting models, most
pavement analyses using uniform contact stresses equal to the studies employ the correlations between strains and load repeti-
inflation pressure might overestimate the pavement’s life. tions. The correlations between the vertical strain on top of the
According to Ford and Yap 共1990兲, for two Goodyear supers- subgrade and the number of equivalent single axle load 共ESAL兲
ingle tires 共385/65R22.5 and 425/65R22.5兲 at constant load, the repetitions are widely used. These types of models assume that
tire inflation pressure variation primarily affected the contact rutting can be minimized by limiting the amount of vertical com-
stresses in the central region of the contact area. This was also pressive strains on top of the subgrade 共Huang 1993兲.
reported by Yap 共1988兲 in a similar study; the higher the inflation
pressure, the greater the contact pressures in this central region. Field Testing
The contact pressures in the outer portions of the tires were es-
sentially not affected. Field testing 共Huhtala et al. 1989; Akram et al. 1992; Bonaquist
A comparison of radial tires with wide base tires by Myers et 1992兲 involving trucks equipped with supersingle tires has rarely
al. 共1999兲 showed that the vertical and transverse contact stresses been performed since it is very expensive. Using an accelerated
loading facility to simulate traffic loading, Bonaquist 共1992兲 con-
are higher for wide base tires because wide base tires have a
cluded that, for the same load and tire pressure, supersingle tires
higher load per tire ratio than any other type of tires. The distri-
induced higher vertical compressive strains in all layers, higher
bution of the vertical contact stresses was also not uniform. The
tensile strains at the bottom of the asphalt concrete layer, and
maximum value was found to occur at the center of the contact around two times more rutting than dual tires. The fatigue life of
area with a value approximately equal to 2.3 times the inflation the pavement was reduced to one fourth of that observed when
pressure. dual tires are used. Akram et al. 共1992兲 found that wide base tires
De Beer et al. 共1997兲 did the most extensive research on tire were approximately 2.8 times more damaging to the thin pave-
contact stresses. The stresses were measured simultaneously with ment and 2.5 times more damaging to the thick pavement based
the vehicle-road surface pressure transducer array. The influence on a design equation using vertical compressive strain. Huhtala et
of the increase of tire loads and inflation pressures on the vertical al. 共1989兲 indicated that supersingle tires are more damaging than
contact stresses was examined. The following general equation dual tires by a factor of 2.3–4 in the same pavement conditions
was proposed to estimate the three components 共normal stress, for dual tires. It was also observed that when tire inflation pres-
longitudinal and transverse shear stresses兲 of the contact stresses: sure was increased by 20%, the damage in terms of the load

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005 / 733

J. Transp. Eng. 2005.131:732-743.


equivalency factor 共the number of equivalent standard axle load F2 = 共I1 − pa兲2 + R2J2 − R2关k + pa tan共tan−1 ␣兲兴2 = 0 共6兲
applications causing the same amount of damage by one passage
of an axle兲 increased by 10–40%.
where R = aspect ratio of the ellipse 共ratio of its horizontal to
vertical axis兲, which was set as 0.3. R = 0.3 was used as vertical
Analysis deformations were not sensitive to different R values and less
computation time was required for this value of R. The analyses
In this research, a commercial finite element program 共ABAQUS兲 showed that larger R values led only to slightly larger permanent
was used to analyze flexible pavement crosssections subjected to deformations due to the decreasing cap size, but that the compu-
wide base tire loadings. Eight-noded and 20-node quadratic solid tation time increased substantially. The parameter pa describes the
elements were used for the plane-strain two-dimensional 共2D兲 and evolution of volumetric plastic strain hardening; it is given by
3D analysis, respectively.

pb − Rk
Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

Typical Pavement Cross Sections pa = 共7兲


关1 + pa tan共tan−1 ␣兲兴
In our analyses, we consider two typical flexible pavement cross
sections found across the United States. The cross sections ana- in which pb = preconsolidation pressure. In ABAQUS 共1997兲, the
lyzed are composed of two traffic lanes, and are 13.3 m wide. hardening rule is a user defined, piecewise linear function relating
They are composed of a 15 cm hot mix asphalt layer, a 17 cm the hydrostatic compression yield stress pb and the corresponding
base layer, a 61 cm compacted clay or sand subgrade layer, and a volumetric strain. In our analysis, the relationship between pre-
607 cm clay or sand subgrade layer, top to bottom. consolidation pressure and plastic volumetric strain 共␧␯p兲 was ob-
tained from the relationship between specific volume ␯ and effec-
tive stress assuming the one-dimensional compression index Cc
Soil Modeling
= 0.3 and the one-dimensional swelling index Cs = 0.045. For ex-
Subgrade soils are usually modeled as either a sand or a clay. The ample, the plastic volumetric strains for the preconsolidation pres-
subgrade would ideally be composed of either denser sands or sures of 20 kPa 共corresponding to OCR= 1兲, 60 kPa 共OCR= 3兲,
stiffer clays, as the subgrade must have enough bearing capacity 100 kPa 共OCR= 5兲, 140 kPa 共OCR= 7兲, and 180 kPa 共OCR= 9兲
to support the asphalt and base layers and the traffic loadings. In were 0, 0.03975, 0.04948, 0.05434, and 0.0574, respectively. All
our analysis, we model subgrade soils using the Drucker–Prager loading with a component normal to the cap will result in plastic
model. For clays, but not for sands, a cap is used. compression, and the cap will expand according to the hardening
The main purpose of the use of the capped Drucker–Prager rule. The number of parameters in the Drucker–Prager model with
model for clay subgrades was to consider the overconsolidation cap is seven, including two elastic parameters 共i.e., E and ␯ or K
ratio 共OCR兲 of the clay subgrade, as behavior of clay is dependent and G兲, ␣ and k for the shear failure surface, and pb, ␧␯p, and R for
on stress history, as expressed by the preconsolidation pressure or the cap.
OCR. The cap plays a role related to the role of preconsolidation
pressure in clays, marking the stress states at which plastic vol-
ume change would develop. The use of the same type of consti- Implication of Nonuniform Contact Stresses Observed
tutive model for sand and clay subgrades allows more consistent for Supersingle Tires
comparison of the behaviors of these subgrades. In sands, plastic
In the FEM, equivalent nodal forces corresponding to the distrib-
volume change does not develop for the range of stresses present
uted loads applied to an element can be formulated as follows:
in the pavement subgrades, and thus no cap is used.
In terms of the stress invariants I1 and J2, the Drucker–Prager

冕 冕
model without cap can be written as 共Desai and Siriwardane
1984; Chen and Saleeb 1994兲 f= NT⌽dV + NTTdS + NT P 共8兲
V S

f共I1,J2兲 = 冑J2 − ␣I1 − k = 0 共3兲


where I1 and J2 = first invariant of the stress tensor and the second where N = shape function; ⌽ = body force; T = surface traction; and
invariant of the deviator stress tensor, respectively, and ␣ and k P = concentrated load. The equivalent nodal force formulation is
= model parameters that can be related to the Mohr–Coulomb based on the requirement that the work done by the nodal forces
strength parameters c and ␾ for triaxial conditions by the follow- f for given nodal displacements d be equal to the work done by
ing equations: the distributed loads ⌽, T, and the concentrated load P for the
displacement field related to d through the shape functions N
共Cook et al. 1989兲.
2 sin ␾ Figs. 2共a and b兲 illustrate the typical configuration of non-
␣= 共4兲
冑3共3 − sin ␾兲 uniform vertical contact stresses for a supersingle tire and the
corresponding equivalent nodal forces in the ␰-␩ coordinate
system 关see Fig. 2共c兲兴, respectively, where ␰ and ␩ are the trans-
6c cos ␾ formed coordinates that are determined by the mapping from x
k= 共5兲 and y 共−1 艋 ␰ 艋 1 , −1 艋 ␩ 艋 1兲. Only if traction forces or vertical
冑3共3 − sin ␾兲
contact stresses as shown in Fig. 2共a兲 are applied to a three di-
The cap used with the Drucker–Prager model is elliptical, with mensional quadratic 20-node element in a nonuniform configura-
equation tion, can Eq. 共8兲 be written as follows:

734 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005

J. Transp. Eng. 2005.131:732-743.


Tire contact stresses are dependent on the tire type 共Tielking
1994; Weissman 1999兲. In order to account for the effects of
various possible contact stress distributions, several linear elastic
analyses assuming typical parameters were done to assess the
contact stress distributions that would be most adverse to the
performance of the pavement system. In these analyses, as men-
tioned before, the tire load, inflation pressure, maximum vertical
contact stress obtained from Eq. 共2兲, and the ratio of the contact
area are 51 kN 共11,400 lb兲, 862 kPa 共125 psi兲, 1,607 kPa
共233 psi兲, and 1:0.85, respectively. Note that the total tire load
and the maximum vertical contact stress remain the same in all
the cases; only contact areas and stress distributions used in the
FEM are changed in each case. The six cases for supersingle tires
are as follows:
Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

1. Uniform 共square兲: uniform contact stress is equal to the in-


flation pressure over the square contact area 共width and
length= 24.26 cm兲.
2. Uniform 共equivalent兲: uniform contact stress is equal to the
inflation pressure over the equivalent contact area 共width
= 20.13 cm, length= 29.23 cm兲 obtained from Eq. 共1兲 关see
Fig. 1共b兲兴.
3. Uniform 共maximum兲: uniform contact stress is equal to the
maximum vertical contact stress over the contact area with
the ratio 1:0.85 共width= 19.27 cm, length= 16.38 cm兲.
4. Trapezoidal 共10%兲: trapezoidal contact stress shape with the
maximum vertical contact stress acts on the middle 10% of
the contact area with the ratio 1:0.85 共width= 25.98 cm,
length= 22.09 cm兲. This means the smaller base of the trap-
ezoid is only 10% of the larger base 关see Fig. 2共a兲兴.
5. Trapezoidal 共30%兲: trapezoidal contact stress shape with the
maximum vertical contact stress acts on the middle 30% of
the contact area with the ratio 1:0.85 共width= 23.90 cm,
length= 20.32 cm兲 关see Fig. 2共a兲兴.
6. Trapezoidal 共50%兲: trapezoidal contact stress shape with the
maximum vertical contact stress acts on the middle 50% of
the contact area with the ratio 1:0.85 共width= 22.25 cm,
length= 18.91 cm兲 关see Fig. 2共a兲兴.
Figs. 3共a and b兲 show the vertical strains on top of the sub-
grade and the horizontal strains at the bottom of the asphalt layer
induced by the various shapes of contact stress distributions. A
comparison of the uniform 共square兲 case with the uniform
共equivalent兲 case shows a very slight difference in vertical and
horizontal strains. It is also seen in Figs. 3共a and b兲 that the
smallest vertical strain is observed in the case of the uniform
共equivalent兲 case. These results suggest that current pavement de-
sign methods assuming the contact stress equal to the inflation
pressure, and using the uniform circular type of contact stress,
appear to be unconservative for pavement systems subjected to
supersingle tire loadings. The largest vertical strain occurs in the
case of the uniform 共maximum兲 case, which is only slightly larger
Fig. 2. 共a兲 Nonuniform vertical contact stresses and 共b兲 than that resulting from the trapezoidal 共30%兲 case.
corresponding equivalent nodal forces in three-dimensional element Conventional dual tires typically induce the maximum hori-
in x-z plane 共c兲 ␰-␩ coordinate system zontal strain at the bottom of the asphalt layer in the longitudinal
direction 共Siddharthan and Sebaaly 1999兲. However, it is worth
noting that, except for the uniform 共square兲 case and the uniform
冕冕 冕冕
1 1 1 1
共equivalent兲 case, the maximum horizontal strains occurred in the
f= NTw1 d␰ d␩ + NTw2 d␰ d␩
−1 −1 −1 −1
transverse direction. This indicates that the maximum horizontal
strain may be related to the shape of the contact area of a tire. As

冕冕
1 1
Siddharthan and Sebaaly 共1999兲 indicated, this would imply that
+ NTw3 d␰ d␩ 共9兲 wider tires initiate fatigue cracking in the longitudinal direction.
−1 −1
As seen in Figs. 3共a and b兲, the greatest vertical and horizontal
The shape functions for Ni=1,2,3,4,5,6,7,8 can be found in Cook et al. strains are not generated for the trapezoidal cases but for the
共1989兲. uniform distribution with stress equal to the maximum vertical

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005 / 735

J. Transp. Eng. 2005.131:732-743.


Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. 共a兲 Vertical strains and 共b兲 horizontal strains for various contact stresses

stress. This implies that as the maximum vertical contact stresses depend on the hydraulic conductivity of the soil and of the mate-
act on a larger portion of the contact area, higher strains occur in rials above it. Both effective stress analysis and total stress analy-
the pavement layers. Therefore, the analysis that considers the sis were done for undrained conditions.
uniform maximum vertical contact stress can be conservative for
increased contact stresses of supersingle tires. In this regard, the
Dynamic Finite Element Analysis
uniform maximum vertical contact stress is used as the contact
stress for supersingle tires in our FE analysis. Since the supersingle tire imposes a moving load on a given road,
a dynamic simulation of this load is the most appropriate treat-
ment of the problem. In ABAQUS 共1997兲, the dynamic analysis is
Static Finite Element Analysis
performed by including the inertial forces in the dynamic equa-
In finite element analysis, the selection of boundary conditions tion of equilibrium. The moving load was modeled as a trapezoi-
and mesh size and fineness are important factors for obtaining dal step load. As the tire load enters a certain element, the load in
reasonable results. In our analyses, we investigated the effects of the element increases linearly. The load then reaches a plateau
these factors thoroughly in both the plane-strain 共2D兲 and 3D during which the load is constant for some time. Then, as the tire
analyses. Analyses varying both the width and the depth of the load exits the element, the load in the element decreases linearly
mesh indicated that the results are much more affected by bound- to zero. Because the tire loading is assumed to be moving at
ary proximity for plane-strain 共2D兲 conditions than for 3D condi- constant speed, the time for the tire loading to move across the
tions. Figs. 4共a and b兲 show typical plane-strain 共2D兲 and 3D element can be expressed as the truck velocity divided by the
finite element geometries used in our analyses. length of the element. We observed the response of the central
The subgrade is modeled as saturated, so that tire loadings elements of the mesh for every layer of the pavement, including
applied to the pavement surface generate porewater pressures the subgrade. All calculations were made for the case in which the
within the subgrade. For the clay subgrade, undrained conditions load travels across the central top element of the mesh. The ele-
may be assumed since there is not enough time for the porewater ment is subjected to the cyclic loading 共i.e., repeated loading兲,
pressures to dissipate. For the sand subgrade, drainage conditions having the shape of a trapezoid.

736 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005

J. Transp. Eng. 2005.131:732-743.


E⬘
G⬘ = 共10兲
2共1 + ␯⬘兲
for drained and

Eu
Gu = 共11兲
2共1 + ␯u兲
for undrained loading.
For undrained conditions, because Gu = G⬘ and ␯u = 0.5,
Young’s modulus is given by

3E⬘
Eu = 共12兲
2共1 + ␯⬘兲
Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

For a sand subgrade, in order to estimate shear modulus more


realistically, the properties of Ticino sand 共Salgado 1995; Salgado
et al. 1999兲 were used, and Young’s modulus was calculated using
Eqs. 共10兲 and 共11兲. The analyses in the paper were done in terms
of small-strain values of modulus. This allows us to retain con-
sistency throughout, but leads to subgrade deformations that are
slightly less than would be observed if modulus degradation were
allowed. The shear modulus was calculated using the empirical
equation of Hardin and Black 共1968兲

G0
pa
= Cg
1 + e0
冉 冊
共eg − e0兲2 ␴m⬘
pa
ng
共13兲

where Cg, eg, and ng = material constants that depend only on the
nature of the soil; e0 = initial void ratio; pa = reference pressure
Fig. 4. Typical 共a兲 plane-strain two-dimensional and 共b兲 = 100 kPa⬇ 1 kgf/ cm2 ⬇ 2,000 psf⬇ 14.5 psi; and ␴m⬘ = initial
three-dimensional finite element geometries mean effective stress in the same unit as pa.
For a clay subgrade, the shear modulus was calculated using
the equation of Hardin and Black 共1968兲, which takes into ac-
Analysis Results count the overconsolidation ratio of the clay. Their empirical
equation is given by

冉 冊
Analysis Parameters
G0 共2.973 − e0兲2 ␴m⬘ ng
For the asphalt and base layers, the values of Young’s modulus of = 323 共OCR兲k 共14兲
Zaghloul and White 共1994兲 and Poisson’s ratios obtained from pa 共1 + e0兲 pa
Jaky’s equation were used 共see Table 1兲. The shear strength pa- where k depends on the plasticity index 共PI兲. The K0 of over-
rameters c and ␾ for the asphalt layer were taken from the results consolidated soil was calculated using 共Mayne and Kulhawy
obtained by Goetz et al. 共1957兲. Values of c and ␾ for the base 1982兲
layer were taken from typical values reported by Zaghloul and
White 共1994兲. The values of K0 were determined using K0共NC兲
K0共OC兲 = K0共NC兲冑OCR 共15兲
= 1 − sin ␾⬘ 共Jaky’s equation兲, and the value of ␯⬘ was found from
␯⬘ = K0 / 共1 + K0兲. Finally, assuming a value for E⬘ for the asphalt The undrained shear strength 共su兲 was estimated using the
layer or base layer, values for G⬘ or Gu can be found using following relationship 共Wroth 1984兲:

Table 1. Material Constants for Asphalt, Base, Sand, and Clay Layers
Young’s modulus Poisson’s ratio Cohesion Friction angle

Material E⬘ 共kPa兲 Eu 共kPa兲 ␯⬘ ␯u c⬘ 共kPa兲 cu 共kPa兲 ␾⬘ 共deg兲 ␾u 共deg兲


Asphalt layer 2,068,000 2,068,000 0.32 0.32 207 207 32.5 32.5
Base layer 344,000 404,000 0.28 0.499 0.1 0.1 38 38
Sand 173,000 198,000 0.31 0.499 0.1 0.1 33 33
Clay 共NC兲 60,000 66,000 0.35 0.499 2.6 6 13 0
Clay 共OCR= 3兲 103,000 114,000 0.35 0.499 8.4 14 13 0
Clay 共OCR= 5兲 134,000 148,000 0.35 0.499 12.5 20 13 0
Clay 共OCR= 7兲 160,000 177,000 0.35 0.499 17.7 27 13 0
Clay 共OCR= 9兲 183,000 203,000 0.35 0.499 22.8 33 13 0
Clay 共OCR= 15兲 242,000 268,000 0.35 0.499 32.4 49 13 0

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005 / 737

J. Transp. Eng. 2005.131:732-743.


Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 5. 共a兲 Horizontal stresses at bottom of asphalt layer and 共b兲 vertical stresses on top of subgrade layer from plane-strain two-dimensional finite
element method, comparison of 共c兲 stresses and 共d兲 strains from plane-strain two-dimensional finite element method, three-dimensional finite
element method and ELSYM5

su ␾⬘ shear strength su calculated from Eqs. 共16兲 and 共17兲. Therefore,


= = 0.28 共16兲 the undrained parameters and drained parameters used in this
␴⬘v 100
analysis are internally consistent. The undrained Poisson’s ratio
for a ␾⬘ of 28°. The undrained shear strength of OC clays was was input as 0.499 in place of 0.5 to avoid numerical problems.
estimated using the following correlation suggested by Ladd et al. For sand, following Lee et al. 共1995兲 and Mohammad et al.
共1977兲: 共1995兲, the same material parameters as shown in Table 1 were

冉 冊
Su
␴⬘v
used for both drained and undrained cases.

冉 冊
OC
= OCR0.8 共17兲 Results of Static Finite Element Analysis
Su
␴⬘v NC Plane-Strain Two-Dimensional and Three-Dimensional
In Table 1, c⬘ and ␾⬘ for clay subgrades represent the values of Elastic Analysis
Hvorslev’s true cohesion and Hvorslev’s true friction angle, re- Most studies 共Hallin et al. 1983; Gillespie et al. 1993; Perdomoet
spectively 共Atkinson and Bransby 1978兲. The e-log p graph was et al. 1993; Bell et al. 1992兲 on supersingle tires done to date
drawn by assuming typical values of Cc and Cs for clays 共i.e., focus on comparing the effects on the pavement structure of su-
Cc = 0.3, Cs = 0.045兲. The input values for ce 共Hvorslev’s true co- persingle tires with those of dual tires both in their standard
hesion兲 and ␾e 共Hvorslev’s true friction angle兲 were determined 80 kN 共18,000 lb兲 axle load configuration and with a load consis-
from the two effective stress failure envelopes 共i.e., both NC and tent with the recommended maximum tire load for supersingle
OC clays兲 for the same void ratio. The effective failure envelopes tires. We analyzed three configurations: 80 kN 共18,000 lb兲 dual
were obtained by subtracting the pore pressures 关assuming values tires, 101 kN 共22,800 lb兲 dual tires, and 101 kN 共22,800 lb兲
for Skempton pore pressure parameter A f consistent with OCRs; 1 single tires. As seen in Fig. 5共a兲, supersingle tires induce com-
共corresponding to OCR= 1兲, 0.2 共OCR= 3兲, 0 共OCR= 5 and 7兲, paratively large horizontal stresses at the bottom of the asphalt
−0.1 共OCR= 9兲 and −0.2 共OCR= 15兲兴 from the total stress failure layer, which can be a significant cause of fatigue failure when
envelopes which were drawn using the values of the undrained repeated many times. The horizontal stresses are highest immedi-

738 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005

J. Transp. Eng. 2005.131:732-743.


Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Pore pressures in 共a兲 sand subgrade and 共b兲 normally


Fig. 6. 共a兲 Displacements versus overconsolidation ratios, 共b兲 vertical
consolidated subgrade analyzed by elastic plasticity
strains versus overconsolidation ratios, and 共c兲 overconsolidation
ratios versus pore pressures

modeling. This means that plane-strain 共2D兲 analysis is much


ately below the center of the tires, regardless of the tire configu- more conservative than 3D analysis of the pavement.
rations. Fig. 5共b兲 represents the vertical stresses generated on top ELSYM5 is one of the widely used linear-elastic programs in
of the subgrade by the three types of configurations. As expected, pavement engineering. This program assumes that the contact
supersingle tires produce the highest vertical stress increases, stress is equal to the inflation pressure. The program was devel-
which are approximately 20% larger than those produced by con- oped based on the solution of Burmister 共1943兲 for the problem of
ventional dual tires. a multilayered elastic medium subjected to circular tire loadings.
As in two-dimensional analysis, comparisons were made in As seen in Figs. 5共c and d兲, the results of ELSYM5 and of our 3D
three dimensions of the effects of two conventional dual tire con- analysis are in good agreement for both dual tires and supersingle
figuration with supersingle tires. The trends observed in the three tires.
dimensional results were similar to those observed under plane-
strain conditions, but the magnitudes of stresses, strains, and dis- Three-Dimensional Elastic-Plastic Analysis
placements were considerably different. In the effective stress analyses of clayey subgrades, the soil was
modeled using the Drucker–Prager model with cap. This model
Comparison of Plane-Strain Two-Dimensional can account for two types of failure: shear failure and yield re-
and Three-Dimensional Elastic Finide Element Method sulting from excessive mean stress. An initial cap is set based on
and ELSYM5 Results the preconsolidation pressure. Fig. 6 shows typical results for
Three-dimensional modeling is more realistic than plane-strain normally consolidated and overconsolidated cohesive soils. The
modeling, in which a load extending to infinity in the traffic di- higher the OCR is, the smaller the displacement. As the OCR
rection is applied on top of the pavement. In 3D modeling, the increases, the displacement decreases due to the increase in soil
loading is applied to the limited number of elements correspond- stiffness. For OCR⬎ 7, the rate of reduction of the vertical plastic
ing to the tire contact area. Figs. 5共c and d兲 show that the plane- strains decreases with increasing OCR. Positive pore pressures
strain 共2D兲 modeling induces higher stresses and strains than 3D are generated in cohesive soil with OCR less than 5. Negative

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005 / 739

J. Transp. Eng. 2005.131:732-743.


Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 8. 共a兲 Vertical displacements on top of subgrade for various speeds, 共b兲 permanent vertical displacements of each layer, 共c兲 plastic strains due
to repeated moving loads, and 共d兲 load repetitions versus vertical plastic strains

pore pressures are generated for OCR⬎ 6. The results from Fig. Results of Dynamic Analyses
6共b兲 indicate that the soil does enter the plastic range for supers-
ingle tire loadings. All the dynamic results are presented for sand subgrade modeled
using the Drucker–Prager model.
Generation of Pore Pressures in Sand and Clay Subgrades
Fig. 7共a兲 illustrates the generation of pore pressures in a dense Effects of Moving Loads
sand subgrade based on elasticplastic behavior. Fig. 7共a兲 shows Fig. 8共a兲 is an illustration of the displacements resulting in an
the portions of the elements located immediately below the su- element from different truck speeds for supersingle tires. The
persingle tires. Negative pore pressures are generated both in the lower the truck speed, the larger the displacements. This implies
base layer and in the subgrade. If we take a closer look, negative that at lower speeds, trucks can do more damage to road systems.
pore pressures are being generated in the upper part of the sub- Fig. 8共b兲 shows the displacements of each pavement layer. The
grade, while positive pore pressures are being generated at greater horizontal axis represents the time required for a tire loading to
depths into the subgrade. pass the given element for which the displacement is plotted. The
Fig. 7共b兲 shows the pore pressures in the normally consoli- figure shows that as the moving tire load enters the element being
dated clay subgrade. The pore pressure trends in the base layer are analyzed, the pavement structure starts to deform. After the tire
similar to those observed in Fig. 7共a兲 because the same material loading is out of the element, only permanent, plastic deforma-
properties were used for the base layer as for the sand subgrade. tions remain. It can be seen that permanent deformation occurs in
In contrast to the sand subgrade, positive pore pressures are pro- every pavement layer, but are particularly large in the subgrade
duced in the NC clay subgrade. layer. As mentioned previously, the permanent strain on top of the
The generation of porewater pressures affects the shear subgrade is the basis for rutting criteria. Fig. 8共c兲 shows the
strength of the subgrade. Positive pore pressures in the subgrade evaluation of permanent vertical strains owing to five passages of
will decrease its shear strength due to the reduction of the effec- moving supersingle tires over the element. The permanent strains
tive stresses. In a pavement subjected to repeated tire loadings, as initially increase gradually, and then stabilize. Fig. 8共d兲 shows the
the number of repetitions of the loadings increases, pore pressure predicted permanent strains by FEM from zero to five repetitions,
buildup is expected, further degrading the strength of the sub- as well as those by the predictive model of Diyaljee and Raymond
grade. In contrast, negative pore pressures in the subgrade will 共1983兲. As seen in Fig. 8共d兲, the results of FEM and the perma-
increase the effective stresses, resulting in increasing shear nent strain model of Diyaljee and Raymond are in good agree-
strength. ment in the 0–5 load repetition range. The time required for FE

740 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005

J. Transp. Eng. 2005.131:732-743.


Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 9. 共a兲 configuration of tridem axle 共b兲 axle configurations versus vertical plastic strains

analysis for a large number of repetitions is prohibitive, so ex- = 304 kN 共68,400 lb兲 for tridem axle, and 202 kN 共45,600 lb兲 for
trapolation of the FE results by a relationship such as that of tandem axle兴. Modeling of tire loadings was not conducted for a
Diyaljee and Raymond 共1983兲 is necessary. As the number of whole vehicle, but for axle loadings. Fig. 9共b兲 compares the ef-
passages of a supersingle tire increases, the permanent deforma- fects of axle configurations based on 3D analysis. It shows that,
tions clearly increase, if such a model is deemed satisfactory for single tires induce higher vertical plastic strains than dual tires.
repetitions beyond 5. Supersingle tires with a single axle produce the highest vertical
plastic strains. This confirms the findings of Kilareski 共1992兲 and
Implications for Design and Rehabilitation Practice Jooste and Fernando 共1995兲. Although the total loads for tandem
and tridem axles are larger than the total load for single axles, the
Effects of Tandem and Tridem Axle Configurations distance between the two or three axles in these configurations is
Fig. 9共a兲 shows the configuration of a tridem axle with supers- large enough and hence there is little superposition of effects.
ingle tires. Each tire load is 51 kN 共11,400 lb兲 关total axle load Therefore, the single axle with supersingle tires is more adverse

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005 / 741

J. Transp. Eng. 2005.131:732-743.


Table 2. Load Equivalent Factors 2. Plane-strain modeling of the pavement is much more conser-
Load equivalent LEF vative than 3D modeling. Although the computational effort
Tire type Axle load factor 共LEF兲 共fatigue兲 共rutting兲 is much larger for 3D analyses, the more realistic results
justify the use of these analyses where greater accuracy is
Dual 80 kN 1 1
required.
共18,000 lbs兲
3. According to the comparison of conventional and supersingle
Dual 101 kN 1.27 1.22
tires under elastic-plastic conditions, supersingle tires induce
共22,800 lbs兲
approximately four times larger permanent strains in the
Super single 101 kN 6.23 4.37
pavement layers than conventional tires. Therefore, design of
共22,800 lbs兲
a pavement using LEF values for dual tires leads to overes-
timation of the pavement design life.
to pavement systems than tandem axles and tridem axles with 4. Single axle loadings with supersingle tires induce the largest
supersingle tires. This is due to the higher second deviatoric ten- vertical plastic strains on top of the subgrade of all the axle
sor J2 produced by the single axle with supersingle tires, as ob- configurations considered 共about 1.5 times larger than tridem
Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

served also in the study done by Jooste and Fernando 共1995兲. axle configuration兲.
5. The analysis done for moving loads shows that the higher the
Load Equivalent Factors speed of the truck, the less the load on the subgrade.
There are two major flexible pavement design methods: the As- 6. For clay subgrades, the higher the OCR, the less the defor-
phalt Institute method 共AI method兲 and the AASHTO method. mation.
Both methods employ the concept of load equivalent factor 共LEF兲 7. Positive pore pressures are generated in normally consoli-
and the ESAL. The LEF is defined as the number of equivalent dated clay subgrades, while negative pore pressures are typi-
80 kN 共18,000 lb兲 single axle load applications causing the same cally generated within heavily overconsolidated clays. There-
amount of damage by one passage of an axle. The ESAL is de- fore, in a NC clayey subgrade, the shear strength is reduced
fined as the LEF⫻ number of passages of an axle. However, nei- as a result of traffic loadings.
ther of these two methods can currently account for the LEF of 8. The estimation of permanent vertical deformations using
axles with supersingle tires. As shown in Table 2, rutting LEF and FEM shows good agreement with the permanent strain
fatigue LEF were estimated using 3D elastic-plastic, dynamic fi- model for 0–5 load repetitions. The FE analysis is prohibitive
nite element analysis considering the repeated, moving tire load- for a large number of load repetitions, making the use of a
ings. In order to obtain the dynamic load equivalent factors, the simplified model a necessity.
80 kN 共18,000 lb兲 single axle load was repeatedly applied up to 9. Repeated supersingle tire loadings induce approximately four
20 times. The LEF for rutting was obtained from the permanent times larger permanent vertical strains in the subgrade for
vertical strains on top of the subgrade layer, while the LEF for existing roads than conventional dual tire loadings. This im-
fatigue was determined from the permanent horizontal strains at plies that either mitigation of permanent strains in the sub-
the bottom of the asphalt layer. The number of repetitions of the grade must be pursued or the number of passages of supers-
80 kN 共18,000 lb兲 single axle was obtained from the magnitude ingle tires must be limited by appropriate regulation.
of the permanent vertical strains on top of the subgrade and the
permanent horizontal strains at the bottom of the asphalt layer.
These strains are created by one passage of a moving supersingle Acknowledgments
tire loading. As seen in Table 2, one passage of the supersingle
tire generates the same damage as about four passages of dual Discussions with Panagiotis Kavouklis, Kumar Dave, Nayyar Zia,
tires. This suggests supersingle tire loadings can be approximately Samy Noureldin, and Tony DeSimone were helpful in defining
four times more damaging to pavements than dual tires. This is research needs and are appreciated. The research presented in this
slightly greater than observed by Huhtala et al. 共1989兲 共2.7 times兲 paper was funded by INDOT and the Federal Highway Adminis-
and Akram et al. 共1992兲 共3.2 times兲. This can be explained by the tration through the Joint Transportation Research Program. This
fact that the magnitude of the tire loading used in this study is support is greatly appreciated.
slightly larger than that used in their studies.

References
Summary and Conclusions
ABAQUS Theory Manual 共1997兲. Hibbitt, Karlsson & Sorensen, Inc.
As a result of the trend in the trucking industry of relying more on Akram, T., Scullion, T., Smith, R. E., and Fernando, E. G. 共1992兲. “Esti-
trucks using supersingle tires, the effects of the loading imposed mating damage effects of dual versus wide base tires with multidepth
by these tires on pavements require careful study. In this paper, deflectometers.” Transportation Research Record, 1355, Transporta-
we studied this problem through extensive finite element analyses tion Research Board, Washington, D.C., 59–66.
American Association of State High Officials 共AASHO兲. 共1962兲, “The
using both elastic and elastic-plastic models, considering both
AASHO road test.” Rep. No. 7, Washington, D.C.
plane-strain 共2D兲 and 3D conditions. The main findings of our
Atkinson, J. H., and Bransby, P. L. 共1978兲. The mechanics of soils: An
study, which must be validated in future work that includes well-
introduction to critical soil mechanics, McGraw-Hill New York.
designed field tests, are as follows: Bell, C. A., Randhaw, S., and Xu, Z. K. 共1996兲. “Impact of high-
1. It was found that the orientation of the maximum tensile pressures and single-tired axles in Oregon.” Transportation Research
stress is dependent on the shape of the contact stress distri- Record, 1540, Transportation Research Board, Washington, D.C.,
bution and that the maximum tensile strain occurring in the 132–141.
transverse direction is larger than that in the longitudinal Bonaquist, R. 共1992兲, “An assessment of the increased damage potential
direction. of wide base single tires.” Proc., 7th Int. Conf. on Asphalt Pavements,

742 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005

J. Transp. Eng. 2005.131:732-743.


Vol. 3, 1–16. 共1977兲. “Stress deformation and strength characteristics.” Proc., 9th
Burmister, D. M. 共1943兲. “The theory of stresses and displacements in Int. Conf. on Soil Mechanics and Foundation Engineering, Tokyo,
layered systems and applications to the design of airport runways.” Vol. 2, 421–494.
Proc., Highway Research Board, Vol. 23, 126–144. Lee, W., Bohra, N. C., and Altschaeffl, A. G. 共1995兲. “Resilient charac-
Chen, W. F., and Saleeb, A. F. 共1994兲. Constitutive equations for engi- teristics of dune sand.” J. Transp. Eng., 121共6兲, 502–506.
neering materials, Vol. 2, Elsevier Science B.V., Amsterdam, The Marshek, K. M., Chen, H. H., Conell, R. B., and Hudson, R. W. 共1986兲.
Netherlands. “Experimental determination of pressure distribution of truck tire-
Cook, R. D., Malkus, D. S., and Plesha, M. E. 共1989兲. Concepts and pavement contact.” Transportation Research Record, 1070, Transpor-
applications of finite element analysis, 3rd Ed., Wiley, New York. tation Research Board, Washington, D.C., 9–13.
De Beer, M., Fisher, C., and Jooste, F. J. 共1997兲. “Determination of pneu- Mayne, P. W., and Kulhawy, F. H. 共1982兲. “K0-OCR relationships in
matic tire/pavement interface contact stresses under moving loads and soil.” J. Geotech. Eng. Div., Am. Soc. Civ. Eng., 108共6兲, 851–872.
some effects on pavements with thin asphalt surfacing layers.” Proc., Michelin, Inc. 共1997兲. Michelin Truck Tire Characteristics May 1997.
8th Int.Conf. on Asphalt Pavements, Univ. of Washington, Seattle, Mohammad, L. N., Puppala, A. J., and Alavili, P. 共1995兲. “Resilient prop-
179–227. erties of laboratory compacted subgrade soils.” Transportation Re-
Desai, C. S., and Siriwardane, H. J. 共1984兲. Constitutive laws for engi- search Record, XXX, Transportation Research Board, Washington,
Downloaded from ascelibrary.org by University of Leeds on 05/04/13. Copyright ASCE. For personal use only; all rights reserved.

neering materials with emphasis on geological materials, Prentice D.C., 87–102.


Hall, Englewood Cliffs, NJ. Myers, L. A., Roque, R., Ruth, B. E., and Drakos, C. 共1999兲. “Measure-
Diyaljee, V. A., and Raymond, G. P. 共1983兲. “Repetitive load deformation ment of contact stresses for different truck tire types to evaluate their
of cohesionless soil.” J. Geotech. Eng. Div., Am. Soc. Civ. Eng., influence on near-surface cracking and rutting.” Transportation Re-
108共10兲, 1215–1229. search Record, 1655, Transportation Research Board, Washington,
D.C., 175–184.
Ford, T. L., and Yap, P. 共1990兲. “Truck tire/pavement interface.” Proc.,
Perdomo, D., and Nokes, B. 共1993兲. “Theoretical analysis of the effects
The Promise of New Technology in the Automotive Industry: 23rd of supersingle tires on flexible pavements using circly.” Transporta-
FISITA Congress, Torino, Italy, 333–340. tion Research Record, 1388, Transportation Research Board, Wash-
Gillespie, T. D., Karamihas, S. M., Sayers, M. W., Nasim, M. A., Hansen, ington, D.C., 108–119.
W., and Ehsan, N. 共1993兲. “Effect of heavy vehicle characteristics on Salgado, R. 共1995兲. “Analysis of penetration resistance in sand.” PhD
pavement response and performance.” NCHRP Rep. No. 353, Trans- thesis, Univ. of California, Berkeley, Calif.
portation Research Board, National Research Council, Washington, Salgado, R., Bandini, P., and Karim, A. 共1999兲. “Shear strength and stiff-
D.C. ness of silty sand.” J. Geotech. Geoenviron. Eng., 126共5兲, 451–462.
Goez, W. H., Mclaughlin, J. H., and Wood, L. E. 共1957兲. “The strength of Siddharthan, R. V., and Sebaaly, P. 共1999兲. “Investigation of asphalt con-
a thin layer of bituminous concrete.” Proc., Association of Asphalt crete layer strains from wide-base tires.” Transportation Research
Paving Technologists, Vol. 26, 237–296. Record, 1655, Transportation Research Board, Washington, D.C.,
Hallin, J. P., Sharma, J., and Mahoney, J. P. 共1983兲. “Development of 168–174.
rigid and flexible pavement load equivalency factors for various Siddharthan, R. V., Yao, J., and Sebaaly, P. 共1998兲. “Pavement strain from
widths of single tires.” Transportation Research Record, 949, Trans- moving dynamic 3D load distribution.” J. Transp. Eng., 124共6兲, 557–
portation Research Board, Washington, D.C., 4–13. 566.
Hardin, B. O., and Black, W. L. 共1968兲. “Vibration modulus of normally Tielking, J. T. 共1994兲. “Force transmissibility of heavy truck tires.” Tire
consolidated clay.” J. Soil Mech. Found. Div., 94共SM2兲, 353–369. Sci. Technol., 22共1兲, 60–74.
Huang, Y. H. 共1993兲. Pavement analysis and design, Prentice Hall, Engle- Weissman, S. L. 共1999兲. “Influence of tire-pavement contact stress distri-
wood Cliffs, N.J. bution on development of distress mechanisms in pavements.” Trans-
Huhtala, M., Philajamaki, J., and Pienimaki, M. 共1989兲. “Effects of tires portation Research Record, 1655, Transportation Research Board,
and tire pressures on road pavements.” Transportation Research Washington, D.C., 161–167.
Record, 1227, Transportation Research Board, Washington, D.C., Wroth, C. P. 共1984兲. “Interpretation of in-situ soil tests.” Geotechnique,
107–114. 34共4兲, 449–489.
Jooste, F. J., and Fernando, E. G. 共1995兲. “Development of a procedure Yap, P. 共1988兲. “A comparative study of the effect of truck tire types on
for the structural evaluation of superheavy load routes.” Research road contact pressures.” Paper 881846, Society of Automotive Engi-
Rep. No. 1335-F, Texas Transportation Institute, College Station, Tex. neers, Inc., 53–59.
Kilareski, W. 共1992兲. “Heavy vehicle evaluation for overload permits.” Zaghloul, S., and White, T. D. 共1994兲. “Guidelines for permitting
Transportation Research Record, XXX, Transportation Research overloads-Part 1: Effect of overloaded vehicles on the Indiana High-
Board, Washington, D.C., 194–204. way Network.” FHWA/IN/JHRP-93-5, Purdue Univ., West Lafayette,
Ladd, C. C., Foote, R., Ishihara, K., Schlosser, F., and Poulos, H. G. Ind.

JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / OCTOBER 2005 / 743

J. Transp. Eng. 2005.131:732-743.

You might also like