Available online at www.sciencedirect.

com
Journal
of
Terramechanics
Journal of Terramechanics 48 (2011) 113–123
www.elsevier.com/locate/jterra

Finite element modeling of tire/terrain interaction: Application

to predicting soil compaction and tire mobility
Kaiming Xia
Division of Engineering, Colorado School of Mines, Golden, CO 80401, USA

Received 6 March 2010; received in revised form 3 May 2010; accepted 14 May 2010
Available online 9 June 2010

Abstract

Tire/terrain interaction has been an important research topic in terramechanics. For off-road vehicle design, good tire mobility and
little compaction on terrain are always strongly desired. These two issues were always investigated based on empirical approaches or
testing methods. Finite element modeling of tire/terrain interaction seems a good approach, but the capability of the finite element
has not well demonstrated. In this paper, the fundamental formulations on modeling soil compaction and tire mobility issues are further
introduced. The Drucker–Prager/Cap model implemented in ABAQUS is used to model the soil compaction. A user subroutine for finite
strain hyperelasticity model is developed to model nearly incompressible rubber material for tire. In order to predict transient spatial
density, large deformation finite element formulation is used to capture the configuration change, which combines with soil elastoplastic
model to calculate the transient spatial density due to tire compaction on terrain. Representative simulations are provided to demon-
strate how the tire/terrain interaction model can be used to predict soil compaction and tire mobility in the field of terramechanics.
Ó 2010 ISTVS. Published by Elsevier Ltd. All rights reserved.

Keywords: Tire/terrain interaction; Finite deformation; Finite element; Soil compaction; Tire mobility

1. Introduction soil properties, tire/terrain interface properties, and vehicle

load. The manner in which traffic load transferred to the
Tire/terrain interaction is a very complex research topic terrain surface and to the entire soil ground is controlled
in terramechanics and has been regarded as a critical issue by the tire/terrain interaction mechanism. The tire inflation
to the design of off-road vehicle. It generally covers the pressure and the tire structure are the two most important
issues on tire–terrain friction, rutting of terrain, soil com- factors that influence the contact area and contact pressure
paction, traction, rolling resistance. Obtaining accurate at the tire terrain interface for a given load magnitude. The
solutions to tire/terrain interaction can directly help us distribution of stresses and deformation of the soil ground
understand how tire types and natural terrain conditions are qualitatively understood to be affected by a number of
affect tire mobility and traction performance. These techni- factors, including terrain stiffness, load magnitude, axle
cal inputs constitute the foundation for engineers to design type and configuration, tire type and inflation pressure.
powertrain suited for the worst terrain conditions. There- In the past, analytical and empirical analyses employing
fore study on tire/terrain interaction helps in many level various prediction models were attempted to identify the
of decision making on off-road vehicle design, in situ oper- critical factor rankings. In situ instrumentations were also
ation, and alleviate negative impact of soil compaction for used to monitor the response of soil ground subjected to
agricultural vehicles. The performance of tire mobility is various combinations of axle loads, axle configurations,
directly related to the inflation pressure, tire contact area, tire types and inflation pressures.
Operating off-road vehicle on natural terrain more or
less generates soil compaction. Soil compaction is a
E-mail address: kxia@mines.edu mechanical process by which soil particles are pressed

0022-4898/$36.00 Ó 2010 ISTVS. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.jterra.2010.05.001
114 K. Xia / Journal of Terramechanics 48 (2011) 113–123

together by the momentary application of loads through and the reason why it can be directly used to predict tran-
rolling tires or wheels and eventually increases the bulk den- sient spatial density is introduced in this paper. Tire is usu-
sity of soils. Soil compaction is strongly desired in civil/pave- ally made of incompressible or nearly incompressible
ment/geotechnical engineering, but is extremely unwelcome rubber material and deforms in the category of finite strain
in agricultural engineering. When used as a construction hyperelastic behavior. A finite strain hyperelastic model is
material, the significant engineering properties of soil are introduced below and implemented in ABAQUS as a user
its shear strength, compressibility, and permeability. Com- subroutine for the tire/terrain interaction. Also hybrid
paction of the soil generally increases its shear strength, finite element formulation should be used to alleviate volu-
decreases its compressibility and permeability. Therefore metric locking for nearly incompressible rubber materials.
soil compaction has been an important construction proce- The right choice of tire size is a matter of great importance
dure in civil/pavement engineering. In agricultural engineer- in the design and operation of off-road wheeled vehicles.
ing, driving heavy tractors on the land during ploughing and Driving tires for off-road are required to provide traction
harvesting is a major cause of soil compaction. Therefore on natural terrain, to support the vehicle and to provide
soil compaction should be alleviated as much as possible a minimum resistance to movement over the terrain in
when operating agricultural vehicles on land during seeding the direction of travel. This paper attempts to discuss sev-
or tilling and should be counteracted in many ways. eral aspects concerning tire/terrain interaction modeling
During the past few decades, either tire mobility or soil and investigate the effects of tire inflation pressure, angular
compaction has received much attention in the field of ter- velocity, and frictional property of tire/terrain interface.
ramechanics and geotechnical engineering. The majority of These issues have not been demonstrated in the previous
research effort was focused on field tests [1,2]. Some analyt- work.
ical and empirical approaches on traction and soil compac- The remainder of this paper is organized as follows. The
tion were employed [3–9] as well as finite element method underpinning principle of large deformation for the calcula-
attempted by the author [10]. The general intention of these tion of density is introduced in Section 2. Section 3 presents a
approaches was to determine the factor rankings (tire type, brief introduction of the fundamental formulations for finite
inflation pressure, base of tire, axle load, soil strength) on strain hyperelasticity. In Section 4, we discuss the finite ele-
rut or tire sinkage and traction based on the tests. These ment model for tire/terrain interaction. Applications of
empirical methods, to some extent, have facilitated the finite element tire/terrain interaction model to predicting
development of soil compaction technology and are benefi- soil compaction and tire mobility are presented in Section
cial to off-road vehicle development. However, these ana- 5. Concluding remarks are presented in Section 6.
lytical models oversimplify the dynamic soil/wheel or
soil/tire interactions and cannot accurately capture the 2. Finite deformation and soil compaction
dynamic compaction process. Compared to the existing
analytical and empirical methods, the finite element has In order to predict the transient density, it is necessary
the capability of modeling the tire/terrain interaction in a to recognize that the deformation of soil medium could
very detailed manner without introducing many simplify- be relatively large during a loading–unloading cycle, which
ing assumptions, especially on the dynamic contact part. in turn necessities the clear distinguishing between the
The tire can be modeled as a flexible body using finite strain undeformed configuration and deformed configuration
hyperelastic material model. As a robust numerical tool, after soil compaction. For a typical time step, the updated
finite element has been widely used for stress and deforma- configuration of the body at step tn + Dt may be written as
tion analysis. The advantage of FEA model for tire/terrain a function of the configuration at step tn and the incremen-
interaction is that impenetrability and traction conditions tal displacement Du during the time step Dt. When the com-
can be better treated at the contact interface. On modeling mon origin and direction for the coordinate configurations
of tire traction and mobility, representative work on finite are used, an updated position vector can be given by
element modeling of tire/terrain interaction are given in
[11–14]. Among these early works, Fervers [12] established xnþ1 ¼ X þ u ¼ xn þ Du ð1Þ
a flexible tire model and applied it to study tire/soil interac- where u is the total displacement vector with respect to the
tion using finite element. Nakashima and Oida [13] devel- original configuration. The deformation gradient or inter-
oped soil–tire contact model based on dynamic finite mediate deformation gradient are defined as follows:
element-discrete element method. A lot of tire testing and
@u @u @xnþ1
modeling work has also been conducted by Shoop, etc., F¼ ¼1þ or f¼ ð2Þ
which is mentioned in the literature [14]. In the past, soil @X @X @xn
compaction was always evaluated in term of surface settle- where F is the deformation gradient, f is the relative defor-
ment or tire/wheel sinkage instead of using density. No mation gradient, and 1 is the identity unit tensor. Based on
paper in the field of terramechanics has discussed how to the mass conservation, the volume relationship between
use finite element to predict soil compaction by calculating reference and current configurations can be established as
the spatial density. For accurately predicting spatial soil
density, finite deformation formulation has to be employed dXnþ1 ¼ det Fnþ1 dX0 ¼ J nþ1 dX0 ð3Þ
 K. Xia / Journal of Terramechanics 48 (2011) 113–123 115

Therefore, the large deformation updated Lagrangian can be directly calculated by taking a derivative with
formulation can be used to predict the density change or respect to a deformation measure. Generally a strain
compaction. In updated Lagrangian formulation, an incre- energy density function is expressed in terms of the right
mental displacement is defined with respect to the configura- Cauchy–Green deformation tensor C instead of Green–
tion at time tn, which is considered as the reference Lagrange strain tensor E. Based on Eq. (6), the Second
configuration for the current load step. The updated Piola–Kirchhoff stress tensor S can be given by
Lagrangian formulation can therefore be visualized as a ser- @W @W
ies of intermediate total Lagrangian formulations. And the S¼ ¼2 ð8Þ
@E @C
relative compaction density can be updated at time tn+1 by
q0 qn The hyperelasticity strain energy density function W is
qnþ1 ¼ or qnþ1 ¼ ð4Þ generally expressed in terms of three invariants of C, which
DetFnþ1 Detf nþ1
are
Here q0 is the initial relative density and qn+1 is the rel- 1
ative density at time step tn+1. Fig. 1 shows the current con- I C ¼ trC; II C ¼ ðI 2C  trC2 Þ; III C ¼ det C ¼ J 2 ð9Þ
2
figurations Xn+1 for updated Lagrangian formulation and
Thus the stress is computed as
their relationship with the reference configuration X0.
Therefore, large deformation analysis has a lot of @W @I C @W @II C @W @J
S¼2 þ2 þ2 ð10Þ
advantage in predicting soil compaction. For the reference @I C @C @II C @C @J @C
configuration, the right Cauchy–Green deformation tensor The derivatives of the invariants are derived as follows:
C and Green–Lagrange strain tensor are defined as
@I C @II C @J
1 ¼ 1; ¼ I C 1  C; ¼ J C1 ð11Þ
C ¼ FT F; E ¼ ðC  1Þ ð5Þ @C @C @C
2
As mentioned earlier, the rubber material is nearly
In the current configuration a common deformation
incompressible in deformation and poses a volumetric
measure is the left Cauchy–Green deformation tensor b
locking issue. The hyperelasticity strain energy density
that is defined as
function is assumed to be composed of two parts,
b ¼ FFT ð6Þ  ðI C ; II C Þ þ
W ðI C ; II C ; J Þ ¼ W U ðJ Þ ð12Þ
ðdeviatoric partÞ ðvolumetric partÞ
For finite deformation problem, the stress measure with
respect to the reference configuration is named by second The deviatoric part and volumetric part are respectively
Piola–Kirchhoff stress S. The Cauchy stress r and the Kirch- given by
hoff stress s are defined with respect to the current configura-  
tion. The relationships among these stresses are as follows:  ðI C ; J Þ ¼ 1 l J 23 trC  3 ; U ðJ Þ ¼ 1 KðJ  1Þ2
W ð13Þ
2 2
s ¼ J r ¼ FT SF ð7Þ Based on Eq. (9), the deviatoric stress can be calculated
as follows:
 
3. Hyperelasticity model for rubber material of tire ^
@W 23 1

S¼2 ¼J 1  trCC 1
ð14Þ
@C 3
Rubber material exhibits a nearly incompressible fea-
ture. For rubber material, we postulate the existence of a Applying push forward operation to Eq. (15) gives the
strain energy density function, W, from which stresses deviatoric stress in the current configuration as [15,16].
 
2 @W ^ T 53 1
s ¼ F F ¼ lJ b  trC1 ð15Þ
J @C 3
The deviatoric modulus in the reference configuration
can be give by
2 ^
 ¼4 @ W
D
@C@C
 
2 2 1 1 1 1 1 @C1
¼ lJ 3 trCðC  C Þ  C  1  1  C  trC
3 3 @C
ð16Þ
where,
@C1
¼ C1  C1 ðinvariant formÞ;
@C
@C1
IJ
¼ C1 1
IK CJL ðindicial formÞ ð17Þ
Fig. 1. Current configurations for updated Lagrangian formulation. @CKL
116 K. Xia / Journal of Terramechanics 48 (2011) 113–123

The indicial form of Eq. (16) can be expressed as used to calculate the incremental strain with respect to the
follows: current configuration for large deformation, the return
 1  mapping algorithm for the finite inelasticity is the same
 IJKL ¼ 2 lJ 23 1 trCðC1 C1 Þ  C1 dIJ  dKL C1  trC @CIJ
D as the small strain formulation. The detailed procedures
IJ KL KL IJ
3 3 @CKL
can be found in [16]. The tangent modulus can be obtained
ð18Þ by taking the derivative of incremental stress tensor due to
Performing the push forward operation of Eq. (18) to Jaumann rate,
the current configuration, the indicial forms are given as @Dr
follows: CJEI ¼ ð24Þ
@De
1 The fourth-order continuum tangent modulus tensor
dijkl ¼ F iI F jJ F kK F lL D
 IJKL
J can be used in a similar way as the conventional elastoplas-
 
2 5 1 tic tangent modulus. For static compression, standard
¼ lJ 3 trCdij dkl  dkl bij  bkl dij  trCdik djl ð19Þ
3 3 implicit finite element can accurately predict the soil com-
paction and needs the tangent modulus to calculate the
Based on Eq. (14), the pressure can be defined as
stiffness matrix. However, standard implicit finite element
follows:
formulation cannot effectively handle the dynamic soil/
dU ðJ Þ compactor interaction. Dynamic explicit finite element
p¼ ¼ KðJ  1Þ ð20Þ
dJ seems the robust choice to simulate dynamic soil/compac-
The total Cauchy stress is expressed by tor interaction and soil compaction. In that situation, tan-
gent modulus is not required.
r ¼ s þ p1 ð21Þ
In order to model soil compaction, soil materials cannot 4. Three dimensional finite element model for tire/terrain
be modeled as elastic. It can be modeled as a hypoplastic interaction
model. The Cauchy stress is used and expressed in terms
of the Jaumann rate, which is denoted by the superscript Fig. 2 is the three dimensional finite element model for
J. The additive decomposition of the rate of deformation tire/terrain interaction. The tire model is created purely
tensor, D, into elastic and inelastic parts is assumed: for research use and does not represent any real tire prod-
uct. The tire is 981 mm in diameter and the tread width is
D ¼ De þ DI ð22Þ
327 mm. The element size of the tire is around 30 mm. Tire
The elastic response is specified by applying the hypo- consists of tread, sidewall and shoulder. In the modeling of
elastic law to the elastic part of the rate of deformation. a tire, three structural components are considered, which
The incremental inelastic strain and Cauchy stress tensor are rim, sidewalls, and tread. The tread and sidewalls are
can be updated as made of rubber and are constructed from fiber-reinforced
rubber composites. Generally, the rubber is modeled as
n þ Dr ¼ Xrn þ rn X þ CJE : ðDe  Dep Þ
rnþ1 ¼ r ð23Þ
incompressible or nearly incompressible hyperelastic mate-
The objectivity requires that elastic modulus CJE
should rial, and the fiber reinforcement is modeled as a linear elas-
be isotropic. Otherwise, it will change as the material tic material. The sidewall is composed of lower, upper and
rotates due to its expression in terms of a fixed coordinate intermediate carcass, which bridges between tread and
system by the push-forward. As the rate of deformation is bead. The sidewall is reinforced with rubber and fabric

Fig. 2. Finite element model of tire/terrain interaction.

 K. Xia / Journal of Terramechanics 48 (2011) 113–123 117

plies that provide for strength and flexibility. The tread is paction and traction applications, which controls the trans-
the part of tire which comes in contact with the road sur- ferring of horizontal force and shear stress. Generally the
face. The tread is made of a thick rubber, or rubber/com- friction between two surfaces has been modeled using Cou-
posite compound formulated to provide an appropriate lomb model, where a friction coefficient and a shear stress
level of traction that does not wear away too quickly. limit are introduced to define the contact interfacial fric-
The tread pattern is characterized by the geometrical shape tional behavior and can be obtained from tests. The draw-
of the grooves, lugs, voids and sipes, which is not consid- back of Coulomb frictional law is its inability to capture
ered detailed in this paper. The tire is discretized using the rate dependent frictional contact behavior. In this paper,
8-node linear hexahedral element. In order to alleviate the the Coulomb’s friction law is temporarily employed to define
volumetric locking, some mixed and hybrid finite element the tangential friction behavior between tire and ground.
formulations, such as reduced integration, enhanced/ The simulation is executed using two steps. The first step is
assumed strain, B-bar approach or mixed displacement– simulated by applying the axle load without rolling. The sec-
pressure formulation have to be used to model the nearly ond step is to simulate the tire rolling over deformable soil
incompressible hyperelastic material in the nearly incom- ground and an angular velocity is applied at the tire axle.
pressible limit. The rim is modeled as a rigid using kinematic
coupling. The kinematic coupling is used to maintain the 5. Application of tire/terrain interaction model
constant distance between the tire center and bead area.
The inflation pressure is applied to the inside surface of tire, In this section, some numerical examples are presented to
which can also be modeled using hydrostatic fluid element demonstrate the application of dynamic tire/terrain interac-
with ABAQUS/Explicit and represents fluid-filled cavities tion on prediction of soil compaction and tire mobility.
under hydrostatic conditions. ABAQUS includes a family
of elements that can be used to represent fluid-filled cavities 5.1. Simulation of tire footprint
under hydrostatic conditions. The vehicle load is applied at
the tire hub and acts only in the vertical direction. In order to verify the tire model performance, a rela-
The terrain part is modeled as two-layer system, which is tively softer ground is selected for modeling. The effect of
12,000 mm in length, 3000 mm in width and 2000 mm in inflation pressure is investigated on tire footprint. In this
depth. The top layer of soil is assumed to exhibit elastoplas- paper, we assume that the material properties for rubber
tic mechanical behavior and modeled using Drucker–Prag- material model are l = 1.62 MPa, and bulk modulus
er/Cap model. The bottom layer soil is relatively stiffer and K = 4000 MPa. Based on the given value of the two model
is assumed to deform elastically. All the terrain part is mod- parameters, the equivalent Poisson’s ration is 0.49998 and
eled using 8-node hexahedral elements. In order to reduce is very close to nearly incompressible material. For the ter-
the number of elements, we use fine mesh for the potential rain part, the Young’s modulus E and the Poisson’s ratio m
tire/terrain interaction area and relatively coarse mesh is are assumed to be 3.0 MPa and 0.3, respectively. The con-
used outside the contact area. The boundary conditions stitutive parameters for the Drucker–Prager/Cap model are
are fully fixed at the bottom of the domain and only fixed assumed to be: cohesion c = 0.03 MPa, soil frictional angle
in the normal direction with other surfaces as shown in / = 30°, Cap eccentricity parameter R = 0.5, initial Cap
Fig. 2. Tire/terrain interaction is defined by the interaction yield surface on the volumetric inelastic strain axis ein
vol ¼
property, which is defined by the tangential friction law. 0:04, and transition surface radius parameter a = 0.02.
The interaction between tire and soil is determined by the The tire axle load is 40 kN, which is loaded vertically.
interfacial frictional contact law. The interfacial behavior Fig. 3 shows the effect of tire inflation pressure on tire
related to frictional response is very important in soil com- hyperelastic deformation under the same vertical load.

Fig. 3. Tire footprinting on soft ground versus inflation pressure (0.3 MPa, 0.5 MPa and 0.7 MPa).
118 K. Xia / Journal of Terramechanics 48 (2011) 113–123

0.150

0.145

Tire contact area (m2)

0.140

0.135

Fig. 4. Stress contour projected onto the deformed tire. 0.130

0.3 0.4 0.5 0.6 0.7 0.8
Tire inflation pressure (MPa)
Lower tire inflation will have a relatively higher contact
area with ground. The contact area is calculated based Fig. 6. Tire contact area versus tire inflation pressure.
on the contact algorithms from the finite element output.
The contact area will decrease with increasing tire inflation
80
pressure (see Fig. 6). Figs. 4 and 5 show the stress contours

Vertical displacement of tire hub (mm)

of tire and soil ground, which are projected onto the
deformed tire and terrain during rolling, respectively. The 70
stress unit is MPa. A stress bulb can be easily seen at the
contact area. For the same axle load, Fig. 7 shows the ver- 60
tical displacement of tire hub versus inflation pressure. The
vertical displacement will decrease when inflation pressure
50
is increased. Theoretically, the total displacement of tire
hub consists of two parts: the deformation of soft ground
surface and the elastic vertical displacement due to tire 40
itself. Generally it is very difficult to tell which deformation
contributes to the major vertical displacement of tire hub. 30
The exact soil condition and tire properties are two impor- 0.3 0.4 0.5 0.6 0.7 0.8

tant factors that should be included to predict. However, Tire inflation pressure (MPa)
for the simulated tire properties and soil condition, infla- Fig. 7. Tire hub displacement versus inflation pressure.
tion pressure seems to play an important role to control
the tire vertical displacement. Fig. 7 shows that increasing
inflation pressure is an efficient way to increase the tire stiff- cultural soil. The numerical test simulates the tire rolling
ness and reduces the tire vertical displacement due to tire over a soft terrain. The tire/terrain interaction model is
deformation itself. exactly the same as shown in Fig. 2. For this example,
the tire axle load is 40 kN. The material properties for rub-
5.2. Application to predicting soil compaction on agricultural ber material model are l = 1.62 MPa, and bulk modulus
soil K = 4000.0 MPa. Based on the given values of the two
model parameters, the equivalent Poisson’s ratio is
The aim of this numerical example is to demonstrate the 0.49998 and is very close to nearly incompressible material.
application of large deformation finite element formulation The initial soil density is assumed to be 1800 kg/m3. The
to predict soil compaction induced by tire rolling over agri- Young’s modulus E and the Poisson’s ratio m are

Fig. 5. Stress contour projected onto the deformed soil ground.

 K. Xia / Journal of Terramechanics 48 (2011) 113–123 119

0.5 2.30

Compaction density (ton/m3)

0.4
Volumetric plastic strain

2.20

0.3 2.10

0.2 2.00

0.1 1.90

0 1.80
0 0.1 0.2 0.3 0.4 0.5 0.6 1.0 1.5 2.0 2.5 3.0 3.5
Hydrostatic pressure (MPa) Tire angular velocity (radian/s)
Fig. 8. Hydrostatic pressure versus volumetric plastic strain. Fig. 10. Compaction density versus tire angular velocity.

3.0 MPa and 0.3, respectively. The constitutive model after each tire compaction. Density variation is an effective
parameters for the Drucker–Prager/Cap model are: cohe- measurement of soil compaction and therefore it is a quan-
sion c = 0.03 MPa, soil frictional angle / = 30°, Cap eccen- tity of practical importance in estimating agricultural soil
tricity parameter R = 0.5, initial Cap yield surface on the compaction. Fig. 10 shows the change of compacted den-
volumetric inelastic strain axis einvol ¼ 0:04, and transition sity versus tire angular velocity. The compacted density will
surface radius parameter a = 0.02. The coefficient of fric- decrease versus angular velocity, which is difficult to be
tion between tire and terrain is assumed to be a constant understood. The reason is that compaction is generated
of 0.6. For the large deformation formulation, the compac- by vertical direction axle load and eventually depends on
tion density can be directly predicted based on the updated the amount of external work done by axle load. For same
configuration of each element. The volumetric hardening soil domain and compacted by same axle load, rolling fas-
curve is shown in Fig. 8. Fig. 9 shows the compaction den- ter will reduce the loading time compared to lower angular
sity versus tire inflation pressure, which shows the compac- velocity if the traction is enough. Therefore relatively smal-
tion density will increase with increasing inflation pressure. ler amount of external work is done on the soil. That is why
The reason might be higher inflation pressure will make the tire rolling faster will reduce soil compaction. The similar
tire stiffer and lead to relatively smaller contact area with explanation has been mentioned in [18]. Fig. 11 shows
terrain, which produces a higher compaction on soil. the vertical displacement will decrease with increasing tire
As we mentioned earlier, large deformation finite ele- angular velocity. Similar results have been observed in
ment formulation can calculate the volume change and soil/asphalt compactor analysis.
therefore can be used to predict the soil compaction pro-
cess in agricultural/geotechnical engineering. In this exam- 5.3. Application to predicting tire mobility
ple study, the initial density is assumed to be 1.8 ton/m3.
After tire rolling over, the spatial density becomes larger In this section, the finite element model of tire/terrain
interaction is used to study tire performance for different
2.00
55
Vertical displacement of tire hub (mm)
Compaction density (ton/m3)

1.95
50

1.90 45

40
1.85

35

1.80
0.3 0.4 0.5 0.6 0.7 0.8 30
Tire inflation pressure (MPa) 1.0 1.5 2.0 2.5 3.0 3.5
Tire angular velocity (radian/s)
Fig. 9. Compaction density of the tire contact area versus inflation
pressure. Fig. 11. Vertical displacement of tire hub versus tire angular velocity.
120 K. Xia / Journal of Terramechanics 48 (2011) 113–123

operating conditions. Driving tires for agricultural tractors 7.0

are required to secure enough traction on tire to move the
vehicle forward. The material properties for both terrain 6.0
and rubber materials are exactly the same as those used

Frictional force (kN)

in the example of soil compaction case. The tire axle load
5.0
is 40 kN. Tire rolling over a deformable terrain is different
from tire rolling over a rigid surface. It has been under-
stood that tire inflation pressure, tire angular velocity, 4.0
and interaction property of the tire/terrain interface impact
a lot on the overall performance of tire mobility, which are 3.0
investigated in this part. The results and discussions are
presented as follows.
2.0
0.3 0.4 0.5 0.6 0.7 0.8
5.3.1. Effect of tire inflation pressure Tire inflation pressure (MPa)
Fig. 12 shows the tire rolling radius increases with
increasing tire inflation pressure. The rolling radius is Fig. 13. Tire traction versus inflation pressure.
defined as the ratio of translation velocity to angular veloc-
ity at the tire hub. For this simulation, an angular velocity relatively lower required torque and power to maintain
is applied at the tire axle. The translation velocity of the tire the vehicle speed, which is shown in Fig. 14. As mentioned
hub will be output directly from the FEA model. Therefore in Section 5.1, the soil condition of deformable terrain is
the rolling radius can be calculated. Once the inflation pres- relatively stiffer. The results show that rolling high inflation
sure is increased, tire structure becomes relatively stiffer tire on the relatively stiffer deformable terrain can reduce
compared to that of low inflation pressure. Under the same the energy cost. For relatively softer soil, the results might
axle load and given rubber material properties and tire turn out to be completely different. Higher inflation pres-
structure, the rolling radius becomes relatively larger as sure will require higher torque to maintain the vehicle
inflation pressure increases and also converges to a con- speed. Therefore, in most time, the selection of a right type
stant rolling radius. Theoretically the rolling radius con- of tire involves a decision making process, in which a bal-
verges to the undeformed tire radius on rigid surface. ance should be made between good traction for accelera-
Inflation pressure impacts the contact area a lot, and also tion stage and energy reduction.
influences tire traction. Fig. 13 shows the relationship
between frictional force acting on the tire contact patch 5.3.2. Effect of tire angular velocity
(traction) and tire inflation pressure. With increasing infla- Fig. 15 shows the trend of tire rolling radius versus tire
tion pressure, the tire becomes stiffer and therefore the con- angular velocity. The rolling radius increases with increas-
tact area decreases. Overall increasing inflation pressure ing angular velocity. This trend seems to be inconsistent
reduces the frictional force acting on tread (tire traction). with the results as expected and agrees well with Fig. 10.
However, once the tire’s speed is increased to the desired The best explanation of this result might be focused on
magnitude, the acceleration drops to zero. The frictional dynamic effect. With increasing speed, the terrain does
force acting on the tire tread will also drop to a lower mag- not get enough time to be fully compacted and deformed
nitude to maintain the vehicle speed for a higher inflation to the maximum level. Also the vertical displacement of tire
tire. Therefore, the higher inflation pressure tire delivers a

500 5.0

490 4.0
Rolling radius (mm)

Torque (kN-m)

480 3.0

470 2.0

460 1.0

450 0.0
0.3 0.4 0.5 0.6 0.7 0.8 0.3 0.4 0.5 0.6 0.7 0.8
Tire inflation pressure (MPa) Tire inflation pressure (MPa)

Fig. 12. Tire rolling radius versus inflation pressure. Fig. 14. Axle torque versus inflation pressure.
 K. Xia / Journal of Terramechanics 48 (2011) 113–123 121

490 250

Acceleration of tire hub (mm/s2)

485
220
Rolling radius (mm)

480
190
475

160
470

465 130

460 100
1.0 1.5 2.0 2.5 3.0 3.5 1.0 1.5 2.0 2.5 3.0 3.5
Tire angular velocity (radian/s) Tire angular velocity (radian/s)
Fig. 15. Rolling radius versus tire angular velocity. Fig. 17. The acceleration of tire hub versus angular velocity.

hub has not reached an equilibrium position in the vertical is the product of angular velocity x and rolling radius R.
direction and is lower than the static displacement. This Fig. 18 shows the relationship between tire torque and
might be the reason why the dynamic finite element model angular velocity. The torque of tire axle increases when tire
of tire/terrain interaction produces a relatively high rolling angular velocity is increased. It indicates a faster rolling tire
radius. Even though the rolling radius has been shown to does need a higher torque to rotate the axle of tire. This
increase when the tire angular velocity is increased, the agrees with the results as expected.
traveling velocity of tire hub seems to be linearly propor-
tional to angular velocity (see Fig. 16). 5.3.3. Effect of frictional coefficient of tire/terrain interface
Fig. 17 reflects the relationship between translation This example is used to verify the effect of frictional
acceleration and applied angular velocity. For the given property of contact interface on the tire mobility. We sim-
tire/terrain interaction properties and tire structure, the ulate the tire rolling over deformable terrain with same
acceleration of tire hub increases with increasing angular angular velocity at the acceleration stage, but with different
velocity and tends to converge with increasing angular frictional coefficient of the contact interface. During the
velocity, which is shown in Fig. 17. The most important acceleration stage of a vehicle, the frictional force between
reason might be due to the limitation of traction in the tire and terrain interface serves as traction to accelerate the
tire/terrain interface. In reality, the traction in the tire/ter- tire translation velocity. Higher frictional coefficient will
rain interface cannot be arbitrarily increased. An upper provide larger traction, which produces a higher accelera-
limit always exists and is controlled by the frictional prop- tion at the acceleration stage. Fig. 19 shows that accelera-
erty of the contact interface. That is why the acceleration of tion increases with increasing frictional coefficient of tire/
the tire hub cannot be exactly equal to the product of roll- terrain interface. This simulation results agrees well with
ing radius and angular acceleration. The velocity will take the statement in [17]. Fig. 20 shows the traveling distance
time to be increased to the desired magnitude of xR, which during the first 7 s for two frictional coefficients of tire/ter-

1500
6.0
Traveling velocity of tire hub (mm/s)

5.0
1200
Torque (kN-m)

4.0

900 3.0

2.0
600
1.0

300 0.0
1.0 1.5 2.0 2.5 3.0 3.5 1.0 1.5 2.0 2.5 3.0 3.5
Tire angular velocity (radian/s) Tire angular velocity (radian/s)

Fig. 16. Traveling velocity of tire hub versus angular velocity. Fig. 18. The tire axle torque versus angular velocity.
122 K. Xia / Journal of Terramechanics 48 (2011) 113–123

200 studied in the past. In order to accurately model tire/terrain

interaction, the tire has to be modeled as deformable body
160 and better modeled as finite strain hyperelasticity. This kind
Acceleration (mm/s2)

of simulation has rarely been conducted in the previous

research. In order to predict soil compaction, finite deforma-
120
tion formulation has to be used to capture the configuration
change, which can be directly used to predict the transient
80 spatial density due to compaction. This paper offers the
novel feature of modeling the spatial density change, which
40
has been an issue of great significance in the application of
finite element techniques to agricultural/geotechnical/pave-
ment engineering. Tire mobility is an important issue in the
0 field of terramechanics for ground vehicle design. The effects
0.3 0.4 0.5 0.6 0.7 0.8
Coefficient of friction of contact interface
of tire inflation pressure, rolling speed and frictional prop-
erty of tire/terrain interface on rolling radius, acceleration,
Fig. 19. The acceleration of tire hub versus frictional coefficient of tire/ torque, traction are obtained respectively from finite element
terrain interface.
tire/terrain interaction model. The right choice of tire size is
a matter of great importance in the design and operation of
rain interface. During the tire acceleration stage, frictional off-road wheeled/tired vehicles. The major effort to predict
force between tire and terrain provides an impetus to accel- tire mobility was focused on empirical approaches or field
erate the tire to the desired velocity, which is approximately tests in the past. It has been proved to be extremely expensive
equal to the product of rolling radius and angular velocity. and time consuming to test tire performance using field tests.
If the traction is too small at the beginning, the acceleration This paper demonstrates numerical model can serve as a
will be small and the tire has a large slipping rate and will robust tool on predicting soil compaction and tire mobility.
spin against ground. Based on Fig. 20, we can clearly see Some technical issues on developing finite element model for
that higher traction will accelerate the tire rolling faster tire/terrain interaction were discussed in this paper. Cer-
than that of smaller traction. All these figures show that tainly, we cannot address all the issues in this paper. A lot
FEA based tire/terrain interaction model can reasonably of efforts need to be further investigated in the future, such
predict the effect of frictional property of contact interface as simulation of cornering behavior, turning, contact surface
on tire mobility. adhesion, tread width effect, and validation with field tests.
The main purpose of this paper is to demonstrate that a lot
6. Concluding remarks of issues relating to soil compaction and tire mobility can
be directly predicted based on the finite element modeling
In this paper, a three dimensional finite element model for of tire/terrain interaction for ground vehicle development
tire/terrain interaction is developed as well as a finite strain in the future.
hyperelasticity model for modeling of rubber materials.
For tire rolling over deformable terrain, two important Acknowledgement
issues involving soil compaction and tire mobility have been
The author is grateful to an anonymous reviewer for the
comments that have helped to enhance the quality of the
1200 paper.

1000 References
Traveling distance (mm)

800 Frctional coefficient: 0.3 [1] Eguchi T, Muro T. Measurement of compacted soil density in
Frictional coefficient: 0.8
a compaction of thick finishing layer. J Terramech 2007;44:
600 347–53.
[2] Nguyen VN, Matsuo T, Inaba S, Koumoto T. Experimental analysis
of vertical soil reaction and soil stress distribution under off-road
400 tires. J Terramech 2008;45:25–44.
[3] Harnisch C, Lach B, Jakobs R, Troulis M, Nehls O. A new tyre–soil
200 interaction model for vehicle simulation on deformable ground.
Vehicle Syst Dynam 2005;43:384–94.
0
[4] Pietzsch D, Poppy W. Simulation of soil compaction with vibratory
1 2 3 4 5 6 7 roller. J Terramech 1992;29:585–97.
Time (sec) [5] Carman K. Prediction of soil compaction under pneumatic tires using
fuzzy logic approach. J Terramech 2008;45:103–8.
Fig. 20. Traveling distance versus time for different frictional coefficient of [6] Canillas E, Salokhe V. Modeling compaction in agricultural soils. J
tire/terrain interface. Terramech 2002;39:71–84.
 K. Xia / Journal of Terramechanics 48 (2011) 113–123 123

[7] Arvidsson J, Ristic S. Soil stress and compaction effects for four [13] Nakashima H, Oida A. Algorithm and implementation of soil–tire
tractor tyres. J Terramech 1996;33(5):223–32. contact analysis code based on dynamic FE–DE method. J Terra-
[8] Braunack MV. A tyre option for sugarcane haulout trucks to mech 2004;41:127–37.
minimize soil compaction. J Terramech 2004;41:243–53. [14] Shoop SA, Richmond PW, Lacombe J. Overview of cold regions
[9] Sharma AK, Pandey KP. The deflection and contact characteristics of mobility modeling at CRREL. J Terramech 2006;43:1–26.
some agricultural tyres with zero sinkage. J Terramech 1996;33(6): [15] Belytschko T, Liu WK, Moran B. Nonlinear finite elements for
293–9. continua and structures. John Wiley & Sons Ltd; 2000.
[10] Xia K, Masud A. A stabilized finite element for finite deformation [16] Simo JC, Hughes TJR. Computational inelasticity. Springer-Verlag;
elastoplasticity in geomechanics. Comput Geotech 2009;36:396–405. 1993.
[11] Liu CH, Wong JY. Numerical simulations of tire–soil interaction [17] Gillespie T. Fundamentals of vehicle dynamics. Society of Automa-
based on critical state soil mechanics. J Terramech 1996;33(5):209–21. tive Engineers, Inc., 1992.
[12] Fervers CW. Improved FEM simulation model for tire–soil interac- [18] Huang YH. Pavement analysis and design. NJ: Prentice Hall; 1993. p.
tion. J Terramech 2004;41:87–100. 31–2.