See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225859898

Prediction of Pavement Fatigue and Rutting Life Using Different Tire Types

Article  in  KSCE Journal of Civil Engineering · September 2008

DOI: 10.1007/s12205-008-0297-4

CITATIONS READS

11 1,979

1 author:

Daewook Park
Kunsan National University
85 PUBLICATIONS   647 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Daewook Park on 26 February 2016.

The user has requested enhancement of the downloaded file.

KSCE Journal of Civil Engineering (2008) 12(5):297-303 Highway Engineering
DOI 10.1007/s12205-008-0297-4
www.springer.com/12205

Prediction of Pavement Fatigue and Rutting Life Using Different Tire Types
Dae-Wook Park*
Received February 15, 2008/Accepted May 20, 2008

···································································································································································································································

Abstract

A uniform circular vertical contact stress is commonly assumed in representing wheel loads in pavement analysis procedures.
However, experimental measurements have shown that the actual loading conditions are non-uniform and depend on the tire
construction, tire load and tire inflation pressure. The objective was to study the effects of different tire types on pavement service life
and to suggest how wheel loads may be modeled in currently implemented layered elastic pavement analysis programs to provide a
better approximation of pavement response parameters for pavement design and evaluation purposes. To carry out this objective,
three types of tires (radial, bias ply, and wide-base) were used to study the effects of tire types on pavement service life. A 3D FE
model of asphalt pavements was developed and used to investigate the influence of 3D tire contact stresses on predicted pavement
life. The results of the 3D FE model were compared with corresponding results obtained using the layered linear elastic program,
BISAR.
Keywords: tire contact stress, non-uniform stress, 3D FEM, layered linear elastic program
···································································································································································································································

1. Introduction design procedures. Wang and Machemehl (2006) conducted the

Advances in computing power and experimental characteriza-
tion techniques have led to more utilization of the finite element
(FE) analysis to predict the response of asphalt and concrete
pavements under actual complex loading conditions. De Beer et
al. (1997, 2002) conducted the intensive studies on pavement
design and analysis of a thin pavement using the 3D measured
tire contact stresses. The results of their study showed that the
shape of tire contact stress distributions is dependent on tire
loading conditions. Under overloaded/under-inflated tire, two
maximum stresses that can cause the downward cracking near
wheel paths are located at the tire edge. The application of
measured vertical and lateral stresses in pavement design and
analysis can provide much more realistic tire pavement loading
condition than the circular uniform tire loading conditions. Also,
use of the lateral tire contact stresses on thin asphalt surface can
cause large differences compared to conventional pavement
analysis methods. Roque et al. (1998) analyzed the near-surface
pavement stresses that may be induced under bias-ply and radial
tires using measured tire contact stresses. The authors noted that
significant transverse tensile stresses were predicted under the
treads of radial tires, which can potentially initiate surface cracks
along the wheel path. However, transverse tensile stresses were
not predicted under bias-ply tires. Based on these results, the
authors pointed to the need for better defining tire-pavement
contact stresses and their consideration in modern pavement
prediction of pavement response using measured tire contact
stresses.
The results of a three dimensional (3D) finite element (FE)
model with measured tire contact stresses were compared with
those of a layered elastic program. The results from the layered
elastic program showed that a circular uniform pressure
distribution overestimated tensile strains at the bottom of the
asphalt layer and overestimated compressive strains at the top of
the subgrade. They also studied the effects of tire pressures on
pavement response, and found that increased tire pressure can
cause more fatigue cracking and rutting. Park et al. (2005)
conducted an analysis of the pavement response of WesTrack.
The profiles of permanent deformation at WesTrack were
simulated by elastic-viscoplastic constitutive model embedded in
3D FE model and measured tire contact stresses. They also
investigated the effect of tire load on pavement response. They
found that the tire pressure distribution was found to be
dependent on the tire load, and it had significant influence on the
distributions of shear stress and plastic strain. Under the fixed tire
pressure, the increased shear stresses and plastic strains were
found at the tire edges. The FE analysis based on the elastic-
viscoplastic relation was able to simulate the measured
permanent deformation profiles.
In this paper, three types of tires (radial, bias, and wide-base)
were used to study the effects of tire types on pavement
responses. A3D FE model of asphalt pavements was developed

*Member, Assistant Professor, School of Civil and Environmental Engineering, Kunsan National University, Kunsan 573-701, Korea (E-mail: dpark@
kunsan.ac.kr)

− 297 −
 Dae-Wook Park

and used to investigate the influence of 3D tire contact stresses components over that area (De Beer and Fisher, 1997). Data on
on predicted pavement life. The results of the 3D FE model were the 11R22.5, 10x20 bias ply, and 425/65R22.5tires were measured
compared with corresponding results obtained using the layered with the VRSPTA in a study conducted in 1997 by the University
linear elastic program, BISAR. Three different analysis methods of California at Berkeley (UCB).
were used: Three different loading conditions were selected to analyze
1) 3D FE with measured tire contact stresses; pavement response. These represent the low, middle and high
2) BISAR using measured tire contact areas to determine combinations of tire load and tire inflation pressure used for a
equivalent uniform circular pressure distributions (BM given tire during the tests. The low load high inflation pressure
analysis); and and the high load low inflation pressure combinations were not
3) BISAR using the conventional procedure of calculating the considered in this study as these conditions represent extreme
equivalent circular loaded area by dividing the tire load with deviations from tire manufacturer recommended tire load/tire
the tire inflation pressure (BC analysis). inflation pressure guidelines. The load and tire inflation pressures
used in this paper are summarized in Table 1.
2. Tire Contact Stress The tire contact area is needed to analyze the interactions
between the tire load and the pavement. Tire imprints were
Published data from tire contact pressure measurements show obtained for the different tires tested and for the matrix of tire
that the vertical stress components at the moving tire/pavement loads and tire inflation pressures used. The BM method used the
interface are not uniformly distributed (De Beer and Fisher, measured tire contact area in the analysis of pavement response
1997). In addition, shear stresses in both the transverse and using the BISAR program. Table 2 shows tire contact areas and
longitudinal directions exist. However, it has been difficult to equivalent tire contact pressures used in the BM and BC
accurately quantify these shear stresses under moving truck tires. analyses.
In this study, relationships for predicting tire contact areas
2.1 Measurements
The VRSPTA and its most recent version, referred to as the
Stress-in-Motion (SIM) Mk IV system, have been used to to
measure the 3D tire contact stresses occuring between the
pavement surface and relatively slow-moving pneumatic tires.
The VRSPTA was used in a study conducted by the University
of California at Berkeley to measure tire contact stresses under
various truck and aircraft tires (De Beer and Fisher, 1997). It was
developed by the Council for Scientific and Industrial Research
(CSIR) of South Africa. The VRSPTA is set within a steel pad
embedded in the pavement such that the instrumented pad is
flushed with the pavement surface. The VRSPTA system
consists of 1041 flat-topped cone shaped hollow steel pins
mounted along 51 rows on a 50 mm steel base plate (De Beer
and Fisher, 2002). Twenty pins in the center row are instru-
mented with strain gauges, thus forming 20 small individual
triaxial load cells transversely distributed across the base plate.
The measured voltages from the strain gauges are transformed to
load values using regression coefficients. An effective diamond
shaped area (250 mm2) is used during post-processing to convert
the various measured load components in the three main
directions (X, Y, and Z) to the respective average contact stress Fig. 1. Tire Imprint Area for 11R22.5 Radial Tire

Table 1. Summary of Load and Tire Inflation Pressure Combinations Used in Analyses
Tire Type
Tire load-inflation 11R22.5 10x20 425/65R22.5
pressure combination
Inflation pressure Inflation pressure Inflation pressure
Load (kN) Load (kN) Load (kN)
(kPa) (kPa) (kPa)
Low 26.0 520 26.0 689 26.0 503
Middle 31.0 690 31.0 820 46.0 696
High 41.0 820 36.0 917 56.0 896

− 298 − KSCE Journal of Civil Engineering

 Prediction of Pavement Fatigue and Rutting Life using Different Tire Types

Table 2. Tire Contact Areas and Equivalent Tire Contact Pressures Used in Analyses
Tire Inflation Pressure Calculated Tire Contact Measured Tire Contact Equivalent Tire Contact
Tire Type Load (kN)
(kPa) Area (mm2) Area (mm2) Pressure (kPa)
26.0 520 50,406 36,585 712
11R22.5 31.0 690 45,000 34,547 898
41.0 820 50,013 43,389 945
26.0 689 37,806 35,264 739
10x20 31.0 820 37,815 36,897 840
36.0 917 39,291 42,406 849
26.0 503 51,790 33,858 770
425/65R22.5 46.0 696 66,113 54,664 842
56.0 896 62,531 54,903 1,020

were developed based on the tire imprints available from the
tests. The tire imprints were digitized and a program was then
used to read the electronic prints and to calculate the contact area
based on the shading intensity of the pixels read. Researchers
then conducted regression analyses to determine relationships
between tire load, tire inflation pressure and tire contact area for
the different tires tested. Tire imprint area for 11R22.5 radial tire
is shown in Fig. 1.
The equations for predicting contact area are given below:
11R22.5: A = 54.4740 + 0.0066 × TL − 0.4258 × TP (1)
10 × 20: A = 66.6831 + 0.0059 × TL − 0.4218 × TP (2)
425/65R22.5: A = 53.64 + 0.0055 × TL − 0.2915 × TP (3)
where, A = Predicted Tire Contact Area, in2
TL = Tire Load, lbs
TP = Tire Pressure, psi
2.2 Analysis
Tire contact stresses were measured along the longitudinal
direction at approximately 2-mm intervals. These measurements
are much too fine to use directly in a 3D finite element analysis.
To reduce computation time to a feasible level, author used a
simple routine to reduce the sampling rate without significant
accuracy loss. When the sampling data is decreased, the process
is called decimation because the original data is decimated
(Samuel and Ruth, 1993). The decimation was accomplished
using a procedure that preserves the frequency content of the
original data, and resulted in a mesh configuration where the
finest mesh size is approximately 5 mm (0.2 inch) in the
direction of wheel travel. Figs. 2 and 3 show the original and
decimated data of vertical tire contact stress, respectively, for the
11R22.5 tire corresponding to a test inflation pressure of 520 kPa
and a tire load 26 kN.

Fig. 2. Original Tire Contact Stress in Vertical Direction (11R22.5 tire at 26 kN load and 520 kPa inflation pressure)

Vol. 12, No. 5 / September 2008 − 299 −

 Dae-Wook Park

Fig. 3. Decimated Tire Contact Stress in Vertical Direction (11R22.5 Tire at 26 kN load and 520 kPa inflation pressure)

3. 3D Finite Element Model Table 3. Modulus of Layer Used in Analyses

Modulus of Layer (MPa)
Researchers established a 3D FE model based on the multi- Case No.
Asphalt Base Subgrade
purpose finite element package, ABAQUS (ABAQUS, 2002).
Initially, a suitable mesh configuration had to be determined for 1 2758 345 52
predicting pavement response under measured tire contact 2 4482 690 103
stresses. For this purpose, researchers evaluated a number of
mesh configurations of varying number of elements and element
types by comparing finite element predictions with correspond-
ing predictions from the BISAR program. To be consistent with
the assumptions of horizontally infinite pavement layers and a
semi-infinite subgrade, the lateral and longitudinal dimensions of
the finite element mesh were varied until predictions of pave-
ment response compared reasonably well with the layered elastic
analysis results. From this analysis, a finite element mesh with
lateral and longitudinal dimensions of 3.81 m × 3.81 m was
found to be appropriate. In addition, the subgrade was modeled
using infinite elements. The thickness used is 152 mm AC layer,
304 mm base layer, and infinite of subgrade. Fig. 4 shows the 3D
finite element mesh used in this study. Elements in the region of
the tire contact stresses are approximately 17 mm (column) × 5
mm (row) in size, with the tire contact area consisting of 20
columns and 68 rows. A coarser mesh was used outside of the
wheel load area, with the element size progressively increasing Fig. 4. 3D FE Model (Top View; y is wheel travel direction)
with distance from the load. With this biased mesh configuration,
computation time was further reduced. The interface between 4. Layered Elastic Analysis
layers was assumed to be fully bonded. Since the measured tire
contact stresses were not symmetrical, a full 3D FE model was Layered elastic analysis is commonly used in pavement design
used. and evaluation programs implemented within transportation
In the analyses, each layer was assumed to be linear elastic to agencies. On the other hand, 3D finite element programs, while
compare with the BISAR predictions. The elastic modulus of more powerful and versatile, are more difficult to implement and
each layer was varied according to Table 3. A Poisson's ratio of are thus, not used on a routine basis, except for in-house research
0.4 for all the layers was used. or specialized investigations. Consequently, a comparison of 3D

− 300 − KSCE Journal of Civil Engineering

 Prediction of Pavement Fatigue and Rutting Life using Different Tire Types

FE and layered elastic analysis results were conducted in an
attempt to establish guidelines on how existing layered elastic
programs may be used to provide a better approximation of
pavement response parameters for pavement design and evalua-
tion purposes.
In this study, researchers compared the results from the 3D FE
analyses with corresponding predictions from the layered elastic
analyses. Two types of layered elastic analysis were used as
noted earlier in the paper. In the BM analysis, the measured tire
contact area for a given tire type, tire load and tire inflation
pressure was used to determine an equivalent uniform circular
pressure distribution. In the conventional analysis, the tire
inflation pressure is assumed as the tire contact stress on the
pavement surface. This approach provides no differentiation
between tire types, even though tire footprints vary with tire
type.
5. Prediction of Pavement Fatigue and Rutting
Life
Researchers compared the predicted AC tensile strains and
subgrade vertical strains from the finite element and layered
elastic analyses. To find the critical point for a given condition,
the tensile strain at the bottom of the AC layer and the
compressive strain at the top of subgrade were evaluated at the
center and along the four edges of the tire contact area.

Fig. 5. Comparisons of Predicted Log Fatigue Life and Rutting Life: (a) 11R22.5; (b) 10×20 Bias Ply; and (c) 425/65R22.5

Vol. 12, No. 5 / September 2008 − 301 −

 Dae-Wook Park

Researchers found the predicted response at the center of the where,

loaded area to be critical. In lieu of comparing the strain
predictions from the 3D FE and layered elastic analyses,
pavement service lives were predicted using the Asphalt Institute
equations for fatigue cracking and rutting (Asphalt Institute,
1983). In this way, the relative differences in the predicted strains
between uniform and non-uniform loading conditions were
assessed in terms of service life. The Asphalt Institute equations
are given by (Asphalt Institute, 1983):
( Nf ) = 0.0796 ( εac )
c –3.291 ( Eac ) –0.854
(4)
( Nf )r = 1.365 × 10 ( εsg )–4.477
–9
(5)
(Nf)c = allowable number of load repetitions based on fatigue
cracking,
(Nf)r = allowable number of load repetitions based on rutting,
εac = predicted tensile strain at the bottom of the AC layer,
Eac = asphalt concrete modulus (psi), and
εsg = predicted vertical compressive strain at the top of the
subgrade.
Fig. 4 compares the predicted fatigue lives and rutting lives
based on the computed AC tensile strains and the computed
subgrade vertical compressive strains from the 3D FE and
layered elastic analyses. For the 11R22.5 radial tire and 465/

Fig. 6. Predicted Log Fatigue Life and Rutting Life for Case 1 Pavement: (a) 11R22.5; (b) 10×20 Bias Ply; and (c) 425/65R22.5

− 302 − KSCE Journal of Civil Engineering

 Prediction of Pavement Fatigue and Rutting Life using Different Tire Types
65R22.5 wide-base tire, the equivalent uniform circular contact
stresses based on the measured tire contact areas are larger than
the tire inflation pressures, which are used as contact stresses in
the conventional layered elastic (BC) analysis. Thus, the
predicted fatigue and rutting lives from the modified layered
elastic (BM) analysis are less than the corresponding predictions
from the BC analysis. For the 10×20 bias ply tire, the equivalent
uniform circular contact stresses based on the measured tire
contact areas are similar to the corresponding tire inflation
pressures used in the BC analysis. Thus, the predicted fatigue
lives from the BM and BC analyses are similar. In terms of the
predicted rutting service lives based on subgrade compressive
strain, the results show no significant differences between the 3D
finite element and layered elastic (BM and BC) analyses, as may
be observed from Fig. 5. This finding indicates that the effects of
differences in tire contact pressure distributions are mainly seen
near the surface and diminish with depth. This result, however,
does not imply that tire contact pressure distributions have no
effect on rutting. This finding indicates that tire contact stresses
will influence the development of rutting in the surface layer as
well as top down cracking.
Fig. 6 compares the predicted fatigue and rutting lives of Case
1 material property for each tire, and shows the similar results
with Fig. 5. The fatigue and rutting lives of the 11R22.5 and 465/
65R22.5 tires by 3D FEM were overestimated by BC analysis in
most load-tire inflation pressure combinations. For the 10×20
bias ply tire, the difference between BM analysis and BC
analysis was not significant.
6. Conclusions
Three types of tires (radial, bias ply, and wide-base) were used
to study the effects of tire types on pavement service life. A 3D
FE model of asphalt pavements was developed and used to
investigate the influence of 3D measured tire contact stresses on
predicted pavement life. The results of the 3D FE model were
compared with corresponding results obtained using the layered
linear elastic program, BISAR. Two types of BISAR analysis
were used in this study. The BM method is a modified analysis
method which can apply tire contact stress that is obtained load
divided by tire footprint and the BC method is a conventional
analysis method which is applying tire inflation pressure. Based
on the results presented, the following findings are noted:
For the 11R22.5 radial tire and the 465/65R22.5 wide-base
tire, the comparisons of the results from the 3D FE, BM and BC
analyses showed that the differences in the fatigue predictions
are significant. For the 10×20 bias ply tire, similar findings were
obtained between the BM analysis and BC analysis.
Vol. 12, No. 5 / September 2008
View publication stats
All three-analysis methods gave similar predictions of
pavement rutting life based on the subgrade strain criterion.
Thus, in terms of the effect of tire contact stresses on pavement
rutting, the permanent deformation properties of the surface
material would be critical.
Differences in the fatigue predictions between the BM and BC
methods showed the effect of tire contact area in the predicted
AC tensile strains. For a given tire load and tire inflation pressure,
the conventional method of analysis ignores the differences in
the tire footprints between various tire types.
Based on the results from comparisons of the 3D FE, BM and
BC methods, the author recommend the application of the BM
method in layered elastic analyses of pavement response.
Acknowledgements
This work was supported by the Korea Research Foundation
Grant funded by the Korean Government (MOEHRD, Basic
Research Promotion Fund) (KRF-2006-003-D00603). The authors
would like to thank the MOEHRD for support.
References
ABAQUS (2002). ABAQUS User’s Manual, Version 6.2, Hibbit,
Karlsson, and Sorenson, Inc., Pawtucket, Rhode Island, 2002.
Asphalt Institute (1982). Research and Development of the Asphalt
Institute's Thickness Design Manual (MS-1), 9th Edition, Research
Report 82-2, Asphalt Institute, 1982.
De Beer, M., Fisher, C., and Jooste, F. (1997). “Determination of
pneumatic tyre/pavement interface contact stresses under moving
loads and some effects on pavements with thin asphalt surfacing
layers.” 8th International Conference on Asphalt Pavements, Seattle,
USA, pp. 179-227.
De Beer, M., Fisher, C., and Jooste, F. (2002). “Evaluation of non-
uniform tyre contact stresses on thin asphalt pavements.” 9th
International Conference on Asphalt Pavements (CD), Copenhagen,
Denmark.
De Beer, M. and Fisher, C. (1997). Contact stresses of pneumatic tires
measured with the vehicle-road surface pressure transducer array
(VRSPTA) system for the university of california at berkeley (UCB)
and the Nevada Automotive Test Center (NATC), Research Report
CR-97/053, Transportek, Pretoria, South Africa.
Park, D.W., Martin, A.E., AND Masad, E. (2005). “Effects of nonuniform
tire contact stresses on pavement response.” Journal of Transporta-
tion Engineering, ASCE, Vol. 131, No 11, pp. 873-879.
Samuel, D.S. and Ruth, A.D. (1993). Signal processing algorithm in
fortran and C, Prentice Hall, Englewood Cliffs, New Jersey.
Wang, F. and Machemehl, R.B. (2006). “Mechanistic-empirical study of
effects of truck tire pressure on pavement: Measured tire-pavement
contact stress data.” The Journal of Transportation Research
Record, No. 1947, pp. 136-145.
− 303 −