S. Oida,1 E. Seta,1 H. Heguri,1 and K.

Kato1

Soil/Tire Interaction Analysis Using

FEM and FVM2

REFERENCE: Oida, S., Seta, E., Heguri, H. and Kato, K., “Soil/Tire Interaction Analysis
Using FEM and FVM,” Tire Science and Technology, TSTCA, Vol. 33, No. 1, January-March
2005, pp. 38-62.

ABSTRACT: Vehicles, such as an agricultural tractor, construction vehicle, mobile machinery,

and 4-wheel drive vehicle, are often operated on unpaved ground. In many cases, the ground is
deformable; therefore, the deformation should be taken into consideration in order to assess the
off-the-road performance of a tire. Recent progress in computational mechanics enabled us to
simulate the large scale coupling problem, in which the deformation of tire structure and of sur-
rounding medium can be interactively considered. Using this technology, hydroplaning phe-
nomena and tire traction on snow have been predicted.
In this paper, the simulation methodology of tire/soil coupling problems is developed for
pneumatic tires of arbitrary tread patterns. The Finite Element Method (FEM) and the Finite
Volume Method (FVM) are used for structural and for soil-flow analysis, respectively. The soil
is modeled as an elastoplastic material with a specified yield criterion and a nonlinear elastici-
ty. The material constants are referred to measurement data, so that the cone penetration resist-
ance and the shear resistance are represented. Finally, the traction force of the tire in a culti-
vated field is predicted, and a good correlation with experiments is obtained.

KEY WORDS: FEM, FVM, soil, traction, tire

Introduction

Vehicles operated in unpaved fields require off-the-road performance, as

well as on-road performance. Numerical simulation is a powerful tool for tire
design; consequently, implicit and explicit Finite Element Methods (FEMs) [1-
5] have been used to improve tire performance on dry roads. Furthermore, a lot
of investigations regarding the interaction between a tire and deformable materi-
al have been carried out [6-8]. One of the most advanced studies is the numeri-
cal simulation solving a soil/tire interaction problem. The recent progress in
computational mechanics enables large-scale simulation considering complex
material behavior, i.e. material nonlinearity, large deformation, contact between
deformable materials. Many researchers have studied the nonlinear elastic or
plastic laws of soil using FEM. However, in most cases a tire is modeled using
1 Tire Research Department, Bridgestone Corporation, 3-1-1, Ogawahigashi-Cho Kodaira-Shi,
Tokyo 187-8531, Japan.
2 Presented at the twenty third annual conference of The Tire Society, Akron, Ohio, September 20-
21, 2004.
38
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 39

two-dimensional computational models as a rigid or an elastic wheel without

lugs. It should be pointed out that such simulation methods are insufficient in
analyzing practical tires. The three-dimensional deformation under the nonlinear
constitutive law of soil should be considered in the soil/tire interaction analyses.
This paper proposes a general purpose analysis method for practical tires that are
operated on deformable soil.
One of the typical applications of the soil/tire interaction problem is the
study of traction forces in agricultural tires. Existing studies of traction forces on
deformable soil can be classified into three groups. The first group is the direct
evaluation, which is based on the experimental method. The second is known as
the empirical or the semi-empirical method; and, the last is the numerical simu-
lation.

Direct Evaluation
In this case, testing tires are set up on a tractor and the tractor drives in test
fields measuring its drawbar pull. Though the test fields are expected to be close
to the market condition of the tires, it is quite difficult to keep soil conditions
constant and uniform because they change with the position in the fields and with
the seasons. To control soil condition easily, the indoor Single Wheel Testers
(SWT) [9-10] are made by many researchers. They evaluate traction performance
with respect to slip ratio or applied loads. Hence, the soil bin used in the SWT is
relatively small and the soil condition is well-controlled. Therefore, performance
can be measured under the stable soil condition. However, the application of
SWT is limited to small tires with small loads, such as passenger car tires. It is
difficult to apply SWT to the practical size of agricultural tires. Many tire design-
ers desire to know the soil motion under tires, which is closely related to the
design of lug geometry, as well as traction performance. Some researchers have
tried to use small devices as tracers, which transmit their positions, in order to
quantify the soil motion [11-14]. Though it is fundamentally applicable to any
kind of soil condition, it is difficult to observe the wide range of soil motion in
detail because of its dimension.

Empirical Method
The well-known empirical method [15-16], which was developed by U.S.
Army Waterways Experiment Station (WES) to assess vehicle mobility on a
“go/no go” basis, is based on some measured indexes, such as the cone index, the
rating cone index, the mobility index, etc. The cone indexes are obtained by a
cone penetrometer, and the mobility index is determined by empirical expres-
sions which consist of vehicle weight, contact area, size of grouser, power of the
engine, type of the transmission, etc. After these indexes and soil strength have
been determined, several performance parameters of a tracked vehicle, such as
drawbar pull coefficient and towed motion resistance coefficient, are determined
empirically.
40 TIRE SCIENCE & TECHNOLOGY

Semi-Empirical Method
In this method [17-22], soil strength parameters are obtained experimental-
ly, and traction performance is predicted using computational calculation in order
to analyze stress distribution. This method is introduced by Bekker to predict the
motion resistance using a pressure-sinkage equation. The pressure-sinkage equa-
tion is obtained by plate sinkage tests, and is expressed by sinkage and soil
parameters. The sinkage is calculated from the equilibrium of vertical load and
soil pressure. The motion resistance is calculated by integrating the horizontal
component of soil pressure in the contact patch. The gross traction is calculated
from the relationship between shear stress and slip displacement, which is
obtained from the result of direct shear box tests. Several expressions of the rela-
tionship are proposed by Bekker [17-19], Janosi and Hanamoto [23], Kacigin
and Guskov [24], Wong and Preston-Thomas [25]. All these expressions are writ-
ten as the function of maximum shear stress, which is determined by the Mohr-
Coulomb failure criterion.
Although the stress distribution is calculated theoretically, the geometry of a
deformed tire must be assumed. The procedure for evaluating traction force is as
follows. The deformation of a contact portion between the tire and the soil is
defined. The sinkage at an arbitrary tire surface is obtained, and then the contact
pressure can be obtained using the pressure-sinkage equation. By applying the
pressure to the shear stress equation, the shear stress at an arbitrary position can
be calculated. By integrating this shear stress on the contact patch in the hori-
zontal direction, the traction of the tire can be evaluated. The tire deformation can
be assumed either as a rigid mode without deflection or as an elastic mode with
deflection. Some commercial codes have the ability to evaluate the traction of
tires using this method [26-27].

Numerical Simulation
The performance of tires used on deformable soil is determined by the inter-
action between the soil and the tire. Though the interaction must be solved in the
numerical simulation, this kind of analysis is still one of the most difficult prob-
lems in the non-linear contact analyses. Many researchers use FEM for soil
deformation analysis only, where the tire is modeled as a rigid body, the soil is
modeled as an elastic or a viscoelastic material, and small deformation of the soil
is assumed. Using this analysis, the stress distribution in the soil can be predict-
ed and a good correlation between the prediction and the experiment is obtained.
In recent years, soil has been modeled as an elastoplastic material to evaluate its
plasticity such as permanent deformation, residual strain, and so on. Though
many applications of FEM considering tire/soil interaction have been studied
[28-33], there are still few studies considering practical tire geometry with lugs.
There are other numerical methods that can be applied to soil deformation analy-
sis, such as Distinct Element Method (DEM) [34-38]. In DEM, soil is modeled
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 41

as an assemblage of small particles where contact with adjacent particles is per-

mitted. Any particles that move following Newton’s Law of Motion and each
contact behavior between particles are analyzed. This approach is considered to
be reasonable because soil is actually composed of small particles. Hence, DEM
does not need “elements” for soil. Unlike FEM, large deformation of soil is eas-
ily analyzed. It is ascertained that the result of DEM with soil and rigid wheel is
in good agreement with experimental data using SWT [37]. One of the disad-
vantages of DEM is the computational cost, because quite a huge number of par-
ticles are required for soil modeling. That is why DEM is mainly applied to two-
dimensional analysis and is hardly ever applied to three-dimensional analysis.

Classical Evaluation of Traction in Soil

It is well known that the traction of a tire can be divided into two parts as
follows.

NT = GT − MR (1)

where NT is the net traction; GT is the gross traction; and MR is the motion resist-
ance. Net traction is a thrust force to bring a tire forward, and it acts as drawbar
pull. In the following sections, typical aspects of MR and GT are briefly
described as discussed in the semi-empirical method.

Motion Resistance
Motion resistance is the force that is caused by compressing the forward soil
or pushing soil aside, as shown in Fig. 1.
The basic property of motion resistance can be explained by Bekker’s
model. In this model, the motion resistance can be written as follows:
θ0 z0
⎛k ⎞
MR = b P r sin θ dθ = b ⎜ c + kφ ⎟ z n dz
∫0
⎝b∫
0
⎠ (2)

where P is the normal pressure; r is the radius of the rigid wheel (shown in Fig.
2); b is the width of the rigid wheel; kc, kφ and n are the soil parameters. These
parameters are obtained from the plate sinking test using at least two plates. It is
assumed that the terrain reaction force at the contact patch is purely radial and is
equal to the normal pressure beneath the horizontal plate at the same depth in a
pressure-sinkage test. The motion resistance of a rigid wheel, which acts as the
compaction resistance, is due to the vertical work done in making a rut of depth
z0. Equation (2) shows that the more a tire sinks, the more motion resistance is
42 TIRE SCIENCE & TECHNOLOGY

generated. When a tire drives on softer soil, larger motion resistance is applied to
the tire. It should be noted that the sinkage z0 at the bottom of the wheel is
obtained from the equilibrium condition of the vertical force.

FIG. 1 — The motion resistance, (upper) a wheel without lugs, (lower) a wheel with lugs.

In the case of pneumatic tires, motion resistance depends on the wheel’s

deformation in operation. If the ground is sufficiently soft and the sum of the
inflation pressure pi and the pressure produced by the stiffness of the carcass pc
is greater than the maximum pressure that the terrain can support at the lowest
point of the tire, the tire maintains the round shape just like a rigid wheel. This
is usually referred to as the rigid mode. In the case of rigid mode, motion resist-
ance can be calculated as mentioned above. On the other hand, if the terrain is
firm and the tire inflation pressure is low, a portion of the tire’s circumference
will be flattened. This is referred to as the elastic mode. In this case, the pressure
on the flattened circumference becomes the sum of pi and pc, and the sinkage at
this portion can be estimated using Bekker’s pressure-sinkage equation. The rest
of the derivation is basically the same as in the case of the rigid mode. The MR
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 43

can be evaluated by integrating the horizontal component of the soil pressure at

the contact patch.

FIG. 2 — Basic diagram for the derivation of the motion resistance.

Gross Traction
In the case of a tire without lugs, the gross traction is generated by the fric-
tional shear displacement developed along the contact patch of the tire and is
determined from the slip velocity. In the case with lugs, the effect of the lugs
should be added. The gross traction can be written as follows.

GT = GTFR + GTSL + GTPF (3)

where GTFR is the frictional traction; GTSL is a traction produced by slip dis-
placement across the lug tips; and GTPF is a traction produced by passive failure
of soil as shown in Fig. 3. GTFR is generated on the contact surface between the
tire and the soil, and is due to the external friction of the soil. On the other hand,
GTSL is a traction force due to the internal friction force of the soil. Though GTSL
and GTFR are different forces, these forces are included as a shear force due to
slip displacement of the soil in the semi-empirical method.
For a tire without lugs, the gross traction is produced as a friction force
between the tire surface and the soil. The slip velocity coincides with the tan gen-
tial velocity relative to the terrain. The magnitude of the slip velocity of a point
on the rigid wheel can be expressed by the angle and the slip. The shear dis-
placement is obtained by integrating the velocity from the leading edge of the
44 TIRE SCIENCE & TECHNOLOGY

FIG. 3 — The gross traction, (upper) Shear stress due to frictional slip on a wheel without lugs, (mid-
dle) Shear stress along with lug tips, (lower) Gross traction due to passive failure behind of lugs.

contact patch to the current position. Several stress-displacement relationships

have been proposed as mentioned above. For example, the relationship proposed
by Janosi and Hanamoto [23] is written as follows:
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 45

τ (θ ) = τ max (1 − e − j / K ) = [c + p(θ ) tan φ ](1 − e − j / K ) (4)

where τ(θ) is the shear stress; τmax is the maximum shear stress determined by
the Mohr-Coulomb failure criterion; j is the shear displacement; K is referred to
as the shear deformation modulus; c is the cohesion; p(θ) is the normal pressure
at angle θ; and φ is the internal friction angle, as shown in Fig. 4. The normal
pressure along the rigid wheel can be estimated by a variety of methods, includ-
ing the simplified method proposed by Bekker, as mentioned above. By integrat-
ing the horizontal component of the tangential stress over the entire contact
patch, the total horizontal component of the gross traction, i.e. GTFR, can be
determined. The gross traction of a wheel of an elastic mode can be basically
determined as well as a rigid mode.

FIG. 4 — Mohr-Coulomb plasticity model.

The effect of lugs is not only the frictional force between the soil and the tire,
but also the soil strength between lugs. The effect can be evaluated as the state of
Rankine passive state, where the stress state prior to plastic flow caused by com-
pression can be determined. The theory of passive earth pressure has found
applications in the prediction of the force acting on a soil-cutting blade and in the
estimation of the traction effort developed by the lug of a wheel. The gross trac-
tion developed by the lug is calculated by integrating the passive earth pressure
over the cutting depth.
46 TIRE SCIENCE & TECHNOLOGY

hl hl

GTPF =
∫σ
0
p dz =
∫ (γ
0
s zNφ )
+ qNφ + 2c Nφ dz (5)

where σp is the earth pressure; hλ is the height of a lug; γs is the density; q is the
pressure applied to the soil surface; Nφ is the soil parameter related to the inter-
nal friction angle, and is referred to as the flow value.
In general, the lugs behave in one of two ways. If the spacing between the
lugs is too small, the space between them may be filled up with the soil, and
shearing will occur across the lug tips. Under these conditions, the major effect
of the lugs would be the increase of the effective diameter of the wheel, in which
the gross traction can be estimated by the shear displacement along the contact
patch, i.e. GTSL would be produced along the surface at the effective diameter.
This GTSL is usually evaluated by adding to the diameter of the wheel by the
height of the lug. It should be noted again that the GTSL is produced by the inter-
nal friction, not by external friction that produces frictional traction GTFR. On the
other hand, if the spacing between the lugs is large, the soil would fail as shown
by Rankine passive failure, and Eq. (5).
The classical evaluation is based on many assumptions, and its applicability
is quite limited. But its fundamental approach for evaluating traction force is
quite reasonable. It is desirable to apply and extend this classical approach and
apply to practical problems. In this paper, we not only focus on the soil failure,
but also the shear stress generated by slip phenomena (except frictional forces).
Numerical simulations are used to predict these complex deformations and inter-
actions. The aim of this study is to apply this simulation technology to the prac-
tical tread pattern design, and to determine the most effective tread pattern design
in the realm of traction performance.

Current Modeling

As mentioned above, soil/tire interaction is the most important problem in

predicting traction performance of a tire. Most research assumes the deformation
shape as the basis of magnitude relationship between a tire and soil pressures.
The assumed deformation shape is simplified in order to integrate the soil pres-
sure mathematically. In this proposal, the deformation is determined completely
by the contact analysis between a deformable tire and deformable soil. The tire
is modeled three-dimensionally with lugs of arbitrary shape as a practical tire.
The soil/tire interaction is computed by using the “general coupling func-
tion” in the commercial code MSC.Dytran [39], which has the capability to auto-
matically calculate the interaction between Lagrangian meshes and Eulerian
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 47

meshes (shown in Fig. 5). Structure is modeled by FEM using Lagrangian mesh.
Soil is modeled by the Finite Volume Method (FVM) using Eulerian mesh. As
mentioned above, large strain analysis would be required for this interaction due
to the existence of lugs. Using Eulerian elements for the soil, the large deforma-
tion and the large strain in the soil can be computed without severe distortion,
collapse or reversion, which often become problematic if Lagrangian elements
are used.

FIG. 5 — General coupling.

In the existing analyses of soil/tire interaction, an elastic or a viscoelastic

model for soil is often used. However, these can not be applied to the deforma-
tion with residual strain, like ruts, though the residual strain may be important for
the analysis of a traveling tire. Furthermore, the deformation of the tire and soil
are deeply related to each other and are subject to the stress and strain distribu-
tion in the contact patch considering plasticity. Thus, the elastoplastic model used
to simulate the permanent deformation of soil is often used in recent studies.
Various elastoplastic material models have been developed for geomaterials
[40-52]. Among them, the Mohr-Coulomb plasticity model is used in this study.
The Mohr-Coulomb failure criterion is well studied in terramechanics, and good
correlations with experimental results have been confirmed as the criterion is
often used in the semi-empirical method. The Mohr-Coulomb model can express
the most fundamental characteristic that soil becomes firm when it is com-
pressed. This characteristic can be observed in Fig. 4.
One of the most important problems is how to determine the soil parameters.
There are two important plastic parameters in the Mohr-Coulomb plasticity
model: the cohesion c, and the internal friction angle φ . The triaxial or shear box
apparatus is usually employed to examine these parameters. Hence, these appa-
ratuses are not suitable for in-situ measurements. The cone penetrometer with a
48 TIRE SCIENCE & TECHNOLOGY

shear ring, which can measure both cone penetration resistance and shear resist-
ance conveniently, is used in this study. The results obtained by the cone pen-
etrometer are used to identify the soil parameters.

Cone Penetration Test

The cone penetrometer consists of a cone, which is inserted into soil, and a
load cell to measure the penetration resistance. The cone has 30 degrees of top
angle and 2×10-4 m2 sectional area. A cone model of the same geometry is also
used in the numerical simulation. As shown in Fig. 6, a quarter soil region is used

FIG. 6 — A numerical simulation model for cone penetration test (Quarter model).
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 49

to consider symmetry. In the numerical simulation, soil parameters are deter-

mined to fit the experimental results. The fitted results obtained by the simula-
tion and the two experimental results are shown in Fig. 7. The experimental
results have considerable variation because soil property differs depending on
testing points, cultivation, roots of plants, etc. The simulated result shows that the
resistance increases along with the penetration depth, as shown in the experi-
ments. The simulation result saturates over 0.1 m depth because the soil materi-
al is uniform unlike the experimental results. The simulated result appears to be
poor in the range of 0.05 m. Even though this might be a serious problem when
the soil deformation in this range is of interest, it seems to be applicable to the
soil and tire interaction analyses. However, it is quite difficult to expect any more
improvements on the simulated curve for the cone penetration resistance. It is
noted that the same phenomenon can be observed by using another commercial
software [53]. Some new soil modeling methods are required for further
improvement on the cone penetration simulation.

FIG. 7 — Comparison of cone penetration resistance between experimental and simulation results.

Direct Shear Box Test

For convenience, the shear ring is used in the in situ soil tests instead of the
shear box apparatus. The direct shear box test is introduced in the simulation in
50 TIRE SCIENCE & TECHNOLOGY

order to avoid the instability of computations due to the thin shearing plate
equipped on the shear ring, and it is often simulated as a verification of FEA [54].
The direct shear box apparatus uses two open boxes with the top box moved in
the lateral direction relative to the bottom box. Soil is loaded by normal pressure
through the top box. The relationship between the normal pressure and the later-
al force per unit area produced by sliding the top box is evaluated. Failure of the
soil occurs in the thin band referred to as the shear zone. The top box is modeled
in the same way as the apparatus and the bottom box is modeled as a soil bin.
Figure 8 shows the equivalent shear strain in the soil, and the concentration in the
narrow band between the top and bottom boxes. Figure 9 shows the comparison
between the numerical simulation and the experimental results, with good agree-
ment between them.

FIG. 8 — Simulation model for shear box test.

Numerical Evaluation of Gross Traction and Motion Resistance

In this numerical simulation, the traction components in Eq. (1) can also be
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 51

FIG. 9 — Comparison of shear resistance between numerical simulation and experimental data.

evaluated. As mentioned in the semi-empirical method, the gross traction is the

force to bring a tire forward, and the motion resistance is the force to interrupt
the driving motion. Basically, the same approaches with semi-empirical method
are introduced using the interaction stress. The stress can be obtained on the con-
tacting surface with soil referred to as coupling surface in the general coupling
function of MSC.Dytran.
The force acting on an element contacting with soil can be computed as fol-
lows:
⎛ 3 ⎞
Ti k =
∫∑ ⎜
⎜
Sk ⎝ j =1
σ ijk n kj ⎟ dS k
⎟
⎠
(6)

where superscript k denotes the target element number; Ti is the force component
in the i-th coordinate axis direction; σij is the interaction stress tensor; nj is the
unit normal vector components; S is the total area of the element. It is noted that
the stress tensor σij is the interaction stress subjected to the tire from the soil, and
that the unit normal vector nj can be computed from the deformed configuration
of the element. If the tire is driving along the X1 coordinate axis direction, the
force of interest would be T1. The judgment criterion whether the force of inter-
est is the gross traction or the motion resistance is only the sign of the force. For
example, T1 would act as gross traction if it is positive, and it would act as motion
resistance if it is negative. Summing it up on all elements contacting with the soil
separately according to the sign of each component of the force, the gross trac-
52 TIRE SCIENCE & TECHNOLOGY

tion and motion resistance can be obtained as follows:

N ⎛ 3 +⎞
GT = ∑∫ ∑ ⎜
k =1 S k ⎝ i =1
σ ik1nik ⎟ dS
⎠
k (7)

N ⎛ 3 −⎞
MR = ∑∫ ∑ ⎜
k =1 S k ⎝ i =1
σ ik1nik ⎟ dS
⎠
k (8)

where N is the total number of elements contacting with the soil. The brackets “<
>” with superscript ‘+’ or ‘–’ form the following operators.

+ ⎧ A, if A ≥ 0 (9)
A =⎨
⎩ 0, if A < 0

− ⎧ 0, if A ≥ 0
A =⎨ (10)
⎩ A, if A < 0

Soil and Tire Interaction Analysis

Predictability of Traction Performance in Different Soil Conditions

First of all, two different soil models are prepared. Hence, the relationship
of soil to soil pressure, shown in the Young’s modulus is also taken into account.
The tire model used in this simulation is shown in Fig. 10. The tire size is
540/65R30, inflation pressure is 240 kPa, and load is 32.86 kN. The number of
Lagrangian elements of the tire model is about 41,000 including coupling ele-
ments that cover the tread surface. The soil bin of 0.326 m depth is modeled by
about 84,000 Eulerian elements. The friction coefficient between the tire and soil
surfaces is ignored.
Figures 11 and 12 show the ideal soil models A and B, respectively. Their
cone penetration resistances obtained by computation are different, as shown in
each figure. It is observed that soil A is much harder than soil B. The rut shapes
created when the tire is driven at nearly 100% slip ratio on each soil model are
also shown in each figure. It can be observed that the rut shapes generated by the
tire are also different and that the tire sinkage in soil A is smaller than in soil B.
The tire sinkage is nearly 24 mm and nearly 110 mm in the case of soil A and
soil B, respectively. It is considered that the result is in good qualitative agree-
ment with common phenomena.
OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 53

FIG. 10 — Tire model.

FIG. 11 — Soil model A, rut shape (left) and CPR (right).

54 TIRE SCIENCE & TECHNOLOGY

FIG. 12 — Soil model B, rut shape (left) and CPR (right).

Soil Motion and Traction Components

Although the motions of soil under rigid wheels without lugs are often stud-
ied, it is well known that the effect of the lugs on traction performance is very
important in practical tires. Soil model A is used in the following simulations.
Figure 13 shows the flow diagram of the soil where the tire is driving at the
0% slip ratio. It is observed that the soil beneath the leading edge of the tire

FIG. 13 — Flow distribution in soil at 0% slip ratio.

 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 55

moves forward and downward, and the soil is acting as motion resistance to the
tire. This condition is quite similar to the case of a rigid or an elastic wheel with-
out lugs. The right figure in Fig. 14 shows the motion resistance distribution per
unit area at this state. The motion resistance arises mainly on the top surfaces of
the lugs. On the other hand, the left figure in Fig. 14 shows the distribution of the
gross traction per unit area. The gross traction arises mainly on the lug wall at the
leading edge, and seems to be less than the motion resistance at 0% slip ratio.

FIG. 14 — Gross traction (left) and motion resistance (right) distributions at 0% slip ratio.

Figure 15 shows the flow diagram when the tire is driving at 30% slip ratio.
Most of the soil beneath the tire flows backward, which is contrary to the case of
0% slip ratio. This means that the gross traction is dominant, and the motion
resistance is very small. In this case, the motion of the soil is limited to the shal-
low part of the soil region aligned to the top surface of the tire. It is considered
that pushing the soil backward between the lugs produces gross traction at 30%
slip ratio. The same tendency can be observed from Fig. 16. In the case of 30%
slip ratio, it can be seen that the acting range of motion resistance on the lug wall
at the leading edge becomes smaller and the gross traction becomes larger.
Figure 17 shows the net traction history in our analyses. The net traction at
0% slip ratio becomes negative because the motion resistance becomes greater
than the gross traction. Driving at 0% slip ratio with negative net traction is the
ideal situation for the effect of motion resistance to be magnified. It is one of the
advantages of numerical simulation that any condition can be simulated.
56 TIRE SCIENCE & TECHNOLOGY

FIG. 15 — Flow distribution in soil at 30% slip ratio.

FIG. 16 — Gross traction (left) and motion resistance (right) distributions at 30% slip ratio.

Verification of Predictability
To verify the accuracy of the prediction, four kinds of tires, such as A to D
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 57

FIG. 17 — Time history of traction components.

(as shown in Table 1), are made and the direct traction tests are carried out. The
size of the tires is 540/65R30, the inflation pressure is 240 kPa, and the load is
32.86 kN. The test fields are covered with grass before cultivation and have a rel-
atively hard soil. These tires have a different number of pitch and lug bar angles
and have the same lug section geometry, as shown in Fig. 18, in order to evalu-
ate the effect of the number of pitch and the bar angles.

FIG. 18 — Four tread pattern models used in the verification of predictability.

The experimental and predicted results are summarized in Table 1. The net
traction is measured twice for each tire in the same field. The result for Tire C in
the second test is that it seems to have too much traction. Since the soil condition
at that time might be harder than those in other tests, this result should be con-
sidered as an irregular datum.
58 TIRE SCIENCE & TECHNOLOGY

TABLE 1- Comparison of traction indexes between experimental and predicted results.

By comparing Tire A and B, or C and D, which have different lug bar angles,
it is confirmed that the smaller bar angle yields better traction. The simulation
results show that Tire A has better traction than B, and Tire C also has better trac-
tion than D. Next, by comparing Tire A and C, or B and D, which have the same
lug bar angle and a different number of pitch, it is shown that the tires with a
smaller number of pitch have better traction. The order in terms of better traction
is completely the same between the predictions and the experimental results.
This demonstrates that the prediction is in good qualitative agreement with the
experiments. It can be concluded that this simulation methodology is applicable
to evaluate the tread pattern difference of practical tires.

Summary and Conclusions

A new procedure to predict soil/tire interaction phenomena is established. A

tire is modeled by the Lagrangian formulation using FEM, and soil is modeled
by the Eulerian formulation using FVM. The interaction between the soil and the
tire is automatically calculated by the “general coupling function” through the
coupling surface that covers the tread surface. The soil is modeled as an elasto-
plastic material based on the Mohr-Coulomb yield criteria, which is described by
the cohesion, the internal friction angle of the soil and the soil pressure. Basic
soil behavior can be predicted using this simulation technology. In order to deter-
mine the soil parameters, cone penetration tests and the direct shear box tests are
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 59

carried out. However, the soil model used in the comparison with the cone pen-
etration tests in the test field may be insufficient to reproduce the complicated
distribution of the actual cone penetration resistance. Some new soil modeling
methods are required for further improvement on the cone penetration simula-
tion.
The motion resistance produced by a tire is mainly governed by compaction
of the soil in front of the tire. The gross traction is governed by the shear slip
along the effective diameter of the tire and the passive failure pressure in front of
the lugs. The effects of these contributions are evaluated. Since the aim of this
study is to apply this simulation technology to practical pattern designs of tires,
the friction force between the tire and the soil surface is ignored.
Finally, the accuracy of this numerical simulation technology is confirmed
by the comparison with the results obtained by direct traction tests. The effect of
the number of pitch and the bar angle of lugs can be predicted qualitatively.
Further study of numerical simulation is required. For example, the friction
between a tire and soil should be considered to improve the accuracy of the sim-
ulation. We hope that this technology will serve to improve tire performance and
make tire design more efficient.

References

[1] Nakajima, Y., Kamegawa, T., and Abe, A., “New Tire Design Procedure Based on Optimization
Technique”, SAE Paper, 960997.
[2] Nakajima, Y., Kamegawa, T., and Abe, A., “Theory of Optimum Tire Contour and Its
Application,” Tire Science and Technology, TSTCA, Vol. 24, No. 3, 1996, pp. 184-203.
[3] Abe, A., Kamegawa, T., and Nakajima, Y., “Optimum Young’s Modulus Distribution in Tire
Design,” Tire Science and Technology, TSTCA, Vol. 24, No.3, 1996, pp. 204-219.
[4] Nakajima, Y., Kadowaki, H., Kamegawa, T., and Ueno, K., “Application of a Neural Network
for the Optimization of Tire Design,” Tire Science and Technology, TSTCA, Vol. 27, No. 2,
1996, pp. 62-83.
[5] Koide, M., Heguri, H., Kamegawa, T., Nakajima, Y., and Ogawa, H., “Optimization for
Motorcycle Tire Using Explicit FEM,” Tire Science and Technology, TSTCA, Vol. 29, No. 4,
2001, pp. 230-243.
[6] Seta, E., Nakajima, Y., Kamegawa, T., and Ogawa, H., “Hydroplaning Analysis by FEM and
FVM: Effect of Tire Rolling and Tire Pattern on Hydroplaning,” Tire Science and Technology,
TSTCA, Vol. 28, No. 3, 2000, pp. 140-156.
[7] Okano, T., and Koishi, M., “A New Computational Procedure to Predict Transient Hydroplaning
Performance of a Tire,” Tire Science and Technology, TSTCA, Vol. 29, No. 1, 2001, pp. 2-22.
[8] Seta, E., Kamegawa, T., Nakajima, Y., “Prediction of Snow/Tire Interaction Using Explicit FEM
and FVM,” Tire Science and Technology, TSTCA, Vol. 31, No. 3, 2003, pp. 173-188.
[9] Pope, R. G., “The Effect of Wheel Speed on Rolling Resistance,” Journal of Terramechanics,
Vol. 8, No. 1, 1971, pp. 51-58.
[10] Wang, Z., Reece, A. R., “The Performance of Free Rolling Rigid and Flexible Wheels on Sand,”
Journal of Terramechanics, Vol. 21, No. 4, 1984, pp. 347-360.
60 TIRE SCIENCE & TECHNOLOGY

[11] Wong, J. Y., “Behavior of Soil Beneath Rigid Wheels,” Journal of Agricultural Engineering
Research, Vol. 12, No. 4, 1967, pp. 257-269.
[12] Windish, E. J., Yong, R. N., “The Determination of Soil Strain-Rate Behavior Beneath a Moving
Wheel,” Journal of Terramechanics, Vol. 7, No. 1, 1970, pp. 55-67.
[13] Ueno, M., Hashiguchi, K., Nohse, Y., Kokubu, T., Uchiyama, K., Shikanai, T., “A Development
of Analyzing System for Strain and Stress of Soil Under the Wheel,” Proceedings of 10th
International Conference of ISTVS, 1990, pp. 265-276.
[14] Hashiguchi, K., Okayasu, T., Ueno, M., Shikanai, T., “Image Processing On-line Measurement
for Soil Displacement,” Journal of the Japanese Society of Agricultural Machinery, Vol. 60, No.
6, 1998, pp. 11-18, (in Japanese).
[15] Crenshaw, B. M., “Soil/Wheel Interaction at High Speed,” Journal of Terramechanics, Vol. 8,
No. 3, 1972, pp. 71-88.
[16] Wong, J. Y., Theory of Ground Vehicles, 3rd Edition, John Wiley & Sons, Inc., 2001.
[17] Bekker, M. G., Theory of Land Locomotion, The University of Michigan Press, 1956.
[18] Bekker, M. G., Off the Road Locomotion, The University of Michigan Press, Ann Arbor, 1960.
[19] Bekker, M. G., Introduction to Terrain-Vehicle Systems, The University of Michigan Press, Ann
Arbor, MI, 1969.
[20] Wong, J. Y, Reece, A. R., “Prediction of Rigid Wheel Performance Based on the Analysis of
Soil-Wheel Stresses, Part I. Performance of Driven Rigid Wheel,” Journal of Terramechanics,
Vol. 4, No. 1, 1967, pp. 81-98.
[21] Wong, J. Y., Reece, A. R., “Prediction of Rigid Wheel Performance Based on the Analysis of
Soil-Wheel Stresses, Part II. Performance of Towed Rigid Wheels,” Journal of Terramechanics,
Vol. 4, No. 2, 1967, pp. 7-25.
[22] Pope, R. G., “The Effect of Sinkage Rate on Pressure Sinkage Relationships and Rolling
Resistance in Real and Artificial Clays,” Journal of Terramechanics, Vol. 6, No. 4, 1969, pp. 31-
38.
[23] Janosi, Z., Hanamoto, B., “Analytical Determination of Drawbar Pull as a Function of Slip for
Tracked Vehicles in Deformable Soils,” Proceedings of the 1st International Conference on
Terrain-Vehicle Systems, Turin, 1961.
[24] Kacigin, V. V., Guskov, V. V., “The Basis of Tractor Performance Theory,” Journal of
Terramechanics, Vol. 5, No. 3, 1968, pp. 43-66.
[25] Wong, J. Y., Preston-Thomas, J., “On the Characterization of the Shear Stress-Displacement
Relationship of Terrain,” Journal of Terramechanics, Vol. 19, No. 4, 1983, pp. 225-234.
[26] AESCO (Automotive Engineering Software & Consulting), ORSIS (Off Road Systems
Interactive Simulation), http://www.english.orsis.de/index.html.
[27] AESCO (Automotive Engineering Software & Consulting), AS2TM, http://www.tire-soil.de/.
[28] Hiroma, T., Abe, K., Kataoka, T., Ota, Y., “Finite Element Analysis for Tractive Performance of
a Rigid Wheel (Part 1), Analysis Using Loading and Unloading Models of Soil,” Journal of the
Japanese Society of Agricultural Machinery, Vol. 61, No. 1, 1999, pp. 115-121, (in Japanese).
[29] Hiroma, T., Abe, K., Ichinohe, Y., Kataoka, T., Ota, Y., “Finite Element Analysis for Tractive
Performance of a Rigid Wheel (Part 2), Tractive Performance and Stress Distribution in Soil,”
Journal of the Japanese Society of Agricultural Machinery, Vol. 61, No. 1, 1999, pp. 123-129,
(in Japanese).
[30] Hiroma, T., Ota, Y., Kataoka, T., “Analysis of the Soil Deformation Beneath a Wheel by Finite
Element Method (Part 3), Analysis Considering the Viscoelastic Properties of Soil,” Journal of
the Japanese Society of Agricultural Machinery, Vol. 56, No. 6, 1994, pp. 3-10, (in Japanese).
[31] Abu-Hamdeh, N. H., Reeder, R. C., “Measuring and Predicting Stress Distribution under
Tractive Devices in Undisturbed Soils,” Biosystem Engineering, Vol. 85, No. 4, 2003, pp. 493-
502.
 OIDA ET AL. ON TIRE/SOIL INTERACTION ANALYSIS 61

[32] Regli, G., Handke, A., Butikofer, M., “Material Laws as a Basis for Simulation Models for the
Calculation of Wheel-Soil Interaction Examination Using the Finite Element Method,” Journal
of Terramechanics, Vol. 30, No. 3, 1993, pp. 165-179.
[33] Liu, C. H., Wong, J. Y., “Numerical Simulations of Tire-Soil Interaction Based on Critical State
Soil Mechanics,” Journal of Terramechanics, Vol. 33. No. 5, 1996, pp. 209-221.
[34] Momotsu, M., Oida, A., Yamazaki, M., “Simulation of Interaction between Soil and Rotary
Blade by Modified Distinct Element Method,” Proc. of 7th European ISTVS Conf. Ferrara,
1997, pp. 572-579.
[35] Oida, A., Schwanghart, H., Ohkubo, S., Yamazaki, M., “Simulation of Soil Deformation Under
a Track Shoe by the DEM,” Proc. of 7th European ISTVS Conf. Ferrara, 1997, pp. 155-162.
[36] Tanaka, H., Momotsu, M., Oida, A., Yamazaki, M., “Simulation of Soil Deformation and
Resistance at Bar Penetration by the Distinct Element Method,” Proc. of 12th Int. Conf. of
ISTVS, Beijing, 1996, pp. 21-28.
[37] Fujii, H., Oida, A., Nakashima, H., Miyasaka, J., Momozu, M., Kanamori, H., Yokoyama, T.,
“Analysis of Interaction between Lunar Terrain and Treaded Wheel by Distinct Element
Method,” Proc. 14th Int. Conf. ISTVS, Vicksburg, MS, USA, 2002, (CD-ROM).
[38] Nakashima, H., Oida, A., “Algorithm and Implementation of Soil-Tire Contact Analysis Code
Based on Dynamic FE-DE Method,” Proc. 14th Int. Conf. ISTVS, Vicksburg, MS, USA, 2002,
(CD-ROM).
[39] MSC.Dytran User Manual, MSC.Software Corporation, 2003.
[40] Duncan, J. M., Chang, C.-Y., “Non-linear Analysis of Stress and Strain in Soils,” Journal of the
Soil Mechanics and Foundation Division, Proceedings of the ASCE, Vol. 96, No. SM5, 1970,
pp. 1629-1653
[41] Wood, D. M., Soil Behavior and Critical State Soil Mechanics, Cambridge University Press,
Cambridge, NY, 1990.
[42] Shoop, S., Darnell, I., Kestler, K., “Analysis of Tire Models for Rolling on a Deformable
Subsurface,” Tire Science and Technology, TSTCA, Vol. 30, No. 3, 2002, pp. 180-197.
[43] Shoop, S. A., “Finite Element Modeling of Tire-Terrain Interaction,” Technical Report,
ERDC/CRREL TR-01-16, 2001.
[44] Meschke, G., Liu, C., Mang, H. A., “Large Strain Finite-Element Analysis of Snow,” Journal of
Engineering Mechanics, Vol. 122, No. 7, 1996, pp. 591-602.
[45] Meschke, G., “A New Viscoplastic Model for Snow at Finite Strains,” 4th International
Conference on Computational Plasticity, Barcelona, April 3-6, Pineridge Press, 1995, pp. 2295-
2306.
[46] DiMaggio, F. L., Sandler, I. S., “Material Model for Granular Soils,” Journal of the Engineering
Mechanics Division, Proceedings of the ASCE, Vol. 97, No. EM3, 1971, pp. 935-950.
[47] Hashiguchi, K., “Subloading Surface Model in Unconventional Plasticity,” Int. J. Solids
Structures, Vol. 25, No. 8, 1989, pp. 917-945.
[48] Hashiguchi, K., Chen, Z. P., “Elastoplastic Constitutive Equation of Soils with the Subloading
Surface and the Rotational Hardening,” International Journal for Numerical and Analytical
Methods in Geomechanics, Vol. 22, 1998, pp. 197-227.
[49] Hashiguchi, K., Okayasu, T., Tsutsumi, S., Ueno, M., “Development of FEM Program as a Soil-
Water Coupled Finite Deformation Introducing the Extended Subloading Surface Model,”
Journal of the Japanese Society of Agricultural Machinery, Vol. 62, No. 1, 2000, pp. 64-72, (in
Japanese).
[50] Okayasu, T., Hashiguchi, K., Tsutsumi, S., Ozaki, S., Ueno, M., “Elastoplastic Finite Element
Analysis for Ground Deformation Behavior Subjected to Cyclic Loading,” Journal of the
Japanese Society of Agricultural Machinery, Vol. 63, No. 3, 2001, pp. 67-73, (in Japanese).
[51] Hashiguchi, K., Tsutsumi, S., Okayasu, T., Saitoh, K., “Elastoplastic Constitutive Equation with
Tangential Stress Rate Effect and Its Application to Soils,” Journal of the Japanese Society of
Agricultural Machinery, Vol. 62, No. 3, 2000, pp. 68-75, (in Japanese).
62 TIRE SCIENCE & TECHNOLOGY

[52] Hashiguchi, K., Tsutsumi, S., Okayasu, T., Saitoh, K., “Consideration of Shear Band Formation
by Elastoplastic Constitutive Equation of Soils with Tangential Stress Rate Effect,” Journal of
the Japanese Society of Agricultural Machinery, Vol. 62, No. 4, 2000, pp. 71-80, (in Japanese).
[53] Susila, E., Hryciw, R. D., “Large Displacement FEM Modeling of the Cone Penetration Test
(CPT) in Normally Consolidated Sand,” Int. J. Numer. Anal. Meth. Geomech., Vol. 27, 2003, pp.
585-602.
[54] Tanaka, T., Abe, T., “Elasto-Plastic Finite Element Analysis of Direct Shear Test,” Transactions
of the Japanese Society of Irrigation, Drainage and Reclamation Engineering, Vol. 70, No. 2,
2002, pp. 137-143, (in Japanese).