RUBBER ABRASION AND TIRE WEAR

K. A. GROSCH
CONSULTANT TO VMI HOLLAND BV, 8161 EPE, THE NETHERLANDS

ABSTRACT
Rubber abrasion is essentially a mechanical tearing process and has been linked with the rate of a cut growth
process as function of the tearing energy dissipated in the contact area between rubber and track which leads inevitably
to considerable temperature rises at the points of highest stress concentration i.e. at the points at which a mechanical
detachment of abraded particles is most likely. This leads to thermal degradation and oxidation as important secondary
mechanisms which may well decide whether a compound performs better or worse than a reference outweighing the
mechanical properties.
A laboratory method has been developed, using the LAT 100 test equipment, which uses the energy dissipation and
slip speeds in the contact area of a rubber sample wheel, rotating under slip, to evaluate the abrasion performance of a
compound over a wide range of these variables. The resulting equations can be used either to test directly the correlation
between laboratory and road test results which is usually good over a limited range, or they are used in a road test sim-
ulation program with well defined road test conditions. The latter gives a set of single compound ratings and tire lives.
These depend however strongly on the chosen test conditions and in that they reflect reality

CONTENTS
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .471
II. Sliding Abrasion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .471
A. The Load Dependence of Friction . . . . . . . . . . . . . . . . . . . . . . . . . .471
B. The Pressure Dependence of the Sliding Abrasion . . . . . . . . . . . . .472
C. The Sliding Abrasion as Function of the Energy Dissipation . . . . .474
D. The Temperature and Speed Dependence of Sliding Abrasion . . . .479
E. The Appearance of Abraded Rubber Surfaces . . . . . . . . . . . . . . . . .480
F. The Temperature Rise in the Contact Area of Rubber Sliding
Over a Hard Rough Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .482
III. The Abrasion Under Limited Slip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .484
A. Forces and Speed Components of a Slipping Wheel . . . . . . . . . . . .484
B. The Energy and Speed Dependence of Abrasion . . . . . . . . . . . . . . .486
C. The Correlation Between Laboratory Abrasion
and Road Test Ratings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .491
IV. Tire Wear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .493
A. The Influence of the Road Surface . . . . . . . . . . . . . . . . . . . . . . . . .493
B. Tire Construction Influences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .494
C. Driving Influences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .494
D. Road Wear Test Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .496
1. Force Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .497
2. Speed and Load Distributions . . . . . . . . . . . . . . . . . . . . . . . . . .497
3. Results of Road Test Simulation Calculations . . . . . . . . . . . . .498
E. Correlation Between Laboratory Road Test Simulation and
Road Wear Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
1. Correlation with a Set of Passenger Car Tires . . . . . . . . . . . . . .500
2. Correlation with Ratings of Truck Tire Road Tests . . . . . . . . . .501
F. The Energy Consumption and Slip Speeds of Road Wear . . . . . . .502

* Ph: 0049 2471 2339; Fax: 0049 2471 2397; email: Karl-Alfred.Grosch@t-online.de

470
 RUBBER ABRASION AND TIRE WEAR 471

V. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .504
VI. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .504
VII. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .505

I. INTRODUCTION
Abrasion or wear occurs whenever two bodies slide against each other under friction.
Material is transferred from one body to the other and this process goes in both directions. Wear
is therefore associated with friction. High friction is required when forces are transmitted across
two flat surfaces in contact. Research tries to find ways to combine high friction with minimum
wear.
This paper deals with the specific case when forces are transmitted between tire and road
and the wear which occurs on the tire and is affecting its useful life. Although the tire construc-
tion plays an important role in force transmission, the maximum transmittable force is deter-
mined by the friction coefficient of the tread compound under the diverse road conditions which
do occur and the resulting wear by its resistance to it. Surfaces may be dry, wet or icy, rough or
smooth sharp or blunt to various degrees.
It is a unique property of highly elastic materials that the friction coefficient does not only
depend profoundly on the nature of the surface on which it slides but also on the temperature and
sliding speed in the contact area. And this is also true of its wear resistance.
First, the factors will be considered which govern the abrasion of rubber in controlled labo-
ratory experiments, then the major contributing factors to tire wear will be discussed and a link
established between laboratory abrasion and tire wear.

II. SLIDING ABRASION

A. THE LOAD DEPENDENCE OF FRICTION

The frictional force is a function of the load, the force which acts normal to the contact sur-
faces pressing the two bodies together. For friction between two hard bodies this force is gener-
ally proportional to the load and can therefore be written as follows

F = μ•load (1)

FIG. 1. — Friction coefficient of different tread compounds as function

of load on a blunt wet Alumina 180 surface. Speed 0.015 m/s [from Ref. 5]/
472 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

where F is the frictional force, µ is the friction coefficient which is independent of the load. For
rubber this is not always the case. For soft rubbers on smooth surfaces the friction coefficient
decreases with increasing load,1,2,3,4 leading to the following general relation
n
⎧⎪ L ⎫⎪
μ = μ0 ⎨ ⎬ (1a)
⎪⎩ L0 ⎪⎭

For tire tread compounds on rough surfaces the load dependence is small and often absent
altogether. The author has carried out extensive experiments with different tread compounds on
tracks of different asperity shape and coarseness and has found that for these hard compounds
the power index of the friction coefficient ranged between -.1 to +.1 but was in many cases very
small so that the load dependence may be neglected,5 as shown in Figure 1 and Table I which
gives the friction coefficients µo at a load Lo of 100 N and the power indices n.

TABLE I
PARAMETER µ AND THE POWER INDEX n OF THE FRICTION LOAD RELATIONSHIP UNDER DIFFERENT SURFACE CONDITIONS

B. THE PRESSURE DEPENDENCE OF THE SLIDING ABRASION

An abrasion experiment is carried out by sliding a rubber sample over a given distance and
load and the volume loss is determined by some suitable method, usually by determining the dif-
ference in weight before and after. It is always assumed that the abraded volume is proportional
to the distance covered between measurements which is reasonable if the sharpness of the track
remains constant. Hence the abraded volume is always referred to unit distance covered. This
basic quantity depends on the pressure between the sliding surfaces, the temperature, speed,
topography such as sharpness and coarseness of the asperities of the track and, of course, on the
rubber compound.
In contrast to the frictional force, the resulting abrasion is generally a nonlinear function of
the pressure p

n
⎪⎧ p ⎪⎫
abr = abrref ⋅⎨ ⎬ (2)
⎪⎩ p0 ⎪⎭
 RUBBER ABRASION AND TIRE WEAR 473

FIG. 2. — Abrasion loss as function of pressure for a BR tread compound on four different abrasive surfaces.
[from Ref. 6],  tarmac,  Akron abrasive disk,  concrete,  concrete II.

with a distinct positive power index n. For a constant apparent area of contact, a similar relation
results for the load dependence. The reference abrasion loss abrref at the reference pressure po and
the power index n depend both on the type of track and the rubber compound. If the abrasion is
referred to the frictional energy dissipation, usually referred to as the abradability of the rubber
compound, the power index would be modified, if there is a load dependence of the friction coef-
ficient and would remain the same if there is not.
If the track is smooth the abrasive loss may be so small that it deludes normal measurement,
although the frictional force may be very high, whilst on rough sharp surfaces the abrasive loss
may be pronounced at a moderate frictional force. It follows that stress concentrations in the con-
tact area enhance the abrasion process at similar external forces. Figure 2 shows the abrasion loss
of a BR tread compound on four surfaces of different sharpness, as function of the pressure both
plotted on logarithmic scales. The straight line graphs differed both for their reference loss as
well as for their power index n.6 The lowest index n was obtained on the flat side of an silicone
carbide grinding wheel. The others were a tarmac road surface and two concrete floor surfaces.
The power index was always significantly greater than 1. It is a general observation that the
power index becomes larger the blunter the surface on which the abrasion takes place. On very
sharp tracks, on the other hand, the power index may approach 1 as shown in Figure 3.7
474 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

FIG. 3. — Abrasion loss as function of pressure for four different compounds

on a sharp silicone carbide track [from Ref .7]:
A SBR+50 HAF black B NR +50 HAF black
C NR+50 thermal black D NR+50 activated CaCO3

C. THE SLIDING ABRASION AS FUNCTION OF THE ENERGY DISSIPATION

The energy dissipation per unit distance is equal to the frictional force. If its (weak) pressure
dependence is expressed as a power function the index of the abrasion pressure relation would
be slightly modified but the dependence of the sliding abrasion on the energy dissipation would
also be a power law which would clearly be more fundamental than its pressure dependence. This
relation reveals the abrasion process as a mechanical tearing process. Particles are detached from
the rubber due to the cyclic deformations occurring when passing over the asperities of the track.
Figure 4 shows the rates of cut growth in cyclic tearing experiments of rubber samples.8
Straight line graphs are obtained over a wide range of tearing energies when plotting the data on
a log-log scale. The slopes of these lines depend on the type of polymer but not much on the
filler content. Only the rate of cut growth is reduced.
Basic abrasion experiments using a razor blade to scrape a rubber surface repeatedly show
that the abrasion loss correlates well with the rate of cut growth as shown in Figure 5 for differ-
ent gum rubbers. Rates of cut growth and abrasion are plotted against tearing- and friction ener-
gy respectively.9 The slopes of the lines are virtually the same, the difference of the ordinate at a
given energy gives the abrasion per unit cut growth length per cycle.
 RUBBER ABRASION AND TIRE WEAR 475

 ) gum rubber, (+) loaded

FIG. 4. – Rate of cut growth for six different rubbers (
with 20 pphr reinforcing black and () with 50 pphr reinforcing black [from Ref. 8].

FIG. 5. — Log abrasion loss by a blade (solid lines) and log cut growth rate (dashed lines)
of non-crystallizing rubber compounds as function of log frictional and log
tearing energy, respectively. Legend: 1 isomerized NR, 2 SBR, 3 ABR [from Ref. 9].
476 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

Figure 6 shows rates of abrasion loss for gum rubbers and for a black filled BR compound
in a similar abrasion experiment.10 In this case straight line graphs are still obtained when plot-
ted on log-log scales, but the slope of the filled BR compound is distinctly smaller than for the
corresponding gum compound. This appears to be the case for all types of filled compounds. The
power indices of abrasion for black filled tire tread compounds range usually between 1 and 2
depending on the sharpness of the track whilst in cut growth experiments they range between 2
and 6 for both gum and black filled compounds.
The rate of abrasion on very smooth tracks can become exceedingly small while on rough
surfaces it may be more than 103 higher. This is due to the stress concentrations which form when
rubber slides over the asperities of rough surfaces under a normal load. Schallamach has demon-
strated in a two- dimensional transparent model rubber,11 shown in Figure 7, how stresses are dis-
tributed in the rubber due to a wedge, under a normal force and when in addition a frictional force
is applied. Under a normal force the lines of equal stress are circles with their centers underneath
the load force. When the rubber slides adhesion friction produces a tangential stress and lines of
equal stress tilt.

FIG. 6. — Log abrasion by a blade as function of log friction energy for

three gum-and a black filled rubber [from Ref. 10].
 RUBBER ABRASION AND TIRE WEAR 477

FIG. 7. — Two-dimensional stress pattern in a transparent rubber block under a line force [from Ref. 11].

Center and point of force application form an angle tan(α) with the surface plane of the rub-
ber. as shown in Figure 8.12 This leads to a strong stress gradient behind the point of the force
application. Figure 9 (upper) shows the elastic stored energy of two line forces acting at an angle
α to the surface of the semi-infinite body simultaneously and Figure 9 (lower) the horizontal
stress component.12 As expected both energy and stress show strong peaks at the points of contact.

FIG. 8. — Calculated lines of equal stress for a line force made up of normal load
and frictional force acting on the surface of a semi-infinite body.
478 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

FIG. 9. — Calculated stored elastic energy and the horizontal stress component due
to two line forces at an angle α to the plane of the rubber surface and a fixed distance x apart for
different depths from the surface of a semi-infinite body [from Ref 12].

The stress is in compression in front of the contact i.e. the friction component pushes. At the
end of the first contact the stress is strongly in tension, decreasing rapidly away from the contact
and passes again into compression at midpoint between the two contacts. If more contact points
are considered this pattern is repeated. The stress is also limited to a small range of depths. If the
distance between the two contact points is 1 mm then at a depth of 0.04 mm the horizontal stress
component has dropped to about 20% to that of the value at 0.01 mm depth. The stress at the sur-
face itself cannot be calculated because the point has an infinitely small contact area and hence
the stress becomes infinitely large. In reality the force is due to an asperity with a small but meas-
urable tip radius which produces a well defined contact area particularly between a hard asperi-
ty and a soft rubber compound, keeping the contact force finite. Nevertheless, the smaller the
contact radius, the higher are the stresses at the contact points. The contact radius defines the
sharpness of the wedge and in a three-dimensional case the sharpness of an asperity. The stored
energy between the line forces does not become zero and a tensile stress acts at the last of the
line forces and a compressive stress at the front. The energy peaks at the contact points produce
a crack growth at each pass, starting from small flaws This is the mechanical contribution to the
abrasion process.
 RUBBER ABRASION AND TIRE WEAR 479

D. THE TEMPERATURE AND SPEED DEPENDENCE OF SLIDING ABRASION

The sliding abrasion of simple unfilled, non-crystallizing gum rubbers as function of tem-
perature and speed follows the WLF transformation as shown in Figure 10, provided that the slid-
ing speed is so low that temperature rises due to friction are small.12 The abrasion loss decreas-
es with decreasing temperature reaching a minimum value.13 Below the temperature of minimum
abrasion it begins to rise sharply. This is accompanied by a drastic change of appearance of the
abraded surface. It appears that the lowest abrasion loss occurs when the rubber has its maximum
extensibility, which is reasonable because this corresponds to the highest energy density at break.
Abrasion temperature curves match with energy density at break curves if the abrasion sliding
speed of 1m/s corresponds to an extension rate of 104/s. i.e. the rubber which is being extended
to break is of a very small dimension.

FIG. 10. — Sliding abradability as function of the variable log aTv for four
non-crystallizing rubber compounds on dusted silicon carbide 180 [from Ref. 13].

FIG. 11. — Sliding abrasion of three different tread- and corresponding gum compounds as function of temperature
at a sliding speed of 0.01 m/s (a) SBR, (b) ANR, (c) NR, ----- tread compound — gum compound [from Ref. 13].
480 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

The WLF transform works only well for abrasion of non-crystallizing gum rubbers. For tire
tread compounds the temperature dependence is smaller. However, in this case, too, the abrasion
loss reaches a minimum at a particular temperature at a constant speed as shown in Figure 11.
for three rubbers (A) SBR, (B) ABR and (C) NR, the solid lines are for the gum rubbers, the dot-
ted one for the same polymer, filled with 50 HAF black.13 A surprising result is that the tread
compounds have a much higher abrasion loss than the gum rubbers. The reason will become
apparent when examining the appearance of the abraded surfaces.

E. THE APPEARANCE OF ABRADED RUBBER SURFACES

At normal temperatures the surface of a rubber sample which has been subjected to sliding
abrasion has a ridge pattern, the direction of the ridges being normal to the sliding direction of
the abrasion generally called an abrasion- or Schallamach pattern.14 Across-sectional view
reveals that the ridges are undercut, the lips pointing against the abrasion direction. During slid-
ing an asperity of the track pulls the lip over and extends it until the adhesion force can no longer
hold it and it snaps back. However, as the tip extends the stress at the bottom of the ridge rises
and a crack can grow to a limited extent. After a number of passes of an asperity the lip has
become long enough for a particle to be detached. During the sliding the pattern moves slowly
in the direction of abrasion. This abrasion mechanism prevails for soft rubbers with a high exten-
sibility and the higher the extensibility, the coarser the abrasion pattern becomes, as seen in
Figure 12 which was obtained when carrying out the abrasion-temperature experiments
described above. Although the samples were rotated through 90° at regular intervals in order to
suppress their formation abrasion pattern did form. Because of the rotation of the sample they
appear as nipples rather than ridges at low log aTv values (high temperatures, low sliding speeds)
(upper photograph of Figure 12). As the minimum of the abrasion is approached, very pro-
nounced ridges formed despite of the rotation (middle photo of fig 12) and the abrasion increased
forming a small hump in the abrasion log aTv curves (Figure 10 above). At still higher log aTv
values scoring marks appear (lower photo, fig. 12). Rubber acts now like a hard solid, in this case
like a plastic material. Because of the high rate of extension both in the abrasion process as also
in the energy to break experiments, this temperature occurs well before the glass transition tem-
perature is reached. The abrasion increases drastically as seen in Figure 10 above in accordance
with the decreasing energy density at break.
 RUBBER ABRASION AND TIRE WEAR 481

FIG. 12. — Abrasion pattern of the ABR gum rubber during the sliding
abrasion experiment at different temperatures.

Tire tread compounds are filled with highly reinforcing black. This means above all that the
compound becomes harder. Non-crystallizing rubbers become also much stronger. Their exten-
sibility, however, is lower than that of the gum compound and accordingly the abrasion pattern
is much more closely spaced and less pronounced. In many cases scoring marks are present at
the same time as abrasion pattern at all temperatures as is shown in Figure 13. The abrasion loss
is then higher for the black filled compound than for the corresponding gum rubber, as shown in
Figure 11 above. Veith has suggested that the presence of scoring marks in black filled rubbers
indicates the presence of regions of plastic behavior.15 This may well be due to the bound rubber
which the strong activity of reinforcing blacks creates in its vicinity.
482 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

FIG. 13. — Abrasion surface appearance of an NR black filled tire tread compound
for sliding abrasion at different temperatures

F. THE TEMPERATURE RISE IN THE CONTACT AREA OF RUBBER SLIDING

OVER A HARD ROUGH SURFACE

Frictional energy and sliding speed produce a temperature rise in the contact area. Figure 14
shows the temperature rise in the contact area as function of speed.16

FIG. 14. — Temperature rise in the contact area of a small steel indentor with an embedded thermo-couple
as function of speed (a) cone tip radius 1 mm, (b) sphere 11 mm diameter. Load on both 2.5 N [from Ref 16].
 RUBBER ABRASION AND TIRE WEAR 483

To obtain these data a thick rubber strip was glued to a tire testing drum and a thermo-cou-
ple imbedded in a small spherical slider to which different loads could be applied was used to
measure the temperature rise It is seen that the temperature rises rapidly with increasing speed
reaching values of over 300 °C. The curves can be described exactly by a simple relation

temprise = const ⋅ v s (3)

where vs is the sliding speed. The constant contains the strength of the heat source, heat conduc-
tivity and dimensions of the heat source.17 The rubber does not burn, although its inflammation
point has been exceeded. This is due to the fact that the temperature decreases rapidly , as soon
as it leaves the contact area. Because of the low heat conductivity the temperature decreases also
rapidly with increasing depth. Figure 15 shows a theoretical calculation of the temperature rise
of a rectangular pad sliding under a constant load and speed over a semi-infinite rubber solid.17
The Figure shows the temperature profile along the pad in the sliding direction. Heat losses
towards the sides have been neglected. The maximum temperature occurs near the end of the pad
and decays rapidly after the pad has passed. At different speeds, the maximum temperature is the-
oretically given by the above Equation (3).

FIG. 15. — Theoretical temperature rise in the contact area of a pad sliding over a semi-infinite
solid for different depths from the surface. Width 2b: 2mm, speed: 3 m/s pressure:
2 Mp, µ=1, Heat conductivity: 0.15 W/m/°K, heat diffusivity: 10-4 m2/s.

These high temperatures have a considerable effect on the rate of abrasion. To some extent
because the rate of cut growth, i.e. the basic physical process responsible for the abrasion increas-
es with increasing temperature. But at such high temperatures chemical action becomes very
important, first due to the thermal degradation (de-polymerization) of the polymer and second
due to the action of oxygen.
484 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

FIG. 16. — Time record of the abrasion loss on a standard Akron grinding wheel in nitrogen and air
of an NR tread compound (a) unprotected and (b) protected with an antioxidant [from Ref.18].

Figure 16 is an abrasion time record which Schallamach obtained on an Akron grinding

wheel when switching from air to Nitrogen and back again. A mixture of Alumina powder and
Fuller’s earth was fed into the nip between sample and track to counteract smearing. The sample
was first run in air until a constant rate of abrasion had been reached. With the introduction of
nitrogen the abrasion loss dropped rapidly, significantly more so for the unprotected compound.
After air was reintroduced the abrasion rose again to approximately the former value. Each time
the atmosphere was changed it took a few readings before a steady state was reached, indicating
that the oxygen required a certain time to influence the surface layer to saturation.18 Smearing,
mentioned above, is a sign of chemical action in the abrasion process. In laboratory abrasion it
is counteracted by feeding a powder into the nib of the contact area between sample and track.
Thermally sensitive compounds, as for instance NR based ones require more powder than insen-
sitive ones, as for instance BR-blends
Practical experience of tire wear under normal summer temperatures ranks the three poly-
mers most used in tires in the following order NR < SBR < BR. This is exactly the same order
in which their decomposition temperatures are ranked.: NR ≈ 250 °C, SBR ≈ 360 °C and BR >
500 °C. In fact, I know of no other reason which could explain the remarkably high abrasion
resistance of BR.

III. ABRASION UNDER LIMITED SLIP

A. FORCES AND SPEED COMPONENTS OF A SLIPPING WHEEL

Figure 17 shows a rubber sample wheel (or a tire) during cornering and the force and speed
components resulting from the slip. When forces are transmitted between an elastic wheel and a
flat track surface part of the contact area sticks to the track until the product of friction coeffi-
cient and pressure exceeds the local stress due to the deformation of the wheel. The wheel slides
 RUBBER ABRASION AND TIRE WEAR 485

back to its neutral position During this process frictional energy is being dissipated, given by

U=F • s. (kJ7km) (4)

FIG. 17. — Forces and velocities on a cornering elastic wheel showing the
two components responsible for abrasion and tire wear.

Where F is the force which is transmitted by the wheel. This can be either a braking-, accel-
erating- or steering force. Braking and accelerating forces act in the plane of the wheel but in
opposite directions, the steering- or side force acts normal to the plane of the wheel. The slip is
defined as the vectorial difference between circumferential speed and forward speed divided by
the forward speed. For steering- or cornering slip s

s = sin(α) (5)

where α is the slip angle for braking slip the slip ranges from 0 to 1 under locked wheel braking,
for acceleration it ranges from 0 to -∞ at spin through. Also, at limited slip abrasion is a function
of the energy dissipation. For cornering the energy dissipation is given by

Us= S • sinα (6)

and for circumferential slip

Uc= B • sl (7)

where S is the side force and B the braking or accelerating force and sl is the slip. The total ener-
gy loss in a slipping wheel under composite slip is then the sum of the two components.
486 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

B. THE ENERGY AND SPEED DEPENDENCE OF ABRASION

The abrasion is then given by

n
⎡U ⎤
A = Ao ⋅ ⎢ ⎥ (8)
⎢⎣ Uo ⎥⎦

Figure 18 shows the log(abrasion loss) as function of log(energy) for two tire tread com-
pounds, based on NR and SBR respectively, on two Alumina surfaces, 60 sharp and 180 blunt-
ed. The data were obtained with the LAT 100 abrasion and traction test equipment.18 The sam-
ple is a solid rubber wheel and can be set at different slip angles, loads and speeds, The generat-
ed side force is measured and hence the energy can be calculated directly. Straight lines graphs
result, The abrasion constant Ao and the slope n, the power index, depend both on the type of
polymer and on the sharpness of the track surface. NR is seen to have a higher power index than
the SBR compound on both surfaces. However on the blunt surface both indices are higher than
on the sharper Alumina 60. On both surfaces there is a cross over of the abrasion loss at low ener-
gies. This agrees with general practical experience and is attributed to the high temperature sen-
sitivity of NR.

FIG. 18. — Log (abrasion) for two tread compounds NR + black and SBR + black on two surfaces
of different sharpness Alumina 60 and Alumina 180 blunt as function of log (energy dissipation) [from Ref.19].
 RUBBER ABRASION AND TIRE WEAR 487

FIG. 19. — Log (abrasion) as function of log speed for three different tire tread compounds.
Load 76 N, slip angle 14.6° [from Ref. 19] Surface: Alumina 60.

Figure 19 shows the abrasion loss of three tire tread compounds as function of speed both
plotted on a logarithmic scale.18 Although the influence is much smaller than that of the energy
and hence the scatter is larger, clear differences emerge between the compounds. Empirically a
straight line graph was applied to the data. The abrasion rises with increasing speed. The slope
of the straight line differs considerably for the three compounds. The SBR has the smallest
dependence on speed and the NR the largest. This is again most easily explained if it is assumed
that the temperature in the contact area rises with speed and both speed and temperature togeth-
er influence the abrasion speed dependence. In this range it would be expected that the abrasion
rises with increasing temperature, both on visco-elastic and thermo-oxidative degradation con-
siderations while it would be expected to fall with increasing speed on visco-elastic reasoning if
the temperature could be kept constant. Clearly the thermo-oxidative effect outweighs the pure-
ly visco-elastic effect. This agrees with the finding that SBR has a much smaller effect than NR.
The SBR/BR blend filled with silica lies between these two. This is not easily explained since
the author knows of no fundamental data for thermal and oxidative resistance of this type of
material, nor its exact visco-elastic properties.
Since both energy and speed give straight line graphs when both variables are plotted on log-
arithmic scales these quantities may be added up. In order to allow for the fact that the slope of
the log abrasion - .log energy lines depend on speed and vice versa a further term has to be
included. The simplest way is to introduce the product of the two variables log (speed) and log
(energy), giving the following equation:

[log(A)] = a + b1[log(U)] + b 2 [log(v)] + b 3 [log(U)] ⋅ [log(v)] (9)

where A is the abrasion volume loss per km, U is the energy dissipation per km and v is the for-
ward speed of the abrasive disk in the contact area relative to the sample wheel.
In order to determine the coefficients of this equation at least four different testing condi-
tions are required, two energy levels given by two slip angle-load combinations and two speeds.
488 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

The slip can also be circumferential instead of a slip angle, important is that the resulting force
arising from the slip is measured in order to obtain a measure of the energy dissipation. In prac-
tice more test conditions are useful, say three energy and three speed levels and for each condi-
tion repeat measurements are required because abrasion is always subject to variation.

TABLE II
REGRESSION- AND CORRELATION COEFFICIENTS BETWEEN ABRASION, SPEED AND ENERGY
DISSIPATION OF FOUR COMPOUNDS ON CORUNDUM 60

The four coefficients are then calculated from the results using the statistical method of least
square deviations from the mean.19,20 Table II shows the coefficients obtained in this way for four
passenger tire tread compounds. These results, too, as well as most of the following, were
obtained using the LAT 100 Laboratory Test Equipment which is commercially available. The
principle is that of the old Akron Standard Abrader but it allows a wide range of slip angles, loads
and speeds. The generated side force is measured and gives therefore a direct value of the ener-
gy . The coefficient a gives the expected abrasion loss at an energy of 1 kJ/km and a speed of 1
km/h. The coefficient b1 is the power index for the abrasion-energy relation at a constant speed
and temperature. It is positive and larger than 1 for all compounds indicating that the abrasion is
governed essentially by a cut-growth process.
The coefficient b2 is related to the abrasion speed relation at constant energy and tempera-
ture. It is often negative showing that at a constant temperature the abrasion would decrease with
increasing speed, emphasizing the visco-elastic nature of abrasion. The coefficient b3 is always
positive and can be taken as the temperature effect due to both energy and speed. It usually out-
weighs the pure speed effect so that the abrasion loss increases both with increasing energy and
speed.

TABLE III
ABRASION LOSSES/km AS FUNCTION OF LOG ENERGY AND LOG SPEED

These coefficients can be used to calculate the best estimate of the abrasion loss over a wide
range of energy- and speed levels on the particular surface on which they were obtained. The
abrasion loss of log(energy) and log(speed) is best presented in tabular form as shown in Table
III. For practical reasons the experimental range is usually limited to the dotted lines of the table.
At low energy values the abrasion rate is very low and experiments are extremely time consum-
ing. Also the repeatability suffers since the influence of environmental fluctuations has a large
 RUBBER ABRASION AND TIRE WEAR 489

impact on the abrasion rate. Notice that the abrasion between the mildest (extrapolated) condi-
tion (upper left) and the most severe condition (lower right) differs by a factor of more than 1000.
Experimentally the factor is still about 200.
More important for practical use is the relative rating of an experimental compound to a
standard reference compound. This is defined as
abrasion standard compound
rel.rating = •100 (10)
abrasion experimental comp.

TABLE IV
RELATIVE ABRASION RATINGS OF FOUR DIFFERENT TREAD COMPOUNDS AS FUNCTION OF LOG ENERGY AND LOG SPEED.

This rates the abrasion resistance of an experimental compound in relation to a reference

compound. There is no single rating but rather one which depends on the severity of the test. This
is often such that reversals in ranking occur. Table IV shows the ratings of the compounds for
which the abrasion coefficients are listed in Table II with compound 1 as the reference. The three
compounds show reversals in ranking with changes in severity which agree closely with road
wear experience. The two compounds based on SBR-BR with silica filler (compounds 2 and 3)
of Table II are better at low energies and speeds than the reference. This type of compound is
known to be better under driving conditions typical for the USA, i.e. speed restriction and long
straight roads lead to low and moderate severity driving, whilst under European traffic conditions
this type of compound is appreciated for its high wet grip under ABS braking but is also known
to be somewhat inferior in wear resistance in relation to the reference.
490 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

TABLE V
RATINGS OF FOUR COMPOUNDS AS FUNCTION OF LOG ENERGY AND LOG SPEED WITH ROAD TEST RATINGS

The NR compound is known to be better under low temperature conditions than the control
but worse under high temperature conditions. Chemically NR has the lowest thermal stability of
the polymers used for tread compounds in tire technology and it has therefore the highest tem-
perature dependence of abrasion and wear. Thus, it is generally accepted that NR has a higher
wear resistance in a moderate climate than, for instance, SBR but a much lower one in hot cli-
mates. In Table IV NR was equal to the control whenever the sample surface temperature was
near room temperature i.e. at low speeds for all energies and at low energies for all speeds. When
 RUBBER ABRASION AND TIRE WEAR 491

both were raised the sample surface temperature rose considerably and the rating of the NR
decreased correspondingly. A further example is shown in Table V. This gives the laboratory rat-
ing as function of log energy and log speed for four passenger tire tread compounds for which
road test ratings were available. They are shown on the left of each table. For compound 1 prac-
tically all laboratory ratings were less than 100 and in the tire road test the compound was also
distinctly poorer than the control. Compound 3 was slightly better than the control. There is a
region of higher ratings as well as one of lower ones. Hence, it is not certain whether the com-
pound is better under all service conditions, whilst compound 4 was much better than the con-
trol over most of the testing conditions and this is reflected in the high road test rating. The com-
pound developer gets therefore a much broader view of the compounds wear potential with this,
admittedly, more time consuming laboratory abrasion method than with a single point test result.
Also differences between different types of black for the same polymer formulation can be made
visible as shown in Table VI. The effects are much smaller than for polymers and the two carbon
blacks are only better than the control over a limited range of energies and speeds

TABLE VI
RELATIVE ABRASION RATINGS OF TWO TIRE TREAD COMPOUNDS BASED ON AN 80 SBR/20BR
POLYMER BLEND WITH TWP DIFFERENT TYPES OF EXPERIMENTAL BLACK FILLER COMPARED
WITH A STANDARD BLACK FILLER AS FUNCTION OF LOG ENERGY AND LOG SPEED

C. THE CORRELATION BETWEEN LABORATORY ABRASION AND ROAD TEST RATINGS

By using regression analysis between the road test ratings and the ratings obtained for any
one cell of the laboratory rating table as shown above (i.e. any particular testing condition) the
correlation coefficient, the regression coefficient and the intercept of a linear regression equation
can be calculated for each laboratory condition. This is shown in Table VII for the data of Table V.
492 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

TABLE VII
CORRELATION BETWEEN LABORATORY RATINGS OF COMPOUNDS OF TABLE V ON ALUMINA 180 AND ROAD TEST RATINGS

If the criterion is used that the regression coefficient should be nearly 1 and the intercept
nearly zero besides a high correlation coefficient the two best testing conditions differ slightly
from those by inspection. The range of a high correlation coefficient is quite large, for a 1:1 cor-
relation, however, it is very limited.
In a project supported by the European Union to improve the quality of re-treaded truck tires
abrasion experiments were carried out on the LAT 100 to evaluate abrasion resistance of the com-
pounds to be used in road tests. These were three basic Polymer blends, NR-SBR, NR-BR, SBR-
BR as well as a 80 SBR/20BR blend and seven different black fillers
The abrasion program carried out on three compounds of different polymer blends with the
same carbon black gave a high correlation with the road test ratings on driven axles of different
use only at high energy levels.

FIG. 20. — Correlation coefficient between road test ratings of three truck tire compounds on re-treaded
tires at two different axles and laboratory ratings as function of log energy and log speed.
 RUBBER ABRASION AND TIRE WEAR 493

Two such correlations are shown in Figure 20, The correlation coefficients are shown as a
three dimensional graph instead of the tabular form used in Table VI above. For seven com-
pounds containing different types of filler with the same polymer formulation of 80 SBR/20 NR
the correlation between road wear and laboratory ratings were narrower as shown in Figure 21.
Notice that the regression coefficient was only about 0.5. This means that the laboratory experi-
ments gave twice as high a discrimination between the compounds than the road wear data.

FIG. 21. — Correlation- and regression coefficients between road test ratings
of seven truck tire compounds 80SBR/20NR differing by type and amount of filler
and laboratory abrasion ratings obtained on LAT 100 testing equipment.

IV. TIRE WEAR

A. INFLUENCE OF THE ROAD SURFACE

Consider first a test car which is driven in an eight hour shift over a prescribed route. This
will have a length of about 600 km. It could be selected to have special features like predomi-
nantly motorway (turnpike or freeway) or lead through windy hilly country. In either case there
will inevitably be a number of changes of road surface structure resulting in an average road sur-
face. By road surface structure is meant the structure which is likely to influence the abrasion
loss significantly, that is its sharpness, and coarseness. The average of a route may still be dif-
ferent for different test routes but the difference between the two averages are already likely to
be smaller than the differences encountered along each route. Moreover, the average of a test
route is likely to reflect the average of a whole geographical region. Differences between the
average road surfaces in highly developed countries are likely to be small whilst larger differ-
ences exist between such countries and less developed ones.
But even the structures of a chosen test route vary with the season. It is well known that road
surfaces in areas of moderate climate are much sharper in winter than in summer. Even in much
shorter periods the sharpness changes often drastically with the weather conditions. Since wear
results are generally obtained over a time span they are inevitably averages over the sharpness
time fluctuations of the road surfaces
This is not too serious if the ranking of test compounds is not strongly influenced by the road
surface structure and in particular by the changes which occur. Nevertheless it is obvious that
repeatability is limited and as the few examples given from laboratory experiments show rever-
sals in ranking are rather the rule than the exception. Hence a single road test result can only have
a limited validity.
494 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

B. TIRE CONSTRUCTION INFLUENCES

Road wear is force controlled. This is a fundamental difference to slip controlled laboratory
abrasion test machines or wear tests with trailers under constant slip or fifth wheel experiments
with a set slip angle. In force controlled events the abrasion loss is inversely proportional to the
stiffness of the tire whilst under slip control the abrasion is proportional to its stiffness. For small
slip which is usually the case for tires the force is proportional to the slip.

Fa,b = Kcsla,b (11)

where Fa,b is either the accelerating or braking component of the force and sl(a,b) the correspon-
ding slip and

Fs = Ks sinα (11b)

where Fs is the cornering force component and α the slip angle. The two stiffness components
have different values, Kc is numerically about twice Ks The corresponding energies for set slip
values are then given by

Ua,b = Kcs2la,b (12a)

and

Us = Ksα2 (12b)

but at a constant force component

2
Fa,b
U a,b = (12a)
Kc

and
Fs2
Us = (12b)
Ks

For a detailed description of a tire model for a complete force slip relation I refer to the lit-
erature12 This has consequences on the results of tire wear tests. In tire tests, a stiffer tire con-
struction reduces the wear. The stiffer construction does not only include the carcass and belt (i.e.
height to width ratio) but also the tread pattern and the stiffness of the tread compound. A cor-
rect road test requires therefore two identical vehicles, each equipped with a test group of four
identical tires. If this is not possible at least one axle has to have identical tires. If this is not the
case an average slip will balance the acting force on the axle. This will be larger for the stiffer
tire than is required if both tires had the same stiffness and smaller for the softer tire. Hence the
wear result is falsified to the advantage of the softer tire. The same argument holds for multi-sec-
tion tires.

C. DRIVING INFLUENCES

The forces acting on a car are due to acceleration and braking, wind-and rolling resistance.
In order to obtain a quantitative estimate of their effects, the route is divided into small sections
 RUBBER ABRASION AND TIRE WEAR 495

over which it may be assumed that the force and speed conditions are constant. Consider first the
cornering acceleration component. Curves will differ in radii and the speeds with which they are
being negotiated. The decisive aspect is the centrifugal acceleration it produces. Together with
the load (mass) on the tire, it determines the force acting on it. These accelerations can be meas-
ured. They are likely to follow a statistical distribution function symmetrical around zero.

FIG. 22. — Measured distribution of (a) cornering and (b) fore and aft accelerations
in a controlled road wear test for passenger car tires.
Figure 22(top) shows the measured cornering force distribution obtained in a controlled road
test for passenger car tire wear. Also shown is a normal distribution adjusted for its width to fit
the data. It is obvious that the real distribution is very similar to a normal one. The same holds
(to a somewhat lesser degree) for the fore and aft acceleration as seen in Figure 22(bottom). It
appears that the maximum of the distribution is slightly shifted towards negative values: The
vehicle is more often accelerated than braked. Some of the necessary braking has been left to
wind- and rolling resistance.
496 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

FIG. 23. — Speed distribution in a controlled road test for passenger car tires
also shown a distribution made up of two super-posed normal ones.

The force due to wind resistance is a function of the speed and the vehicle dimensions. The
rolling resistance is due to the tires and the vehicle. Both are primarily functions of the load. In
controlled road tests the load is usually kept constant, the speed, however is also distributed.
Figure 23 shows a speed distribution obtained on a controlled road test measured over 600 km.
Two super-posed normal distribution functions gave a good agreement with the data. For the gen-
eral user the load is also not constant and can be represented by distribution functions.

D. ROAD WEAR TEST SIMULATION

Using distribution functions for accelerations, speeds and loads it is possible to calculate the
necessary slip for a large number of driving events with the help of the brush tire model, each
event being defined by a cornering acceleration, a longitudinal acceleration (speed up or brak-
ing), the mass of the vehicle determined by the load and weight of the vehicle and the forces to
overcome the wind- and the rolling resistance of tires and vehicle. All these are assumed to be
constant for a short distance, hence the energy dissipation in the contact area due to the resulting
slip can be calculated. Using the laboratory abrasion equation9 the volume loss for each event is
obtained. With distribution functions describing the frequency of their occurrence the volume
loss over the whole distance is obtained. From the available tread volume of the tire estimated
tire lives are obtained.
 RUBBER ABRASION AND TIRE WEAR 497

FIG. 24. — Theoretical acceleration (force) distribution function in a computer simulation,

the resulting energy dissipation multiplied by the frequency of occurrence and the expected abrasion loss,
using the Equation 18 with parameter for a passenger tread compound.

1. Force Distributions. — Normal Distributions around zero are used for the acceleration
components. Their width is defined by 3 times their standard deviations which includes 99.9%
of all events. They are a major indication of the severity of the driving style. Figure 24 shows a
Normal Distribution of the cornering acceleration, the resulting energy dissipation multiplied by
the frequency of occurrence and the abrasion multiplied by the frequency, using laboratory abra-
sion equation9 with data obtained for a test compound on the LAT 100 abrasion tester. The ener-
gy and abrasion curves were normalized so that their heights correspond approximately to that
of the distribution function. The curves are shown as continuous functions for clarity, in a prac-
tical simulation program a limited number of steps has to be used. The maximum energy con-
sumption occurs at positive and negative cornering accelerations corresponding to clockwise a
anti-clockwise cornering Because the abrasion is a nonlinear function of the energy its maximum
is displaced towards higher accelerations.

2. Speed and Load Distributions. — For the speed- and load distribution three super-posed
normal distributions around three fixed mean values are used, corresponding for the speed to
town-, country road- and motorway (turnpike or freeway) traffic. A maximum speed fixes the
total width of the curve from zero to that maximum. This corresponds to 10σ, where σ stands for
standard deviation of the three superposed distributions. The three mean values are fixed at 3, 5
and 7σ. Their heights can be varied according to the frequency with which the three distributions
occur, their sum has to add up to 1. Figure 25 shows two such distribution functions for two dif-
ferent truck tire uses, both with a maximum speed of 110 km/h (which is already outside the legal
limit of 80 km/h in Europe, but occurs occasionally). A similar distribution is also used to
describe the different load conditions with low, medium and high loads, starting with the mini-
mum load corresponding to the empty vehicle weight.
498 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

FIG. 25. — Two theoretical speed distributions made up of three normal

distributions corresponding to town, country road an freeway traveling

With the assumption that the frequencies of the different distributions determining one
event, the frequency of each event is defined by the product of the individual frequencies..
Having calculated the force for a particular event the slip is calculated using the tire brush model
and hence the energy dissipation is obtained. The speed distributions are used to calculate the
wind resistance contribution and the slip speeds due to the total force on the tire.
The forces are different for a driven and a non-driven axle and accordingly different abra-
sion rates will result. The load transfer which occurs during cornering averages out and is there-
fore not included. With the data of the tires, i.e. dimension, tread width, net/gross of the tread
pattern and tread depth, expected tire lives under the above specified conditions are calculated.
This approach has the merit that with an existent set of laboratory abrasion data a large num-
ber of road test simulations with different boundary conditions can be run in a very short time
and their effect on tire life and compound rating can be estimated.

3. Results of Road Test Simulation Calculations. — Table VIII gives a list of the boundary
conditions which define a tire wear test simulation and in fact also an actual road test. The road
surface is the laboratory surface on which the abrasion data for the simulation were obtained.
There is as yet no definition of a road surface and even if there were one, it would be of little use
 RUBBER ABRASION AND TIRE WEAR 499

since road surface structures change frequently along the road surface as pointed out above. The
tire construction influences the wear by both cornering and longitudinal slip stiffness. This
includes the tire carcass, breaker construction, inflation pressure and tread pattern design.
However, since the two stiffness components can be measured, a knowledge of the construction
details is not necessary.

TABLE VIII
LISTING OF THE BOUNDARY CONDITIONS OF A ROAD TEST SIMULATION

The vehicle geometry influences the tire wear through the air resistance which it creates and
through the load distribution between front and rear axle. The driving parameters are determined
by the maximum values of their distribution functions and in case of load and speed by the ratios
between low, medium and high loads and town, country road and high-way traffic respectively.
The tread pattern influences the tire life through its available compound volume. Table IX shows
the results of the simulation with parameters of Table VIII and the abrasion constants of Table II

TABLE IX
RESULTS OF A ROAD TEST SIMULATION WITH THE BOUNDARY CONDITIONS
OF TABLE VII AND THE ABRASION PARAMETER OF TABLE II

The simulation yields a single raring for each compound, which does, however depend on
the boundary conditions of the test simulation. In addition it gives tire lives as the total distance
achievable between new and the minimum legal tread depth of 1.5 mm. Real road test results
were not available for these compounds but the achievable mileages are certainly close to expec-
tation under the assumed boundary conditions.
500 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

E. CORRELATION BETWEEN LABORATORY ROAD TEST SIMULATION AND ROAD WEAR TEST RESULTS

1. Correlation with a Set of Passenger Car Tires. — A road test simulation reduces the com-
parison of the wear resistance of two compounds again to a single number, which is usually
favored by managers and chemists. However this number is now based on a range of boundary
conditions defining the road test.
For the available road test data the boundary conditions were not known.. However because
a road test simulation takes only a very short time and reasonable conditions are not hard to
guess, good and credible correlations can be shown to exist also between laboratory road test
simulation and actual tire wear road test ratings without the exact knowledge of the boundary
conditions.
TABLE X
COMPARISON OF ROAD TEST RATINGS OF TABLE V WITH RATINGS OBTAINED
FROM ROAD TEST SIMULATIONS UNDER THREE DIFFERENT SEVERITIES
 RUBBER ABRASION AND TIRE WEAR 501

Table X shows three road test simulations of increasing severity by raising the maximum
acceleration components in three steps. The remainder of the boundary conditions were only
loosely known and were therefore guessed. The calculated ratings are compared with the road
test ratings. At a maximum acceleration for both components of 0.35 g a very good correlation
is achieved. This corresponds to a hard driving style for the test tires on the driven axle, but one
which is common for tire test drivers and the achieved tire lives are accordingly short and are in
agreement with this style of driving. It is not known what they really were but they are realistic.

2. Correlation with Ratings of Truck Tire Road Tests. — Table XI shows the ratings of a lab-
oratory wear simulation and the average road test ratings obtained for the three re-tread truck tire
compounds used in a European Commission project to improve retread tire quality. Because on
all vehicles either multi-section tires were mounted or two groups were mounted on the same
axle, in each case the comparison occurred under equal slip conditions rather than equal force.
In the first instance a typical cornering stiffness for the truck tire size 275/80 R 22.5 was used to
calculate tire lives and ratings in relation to compound 1. This compares the compounds under
equal energy conditions.
Considering the tire to act like two shear springs in series the following relation can be used
to estimate the relative compound contribution to the total slip stiffness (cornering or
braking/acceleration) of the tire) if there are significant differences between the stiffness of the
experimental compounds as was the case with these three compounds.

ρ(1 + ϕ )
K x = Ko (20)
(1 + ϕρ)

where Ko is the Slip stiffness component of the control tire, ϕ is the ratio between the tread –and
carcass stiffness of the control and ρ is the stiffness ratio of the experimental to the control tread.
The factor ρ was estimated from side force measurements at small slip angle and high speeds.
The factor ϕ is unknown but was adjusted to give the highest correlation between road test and
laboratory simulation.
502 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

TABLE XI
COMPARISON OF ROAD TEST RATINGS OF THREE TIRE RETREAD COMPOUNDS
WITH A ROAD TEST SIMULATION UNDER EQUAL FORCE-, ENERGY- AND SLIP CONDITIONS,
RESPECTIVELY, BECAUSE THE ROAD TEST RESULTS WERE OBTAINED WITH MULTI-SECTION TIRES

In the second case, the stiffness was calculated with Equation (20) and a factor ϕ = 0.3 This
was the simplest way to estimate what would have happened under equal force conditions.
Thirdly, The basic slip stiffness components Ko were multiplied with the inverse of the ratio of
Equation (20). This corresponds to running the tires under equal slip conditions.
Since no data were available apart from a description of the general use of the truck (most-
ly short distance haulage) and the axle position on which the tires were mounted, again reason-
able assumptions on maximum accelerations and speeds as well as load distributions had to be
made. They are listed in Table XI together with the three comparisons.
The boundary conditions chosen are reasonable for the kind of application under which the
tires operated and the achieved mileages agree well with the ones obtained from the simulation.
It is seen that a good agreement is only reached if equal slip conditions are assumed, with a cor-
relation coefficient of 0.977.

F. THE ENERGY CONSUMPTION AND SLIP SPEEDS IN ROAD WEAR.

The road test simulation program also lists the energy consumption and the slip speeds
which occur for each event taking part in the simulation. These quantities can be presented in
form of frequency distributions. The two energy distributions for a typical passenger tire- and a
truck tire simulation are shown in Figure 26, Notice that they are almost identical for passenger
and truck tires, although the boundary conditions are very different. The reason is twofold: First,
 RUBBER ABRASION AND TIRE WEAR 503

the truck tires are much stiffer and hence the resulting slip angle for the same force is much
smaller. Second accelerations in truck tire use are distinctly lower than for cars. Apart from
engine power/kg mass being much lower (7 to 10 kW / to for trucks compared with about 70 kW
/ to for passenger cars), too high an acceleration could well damage the goods carried.

FIG. 26. — Energy distributions in a typical passenger car- and truck tire road test simulation.
Also shown is the range of the energies used in laboratory abrasion experiments.

In addition the laboratory range of energies, marked by the blue box, is also close to the
range of the energies in practical tire use. The reason for this is that much larger slip angles are
used at a much smaller force. A similar situation exists for the slip speeds, as shown in Figure
27. They are lower for trucks than for passenger tires but both are within the slip speed range of
the laboratory abrasion tests.

FIG. 27. — Slip speed distributions in a typical passenger car- and truck tire road test simulation.
Also shown is the range of the slip speeds used in laboratory abrasion experiments.

Hence the extrapolation between a laboratory abrasion test and tire wear in general use is
much smaller than would be expected at first sight. It is therefore not surprising that the labora-
tory test procedure, which admittedly is much more complex than a single point reading, gives a
close correlation for ratings and indeed also with mileages.
504 RUBBER CHEMISTRY AND TECHNOLOGY VOL. 81

V. CONCLUSION
Basically abrasion is a cut growth process. This process is understood to be governed by the
tearing energy as defined by Rivlin and Thomas21 50 years ago. Schallamach was the first to con-
ceive abrasion as a function of the energy dissipation in the contact area of rubber and track.
Basic sliding abrasion experiments demonstrate the close relation between the rate of cut growth
- tearing energy relation on the one hand and the abrasion - energy consumption relation on the
other. The relation between the abrasion and the energy consumption as a power law is now firm-
ly established and holds over a wide range of energies. Power index and constant depend both
on compound and on the abrasive track. The prime effect is thereby due to the sharpness of the
track rather than its coarseness This influences both the level of abrasion as well as the power
index of the abrasion energy relation of the polymer filler system. Both processes, abrasion and
cut-growth, are not purely physical phenomenon but are also subject to oxidation.
For abrasion this is, however, a much more dominating process than for cut growth. The
main reason is that the energy consumption in the abrasion process raises the temperature in the
interface between rubber and track and thereby modifies this process. The temperature in the
contact patch is a function of the power consumption and depends, therefore, also on the sliding
speed. The temperature does not only influence the oxidation and cut growth process but also
causes thermal degradation.
The physical and the chemical processes are so strong that they lead to different behavior of
compounds under different service conditions and reversals in compound ranking are common.
It is therefore in principle not possible to design a single laboratory abrasion test which could
reflect the practical experience of service conditions. A range of energies and speeds are a min-
imum requirement. For this, complete sliding abrasion experiments are too severe and methods
which operate under limited slip are essential to cover the wide range of severities necessary for
a comprehensive understanding of the abrasion resistance of tire tread compounds.
Computer simulation of road test conditions has shown that both these quantities can be set
in laboratory experiments in the same range as occur in practical service, increasing the likeli-
hood of producing compound ratings which reflect the practical experience.
The question of the influence of the road surface on tire wear cannot be answered unequiv-
ocally because of the large number of different compositions, state of use and weather influences
on their abrasive power. Road surfaces are also not durable enough for laboratory use. Hence
reliance has to be placed on the correlation between laboratory results on a laboratory abrasive
surface and road test experience. Alumina of different grain size (but primarily 60) has proved to
be the most useful. Even its sharpness changes with time of use and disks are limited in their use-
ful life.
The link between laboratory abrasion methods and road tests is now well established so that
further research can concentrate on further elucidation of the basic underlying processes. This
concerns particularly the interaction between filler systems and polymers which is continuously
expanding through both new polymers and new fillers. But now the link has not to extend from
a basic experiment to tire road testing. It can be limited to a laboratory abrasion testing method
and suitable basic experiments.

VI. ACKNOWLEDGMENTS
This paper is based on Chapter 13 the author wrote for “The Pneumatic Tire” ed. by A.N.
Gent and J.D. Walter. Together with the paper “Rubber Friction and its Relation to Tire Traction,”
in Rubber Chemistry and Technology (July-August 2007) which itself is based on Chapter 11 of
the above mentioned book, a short path is traced from basic research to the complex applications
of tire traction and tire wear.
 RUBBER ABRASION AND TIRE WEAR 505

VII. REFERENCES
1F. P. Bowden and D. A. Tabor, "Friction and Lubrication of Solids," Oxford University Press (London) 1954.
2P. Thirion, Rev. Gen. Caoutch. 23, 101 (1946).
3A. Schallamach, Proc. Phys. Soc. London B65, 657 (1952).
4A. Schallamach, Wear 1, 384 (1958).
5K. A. Grosch, RUBBER CHEM. TECHNOL. 69, 495 (1996).
6K.A. Grosch and A. Schallamach, Kautsch. Gummi Kunstst. 22, 288 (1969).
7A. Schallamach and K.A. Grosch, “Tire Traction and Wear” in Mechanics of Pneumatic Tires, S. K. Clark, Ed.,
Washington, D.C: U.S. Dept. of Transportation (NHTSA), 1981, ch. 6, p. 365-474.
8G. J. Lake and P. B. Lindley, Rubber J. 146, 24 (1964).
9D. H. Champ, E. Southern, and A. G. Thomas, "Fracture Mechanics Applied to Rubber Abrasion” in Advances in
Polymer Friction and Wear, Lieng-Huang Lee, Ed., Plenum, New York and London, 1974, pp. 133-146.
10A. N. Gent and C. T.R. Pulford, J. Appl. Polym. Sci. 28, 943 (1983).
11A. Schallamach, Wear 13, 13 (1969).
12K.A. Grosch, “Rubber Abrasion and Tire Wear” in The Pneumatic Tire, A. N. Gent and J. D. Walter, Eds., U.S. Dept.
of Transportation (NHSTA), 2005, ch. 13, p. 545.
13K. A. Grosch and A. Schallamach, Trans. Inst. Rubber Ind. 41, T80 (1965); RUBBER CHEM. TECHNOL. 39, 287 (1966).
14A. Schallamach, Trans. Inst. Rubber Ind. 28, 256 (1952).
15A. Veith, Polym. Test. 7, 177 (1987).
16K. A. Grosch, "The Speed and Temperature Dependence of Rubber Friction and its Bearing on the Skid Resistance of
Tires" in The Physics of Tire Traction, Theory and Experiment, D.F. Hays and A.L. Browne, Eds., Plenum Press: New
York, 1974, p. 143-166.
17H. S. Carslaw and J. C. Jaeger, Conduction of Heat in Solids, Oxford University Press, 1959, p. 279.
18A. Schallamach, Appl. Polym. Sci. 12, 281 (1968).
19K. A. Grosch, H. Moneypenny, I. H. Wallace, A Correlation Between Truck Tyre Road Wear Performance and a System
of Laboratory Abrasion Tests, IRC 2001, Birmingham: IOM Communications, 2001, pp. 307-326.
20K. A. Grosch, Correlation Between Road Wear of Tires and Computer Road Wear Simulation Using Laboratory
Abrasion Data, Rubber Division, ACS Meeting, Cleveland, OH, Fall 2001, Paper 48.
21R. S. Rivlin and A. G. Thomas, J. Polym. Sci. 10, 291 (1953).

[ Paper 88 presented at the Fall Rubber Division, ACS, (Cleveland, OH)

October 16-18, 2007, revised May 2008]