Tyre Modelling for Use in Vehicle Dynamics Studies

Author(s): Egbert Bakker, Lars Nyborg and Hans B. Pacejka

Source: SAE Transactions, Vol. 96, Section 2 (1987), pp. 190-204
Published by: SAE International
Stable URL: https://www.jstor.org/stable/44470677
Accessed: 25-04-2020 16:41 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

SAE International is collaborating with JSTOR to digitize, preserve and extend access to SAE
Transactions

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

Tyre Modelling for Use in Vehicle

Dynamics Studies
Egbert Bakker and Lars Nyborg
Chassis Engineering
Volvo Car Corp.

Hans B. Pacejka
Vehicle Research Laboratory
Delft University of Technology

ABSTRACT first results of this project have been presented in

this paper.
A new way of representing tyre data obtainedThe description and understanding of tyre
from measurements in pure cornering and pure behaviour should be easily obtainable from
braking conditions has been developed in order measured
to data. Furthermore, it is desirable that
further improve the Dynamic Safety of vehicles. its parameters characterize in some way the typi-
The method makes use of a formula with coeff- fying quantities of the tyre such as slip stiffnesses
icients which describe some of the typifying at zero slip and peak values of forces and torques.
quantities of a tyre, such as slip stiffnesses at zero This feature would make it possible to investigate
slip and force and torque peak values. The formula the effect of changes of these quantities upon the
is capable of describing the characteristics of side handling and stability properties of the vehicle.
force, brake force and self aligning torque with That is, one might then easily model a hypotheti-
great accuracy. This mathematical representation cal tyre.
is limited to steady-state conditions during either
pure cornering or pure braking and forms the REPRESENTATION OF TYRE DATA
basis for a model describing tyre behaviour during
combined braking and cornering. Three possible ways to represent measured
tyre data are in use:
a) representation by tables,
b) representation by graphs,
SAFETY IS A FIELD which has been given top c) representation by formulae.
priority in developing Volvo cars, because people The first two possibilities are difficult to handle
should be able to travel safely and securely. For theoretical studies and therefore they do not fulf
this reason, the vehicle properties have to be the requirements mentioned earlier.
adapted to the driver, so the vehicle will be Two possibilities may be distinguished for the
controllable and its behaviour predictable in all representation by formulae:
situations. This is called Dynamic Safety and is a) formulae containing series (Fourier,
obtained after optimizing stability, steering and polynomials),
brake performance by tuning chassis and tyre b) formulae containing special functions.
design to each other. Here, the tyre plays a vital The use of series has some disadvantages:
role and thorough knowledge of its properties is - relatively many coefficients are needed to get a
required. It is therefore necessary to have a proper close curve fit,
description of tyre behaviour available. In order to - the resulting curves have a wavy look, which
achieve this, a joint project between Volvo and the implies that the variation of the slope along the
Delft University of Technology was set up. The curve differs considerably from that of the

2.190
0096-736X/88/9602-01 90$02.50
Copyright 1988 Society of Automotive Engineers, Inc.

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

measured data, Y = Dsin(BX) (1)

- extrapolation beyond the fitted range often
yields large deviations, with Y standing for eit
- in general, the coefficients do not describe the torque or brake force a
typifying quantities in a recognisable way, or longitudinal slip (k
which makes it impossible to change these defined as the ratio of the difference between the
quantities in a controlled, simple way. speed of rotation of the driven or braked tyre and
A representation with the aid of polynomials of the straight free rolling tyre, and the speed of
is described by Sitchin [1]*. This investigator rotation of the straight free rolling tyre, expressed
attempts to solve some of the problems connected as a percentage. A negative value results from a
with the use of polynomials by dividing the data braking torque. In Eq. (1) D is the peak value and
into five regions and using a variable format the product DB equals the slip stiffness at zero
regression equation for each of them. The first two slip.
disadvantages could be solved in this way, but the
others still remain. A problem appears to show up
for the derivative, which is discontinuous at the f Side force
boundaries of the regions.
The best way to fulfil the requirements is to f / Brake force
find a special function, which through its partic-
ular structure is capable of describing the meas-
ured data with great accuracy and which has / ' / Self aligning torque^*»

parameters related to the typifying quantities in a

°/' Slip (angle)
simple manner.
The formula should, if possible, be able to
describe:
- the side force as a function of slip angle,
- the brake force as a function of longitudinal
slip,
- the self aligning torque as a function of slip
angle.
The side force and self aligning torque are Fig 1. - Steady-state tyre characteristics.
measured during pure cornering (cornering with-
out braking) and the brake force during pure
braking (braking without cornering). Both corner- Eq. (1) does not give a good representation for
ing and braking are in steady-state conditions. larger values of X. A gradually increasing
A possible special function used by Ruf [2] is extension of the X axis appears to be necessary. To
a combination of tanh functions. The resulting accomplish this, the arctan function has been
formula can match side and brake force charac- used. The formula (1) now changes into:
teristics rather well. However, the coefficients do
not describe the typifying quantities of the tyre in Y = D sin (C arctan (BX)) (2)
an easy way and therefore these quantities are
difficult to recognize and to vary. In Eq. (2) D is still the peak value, th
stiffness at zero slip is now equal to the
PROPOSED TYRE FORMULA BCD (from now on called the stiffness). The
coefficient C governs the shape of the curve. For
The basic form of each of the characteristics of large values of X, Eq. (2) reduces to:
the tyre (cf. Fig.l) suggest the use of the sine
function as a first step in developing the final Y = D sin (ļnC) (3)
formula as summarized in the Appendix.
Consequently, C defines the
will be used and therefore d
* Numbers in brackets [ ] designate references at the curve. The value of C makes the curve look
the end of the paper. like a side force, brake force or a self aligning

2.191

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
torque characteristic (we find for C in those cases
successively: 1.3, 1.65 and 2.4). With C determined
by the shape and D determined by the peak value,
only B is left to control the stiffness.
Still Eq. (2) is not good enough to describe
every possible measured characteristic. There may
be a need for an additional coefficient which
makes it possible to accomplish a local extra
stretch or compression of the curve. The coefficient
E has been introduced into the formula in such a
way that stiffness and peak value remain
unaffected.
Y = D sin (C arctan (B<ļ>))
with (4) characteristics, these shifts should be included in
<ļ) = (1-E) theXequation. + (E/B)
We obtain:
icients, which is able to describe all the measured
characteristics. The four coefficients are:
B = stiffness factor
C = shape factor
D = peak factor
E = curvature factor
So far, the characteristics are assumed to pass
through the origin. In reality, however, this will
not always be the case. Due to ply steer, conicity
and rolling resistance, the characteristics will be
shifted in the horizontal and/or vertical directions
(Fig. 2). In order to fit the measured
ar

The influence of E on the side force characteristic Y = D sin (C arctan (B<ļ>)) + Sv (5)
has been shown in Fig. 2. Similar effects occur with
with the self aligning torque and brake force <ļ> = (1-E) (X + Sh) + (E/B) arctan (B(X + Sh))
characteristics. Sh = horizontal shift
The result is an equation with four coeff- Sv = vertical shift

Side force I
Fy [N] - I
6000 - I Sh

- 1

4000 - I

2000- I F F F F D

w ļ/BCP
-15 -10 II -5 0 F 5 10 15
y I Slip angle a [deg.]
SV -20 J I

yi/4000 - '

_ J ~o<°<i i .... E < 0

Fig. 2 - Coefficients

2.192

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

THE MEASUREMENTS - Full scale tyre quadratical correction curves around the meas-
measurements have been carried out on a dry ured points. Finally, the average was taken of two
asphalt road. The experiments were conducted or more a-sweeps.
with a specially built test trailer under steady- The measurements during pure braking were
state conditions. The measured quantities are: carried out by gradually increasing the brake
- the side force (Fy), pressure until wheel lock occurs. The data was
- the brake force (Fx), collected while the longitudinal slip was increased
- the self aligning torque (Mz). from 0 to 100% (K-sweep). Here also a
These forces and torque were measured during condensation, correction and averaging process
pure cornering, pure braking or during combin- has been carried out. The experiments on
ations of cornering and braking. During cornering, combined cornering and braking were conducted
the speed was 70 km/h and during braking 60 by doing K-sweeps at fixed values of the slip angle.
km/h. The tests included variations of slip angle The camber angle was fixed at zero degrees.
(-10 to 14 deg.), vertical load (2, 4, 6, 8 kN) and FITTING MEASURED DATA - The data
camber angle (-5, 0, 5 deg.). obtained after processing in the way described
The pure cornering tests were carried out by above has been used as raw input data for the
slowly varying the slip angle (a-sweep) from zero fitting process. An optimization program from th
to the maximum value, from there to the NAG library has been used to obtain the four
minimum value and back to zero again. The coefficients and two shifts for every measure
measured data was condensed per 1/4 degree characteristic. For one of the measured tyres, th
intervals to one measured point per interval. Afterresulting coefficients have been tabulated in Table
that, corrections were made for load variations, 1. The vertical shift obtained for the brake forces
which occurred during the tests, by making use ofrepresents the value for the rolling resistance.

Table 1
Coefficients for tyre formula (first fit)
(a [deg.] , k[%] )

B C D E Sh Sv BCD

2 0.244 1.50 1936 -0.132 -0.280 -118 780.6

Fy 4 0.239 1.19 3650 -0.678 -0.049 -156 1038
6 0.164 1.27 5237 -1.61 -0.126 -181 1091
8 0.112 1.36 6677 -2.16 0.125 -240 1017

2 0.247 2.56 -15.53 -3.92 -0.464 -12.5 -9.820

Mz 4 0.234 2.68 -48.56 -0.46 -0.082 -11.7 -30.45
6 0.164 2.46 -112.5 -2.04 -0.125 -6.00 -45.39
8 0.127 2.41 -191.3 -3.21 0.009 -4.22 -58.55

2 0.178 1.55 2193 0.432 0.000 25.0 605.0

Fx 4 0.171 1.69 4236 0.619 0.000 70.6 1224
6 0.210 1.67 6090 0.686 0.000 80.1 2136
8 0.214 1.78 7711 0.783 0.000 104 2937

2.193

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 Side force
Self aligning torque
Fy [N| Mz [Nm]
6000 - 150-

4000 -

2000 - ļ

-15 -10 -5 oj 5 10 15 -15 >/-10 -5 0 1 5 yr 15

J' Slip angle o [deg.] " a Jr Slip angle
-60 V / *|d*8-1

7^-4000 -

Measurements

- 1 50 -
~ 6000 " ! * Fz = 6000 N I I Measurements

I ♦ Fz = 600

Fig. 3 - Side force characteristic Fig.

fitted using4
the - Self
tyre formula compared with measured
using the t
data. measured data.

Brake force
-Fx [N]
6000

4000 - * * 'IM

2000- .

-75 -50 -25 O 25 50 75

- Longitudinal slip
-2000- -"%1

-4000 -

Measurements

~6000" I » Fz = 6000~Ñ~~| I-

Fig. 5 - Brake force characteristic fitted usin

tyre formula compared with measured
data.

2.194

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

Figures 3, 4 and 5 show comparisons between (aligning stiffness) characteristics, we use the
measurements (after processing) and fitted approximation:
characteristics for one specific load. The result is
very satisfactory: the formula is able to describe a3Fz2 + a4Fz
all the measured characteristics. BCD =

At the moment, the shifts are not of particular e a5^z

interest for further examination and are therefore
put equal to zero (i.e. ply steer, conicity and roll- The shape factor (C) appears to be practically
ing resistance will not be taken into account). If independent of Fz. We take for:
so required, it is possible to add the shifts later on. the side force C = 1.30
INFLUENCE OF VERTICAL LOAD - To the brake force C= 1.65 (9)
reduce the total number of quantified coefficients the self aligning torque: C = 2.4
(Table 1) and to be able to calculate forces and
torques at vertical loads which are different from The stiffness factor (B) is found by
the values used in the measurements, it is necess- stiffness by the shape and the peak fa
ary to include the vertical load explicitly in the
formula. To do so, the coefficients have to be B = BCD/CD (10)
written as a function of the vertical load (Fz).
The peak factor (D) as a function of Fz may be Finally, the curvature f
approximately represented by the relationship: is given by:

D = aļFz2 -I- a2Fz (6) E = aßFz2 + a7Fz + as (11)

For the stiffness (BCD) of the side forceAftercharacter-

processing the raw data again (corre
for the earlier
istic (cornering stiffness), we take the formula: determined horizontal and vertical
shifts) with the aid of the optimization program,
BCD = a3 sin (a4 arctan (a5Fz)) (7) the new coefficients are obtained. Table 2 gives the
coefficients for the same tyre as in Table 1, but
and for the stiffness of both brake force (longitud- now including the influence of the vertical load.
inal slip stiffness) and self aligning torque

Table 2
Coefficients for tyre formula (with load influence)
(FJkN])

al a2 a3 a4 a5 a6 a7 ag

Fy -22.1 1011 1078 1.82 0.208 0.000 -0.354 0.707

Mz -2.72 -2.28 -1.86 -2.73 0.110 -0.070 0.643 -4.04
Fx -21.3 1144 49.6 226 0.069 -0.006 0.056 0.486

Table 3
Coefficients for tyre formula connected with camber influence
ÍY [deg.])

a9 al0 an ai2 ai3

Fy 0.028 0.000 14.8 0.022 0.000

Mz 0.015 -0.066 0.945 0.030 0.070

2.195

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 Side force
Self aligning torque
FV l~l ¿ Mz INm]
; ' Fits
6000 fĻ •* • 150
... Fz = 8000 N
ai i i i i i i i i i » - Fz - 6000 N
'

4000 jr ^ ¿ /*N'00 - F^ 2000N

2000 j 1 M M t'-l

15 10 5 gř 5 10 15
/Ļa Slip angle o (deg. ] Vą J * •*' Slip angle
^ 5o!VW / / <,|d,fll
3/ 4000 100 i ' ' J Measurements
¿f/ Measurements Fits
t J / • F* " 8000 N
. yfr'o ' • Fz - 8000 N ' 1 Fz 8000 N ! ' * * Fz =- 6000 N
^ 1 11 d' * Fz 6000 N - Fz 6000 N ' / X Fz - 4000 N
4 ' 6000 X Fz 4000 N -- Fz 4000 N 1 50 , * ¿ + Fz = 2000 N
ý

Fig. 6 - Side force characteristics

Fig fi
tyre formula which us includes the
of the vertical load, in
compared w
ured data. with measured data.

Brake force .*^bw

-Fx IN] ' 0
6000

4000 •

f "* ,

2000

75 50 25 0 25 50 75

- Longitudinal slip
-X |%| 1
2000 • 1

4000 ■
Measurements Fits

' • Fz =- 8000 N1 • Fz - 8000 N!

* Fz - 6000 N - Fz - 6000 N
6000 X Fz = 4000 N

+ Fz = 2000 N

Fig. 8 - Brake force characteristics f

tyre formula which includes the
of the vertical load, compared w
ured data.

2.196

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

dFy/da 1400 dFV/da

[N/dog. ] (N/deg.| 1 s 1
1 s 1

- Fz 6000 N
1 200
1200 H i

- • Fz = 8000 N ļ ļ
- - Fz = 6000 N
/Z. tx. - Fz = 4000 N 1000 ""
' ' ' i - fz = 2000 N

/ / /oo ''' 800 ****♦.

600 ' '

/ if] ' ' ''
400 ' ' ' ' ' ' '
' ' '
////.: 1 W''
//// 1 '''' ' ' ' ' ' ' '
/J;j //// -I '''' m, ' ' ' '
-15 -10 -5 0 5 10 15 2000 4000 6000
Side force
200 Slip angle o [deg.] -200- FV INI

Fig. 9 - Calculated slope of side force characteris-

Fig. 10 - Cal
tics as a function of slip angle for different tics as a function of side force for different
vertical loads. vertical loads.

Pneumatic trail Pneumatic trail

t[m]0.08I^ t [m]
0.08

- Fz 6000 N

I ' -••• Fz = 8000 N!

[ - - Fz = 6000 N
0.06
0 06
........

/ i '

0.04 - - -

0.02 '

2000
' ' '
0 2000 / 4000 /y } 4000
6000 ; , / ļ
/ » Side , force
j Slip angle a [deg.)
Fy |N]

Fig. 11 - Calculated pneumatic

Fig. 12- Calculated pneumatic trail as a function trail a
of slip angle for of side force for different loads.loads.
different

2.197

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
The new parameter a3 for the side force indicates pure cornering and braking conditions, will be
the maximum possible cornering stiffness gener- used now in a mode which describes the behaviour
ated by the tyre. of the tyre at zero camber angle when a combin-
A comparison between measurements and the ation of cornering and braking takes place. The
newly obtained characteristics has been shown in modelling process implies the use of the basic tyre
Figs. 6,7 and 8. The influence of the vertical load is modelling theory as described in [3].
now included without much loss in accuracy. Figure 13 shows a top view of a tyre during
INFLUENCE OF CAMBER - The total num combined cornering and braking. The force and
ber of coefficients to be quantified may be reduced torque acting on the tyre and the velocity of the
even more by including camber in the formulawheel as have been indicated:
well. The influence of camber (y) is induced -by the total force with components in lateral and
adding two shifts and a possible change in stiff-longitudinal direction (Fy and Fx),
ness. The two additional shifts are: - the self aligning torque (Mz), _
- the speed in the direction of travel_V which is
ASh = a9y composed of the slip_speed vector (Vs) and the
(12) rolling speed vector (Vr) of the tyre.
ASV = (aioFz2 + anFz)y

The change in stiffness is obtained by multi-

plying B by (l-ai2l yl). We have for the increment
of B: ' Vr í
1 w I-

_vsy
AB = -ai2l ylB (13)
_ ~Fx/ " kVT -
The values of the self aligning torque at
higher slip angles will change due to this change
" I; Mz
l Ą /v¡í - vx
in stiffness. To compensate for this effect, the
curvature factor E for Mz must be divided by
(l-aißlyl). For the side force, this compensation is Iy
not necessary because the values at higher slip
angles are not changed appreciably by changing Fig. 13- Top view of tyre during combined
the stiffness. The values obtained have been braking and cornering showing velocity
shown in Table 3. and force vector diagrams.
At this stage about five thousand measure-
ment points have been reduced to three equations BRAKE AND SIDE FORCE DURING COM-
with a total of thirty-one coefficients (coefficients BINED CORNERING AND BRAKING - If the
with value zero not included). It is now possible to tyre is assumed to have isotropic properties then
calculate forces and torques which would arise in the direction of the total force is opposite the slip
circumstances not included in the measurement speed vector. The length of the total force vector is
programme. The equations are summarized in athe function of the total newly introduced slip quan-
Appendix. tity o [3]. This theoretical slip has components
Beside side force, self aligning torque and which are defined as follows:
brake force, other information may be obtained
with the aid of the formula. For instance dFy/da as ox = Vsx/Vr
a function of slip angle or as a function of side force Oy = Vsy/Vr (14)
(Figs. 9 and 10). It is also possible to calculate the
pneumatic trail as a function of slip angle or side O = Vox2 + Oy2
force, merely by dividing self aligning torque by
the side force (Figs. 11 and 12). The vector õ has the same direction as the slip
speed vector Vs. Therefore we find:
MODEL FOR COMBINED CORNERING AND
BRAKING F = -(ö/o)F(o)
Fx = -(ox/o)F(o) (15)
The formulae obtained which describe the Fy = -(oy/o)F(o)

2.198

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

The relations between the theoretical slip and the

Fyo< Fxo
practical slip quantities used earlier (a and k) are: [N]
8000

- Vsx/Vr = -k/(1 + K) / '

I
(16) / ^ - - FVO

Oy = Vsy/Vr = -tana/(l +k) I Fz = 8000 N

6000 I

where we had:
1 1 ļ Fz - 6000 N
tana ^-Vgy/Vx
(17) 4000 l'i T -
h' /I T - - Fz - 4000 N
- -vsx/vx 11 #7/
« ¡If -

As stated before, the theory is valid for a tyre 'v ■■

II I I .
2000 I It S . ,
with isotropic properties (the same characteristic
for side and brake force). But in reality a tyre is W
not isotropic, so the theory must be adjusted.
Instead of a single characteristic F(o), two
different characteristics should be used (cf. Lugner O 0.1 0.2 0 3 04 0.5 0.6 0.7 O.B 0.9 1.0

[41): Theoretical slip

- for pure cornering: Fyo (o)

- for pure braking: Fxo (o) Fig. 14- Original cal
These characteristics are called the original basic
curves. The total force may now be derived as
follows:

Fx = "(Px/o) F xo (o)
Fyo. Fxo
Fy = -(Oy/o) Fy0(o) (18) |N]

8000

f
F = VFx2 + Fy2 j '

The original basic curves are obtained by calcula-

6000
ting Fx and Fy in pure conditions and writing / Ls"
them ļ" - Fz = 8000 N
as a function of the theoretical slip with the aid of
Eq. (16). Figure 14 shows some of these character-
istics for different vertical loads.
j fi / ! Fz - 6000 N
A problem of physical nature arises when
4000
the I
values of the slip connected with the peak values
of Fxo and Fyo differ considerably from each other.j F z = 4000 N
At a slip o in between these two peaks, we have the
2000 HJ ^ ^ I
situation that total slip would occur on one curve
Ijj
and partial sliding on the other. In reality, the 1
tyre Fz = 2000 N
has one total condition of sliding. To solve this
problem a new slip quantity is defined which
normalizes the theoretical slip. This new slip is 0 1 2 3 4 5

Norm
called the normalized slip a*. The normalized slip
is obtained by dividing o by the slip value which
occurs at the peak of the characteristic. We desig- Fig. 1
nate these values of slip as oxm and oym for the Fx
and the Fy characteristics respectively. The
components of the normalized slip read:

2.199

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 If /I /I
¡rX0 k°*

a~Fx

'.
Nss
! ' F h

,Fy

Fig. 16 - Visualization of determination of magnitude and direction of the total horizontal force F using
the normalized basic characteristics, starting out from a given normalized slip vector (cf. Figs. 17,
18 for the factor e^).

5
Direction factor
i J , 5 i
Direction factor
Ed £d

a = 2 deg.

j a 5 deg
4 - ! 4

Fz = 8000 N
3

Fz - 8000 N

yT Fz 6000 N ^ ^

0
o

0.1 1 10 100 1000 0.1 1 10 100 1000

Normalized slip Normali

o*

Fig. 17 Fig.
- Directi 18-
derived derived from
frommeasured data for a slip
angle ofangle of 52
degrees. degre

2.200

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

Ox* - ox/oxm tion factor (ed), when introduced, yields:

Oy* - Oy/OyjYi (19)
Fx = -(ox*/o*) Fxo (o*) (25)
O* = V0X*2 + Oy*2 Fy - -£d (OyVO*) Fy0 fa*) (26)

The basic curves, but now as a function Figure 16of o* the contruction of the total
shows
have been shown in Fig. 15. On both characteris-
horizontal force. The magnitude of ed will be 1 for
tics (Fxo and Fyo), partial sliding occurs small values
forofvalues
o* and oym/oxm for large values of
of o* smaller than 1 and total sliding o*. For intermediate
for values values of o* the value for ed is
larger than 1. With o* as the new slip the determined from the measured relationships
equations (18) become: during combined cornering and braking. Figures
17 and 18 show examples of the variation of ed
Fx = -(ox*/o*) Fxo (o*) (o*,Fz, a) for different loads and slip angles.
Fy = -(oy*/o*) Fyo (o*) (20) Because of the action of the brake force, the
resulting cornering stiffness of the tyre appears to
F = VFX2 + Fy2 become larger. For this reason, the basic normal-
ized side force characteristic has been given an
The direction of the total force is now based on initial slope which is a function of the brake force.
the direction of o*. For small values of o*, Eq. (20) This can easily be achieved by adjusting the
is correct since then the forces will be independent coefficient B for the side force. The increment of B
and the linear relations hold: is:

Fxo ~ CpK oxm o* AB = -ßFxB (ß>0) (27)

Fx = -(ox*/o*) Fxo (o*) (21)
= -(OX*/O*)CFK with ß denoting the longitudinal f
= -Cfk °x of the cornering stiffness.
Further study is required to e
and for Fy after a similar calculation: physical background of ß and of t
specially at high loads.
Fy = -CFq Oy (22) Figures 19 until 22 show comparisons
between measurements and calculations for a tyre
with Cfk and Cfq the longitudinal slip and
in various situations duringthe
combined cornering
and braking. The graphs depict the variations of
cornering stiffness respectively.
However, for large values the
of o*
brake and
force equal
and the side force as a function of
longitudinal
friction coefficient (p) in x and slip and also the
y direction, side force as a
the
function direction
total force should be in the opposite of the brake force.
of To be able to compare
the
slip speed vector (and thus õ). calculations and measurements, the measured
brake forces were corrected for the rolling
Fxo = pF 2 resistance (vertical shift). With the aid of ed and
Fx = -(ox/o)FX0 (o*) (23) adjustment of B (side force), the model gives a good
= -(ox/o)pFz description of the measurements.
SELF ALIGNING TORQUE DURING COM-
and similarly for Fy: BINED CORNERING AND BRAKING - To obtain
the self aligning torque during combined
Fy = -(oy/o)pFz (24) cornering and braking, the side force has first to be
multiplied by the pneumatic trail (t). The
pneumatic
So with small slip, o* should be used trail has been derived as a function of a
to deter-
mine the direction of the total horizontal force and (or tan a) during pure cornering. The deflection
with large slip, o should be used. Obviously, a need distribution over the contact length during
exists for an additional factor which controls the combined cornering and braking is equal to the
direction of the total horizontal force. This direc- distribution during pure cornering if tan a = a (cf.

2.201

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 8000 8000

Side
Side force Fy [N] o 5 deg 1 Fz 8000 N
Brake force -Fx |Nļ 2 Fz 6000 N

0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100

- Longitudinal slip - k [%ļ -Longitudinal

Fig. 19 - Side Fig.and

forGe 20 - Sidefor
brake fo
slip at 2 degrees slip
slip angle.
at 5 degr

8000
8000-1
Side force Side force
Fy [N] a = 2 deg. Fy [N] Q = 5 deg
..... Measurements

- - Calculations Calculations

6000
6000

Fz = 8000 N

Fz - 6000 N ' '

4000-
4000 ^s/'

Fz = 8000 N ļ '
mmm

0 2000 4000 6000 8000 0 2000 4000 6000 8000

Brake force - Fx [N]

Brake fo

Fig. 21 - Fig.
Side 22
force- Sid
vs
angle. angle.

2.202

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
 870421

[3]). Consequently instead of t (tan a) we may

write t (o) which can be converted into t(o*). The
self aligning torque now becomes:

Mz' = -t(o*) Fy (28) 50 -|

Self
Mz [Nm] 1
Due to carcass compliance, the lines of action of
both the side force and the brake force will be
0
shifted which causes a contribution to the self
i# ••**'*• fļ i
aligning torque:
Fz = 4000 N / /
Mz = Mz'-sFxFy (29)
with

s = J- - J- (30) Fz = 6000 N
v^cy ^cx
-150- -

Ccx and Ccy denoting the longitudinal and Fzlateral

= 8000 N
carcass stiffnesses respectively.
-200 - Calculations
Figures 23 and 24 show comparisons between .... Measurements

calculations and measurements. The measured

a = 2 deg. ^

characteristics have been corrected for an appar- -250 J

ently constant offset of the line of action of the

brake force of approx. 15 mm. This means that Fig
even with a slip angle of zero, a large self aligning deg
torque is present. The overturning couple (Mx) was
not measured, so no information is available about
the offset of the vertical force which may have
caused this. The measured values shown in Figs.
23 and 24 have been corrected for this phenome-
non, which is not yet fully understood. 50

Self aligning
-j

torque :
POSSIBLE EXTENSIONS - The model so far Mz [Nm] •

developed, including interaction between side and

brake force, may be extended with camber and o 11

F
positive longitudinal slip (driving), for instance, if
measurements in these conditions are available. -50- Fz = 4000 N

The measurements described earlier did not

include these combinations of input.
Fz = 6000 N

CONCLUSION AND SUMMARY

-150-

The proposed formula turns out to be not only

very accurate in describing the measured data. It Calculations

-200- ^ ~ ^ Measurements
also characterizes some of the typifying quantities
a = 5 deg.
of the tyre such as slip stiffnesses and peak values
and permits the calculation of forces and torque in -250 -I

conditions which deviate from those imposed

during the actual measurements. Compared with Fig
the conventional way of representing tyre data by deg
polynomials with coefficients which have no direct
physical meaning, this way of representing tyre
data seems to be a big step in the right direction.
The results of the model which describes the
behaviour of the tyre during combined cornering

2.203

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms
and braking are very similar to the measurements APPENDIX
which were carried out in these conditions. This
means that the representation of the tyre dataPROPOSED TYRE FORMULAE
which is valid in pure conditions can also be used
to provide information for a combined condition ifSIDE FORCE:
a new slip quantity is used: the normalized slip.
Because of the physical meaningful coefficients Fy = D sin (C arctan (B<1>)) -I- ASV
of the representation, it is easy to vary some of the with
typifying quantities of the tyre. Simulations may <ļ) = (1-E)(a -I- ASh) + (E/B) arctan(B(a + ASh))
then be carried out with a hypothetical tyre. In
this way, the performance of a specific vehicle can D = aiFz2 + a2Fz
be predicted. C = 1.30
The described mathematical representation is a3sin (a4arctan (asFz))
just the beginning of an effort to describe tyre B = (

behaviour in a way which is accurate, compact, CD

physically meaningful and easy to use. By contin- E = aßFz2 +a7Fz + a8

uing this effort, hopefully, the tyre will be given ASh = a9Y
a sufficiently complete representation in vehicle ASy = (aioFz2 + anFz)y
models as accurate as justified by the great
importance the tyre has in controlling the horizon- SELF ALIGNING TORQUE:
tal motion and therefore in the Dynamic Safety of
the vehicle. Mz = D sin (C arctan (B<1>)) + ASV
with
REFERENCES <t> = (1-E)(a + ASh) + (E/B) arctan(B(a + ASh))

D = aļFz2 + a2Fz
1. A. Sitchin, "Acquisition of Transient Tire
C = 2.40
Force and Moment Data for Dynamic Vehicle
Handling Simulations", SAE Paper No. äßF z2 4- a4F z
831790, November 7-10, 1983 B = (

CDeasFz
2. G. Ruf, "Theoretische Untersuchungen des E = (aßFz2 + a7Fz + a8)/(l-ai3lYl)
Federungsverhaltens von Vierrad-Strassen- ASh = a9Y
fahrzeugen", Fortschr.-Ber. VDI-Z Reihe 12 ASV= (aioFz2 + anFz)y
No. 44 104-105, Dusseldorf 1983
BRAKE FORCE:
3. H.B. Pacejka, "Modelling of the Tyre as a
Vehicle Component with Applications", Lehr- Fx = D sin (C arctan (B<1>))
gang V 2.01, Carl-Cranz-Gesellschaft, 1982 with
<ļ) = (1-E) K + (E/B) arctan (Bk.)
4. P. Lugner, R. Lorenz, E. Schindler, "The
Connection of Theoretical Simulation and D = aiFz2 + a2Fz
Experiments in Passenger Car Dynamics", C = 1.65
Proc. 8th IAVSD Symp. on the dynamics of a8Fz2 +a4Fz
vehicles, Cambridge, Mass., August 15-19, B =

1983, ed. J.K. Hedrick, Swets & Zeitlinger CDeasFz

B.V., Lisse Neth., 1984 E = a8Fz2 + a7Fz + a8

The values for the coefficients ai until a8 are

tabulated in Table 2 and for ag until ai3 in Table 3.

2.204

This content downloaded from 151.100.101.44 on Sat, 25 Apr 2020 16:41:21 UTC
All use subject to https://about.jstor.org/terms