This article was downloaded by: [University of California Davis]

On: 15 August 2014, At: 00:23

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Vehicle System Dynamics: International

Journal of Vehicle Mechanics and
Mobility
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/nvsd20

The PAC2002 advanced handling tire

model
a a
E. Kuiper & J. J. M. Van Oosten
a
MSC Software Benelux B.V. , Groningenweg 6h, 2803 PV, Gouda,
The Netherlands
Published online: 11 Mar 2008.

To cite this article: E. Kuiper & J. J. M. Van Oosten (2007) The PAC2002 advanced handling tire
model, Vehicle System Dynamics: International Journal of Vehicle Mechanics and Mobility, 45:S1,
153-167, DOI: 10.1080/00423110701773893

To link to this article: http://dx.doi.org/10.1080/00423110701773893

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
 Vehicle System Dynamics
Vol. 45, Supplement, 2007, 153–167

The PAC2002 advanced handling tire model

E. KUIPER and J. J. M. VAN OOSTEN*
MSC Software Benelux B.V., Groningenweg 6h, 2803 PV Gouda, The Netherlands
Downloaded by [University of California Davis] at 00:23 15 August 2014

The history of ‘Magic Formula’ tire models is shortly described and the PAC2002 tire model is the latest
generation of this type of tire models. PAC2002 is a semi-empirical tire model because it has both the
empirical (mathematical formulas) and physical modeling components. The PAC2002 tire model can
describe the behavior of tires traveling over relatively smooth road surfaces and its dynamical behavior
is valid for frequencies up to 12 Hz. The structure of the model is explained and it is shown what the
influences of the different Magic Formula-factors have on the characteristic. The classic tire model
input quantities are extended with turn-slip, which is one of two components of tire spin. The functional
expressions for steady-state cornering and braking/driving are given. In addition to the stretched-string
(first-order relaxation effects) model to describe tire transient behavior, the newly added alternative
contact-mass model is explained. The new formula set for tire spin and parking is given, which extends
validity of the PAC2002 especially for low speed. Some typical advanced handling applications are given
for which the tire model can be used. It is described what measurement program is needed to obtain
the tire model parameters. A reduced measurement program can be used that is more cost-effective
than and gives comparable accuracy to the full program. It is shown that the additions to PAC2002
(contact-mass transient model, tire spin and parking) result in considerable improvements compared
to Magic Formula tire models with the traditional ‘first-order relaxation’ transient tire modeling. Next
to the common applications of a Magic Formula tire model, the PAC2002 can be used for analyzing
parking behavior, low- and zero-speed applications, changing friction properties as well as in more
dynamic analysis such as ABS braking.

Keywords: Tires; Vehicle dynamics; Multi-body simulation

1. Introduction

The tire behavior plays a crucial role in vehicle dynamics since almost all forces and moments
acting on a vehicle are transferred by the tires. Therefore, the vehicle behavior strongly depends
on the characteristics of the tire and tire–road contact. For vehicle dynamics analysis and sim-
ulation, different types of mathematical models have been developed during the last 60 years,
each type for a specific application. These types are different in complexity and accuracy
depending on the area of application.
Tire models can be classified by modeling approach and application area:
(1) Simple tire models. The vertical behavior consists of a linear or non-linear spring-damper
model. The horizontal tire forces are accounted for by linear relationships between slip
and resulting forces and combined slip is not taken into account. Simple tire models are

*Corresponding author. Email: [email protected]

Vehicle System Dynamics

ISSN 0042-3114 print/ISSN 1744-5159 online © 2007 Taylor & Francis
http://www.informaworld.com
DOI: 10.1080/00423110701773893
 154 E. Kuiper and J. J. M. van Oosten

used for static and quasi-static vehicle dynamics analysis and in the design of vehicle
control systems.
(2) Empirical models. This kind of tire model is based on non-linear mathematical approx-
imations of tire forces and moments or interpolation of test data. Empirical tire models
require full-scale tire measurements, data processing and parameter identification. These
tire models are in general very accurate, and are used for vehicle dynamics analysis in a
broad sense, ranging from non-linear handling to ride simulations.
(3) Physical models. These models describe the kinematics and dynamics of the tire contact
patch in detail. The parameters that describe the tire behavior are of physical and geomet-
rical nature and full-scale tire measurements are not necessary. The application of physical
tire models is unlimited: they are used for quasi-static behavior, non-linear handling, ride,
Downloaded by [University of California Davis] at 00:23 15 August 2014

comfort, durability, NVH and acoustics. Physical tire models can be become very complex
and are often tailored for a specific area of application.
(4) Finite-element tire models. The tire is modeled by a detailed finite-element mesh for the
complete tire structure including the compressed air. Almost any physical phenomenon
can be taken into account and the resulting computational effort is very high. Their use
is mostly restricted to detailed structural analysis with high non-linear deformations,
hydroplaning and acoustic analysis. For vehicle dynamics analysis, the computational
effort is yet too high for application on a regular basis.

The PAC2002 tire model is based on the so-called ‘Magic Formula’. The Magic Formula was
developed during the second half of the 1980s by the Delft University of Technology in
cooporation with Volvo Car Corp. [1–3]. Over the years, the Magic Formula was subsequently
enhanced. A new approach to describe the horizontal forces at combined slip conditions with
an empirical method was added by Bayle et al. [4]. Pacejka [5] introduced the pneumatic trail
to calculate the aligning moment, and thus simplified the calculation of this moment under
combined slip conditions. The latest enhancement is the addition of turn-slip by Pacejka [6].
Different versions of the Magic Formula are offered with MSC ADAMS motion simulation
software. The range starts with PAC89, which is an implementation of the tire model presented
in [2]. The PAC94 corresponds with the tire model in [3]. The latest Magic Formula model is
PAC2002 and is kept continuously up-to-date and therefore is the recommended handling tire
model in MSC ADAMS.
PAC2002 can be classified as a semi-empirical tire model. The tire forces and moments
are approximated by mathematical formulae, the ‘Magic Formula’ descriptions. The model
parameters are obtained from the data fitting of tire measurements. Because PAC2002 is a
parametrized model it is capable of interpolation and extrapolation to approximate the tire
operating conditions not available as measurement data. In addition to this empirical part,
PAC2002 contains physically modeled components as well and therefore PAC2002 is a semi-
empirical tire model. The transient tire behavior is modeled by a physical approach, i.e. the
stretched-string model or by explicit carcass compliances. PAC2002 can also be classified
semi-empirical because the Magic Formula has indeed a physical interpretation, for example,
cornering stiffness or friction coefficient.
The PAC2002 tire model describes the behavior of the tire traveling over relatively smooth
road surfaces where the road wavelengths are typically larger than the tire circumference.
The model’s dynamic behavior is valid for frequencies up to ∼12 Hz. This makes PAC2002
applicable for all general generic vehicle handling and stability simulations, including:

(1) steady-state cornering;

(2) single or double lane change;
(3) braking or power-off in a turn;
 PAC2002 advanced handling tire model 155

(4) μ-split braking;

(5) J-turn or other turning maneuvers and
(6) ABS braking.
The PAC2002 tire model has proven to be applicable for car, truck and aircraft tires for
inclination angles not exceeding 15 ◦ .

2. Model description

2.1 Magic Formula basics

Downloaded by [University of California Davis] at 00:23 15 August 2014

The Magic Formula is a type of mathematical formula that is capable of describing basic tire
characteristics surprisingly well. There exist two basic forms of Magic Formula, the sine and
cosine form.

y(x) = D sin[C arctan{Bx − E(Bx − arctan Bx)}], (1)

y(x) = D cos[C arctan{Bx − E(Bx − arctan Bx)}]. (2)

In these two formulas, y denotes the dependent variable, x the independent variable and B, C,
D and E are parameters. The parameters, or factors, have a specific influence on the Magic
Formula characteristic:
(1) B, the stiffness factor, stretches the curve;
(2) C, the shape factor, mainly influences the shape of the curve;
(3) D, the peak factor, determines the peak value of the curve and
(4) E, the curvature factor, influences the characteristic around the peak of the curve and the
location of the peak.
The shape factor C controls the limit value of the curve and thereby the shape of the curve,
according to:
2 y∞
C = 2 − arcsin . (3)
π D
The product BCD corresponds to the slope of the curve at the origin. With known C and D,
the stiffness factor B is left to control the slope of curve at the origin.
The location of the peak value is directly determined by the curvature factor E, which is
related to the peak location xm as defined in equation (4).

Bxm − tan π/2C

E= . (4)
Bxm − arctan Bxm
Similarly, the Magic Formula factors of equation (2) can be interpreted as defined below,
where x0 denotes the location where the curve becomes zero.
2 y∞
C= arccos , (5)
π D
Bx0 − tan π/2C
E= . (6)
Bx0 − arctan Bx0
The equations above illustrate that the Magic Formula factors can indeed have a physical
interpretation.
 156 E. Kuiper and J. J. M. van Oosten

2.2 Tire model input quantities

In order to derive the tire model input quantities, the tire is assumed to behave like a thin
rigid disk as depicted in figure 1. In this figure, three planes are drawn: the road tangent
plane, the wheel plane and the wheel center plane. The wheel–road contact point C is the
intersection point of the three planes. Wheel inclination angle (or wheel camber) γ is defined
as the angle between the wheel plane and the normal to the road. Tire deflection ρ is defined
as the difference between the free R0 and loaded tire radius Rl . The loaded tire radius is the
distance between the wheel center point and the wheel–road contact point:

ρ = R0 − Rl . (7)
Downloaded by [University of California Davis] at 00:23 15 August 2014

The longitudinal slip velocity Vsx in the tire–road contact point is defined by the longitudinal
wheel center velocity Vx , the wheel rotational velocity  and the effective rolling radius re :

Vsx = Vx − re . (8)

The lateral slip velocity Vsy is equal to the lateral velocity of the wheel–road contact point
with respect to the road plane:
Vsy = Vy . (9)
In addition to the slip velocities, the rolling velocity Vr is determined by the wheel angular
velocity and the effective rolling radius:

Vr = re . (10)

In figure 2 the wheel velocities are depicted for clarification.

Different definitions of slip exist in the literature and since tire slip is the input to the
tire model, the slip quantities and model description cannot be separated. PAC2002 uses the
longitudinal slip κ
Vsx
κ=− (11)
Vx

Figure 1. Definition of the position and orientation of the wheel and forces and moments acting on the wheel.
 PAC2002 advanced handling tire model 157

Figure 2. Wheel velocities at combined cornering and braking/traction.

and the slip angle α

Vsy
tan α = . (12)
|Vx |
Downloaded by [University of California Davis] at 00:23 15 August 2014

In addition to longitudinal slip and slip angle, a third slip quantity is used as input for PAC2002,
turn-slip. Turn-slip is one of the two components that contribute to the ‘spin’ of the tire. Turn-
slip ϕt is calculated using the wheel yaw velocity ψ̇, which is the angular velocity of the wheel
around the road normal.
ψ̇
ϕt = − . (13)
Vx

2.3 Steady-state cornering and braking/driving

The so-called steady-state cornering and braking/driving condition is the condition where the
time rate of the slip quantities, i.e. slip and slip angle velocity, can be neglected. Under pure
slip conditions, the tire is either cornering or braking/driving. The longitudinal force for pure
slip conditions (no slip angle) reads:

Fx = Fx0 (κ, Fz , γ ), (14)

Fx0 = Dx sin(Cx arctan(Bx κx − Ex (Bx κx − arctan(Bx κx )))) + SVx , (15)
κx = κ + SHx . (16)

The coefficients Cx , Dx and Ex are function of tire load Fz and camber angle γ . The full set
of equations can be obtained from ref. [6]. The stiffness coefficient Bx is calculated from the
slip stiffness Kx , which is defined explicitly, according to:

Kx = Fz (pKx1 + pKx2 dfz )epKx3 dfz . (17)

The p coefficients are the tire model parameters for pure slip conditions, and dfz is
defined as:
Fz − Fz0
dfz = . (18)
Fz0
The lateral force for pure slip (free rolling) reads:

Fy = Fy0 (α, γ , Fz ), (19)

Fy0 = Dy sin(Cy arctan(By αy − Ey (By αy − arctan(By αy )))) + SVy , (20)
αy = α + SHy . (21)

The cornering stiffness is defined explicitly: it reads:

  
Fz
Ky = pKy1 sin 2 arctan . (22)
pKy2 Fz0
 158 E. Kuiper and J. J. M. van Oosten

The aligning moment is accounted for by the pneumatic trail t and residual moment Mzr ,
according to:

Mz = Mz0 (α, γ , Fz ), (23)

Mz0 = −t · Fy0 + Mzr . (24)

The expressions for the pneumatic trail and residual moment are based on the cosine form of
the Magic Formula:

t = Dt cos(Ct arctan(Bt αt − Et (Bt αt − arctan(Bt αt )))) cos α, (25)

αt = α + SH t , (26)
Downloaded by [University of California Davis] at 00:23 15 August 2014

Mzr = Dr cos(Cr arctan(Br αr )) cos α, (27)

αr = α + SHf , (28)
SVy
SHf = SHy + . (29)
Ky

Under combined slip conditions, the tire is driven or braked while it is cornering. The force
and moment expressions for combined slip conditions are based upon the expressions for pure
slip. The longitudinal force reads:

Fx = Fx0 · Gxα (α, κ, Fz ). (30)

Here, Gxα is the so-called weighting function of the longitudinal force. The weighting function
is the cosine form of the Magic Formula so that the longitudinal force decreases with increasing
slip angle. The expression reads:
cos(Cxα arctan(Bxα αs − Exα (Bxα αs − arctan(Bxα αs ))))
Gxα = . (31)
cos(Cxα arctan(Bxα SHxα − Exα (Bxα SHxα − arctan(Bxα SHxα ))))
The lateral force under combined slip conditions is accounted for in a similar way:
cos(Cyκ arctan(Byκ κs − Eyκ (Byκ κs − arctan(Byκ κs ))))
Gyκ = . (32)
cos(Cyκ arctan(Byκ SHyκ − Eyκ (Byκ SHyκ − arctan(Byκ SHyκ ))))

The expressions for aligning moment under combined slip conditions read:

Mz = −t · (Fy − SVyκ ) + Mzr + s · Fx . (33)

The pneumatic trail and residual moment are evaluated with equivalent slip angles:

t = t (αt,eq ), (34)
Mzr = Mzr (αr,eq ). (35)

The equivalent slip angles αt,eq and αr,eq read:

  2
Kx
αt,eq = arctan tan2 αt + κ 2 · sgn(αt ), (36)
Ky
  2
Kx
αr,eq = arctan tan2 αr + κ 2 · sgn(αr ). (37)
Ky
 PAC2002 advanced handling tire model 159

2.4 Transient tire behavior

The Magic Formula equations are only valid for steady-state operating conditions. Under
realistic vehicle driving conditions however, the influence of the input velocities cannot be
neglected. This low- frequency behavior is called transient tire behavior and covers frequencies
up to 15 H z. PAC2002 provides two methods to account for tire transient behavior:
(1) Stretched-string model.
(2) Contact-mass model

2.4.1 Stretched-string model. The first method to accurately describe the tire transient
Downloaded by [University of California Davis] at 00:23 15 August 2014

behavior is the so-called stretched-string model. The tire belt is modeled as a stretched-string,
which is suspended to the rim by longitudinal and lateral springs, as depicted in figure 3.
When the tire is rolling, the first point that is in contact with the road sticks to the road.
A lateral deflection of the string arises that depends on slip angle size and the deformation
history that were previously in contact with the road. The deflection of the leading point in
contact with the road v1 can be calculated with the following differential equation:
dv1
σα + |Vx |v1 = σα Vsy . (38)
dt
When the tire is rolling, the lateral spring deflection depends on the lateral slip velocity but
at standstill the deflection depends solely on the relaxation length σα , which is a measure
for the lateral stiffness of the tire. The string deflection in longitudinal direction is given in a
similar way:
du1
σκ + |Vx |u1 = −σκ Vsx . (39)
dt
The longitudinal σκ and lateral σα relaxation lengths are functions of the tire load. The
longitudinal slip and slip angle can be calculated based on the string deformation:
u1
κ = · sgn(Vx ), (40)
σκ
 
v1
α  = arctan . (41)
σα
With these transient slip quantities, the Magic Formula equations can be evaluated to calculate
the transient forces and moment:

Fx = Fx (α  , κ  , Fz ), (42)

Figure 3. Top-view of the stretched-string model.

 160 E. Kuiper and J. J. M. van Oosten

Fy = Fy (α  , κ  , γ , Fz ), (43)
 
Mz = Mz (α , κ , γ , Fz ). (44)

2.4.2 Contact-mass model. The contact-mass model is based on the separation of the
contact patch slip properties and the tire carcass compliance. This is in contrast with the
stretched string model where carcass compliance and contact patch slip properties are bundled
in the relaxation length. In the contact-mass model, the carcass compliance is defined explicitly
in the model by springs.
The differential equations for the contact body read:
 
dVcx dψc du
Downloaded by [University of California Davis] at 00:23 15 August 2014

mc − Vcy + kx + cx u = Fsx , (45)

dt dt dt
 
dVcy dψc dv
mc + Vcx + ky + cy v = Fsy , (46)
dt dt dt
d 2 ψc dβ
Jc 2
+ kψ + cψ β = Msz . (47)
dt dt
The contact patch is modeled as a contact body with mass mc and inertia Jc . This contact body
is connected to the wheel by springs with stiffness cx , cy and cψ and dampers with damping
constants kx , ky and kψ in longitudinal, lateral and yaw directions, respectively.
The contact body deflections in longitudinal u, lateral v and yaw β directions are given by:
du
= Vcx − Vsx , (48)
dt
dv
= Vcy − Vsy , (49)
dt
dβ dψc dψ
= − , (50)
dt dt dt
where Vcx , Vcy and dψc /dt are the sliding velocities of the contact body in longitudinal, lateral
and yaw directions, respectively. Vsx , Vst and dψ/dt are the corresponding velocities of the
lower part of the wheel.
The contact body is in contact with the road surface and therefore is slipping with respect
to the road and slip forces and moment Fsx , Fsy and Msz act on the body. The transient slip
properties for the contact body read

dα 
σc + |Vx |α  = Vcy − Vx β + |Vx |βst , (51)
dt
dκ 
σc + |Vx |κ  = −Vcx , (52)
dt
where the static deflection angle βst has been used:
Mz
βst = . (53)
cψ
The slip properties of the contact body are dominated by the relaxation length σc of the tire
tread. The expression for the relaxation length reads:

σc = a(1 − θ ξ  ). (54)
 PAC2002 advanced handling tire model 161

The relaxation length is a function of half contact length a and diminishes with increasing
equivalent slip ξ . The composite tire parameter θ reads:
Ky
θ= , (55)
3μy Fz
and the equivalent slip:
  2
1 Kx
ξ = |α  |2 + |κ  |. (56)
1+κ Ky

2.5 Tire spin and parking

Downloaded by [University of California Davis] at 00:23 15 August 2014

The latest enhancement of PAC2002 is the addition of turn slip to the known slip quantities
side slip, longitudinal slip and wheel camber. The tire spin ϕ has two components: turn slip
ψ and wheel camber γ :
1
ϕ = − (ψ̇ − (1 − γ ) sin γ ). (57)
Vx
The influence of camber is accounted for by the camber reduction factor γ . The camber
reduction factor is an additional parameter; for an ideal solid ball, the reduction factor will be
zero and for a passenger car tire it will be between 0.5 and 1.0 approximately. The camber
reduction factor is assumed to be varying with load and is defined according to ref. [6].
γ = pγ ϕ1 (1 + pγ ϕ2 dfz ). (58)
Camber effects were already accounted for in the expressions for lateral force and aligning
moment. These effects are considered to be a result of a changed contact pressure distribution
for a cambered tire. The expression for longitudinal force, lateral force and aligning moment
were extended for tire spin by introducing weighting functions ζ1 , ζ2 , . . . , ζ8 .
The longitudinal force Fx diminishes with increasing spin by scaling the peak factor Dx .
Dx = μx Fz ζ1 . (59)
The scaling factor ζ1 is given by
ζ1 = cos(arctan(Bxϕ R0 ϕ)). (60)
Similarly, the lateral force Fy diminishes with increasing spin by scaling of Dy and the vertical
shift SVy .
Dy = μy Fz ζ2 . (61)
SVy = Fz (pVy1 + pVy2 dfz + (pVy3 + pVy4 dfz )γ )ζ2 . (62)
Here, the scaling factor ζ2 reads:

ζ2 = cos(arctan(Byϕ (R0 |ϕt | + pDyϕ4 R0 |ϕt |))). (63)
Instead of tire spin ϕ, the turn-slip ϕt is used in the equation above. The cornering stiffness
Ky is reduced with increasing spin by scaling factor ζ3
  
Fz
Ky = pKy1 sin 2 arctan (1 − pKy3 γ 2 )ζ3 , (64)
pKy2 Fz0
ζ3 = cos(arctan(pKyϕ1 R02 ϕt2 )). (65)
 162 E. Kuiper and J. J. M. van Oosten

The horizontal shift SHy is also influenced by spin.

SHy = pHy1 + pHy2 dfz + pHy3 γ ζ0 + ζ4 − 1. (66)

The horizontal shift is affected by two scaling factors ζ0 and ζ4 . They read

ζ0 = 0, (67)
SVyγ
ζ4 = 1 + SHyϕ − . (68)
Ky

The vertical shift as function of camber SVyγ reads

Downloaded by [University of California Davis] at 00:23 15 August 2014

SVyγ = Fz (pVy3 + pVy4 dfz )γ ζ2 . (69)

The horizontal shift due to spin SHyϕ reads

SHyϕ = DHyϕ sin(CHyϕ arctan(BHyϕ R0 ϕ − EHyϕ (BHyϕ R0 ϕ − arctan(BHyϕ R0 ϕ)))). (70)

The aligning moment is modeled by two terms, the product of pneumatic trail and lateral force
and the residual moment. The pneumatic trail t diminishes with increasing spin by scaling of
the peak factor Dt with ζ5 .
ζ5 = cos(arctan(qDtϕ1 R0 ϕ)). (71)
The residual moment is adapted by the modification of the residual moment peak factor Dr ;
it reads:
Dr = Fz ((qDz6 + qDz7 dfz )ζ2 + (qDz8 + qDz9 dfz )γ ζ0 ) + ζ8 − 1. (72)
The scaling factor ζ8 is given by:
ζ8 = 1 + Drϕ . (73)
In this equation, the peak spin torque Drϕ reads:

Drϕ = DDrϕ sin(CDrϕ arctan(BDrϕ R0 ϕ) − EDrϕ (BDrϕ R0 ϕ − arctan(BDrϕ R0 ϕ))). (74)

In addition to the peak factor, Dr of the residual moment, the stiffness factor Br is modified
for diminishing stiffness with increasing spin.

Br = (qBz9 + qBz10 By Cy )ζ6 . (75)

The scaling factor reads:

ζ6 = cos(arctan(qBrϕ1 R0 ϕ)). (76)
The shape factor Cr of the residual moment is given by scaling factor ζ7

Cr = ζ7 . (77)

The scaling factor reads:

 
1 Mzϕ90
ζ7 = arccos . (78)
π |Drϕ |
Here, Mzϕ90 is the moment generated at 90◦ slip angle. For further details on the modeling of
spin, see [6, 7].
Tire parking behavior is dominated by the moment acting on the tire at very low or zero
speed. This moment can become very large and the tire response to spin is dominating under
 PAC2002 advanced handling tire model 163

parking conditions. The maximum parking moment is equal to the moment that arises when the
tire is traveling with a constant turn rate over a circular path with the radius of turn decreasing
to zero.

According to the approach in ref. [6], the transient spin ϕM is calculated by the integration
˙ 
of the transient spin velocity ϕ̄M .
   
 Mzϕ 2

ϕM = 1−p   ϕ̄˙  dt. (79)
Mzϕ∞  M

In this expression, p = 0 if sgnϕM

 = sgnϕ̄˙M

and p = wV low elsewhere. The factor wV low
depends on speed and is 1 at zero speed. The parking moment Mzϕ is given by:
Downloaded by [University of California Davis] at 00:23 15 August 2014


Mzϕ = Drϕ (ϕM ). (80)

3. Using PAC2002 with MSC Adams motion simulation software

The PAC2002 tire model is available for MSC ADAMS motion simulation software. In addition to
standard MSC ADAMS functionality, MSC Software offers specialized modules for automotive
industry. These modules offer specialized vehicle dynamics simulations and templates for
building vehicle models. PAC2002 works with a so-called road property file in which the tire
model behavior can be adapted by modifying variables and specifying parameters.

3.1 Advanced handling application

Tire models based on the Magic Formula are widely considered as state-of-the-art for mod-
eling tire-road interaction forces in vehicle handling applications. Over the years, this type of
model was continuously improved and PAC2002 forms the latest standard of Magic Formula
models.
The classical application of PAC2002 is in the field of vehicle handling. Typical handling
maneuvers include:
(1) Straight-line acceleration and braking;
(2) constant-radius turn;
(3) braking in turn;
(4) step steer;
(5) J-turn;
(6) lane change and double lane change and
(7) roll-over.
In addition to the maneuvers mentioned above, some low-speed situations can be reliably
simulated with PAC2002:
(1) Starting from standstill.
(2) Parking.
(3) Standing on moving surface.
The low-speed behavior was currently improved significantly by the addition of the contact-
mass model and turn-slip/parking. The typical operation conditions mentioned thus far are
steady-state or transient conditions. However, it was shown that some dynamical conditions
can be handled by PAC2002 as well. This is particularly useful for the analysis and design of
 164 E. Kuiper and J. J. M. van Oosten

vehicle control systems, which is increasingly gaining in importance. Some related areas are
as follows:
(1) ABS-braking.
(2) Steering system vibrations.
(3) Active suspension systems.
(4) Real-time simulation.
The most notable enhancements for the PAC2002 tire model are (see also section 4):
(1) Improved tire transient modeling using a contact-mass model. This results in an improved
dynamic tire response at large slip, low speed and standstill.
(2) Addition of parking moment and turn-slip.
Downloaded by [University of California Davis] at 00:23 15 August 2014

(3) Extended loaded radius modeling for handling extreme driving conditions such as roll-over
and racing applications.
(4) Optionally, a non-linear spline for the load-deflection characteristic can be used instead
of a linear stiffness.
(5) Modeling of bottoming of the wheel to the road with a spline to represent the rim stiffness.
(6) On-line scaling of tire properties during a simulation. The PAC2002 scaling factors can be
changed as a function of time, position or any other variable.

3.2 Obtaining model parameters

The parameters for the PAC2002 tire model are obtained directly from measurement data. Since
it is a semi-empirical tire model the parameters are not physical properties, but coefficients of
a mathematical formula. Although the Magic Formula offers good interpolation and extrap-
olation properties, it is important to use measurement data that covers the intended range
of application of the model. Here, it is assumed that classical ‘sweep’ tests are used. With
sweep tests only one input variable is varied, whereas the other variables are kept constant.
For example, the slip angle is changed for constant load and inclination angle or the tire is
braked (changing longitudinal slip) with constant load, slip and inclination angle.
The following signals must be measured:
(1) Translational velocity.
(2) Rotational velocity.
(3) Wheel center height with respect to the road surface.
(4) Forces and moments Fx , Fy , Fz , Mx and Mz .
(5) Steer angle.
(6) Inclination angle.
The measurement program mentioned here is a generic guideline for commonly used passenger
car tires. The program depends on the so-called nominal load Fz0 , which is chosen equal to
the static load of the tire under the vehicle or equal to the load that corresponds to the tire load
index (rated load).
Two measurement programs are given, the full measurement program 1 (table 1) and the
reduced measurement program 2 (table 2).
The reduced test program has much fewer measurements than the full program, but the
effect on accuracy, interpolation and extrapolation properties of the resulting tire model will
be very small. Post-processing of the raw measurement data will normally include low-pass
filtering, trimming and probably averaging of the measurement signals.
The post-processed data is used for parameter identification. In the first step, the parameters
for longitudinal and lateral force under pure slip conditions are identified. This is because the
 PAC2002 advanced handling tire model 165

Table 1. Full tire measurement program.

Slip condition Load Slip angle Inclination angle Longitudinal slip (%)

Pure lateral slip 0.5Fz0 , Fz0 , 1.5Fz0 −12◦ , . . . , 12◦ −5◦ , 0◦ , 5◦ free rolling
Pure longitudinal slip 0.5Fz0 , Fz0 , 1.5Fz0 0◦ −5◦ , 0◦ , 5◦ −100, . . . , 100
Combined slip 0.5Fz0 , Fz0 , 1.5Fz0 −2◦ , 2◦ , 5◦ , 8◦ −5◦ , 0◦ , 5◦ −100, . . . , 100

Table 2. Reduced tire measurement program.

Slip condition Load Slip angle Inclination angle Longitudinal slip (%)

Pure lateral slip 0.5Fz0 , Fz0 , 1.5Fz0 −12◦ , . . . , 12◦ −5◦ , 0◦ , 5◦ Free rolling
Downloaded by [University of California Davis] at 00:23 15 August 2014

Pure longitudinal slip 0.5Fz0 , Fz0 , 1.5Fz0 0◦ 0◦ −100, . . . , 0

Combined slip 0.5Fz0 , Fz0 , 1.5Fz0 −2◦ , 2◦ , 5◦ , 8◦ 0◦ −100, . . . , 0

remaining parameters depend on these pure slip parameters. Next the parameters for aligning
moment under pure slip are identified. Finally, the parameters for combined slip are calculated.
There exist special software packages to calculate Magic Formula tire parameters. However,
generic curve fitting algorithms for non-linear least-squares optimization problems are also
well suited to perform this task.

4. Specific PAC2002 advanced transient properties

Because the PAC2002 is not a traditional steady-state Magic Formula tire model, in this section,
the more advanced aspects of the tire model will be discussed in two sections:
(1) Handling properties.
(2) Higher frequency behavior.

4.1 Specific handling properties

4.1.1 Parking maneuver. The addition of tire spin and parking moment to PAC2002 makes
it possible to simulate a parking maneuver as demonstrated in the TMPT test 1 successfully.
When analyzing the parking moment as a function of the sinusoidal steer angle motion, the
characteristic hysteresis is shown. At the start of this simulation, the parking moment shows
a transient response of the tire as well. This effect is due to the input motion which is a pure
sine wave, therefore the steer angle velocity is not equal to zero. The transient tire properties
are accounted for by the contact-mass model which contains relaxation length effect for spin
as well. The most notable for PAC2002’s parking maneuver simulation is that this result was
obtained without the need for identification of additional parameters. From the pure lateral
slip parameters, the aligning moment dependency on camber angle is known. Camber is one
of two components for tire spin, therefore the aligning moment as function of tire spin is
already accounted for. The parking maneuver is just a special case of tire spin, namely that
the maneuver is performed at zero velocity.

4.1.2 Wheel spin-up from standstill. The spin-up of the wheel, as performed in TMPT
test 2, is often a critical situation for tire models, because the main longitudinal and lateral
slip quantities become singular at zero speed. In addition, the tire wheel system is badly
 166 E. Kuiper and J. J. M. van Oosten

damped therefore this kind of simulation can show instability of the system. During the
simulation, the systems remain stable and are sufficiently damped. The tire model behavior
is obtained without artificially adding damping for lower velocities. The contact-mass takes
care of sufficient damping implicitly.

4.1.3 Constant increase of side-slip angle. The tire model behavior at very large slip
angles (i.e. slip angle of 180◦ ) is not very realistic, but can occur during vehicle dynamics
simulation. For avoiding solver instability problems, the tire model must show consistent
behavior in this case. TMPT test 4 shows the lateral force and aligning moment generated by
PAC2002 for a slip increasing from 0◦ to 180◦ .
With this test, there are two important issues: first, the forward velocity of the wheel
Downloaded by [University of California Davis] at 00:23 15 August 2014

becomes zero at 90◦ slip angle, and second the forward velocity changes sign. The lateral
force shows a symmetric characteristic with respect to 90◦ slip angle (halfway the simulation
time) while the aligning moment has a point-symmetric characteristic. PAC2002 continues
delivering consistent tire behavior under these extreme conditions.

4.1.4 Changing μ-conditions. Changing friction conditions (TMPT test 5) for a tire under
constant slip angle will result in a change or scaling of the lateral force. PAC2002 not only
scales the lateral force, but also shows tire transient behavior as expected for a step change in
friction level.

4.2 Specific higher frequency behavior

In addition to the handling behavior, the PAC2002 tire model is also applicable for a number
of dynamic tire aspects.

4.2.1 Dynamic brake cycle. A dynamic braking cycle as used in the TMPT test 8 gives
good insight into the longitudinal transient tire behavior. The response by PAC2002 shows
that the transient behavior changes significantly for larger values of slip. Although PAC2002
does not incorporate tire structural dynamics, the dynamic response is well suited for vehicle
ABS-braking maneuvers or yaw stability analysis.

5. Conclusions

The extensions of the PAC2002 tire model, in terms of the contact-mass transient model and
the modeling of tire spin and parking, has been described as well as the required measurement
program to obtain the tire model parameters. A reduced measurement program can be used,
which is more cost-effective and gives acceptable accuracy compared to the full program.
These extensions of the PAC2002 result in favorable tire model behavior at a number of
aspects:

(1) additional aligning torque due to turn-slip and parking is modeled without adding a
considerable amount of new parameters;
(2) improved validity at zero and low speed range including spinning up wheels and
(3) the model is capable of describing dynamic tire response at higher frequencies up to 12 Hz
such as ABS-braking cycles.
 PAC2002 advanced handling tire model 167

In fact, the PAC2002 model can deal with a wide range of the handling and dynamic applications,
as long as no short wavelength road obstacles are involved.

References
[1] Bakker, E., Nyborg, L. and Pacejka H.B., 1987, Tire modeling for use in vehicle dynamics studies. SAE Paper
870421 (New York: Society of Automotive Engineers).
[2] Bakker, E., Pacejka, H.B. and Lider, L., 1989, A new tire model with an application in vehicle dynamics studies.
SAE Paper 890087 (New York: Society of Automotive Engineers).
[3] Pacejka, H.B. and Bakker, E., 1993, The Magic Formula tire model. Proceedings of 1st Colloquium on Tyre
Models for Vehicle Analysis, Delft, 1991; Vehicle System Dynamics, 21(Suppl.).
[4] Bayle, P., Forissier, J.F. and Lafon, S., 1993, A new tire model for vehicle dynamics simulation. Automotive
Technology International.
[5] Pacejka, H.B., 1996, The tire as a vehicle component. Proceedings of XXVI FISITA Congress, edited by
Downloaded by [University of California Davis] at 00:23 15 August 2014

M. Apetaur, Prague.
[6] Pacejka, H.B., 2002, Tire and Vehicle Dynamics (Oxford: Butterworth-Heinemann).
[7] ADAMS/Tire Manual, 2005, MSC Software Corporation.