Available online at www.sciencedirect.

com

ScienceDirect
IFAC PapersOnLine 53-2 (2020) 14456–14461
Optimal
Optimal Tyre
Management
Management of
Tyre of a
a Formula
Formula
Optimal
Optimal Tyre
Management
Tyre
Management
One car of
of a
a Formula
Formula
Optimal One
Tyre
One car
Management
car of a Formula
One car
Wilhelm
Wilhelm J. One
J. West
West
∗
carJ.N. Limebeer ∗,∗∗
∗ David ∗,∗∗
∗ David J.N. Limebeer ∗,∗∗
Wilhelm
Wilhelm J. West ∗ David J.N. Limebeer ∗,∗∗
∗ Wilhelm J. J. West
West ∗ David David J.N. J.N. Limebeer
Limebeer ∗,∗∗
∗ Department
DepartmentWilhelm of Electrical, Electronic
∗
J. WestElectronic
of Electrical, David J.N. and
and Computer
LimebeerEngineering,
Computer
∗,∗∗
Engineering,
∗ University of Pretoria, Pretoria, South Africa (e-mail:
∗
∗ Department
Department
University of Electrical,
of Electrical,
Electrical, Electronic
Electronic
Pretoria,and
of Pretoria,Electronic and
South Computer
Computer Engineering,
Engineering,
Africa (e-mail:
Department of wilhelm.west@tuks.co.za). and Computer Engineering,
∗ University
University
Department of Pretoria,
of Pretoria,
of Electrical, Pretoria,and
Pretoria,
Electronic
wilhelm.west@tuks.co.za). South
South Africa (e-mail:
Africa
Computer (e-mail:
Engineering,
∗∗ University of Pretoria, Pretoria, South Africa (e-mail:
∗∗ Department
University of
∗∗ Department ofof Electrical
Electrical and
and Electronic
wilhelm.west@tuks.co.za).
wilhelm.west@tuks.co.za).
Pretoria, Pretoria,
Electronic
wilhelm.west@tuks.co.za). SouthEngineering
Africa (e-mail:
Engineering Science,
Science,
University
∗∗ Department
∗∗ Department of Johannesburg,
of
of Electrical Johannesburg,
and Electronic South
Engineering Africa
South AfricaScience, (e-mail:
University
Department of Electrical
Electrical and Electronic
wilhelm.west@tuks.co.za).
of Johannesburg, Johannesburg,
and Electronic
david.limebeer@gmail.com)
Engineering
Engineering Science,
(e-mail:
Science,
University
∗∗
University of
of Johannesburg, Johannesburg, South Africa (e-mail:
Department
University of Johannesburg,
of Electrical
Johannesburg, Johannesburg,
and Electronic Engineering
david.limebeer@gmail.com)
Johannesburg, South
South Africa (e-mail:
AfricaScience,
(e-mail:
University of Johannesburg, david.limebeer@gmail.com)
david.limebeer@gmail.com)
Johannesburg, South Africa (e-mail:
david.limebeer@gmail.com)
Abstract: david.limebeer@gmail.com)
Abstract: Optimal Optimal control
control calculations
calculations are are used
used to to study
study the the effect
effect of of tyre
tyre wear wear on on racerace
car performance.
Abstract:
Abstract:
car performance. Optimal
Optimal This is
control
control
Thiscontrol achieved by
calculations
calculations
is achieved solving
by solving are
are a
used
used minimum
a minimum to
to study
study lap the
the
lapthe time
effect
effect
time optimal
of
of
optimal tyre
tyre control
wear
wear
control problem
on race
on
problemrace
Abstract:
over multiple Optimal
laps using a calculations
dynamical modelare used
of to study effect of tyre wear on race
car
car performance.
performance.
Abstract:
over multiple Optimal
laps This
This is
is
control
using a achieved
achieved by
by
calculations
dynamical solving
solving
modelare aa aaminimum
used
of Formula
minimum
to
Formula study One
lap
lap
Onethe car.
time
time
effect
car. A
A previously
optimal
optimal
of tyre control
control
previously wear developed
problem
problem
on
developed race
car performance.
thermodynamic This
model is
is achieved
enhanced by solving
by adding ana minimum
additional lap
state timefor optimal
the carcass control problem
temperature
over
over
car multiple
multiple
performance.
thermodynamic laps
laps using
using
This
model is
is a
a dynamical
dynamical
achieved
enhanced by model
model
solving
adding of
of
ana aaminimum
Formula
Formula
additional One
One
lap
state car.
car.
timefor A
A
the previously
previously
optimalcarcass control developed
developed
problem
temperature
over
of the multiple
tyres. lapstyreusing a isdynamical model of a Formula One car. Aandpreviously developed
over
of tyres. The
thermodynamic
thermodynamic
themultiple
thermodynamic laps
The model
model
using
tyre
model
grip
is
is
grip
isworn,
modelled
aenhanced
enhanced
isdynamical
modelled
enhanced
by
by as
as aa function
adding
adding
model
bytyres
adding
an
an
of
function
an
of
of the
aadditional
additional
Formula
additional the tyres’
state
state
One wear
state car.
tyres’ for
for
wear
for
the
the
Aand
the
temperature.
carcass
carcass
previously temperature
temperature
temperature.
carcass developed
temperature
Grip
Grip
reduces
of
of the
the
thermodynamic
reduces when
tyres.
tyres.
whenThe tyres
The
The
tyres get
tyre
tyre
model grip
grip
isworn,
getgrip is
is or the
modelled
modelled
enhancedor thebytyres as
as
addinga
a are not
function
function
are an operated
notadditionalof
of
operated the
the within
tyres’
tyres’ wear
wear
state wear
within fortheirand
and
the
their optimal temperature
temperature.
temperature.
carcass
optimal temperature Grip
Grip
of the
window. tyres.
Overheating tyre the is
tyres modelled
can as
accelerate a function
wear, of
which the tyres’ and temperature. Grip
reduces
reduces
of when
when
the tyres.
window.
reduces when
tyres
tyres
The
Overheating
tyres
get
get
tyre
get
worn,
worn,
grip
the worn,
or
or
is modelled
tyres orcanthe
the tyres
as aare
tyres
theaccelerate
tyres are not
not
function
are wear,
notthe of thein
operated
operated
which
operated
turn,
within
within
intyres’
turn,wear
within
degrades
their
theirand
degrades
their
optimal
optimalperformance.
temperature
temperature
temperature.
optimalperformance.
temperature
The
Grip
The
optimal
window.
window.
reduces
optimal control
Overheating
Overheating
when
control problem
tyres get
problem the
the solver
tyres
tyres
worn,
solver needs
orcan
canthe
needs to
tyres
to ‘manage’
accelerate
accelerate are
‘manage’ wear,
wear,
not the state
which
which
operated
state of
in
in
of tyres
turn,
turn,
within
tyres throughout
degrades
degrades
their
throughoutoptimal a race
performance.
performance.
a (not
temperature
race (not just
The
The
just
awindow.
single Overheating
lap) to ensure the
that tyres
optimal can race
accelerate wear, which
performance is in turn,
achieved. At degrades
some point performance.
during a raceThe
optimal
optimal
awindow. control
control
single lap)Overheating problem
problem
to ensure the
thatsolver
solver
tyres
optimalneeds
needs
can race to
to ‘manage’
‘manage’
accelerate wear,
performance the
the state
state
which
isstate of
of
in
achieved. tyres
tyres
turn, throughout
throughout
degrades
Atthroughout
some point a
a race
race
performance.
during (not
(not
a race justaa
just
The
optimal
pit stop control
may be problem
required solver
to change needs wornto ‘manage’
tyres so the
that tyre of
grip tyres
can be restored. a
It race
is (not
essential just
to
a
a single
single
optimal
pit lap)
lap)
control
stop lap)
may to
to ensure
ensure
problem
beensure
required that
that optimal
optimal
solver
to optimal
change needs race
race
worn performance
performance
to performance
‘manage’
tyres so that the is is
is achieved.
achieved.
state
tyre of tyres
grip At
At
canAtbe some
some
throughoutpoint
point
restored. during
during
aduring
It race a
a
(not
is essential race
race toaaa
just
a single
understand to
the wear that
characteristics race
of various tyre achieved.
compounds in some
order to point
determine a
the race
point
pit
pit stop
stop may
may be
be required
required to
to change
change worn
worn tyres
tyres so
so that
that tyre
tyre grip
grip can
can be
be restored.
restored. It
It is essential toa
a single
understand
pit stop
when the
lap)
maytime
to
the ensure
wear
beneeded
required that
for
optimal
characteristics
toa change
pit stop
race
of
worn
is
performance
various tyre
tyres sointhat
justified terms
is achieved.
compounds
tyreofgrip in At some
order to
can be restored.
subsequent
point
determine
racing It is essential to
is
during essential
performance.
a
the race
point to
understand
understand the
the wear
wear characteristics
characteristics of
of various
is various tyre
tyre compounds
compounds in
in order to determine the point
pit
when stop
themay
understand time
thebeneeded
required
wear fortoa change
pit stopworn
characteristics of tyres so
justified
various inthat
tyre terms tyreofgrip
compounds in order
can
subsequent order to
to determine
be restored.
racing performance.
determine the
the point
It is essential point to
when
Copyright
when the time
the time
© the
2020needed for a
The Authors.
needed for a pitThis stop
is anis open
justified
access inarticle
terms of subsequent
under subsequent
thein racing
BY-NC-ND
CCorder performance.
license the point
understand
when the time wear
needed for a pit
characteristics
pit stop
stop is
of justified
is various
justified in
in terms
tyre compounds
terms of
of subsequent racing performance.
to determine
racing performance.
(http://creativecommons.org/
Keywords:
when
Keywords: the time Thermal
needed
Thermal tyre
tyre alicenses/by-nc-nd/4.0)
formodel,
model, tyre
tyre iswear,
pit stop Formula
justified
wear, Formula in terms One
One of car modelling
carsubsequent
modelling racing performance.
Keywords:
Keywords: Thermal Thermal
Thermal tyre tyre model,
tyre model,
model, tyre tyre wear,
tyre wear, Formula
wear, Formula
Formula One One
One car car modelling
car modelling
modelling
Keywords:
Keywords: 1. INTRODUCTION
Thermal
1. INTRODUCTION tyre model, tyre wear, Formula One 2.
car TYRE
2. TYREmodelling THERMODYNAMIC
THERMODYNAMIC MODEL MODEL
1. INTRODUCTION
1. INTRODUCTION
INTRODUCTION 2. TYRE
2. TYRE THERMODYNAMIC
THERMODYNAMIC MODEL
TYRE THERMODYNAMIC MODEL
1. 2. MODEL
1. INTRODUCTION The
The tyre tyre2. tread
TYRE
tread is aa thin
thin layer
is THERMODYNAMIClayer (≈ (≈ 55 mm) mm)MODEL of
of rubber
rubber that that
makes
The
The tyre
makes tyre
tyrecontact
tread
tread with
contact with
is a
is aa thinthe
thin
thin track,
layer
layer
the layer
track,(≈ (≈ while
(≈while5 mm)
5 mm)
mm) the
theof of carcass
rubber gives
rubber
ofcarcass that
that
gives
The
the tyre tread
its is
mechanical 5 rubber that
makes
makes
The
the
makes tyre
tyre contact
contact
tread
its
contact
with
with
is
mechanical
with a thestrength.
the
thin
the
track,
track,(≈while
layer
strength.
track,
The
while
5
The
while
carcass
mm) the
theofcarcass
carcass
the
weighs
carcass
rubber
weighs
carcass
ap-
gives
gives
that
ap-
gives
Racing
Racing teamsteams perform
perform optimal
optimal control
control simulations
simulations to to tai-
tai- the proximately
tyre its ten times
mechanical more
strength. than Thethe tread.
carcass A lumped
weighs ap-
the
makes tyre
proximately
the tyre its
contact
its mechanical
ten with
times
mechanical thestrength.
track,
more
strength. than The
while
Thethe carcass
the
tread.
carcass weighs
carcassA lumped
weighs ap-
gives
ap-
lor
Racing vehicle
teams set-up
performparameters
optimalspecific
optimal control to
control each
each track,
simulations to and
to tai- parameter
Racing
lor
Racing
to refine
teams
vehicle
teams
their
perform
set-up
performparameters
racing optimal
strategy
specific
control
(Perantoni
simulations
to
simulations
and
track, to
Limebeer,
tai- proximately
and
tai- the
parametertyre itsmodel
proximately
proximately
ten
ten times
mechanical
model
ten
is
is used
times
times
to
to describe
usedmore
more
strength.
more
than
thanThe
describe
than
the
the
the
the
carcass
the thermodynamic
tread.
tread. A
A lumped
lumped
weighs
thermodynamic
tread. A lumped ap-
lor
lor vehicle
vehicle
Racing
to refine teams set-up
set-up
their performparameters
parameters
racing optimal specific
specific
control
strategy specific
(Perantoni to
to each
each
simulations track,
track,
and track, to
Limebeer, and
and
tai- behaviour
parameter
parameter of of the
model
model tyres
is
is usedof
used
used which
to
to describethe
describe
describe states
the are
thetread. the
thermodynamic
thermodynamic tempera-
lor
2014).vehicle
An set-up
important parameters
component of to each and proximately
behaviour
parameter ten
the
model times
tyres
is of more
which
to than
the the
states
the are the A
thermodynamic lumped
tempera-
toto refine
lor refine
2014).
to refineAntheir
vehicle their
set-up racing
racing
important
their racing
strategy
strategy
parameters
component
strategy of race
(Perantoni
(Perantoni
specific
(Perantoni to each
race strategy
and
and track,
strategy
and
is
Limebeer,
Limebeer, the
is and
Limebeer, tures
the behaviour
behaviour
parameter
tures
behaviour
of
of the
theofoftyres’
the
the tyres
model
oftyres’
the
tread
tyres
is
tread
tyres ofand
of
used
of to carcass.
which
which
and
which
the
the states
describe
carcass.
the
Although
the are
states
Although
states are
are
the
the the
thermodynamic
the the carcass
tempera-
tempera-
carcass
tempera-
understanding
2014). An of
important optimal tyre
component management
of race and
strategy how
is theit comprises rubber and a metal core, it is modelled as
2014).
to
2014).
can
refineAntheir
understanding
An
yield
understanding a
important
racing
of optimal
important
competitive
of optimal
component
strategy
component
advantage.
tyre
of race
tyre (Perantoni
of race strategy
management
The
management
strategy
andand
forces and
is the
Limebeer,
how
is
produced
how
theit tures
it
tures
behaviour
comprises
tures
of
of
of
the
the
the oftyres’
tyres’
the
rubber
tyres’
tread
tread
tyres
and
tread ofaand
and
which
metal
and
carcass.
carcass.
the
core,
carcass.
Although
Although
states
it
Although are
is the
modelledthe
the
the as a
carcass
carcass
tempera-
carcass a
understanding
2014).
can Anaimportant
yield
understanding of optimal
competitive
of optimal tyre
component
advantage.
tyretyremanagement
of The
race forces
management and
strategy
and how
is the
produced
how it homogenous
comprises
comprises
tures
it comprises
homogenous substance
rubber
rubber
of therubber
tyres’
substance and with
a
and aawith
tread metal
metal
and physical
core, properties
it
core,Although
carcass.
physical it is is modelled
is modelled
properties modelledtherepresen-
as
as a
carcass
represen- a
at
can the contact
yield a patch
competitive of each
advantage. determine
The forces the motion
produced tative of both and
materials. metal
The two core,
thermalit states, as
together a
canthe
at
can yield
understanding
yield a competitive
contact
a competitive
of optimal
patch advantage.
tyretyre
of advantage.
each The forces
management
determine
The forces and
the how it homogenous
produced
motion
produced homogenous
comprises
tative
homogenous of substance
substance
rubber
both and
materials.
substance
with
with
a metal
The
with
physical
physical
two core,
physical thermalproperties
properties
it is modelled
states,
properties
represen-
represen-
as
together
represen- a
(and
atat the therefore
contact handling
patch
patch of of characteristics)
each
each tyre Theof
tyre determine the car
the the (Kelly,
motion
(Kelly, tative with the
of dominant heat flows, are shown in Fig. 1, with
atcan
(and the
the contact
yield
thereforea competitive
contact handling
patch of advantage.
characteristics)
each tyre determine
offorces
determine the
car
the motion
produced
motion tative
homogenous
with
tative of both
the
of both
dominant
both
materials.
materials.
substance
materials.heat The
The
with
flows,
The
two
two
physical
two are thermal
thermal states,
states,
properties
shown
thermal in
states,Fig. together
together
represen-
1, with
together
2008).
(and Vehicle
thetherefore performance
contact handling eachis
patch of characteristics)in
in turn dependent
of
of the car on
on the
(Kelly, the
(and
at
2008).
(and therefore
Vehicle
therefore handling
performance
handling characteristics)
is tyre
characteristics) determine
turn of the the
dependent
the car motion
car the with
(Kelly,
(Kelly, withtemperatures
tative
the
with
the
the dominant
dominant
of both
temperatures
the dominant
of
materials.the
the tread
of heat
heat
heat
flows,
flows,
The
tread
flows,
and
two are carcass
shown
arethermal
and
are shown states,
carcass
shown
given
in
given
in Fig.by
in Fig.
Fig. by1,T
1,
1,
with
with
Ttread
together
with
tread
thermal
2008).
2008). and
Vehicle
Vehicle
(and therefore
thermal and wearwear characteristics
performance
performance
handling is
is in
in of
turn
turn
characteristics)
characteristics the
of the tyres.
dependent
dependent
oftyres.
the car The
The on
on fric-
the
the
(Kelly,
fric- and
the T
temperaturesrespectively. of the tread and carcass given by T
2008). Vehicle performance is in turn dependent on the the
with
and temperatures
T carc
the
the temperatures
carc dominant
respectively. of the
heat tread
flows, and
are carcass
shown
of the tread and carcass given by Ttread given
in Fig. by1,T with
tread
tread
tional
thermal
thermal
2008).
tional properties
and
and wear
Vehicle
properties wear of racing
racing tyres
of characteristics
characteristics
performance is inare
tyres ofespecially
of
areturn the
the tyres.sensitive
tyres.
dependent
especially The
The on fric-
sensitive to
to and
fric-
the T respectively.
thermal and wear characteristics of the tyres. The fric- and
the
and T
temperatures
T carc
carc respectively.
respectively. of the tread and carcass given by T tread
temperature,
tional
tional properties and
andofofmaximum
racing
racing tyres performance
are
areofespecially is
is only achiev-
sensitive to
to 2.1 carc
thermal
able in
properties
temperature,
tional and wear
properties
a relatively of maximum
racing
small
tyres
characteristics
tyres are
temperature
especially
performance the tyres.
especially
band
sensitive
only The
sensitive
(Kelly
fric-
achiev- to
and and Thermodynamic
2.1 Tcarc respectively.
Thermodynamic model
model
temperature,
temperature,
tional
able inproperties and
a relatively maximum
andofmaximum
maximum
racing
small tyres performance
performance
are especially
temperature is
band only
is only
only achiev-
achiev-
sensitive
(Kelly to 2.1
and 2.1 Thermodynamic
Thermodynamic model model
temperature,
Sharp, 2012). and
In the tyre model performance
presented is
here, achiev-
friction is 2.1 Thermodynamic model
able
able in aa relatively
in 2012).relatively small
small temperature
temperature band
band (Kelly
(Kelly and
andis The six-parameter model described
temperature,
Sharp,
able
modelled
Sharp,
in a as
2012). a
and
In the
relatively maximum
function
In the
tyre model
small
tyre of tread
model
performance
presented
temperature
temperature.
presented
is only
here,
band
here, The
achiev-
friction
(Kellycarcass
friction
and
is 2.1 Thermodynamic
The six-parameter model model described by by Tremlett
Tremlett and and Lime-
Lime-
Sharp,
able
modelled
Sharp, in 2012).
a as
2012). a In
relatively the
function tyre
small
In theastyre of model
tread
modelstore presented
temperature
temperature.
presented here,
band
here, friction
(Kelly
The carcass
friction andis
is The beer
The (2016)
six-parameter
six-parameter forms the
model basis
described of this
by work.
Tremlett The and vehicle
Lime-
can
modelledbe considered
as a function aof heat
tread for
temperature. the tyres
The (Kelly,
carcass beer
The (2016) formsmodel
six-parameter model described
the basis
described of this by
by Tremlett
work. The
Tremlett and
and Lime-
vehicle
Lime-
modelled as a function of tread temperature. The carcass dynamics model used here is described in Appendix A
Sharp,
can
modelled
2008),
can
be
be
2012).
considered
and as a In
plays
considered
the
function
an
as
as
tyreaof model
heat
tread
important
a heat
presented
store
role
store
for
temperature.
in
for
the here,
keeping
the
tyres
The
tyres
friction
the
(Kelly,
carcass
tread
(Kelly,
is beer
beer
The
dynamics
beer
(2016)
(2016)
six-parameter
(2016) modelforms
forms
forms
the
the
model
used the
basis
basis
heredescribed
basisis of
of this
this
described
of by
this
work.
work.
Tremlett
in
work.
The
The
Appendix
The and A of
vehicle
vehicle
Lime-
vehicleof
can
2008),
can be
modelled
be considered
and asplays
a function
considered as a
anasimportant
a heat
ofheat store
treadstore for
temperature.
role for the
in keeping
the tyres
The
tyres (Kelly,
thecarcass
tread dynamics
(Kelly, Tremlett
dynamics and
model
model Limebeer
used
used (2016),
here
here is
is and
described
describedthe interested
in
in Appendix
Appendix reader
A of
A is
of
in
2008),the optimal
and plays temperature
an window. Tyre wear also plays beer
Tremlett
dynamics (2016)and
modelforms
Limebeer
used the basis
(2016),
here is of
and this
describedthe work.
interested
in The
Appendix vehicle
reader
A is
of
2008),
can
in
2008),the and
beoptimal
and plays
considered
plays anasimportant
important
temperature
an a heat
important window.role
store
role
in
role for
in
Tyre
in
keeping
keeping
the wear
keeping tyresthe
also
the
tread
the(Kelly,
tread
plays referred
tread Tremlett
Tremlett there.
and
and This
Limebeer
Limebeer model is
(2016),
(2016), nowand
and expanded
the
the interested
interested by adding
reader
reader an
is
is
a
in big
the part
optimal in handling
temperature performance.
window. As
Tyre a tyre
wear is
also worn,
plays dynamics
referred
Tremlett model
there.
and Thisused
Limebeer here
model isis now
(2016), described
and expanded
the in Appendix
interested by adding A an
reader of
is
inin big
the part
a2008),
the optimal
and in
optimal plays temperature
an important
handling
temperature window.
performance.
window.role in Tyre
As
Tyre wear
keeping
awear also
the
tyrealso plays additional
tread
is worn,
plays referred
referred carcass
there.
there. This
This temperature
model
model is
is now
now state, with
expanded
expanded its
by
by associated
adding
adding an
an
aits ability
big part toinprovide
handling grip is compromised,
performance. As thereby
a tyre iscausing
worn, Tremlett
additional
referred and
there. Limebeer
carcassThis (2016),
temperature
model is nowand the
state, interested
with
expanded its
by reader
associated
adding is
an
a big
in
aits big
the part
optimal
ability
part inprovide
toin handling
temperature
handling performance.
window. Tyre
gripperformance.
is compromised, As thereby
As awear
a tyrealso
tyre iscausing
is worn,
plays heat
worn, additional
additional transfer equations,
carcass
carcass to
temperature
temperature each wheel.
state,
state, with
with Allits
its parameters
associated
associated
increases referred
heat
additional there.
transfer carcassThis
equations,model to
temperature is now
each expanded
wheel.
state, with All by
its adding
parameters
associated an
its
its
a big
increases
its partin
ability
ability
ability
to
in
to
lap
toin
lap time
provide
handling
provide
(Farroni
time grip
provide grip is
(Farroni
grip is
et
et al.,
al., 2017).
is compromised,
compromised,
performance.
compromised, As thereby
2017). a tyre iscausing
thereby
thereby causing
worn, heat
causing and
heat
additional
and variables
transfer
transfer
variables are defined
equations,
equations,
carcass
are defined to
to
temperature in Appendix
each
in each wheel.
wheel.
state,
Appendix A.
A. All
with While
All the
the new
parameters
itsparameters
While associated
new
increases
increases
its ability in
in
to lap
lap time
time
provide (Farroni
(Farroni
grip is et
et al.,
al.,
compromised, 2017).
2017). thereby causing heat
model transfer
is apparentlyequations,more to each
complicated, wheel. the All
number parameters
of free
Section
increases 2 gives
in lap a thermodynamic
time (Farroni et tyre
al., model,
2017). which keeps and
and
heat
model variables
variables
transfer
is apparentlyare
are defined
defined
equations, toin
in Appendix
Appendix
each
more complicated, wheel. A.
A. While
While
All the
the new
of new
parameters
Section 2 gives a thermodynamic tyre model, which keeps and variables are defined in Appendix 1 the
A. number
While the free
new
increases
track of in lap time (Farroni et al., 2017). parameters
model is has
apparently been reduced
more to
complicated, three .
the number of free
Section
Section
track
Section 2 the
of2
2
gives
gives
the
gives
tyre
tyre
a
tread
tread and and carcass
aa thermodynamic
thermodynamic
thermodynamic carcasstyretemperatures.
tyre
tyre
model,
model, which
temperatures.
model, which These
which
keeps
keeps and
These
keeps
model
parameters
model is
is apparently
variables hasare
apparently been more
defined
reduced
more complicated,
to three 1 the
in Appendix
complicated,
1 A.
. While
the number
number theof
of free
new
free
temperatures,
track
track of
of the
the in
tyre
tyre turn,
tread
tread influence
and
and both
carcass
carcass the tyres’
temperatures.
temperatures. grip and
These
These parameters
parameters
model
1 John is
von has
has
apparently
Neumann’sbeen
been reduced
reduced
more to
to
complicated,
warning bearsthree
three 1 .
.
1 the
repeating: number
‘· · · of
with free
four
Section
track of2
temperatures, gives
the tyrea thermodynamic
in turn,
tread influence
and carcasstyre model,
bothtemperatures. which
the tyres’ grip keeps
and
These parameters
1 John von has
Neumann’sbeen reduced
warning to
bearsthree .
repeating: ‘· · · with four
wear
track
wear properties.
temperatures,
temperatures,
of the
properties. in
in
tyre Simulation
turn,
turn,
tread
Simulationinfluence
influence
and results
resultsboth
both
carcass are
are given
the
the
given in
tyres’
tyres’
temperatures. in Section
grip
grip
Section and
and
These 3,
3, parameters
parameters
1
1 John von has
I can fitbeen
Neumann’s reduced
an elephant,
warning and to
bearsthree
with five1I.can make
repeating: ‘· · · him
withwiggle
four
temperatures,
with in turn, influence both the tyres’ grip and John
parameters
1 John von
von Neumann’s
I can warning
fit an elephant,
Neumann’s warning and bears repeating:
with repeating:
bears five I can make ‘·
‘· ·
· ·
· with
him
with four
wiggle
four
wear conclusions
wear
with
wear
properties.
properties.
temperatures,
conclusions
properties.
drawn
in Simulation
Simulation
drawn
Simulation
in
in Section
turn, influence results
results
Section
results
4.
4.are
botharethe
are
given
given
given
in
in Section
tyres’
in Section
grip and
Section
3,
3, 1his
3,
trunk (Dyson,
parameters
parameters
his trunk
John
II can
can
(Dyson,
von
2004).’
fit
fit an
an
2004).’
Neumann’s
elephant,
elephant,
warning
and
and with
with
bears
five
five II can
can
repeating:
parameters I can fit an elephant, and with five I can make him wiggle
make
make‘· · ·
him
him
with
wiggle
wiggle
four
with
wear conclusions
with conclusions drawn
drawn in
properties. Simulation in Section
results4.
Section 4.are given in Section 3, his his trunk
trunk (Dyson, 2004).’ 2004).’
with conclusions drawn in Section 4. his trunk (Dyson,
parameters I can fit2004).’
(Dyson, an elephant, and with five I can make him wiggle
with conclusions
2405-8963 Copyrightdrawn© 2020 The in Section
Authors.4.This is an open access article under his trunkthe CC BY-NC-ND
(Dyson, 2004).’ license.
Peer review under responsibility of International Federation of Automatic Control.
10.1016/j.ifacol.2020.12.1446
 Wilhelm J. West et al. / IFAC PapersOnLine 53-2 (2020) 14456–14461 14457

Table 1. Thermodynamic constants

Parameter Front left Front Right Rear left Rear right
p1 0.7291 0.5670 0.5071 0.6124
p2 0.0060 0.0000 0.0125 0.0000
p3 252.53 263.48 174.93 181.38

where htt is the heat transfer coefficient of conduction and

Acp is the non-sliding area of the contact patch, which is
given by:
Acp = cw cs (α)cl . (9)

The function cs (α) controls the relative proportion of

Fig. 1. Thermodynamic model with dominant heat flows
The heat generated by tyre friction is given by Equation
(2) in Tremlett and Limebeer (2016)
Qf rict = p1 un (|Fx κ| + |Fy tan α|), (1)
where p1 is the fraction of the frictional power lost to the
track, and un , Fx , κ, Fy and α have their usual meaning;
see Tremlett and Limebeer (2016). The heat transfer due
to convection between the tyre tread and the surrounding
ambient air is given by:
Qconv.T A = hf orc Aconv (Ttread − Tamb ), (2)
where
Aconv = Atot − cw cl (3)
is the area available for convection. The first term on the
right-hand side is the surface area of the tread
Atot = πDW, (4)
while the second term is the area of the tread in contact
with the road. The tread contact patch width of the front
and rear tyres are given by cwf and cwr respectively; see
Table A.4. The length of the contact patch cl is given by:
cl = acp Fz0.7 . (5)
where acp is a constant reflecting the difference between
the front and rear contact patch lengths, and Fz is the
normal force acting on the wheel. The heat transfer
coefficient hf orc is determined using an empirical formula
that correlates well with CFD simulation (Farroni et al.
(2014) Equation 10):
Kair   uL 0.805  telemetry data was available for the carcass temperature,
hf orc = 0.0239 , (6)
L vair
where Kair is the conductivity, and vair the kinematic
viscosity, of air. Both are evaluated at the average temper-
ature between the tread and ambient air using a material
physical property database–see Table A–9 (Çengel and
Cimbala, 2014); u is the forward velocity of the vehicle.
The characteristic length of the heat exchange surface is
given by:
1
L= 1 1 , (7)
D + W
where D is the tyre diameter and W the tread width. The
conduction between the track and tyre tread is given by:
Qcond.T T = htt Acp (Ttread − Ttrack ) (8)
sliding and non-sliding regions as the vehicle approaches
its cornering limit and is given by
α
cs (α) = (cs2 − cs1 ) + cs1 ; (10)
αc
with cs1 and cs2 being reference values, while αc is the
reference slip angle. Conduction between the tread and
carcass layers is given by:
Qcond.T C = p3 Atot (Tcarc − Ttread ), (11)
where p3 is the heat transfer coefficient of conduction, and
Atot is given in (4).
Heat is generated as the carcass deflects under load. This
is assumed to be proportional to the square of longitudinal
force acting on the tyre as per Clark and Dodge (1985):
u F2 
n x
Qdef l = p2 . (12)
|Fz |
We can now describe the thermal dynamics of the tyre
tread and carcass layers as follows
d
mt ct Ttread (t) = Qf rict − Qcond.T T
dt
− Qconv.T A + Qcond.T C (13)
d
mc cc Tcarc (t) = Qdef l − Qcond.T C (14)
dt
where mt and mc are the masses, and ct and cc the heat ca-
pacities of the tread and carcass respectively. Parameters
p1 to p3 are fitted using a least-squares minimisation of the
error between telemetry data taken around a lap of ’Circuit
de Catalunya’ in Barcelona and the tread temperature
calculated by the thermodynamic model.
Fig. 2 shows a best-fit parameter set for the front-left
tyre, which has a root mean square error of 6.8◦ C. No
which shows only small variations. This is attributable
to its greater thermal inertia as compared to the thin
tread layer. In essence, the carcass acts as a heat store
that prevents the tread from cooling down to the track
temperature. Table 1 shows the parameters that resulted
in the best fit for each tyre.
2.2 Tyre wear model
The model developed in Tremlett and Limebeer (2016)
forms the basis of this aspect of the work. Wear rate (ẇ) is
added as an extra state to be associated with each tyre, so
that the total tyre wear can be calculated and minimised
as part of the optimal control problem. The mechanical
14458 Wilhelm J. West et al. / IFAC PapersOnLine 53-2 (2020) 14456–14461
Fig. 2. Parameter fitting using telemetry data
abrasion between a tyre and track asperities is modelled
using the power law relationship:
Q wp2
f rict
w˙p = wp1 , (15)
Qref
where wp1 , wp2 and Qref are constants (Tremlett and
Limebeer, 2016). If a tyre is overworked while cold, parti-
cles break away from the surface in a phenomenon called
graining and is modelled by 2 :
w˙g = wg1 (max(ttp − Ttread ), 0)wg2 ; (16)
wg1 and wg2 are constants. If a tyre gets too hot, local hot
spots will form and tyre wear is again accelerated. This
phenomenon is called blistering and is modelled by 3 :
ẇb = wb1 (max(Ttread − ttp ), 0)wb2 ; (17)
wb1 and wb2 are constants. The transition temperature
ttp distinguishes between these two mechanisms: graining
occurs below the transition temperature and blistering
occurs above it. The cumulative wear rate is assumed
to be a simple superposition of the above three wear
mechanisms.
2.3 Performance alteration
Tyre degradation has an impact on race performance and
is thus included in the optimal control problem. Both
temperature and wear characteristics will be considered.
Temperature-grip characteristic The first mechanism
that alters the car’s performance is the tyres’ temperature-
grip characteristic. The temperature window in which
the maximum achievable grip is obtained is shown in
Fig. 3; this curve is representative of a super-soft tyre
2
Most optimal control solvers deal only with ‘smooth’ functions. To
that end (16) is approximated by
  wg2
w˙g = wg1 (ttp − Ttread + (ttp − Ttread )2 + )/2
for ‘small’  > 0. See page 417 of Limebeer and Massaro (2018) for
further details.
3 A usable smooth approximation to (17) is
  wb2
w˙b = wb1 (Ttread − ttp + (Ttread − ttp )2 + )/2
Fig. 3. Optimal temperature window for super-soft tyre
compound
(the compound that will be used for the simulations in
the following section). This window must be adjusted for
each tyre compound. The optimal controller aims to get
the tyre surface temperatures into this optimal operating
window and hold it there as nearly as possible. An example
of such a curve is given in Limebeer and Massaro (2018).
In this case the optimal grip is achieved when the tread
temperature lies between 90◦ C and 105◦ C.
Carcass temperature The tread-only model used in
Tremlett and Limebeer (2016) is limited by the fact that it
exaggerated the extent to which it is possible to cool the
tread once it overheats. The optimal controller achieves
maximum grip by raising the tread temperature into the
optimal operating window. If one wants to reduce the
temperature of an over-heated tread, one has to reduce
the temperature of the tyre carcass—this process is slowed
by the relatively high thermal inertia of the carcass as
compared to that of the tyre tread. For this reason it
is important to introduce a carcass temperature state
that represents this thermal-inertia-related lag. The tread
undergoes rapid temperature changes throughout a lap,
while the carcass temperature varies more slowly. Once the
carcass has heated up, and possibly over-heated, cooling
it down again is also a slow process.
Tyre wear The tyre tread temperature alter perfor-
mance, but in isolation, it is a fully reversible effect (as
it is modelled here). This reversibility is not a tyre-wear
property—once the tyre is worn out it stays worn out!
It is thus necessary to add into the model a mechanism
that recognises tyre degradation due to wear, and prevents
the optimal controller from regaining maximum grip once
degradation has occurred.
Examples of how different tyre types lose performance, as
the tyre degrades, is given in Farroni et al. (2017). Pas-
senger tyres are typically designed for longevity, whereas
racing tyres are designed to offer enhanced grip, but for a
shorter time. The wear curve of a super-soft tyre com-
pound is shown in Fig. 4. The grip initially increases
with wear before it starts to degrade, because racing tyres
. are not fully vulcanised during the manufacturing process
(Haney, 2003), and go through a curing process during
 Wilhelm J. West et al. / IFAC PapersOnLine 53-2 (2020) 14456–14461 14459
Fig. 4. Wear characteristics of super-soft tyre compound
Fig. 5. Performance altering grip curve
the initial wear phase, which form cross-links between
sections of the polymer chains. This increases the rigidity
and durability of the rubber (Mark et al., 2013).
Tyre wear is introduced into the tyre model as a third
dimension to the temperature-grip sensitivity curve. This
allows the tyre grip to be modelled as a function of the
tread temperature (a reversible process) and tyre wear (an
irreversible process). This forces the optimal controller to
manage the tyres usage throughout a race (assuming no pit
stops); the three-dimensional grip-sensitivity curve used
here is shown in Fig. 5. Slices along constant wear planes
resemble Fig. 3, whereas slices along constant temperature
planes resemble Fig. 4.
3. SIMULATION RESULTS
Two simulations are presented that illustrate the signifi-
cance of the additions to the model as presented in Trem-
lett and Limebeer (2016). Firstly, a multi-lap simulation
was performed (in the absence of tyre wear) that investi-
gates the effect of the proposed tyre thermodynamic model
on a multi-lap optimal control problem. After that, we
investigate the influence of wear on the optimal control
solution. The general-purpose optimal control problem
Fig. 6. Output of simulation without wear component
solver (GPOPS-II) by Patterson and Rao (2014) is used
to perform the simulations.
3.1 Addition of tread and carcass temperatures
In this section the optimal control solver is used to min-
imise the time taken to complete fifteen laps around the
Circuit de Catalunya in Barcelona. The initial tyre tem-
peratures were set to 60◦ C, which is well below the optimal
tyre operating temperature. For this simulation the super-
soft compound was used that will achieve optimal grip
when the tread temperature is between 90◦ C and 105◦ C.
The temperature profiles for the rear-right tyre are shown
in Fig. 6. The first lap is slow because the tyres are cold
and thus unable to provide optimal grip. In the second
lap the tread temperature is within the optimal operating
window and the fastest lap of 81.08 s is recorded. The
time difference relative to fastest lap (lap two) is shown as
the orange graph. As the various laps elapse, the carcass
temperature slowly increases, making it difficult for the
optimal controller to keep the tread temperature in the
optimal window. As a result thermal influences cause a loss
of performance with each subsequent lap. The time delta
curve settles out after approximately 10 laps whereafter
the lap times remain roughly constant (≈0.18 s slower than
the fastest lap).
3.2 Addition of tyre wear
Tyre wear influences are now introduced. To exaggerate
the effect of tyre wear, the wear rate model was ‘tweaked’
so that a tyre is completely worn out within a single lap.
The velocity profiles for various wear cases are shown in
Fig. 7. There is no wear in the base case, which inevitably
leads to the fastest lap. When the wear rates are set to
their nominal values, only slight deviations relative to the
base case are observed. This is expected seeing that racing
tyres are designed to last for multiple laps. On new tyres
the car’s maximum speed builds up until the peak of the
wear curve given in Fig. 4 is achieved. As the nominal wear
rate is exaggerated, the vehicle speed is reduced so as to
prevent the tyres from being fully worn before reaching
the end of the lap.
14460 Wilhelm J. West et al. / IFAC PapersOnLine 53-2 (2020) 14456–14461

Fig. 8. Temperature profile for simulation with wear char-

acteristics of the super-soft tyre compound
Fig. 7. Output of simulation with artificial tyre wear
3.3 Multi-lap simulation using wear model

The tyre wear model is now adjusted back to its nominal

values to facilitate multiple laps. The results of a multi-lap
simulation is shown in Fig. 8 and Fig. 9. Initially, the tyres
are purposefully worn and heated so that they can quickly
reach the peak of the friction factor curve. Over the first
two laps the grip increases as the tyres wear and the tread
temperatures are brought into the optimal window; as in
Fig.6, the second lap is the fastest lap. Between laps three
and eight, the lap times increase at a roughly constant
rate. This is the result of a roughly constant decline
in the friction factor characteristic over this tyre usage
range. After lap eight, tyre wear affects performance more
severely, and the lap times start to degrade more rapidly.
This can be ascribed to the slope changes in the friction
factor curve. Another consequence of having increasingly Fig. 9. Performance metrics for simulation with wear
worn tyres, is that is becomes more difficult to keep the characteristics of the super-soft tyre compound
tread in the optimal temperature window. It is clear that
there are increasing departures from the treads’ optimal performance will degrade with use. Each tyre compound
temperature operating window as the race progresses. In has its own characteristics, and will react differently to
contrast to Fig.6, the lap time ∆T does not plateau, but tyre use. The tyre model can be used to determine the
continues to increases more rapidly as the tyres are worn. point where tyre wear compromises performance to such
This speaks to the irreversible nature of tyre wear that an extent that the tyres are no longer competitive—this
has not been captured before. At some point the drop-off is a process not an instantaneous event. Our intention is
in performance degrades to the extent that the car is no that race teams can use the techniques introduced here
longer competitive which will warrant a pit stop to change to gain competitive advantage and improve their tyre
the tyres. management strategies.

4. CONCLUSION
A thermodynamic tyre model has been presented in which
the tread and carcass temperatures of the tyres are cal-
culated. The tread temperature reacts quickly to driver
inputs and can be easily manipulated so that it comes into
the optimal temperature window, where maximum grip is
achieved. The carcass on the other hand, has a higher ther-
mal inertia, and its temperature increases only slowly over
time, making its temperature more difficult to manage. A
wear model is also introduced into the tyre model that
accounts for the loss in performance due to tyres usage.
Tyre wear is modelled as a function of temperature and
frictional power, and describes qualitatively the way tyre
REFERENCES
Çengel, Y. and Cimbala, J. (2014). Fluid Mechanics -
Fundamentals and Applications. McGraw-Hill.
Clark, S.K. and Dodge, R.N. (1985). Heat generation in
aircraft tires. Computers & Structures, 20(1), 535–544.
Dyson, F. (2004). A meeting with Enrico Fermi. Nature,
427, 297.
Farroni, F., Sakhnevych, A., and Timpone, F. (2017).
Physical modelling of tire wear for the analysis of the
influence of thermal and frictional effects on vehicle per-
formance. Proceedings of the Institution of Mechanical
Engineers, Part L: Journal of Materials: Design and
Applications, 231(1-2), 151–161.
 Wilhelm J. West et al. / IFAC PapersOnLine 53-2 (2020) 14456–14461 14461

Farroni, F., Giordano, D., Russo, M., and Timpone, F.

(2014). TRT: thermo racing tyre a physical model to
predict the tyre temperature distribution. Meccanica,
49(3), 707–723.
Haney, P. (2003). The Racing & High-performance Tire.
SAE International and TV Motorsports.
Kelly, D.P. and Sharp, R.S. (2012). Time-optimal control Table A.3. Tyre wear model parameters
of the race car: influence of a thermodynamic tyre
model. Vehicle System Dynamics, 50(4), 641–662.
Kelly, D.P. (2008). Lap Time Simulation with Transient Parameter Description Units Value
wg1 Graining gain factor mm/◦ C s 0.4 × 10−5
Vehicle and Tyre Dynamics. Doctor of philosophy,
wg2 Graining exponent - 2
Cranfield University. wb1 Blistering gain factor mm/◦ C s 0.8 × 10−5
Limebeer, D.J.N. and Massaro, M. (2018). Dynamics and wb2 Blistering exponent - 2
Optimal Control of Road Vehicles. Oxford University wp1 Frictional power gain mm/s 0.09
Press. factor
Mark, J.E., Erman, B., and Roland, C.M. (2013). The wp2 Frictional power expo- - 1.6
Science and Technology of Rubber Fourth Edition. Aca- nent
demic Press, Boston. Qref Reference frictional kW 150
Patterson, M.A. and Rao, A.V. (2014). GPOPS-II: A power
ttp Minimum wear tempera- ◦C 100
MATLAB Software for Solving Multiple-Phase Optimal
ture transition
Control Problems Using hp-Adaptive Gaussian Quadra- gs Temperature gain factor - 0.5
ture Collocation Methods and Sparse Nonlinear Pro-
gramming. ACM Trans. Math. Softw., 41(1), 1:1—-1:37.
Perantoni, G. and Limebeer, D.J.N. (2014). Optimal
control for a Formula One car with variable parameters.
Vehicle System Dynamics, 52(5), 653–678.
Tremlett, A.J. and Limebeer, D.J.N. (2016). Optimal tyre
usage for a Formula One car. Vehicle System Dynamics,
54(10), 1448–1473.

Appendix A. DEFINITION OF MODEL

PARAMETERS

Table A.1. Vehicle dynamics variables

Parameter Description Units Table A.4. Thermodynamic model parameters
un Wheel longitudinal velocity m/s
u Car longitudinal velocity m/s Parameter Description Units Value
Tamb Track air temperature ◦C 25
Fx Longitudinal tyre force N
Ttrack Track surface temperature ◦C 35
Fy Lateral tyre force N
Fz Normal tyre force N mt Tyre tread mass kg 0.5
α Lateral slip angle rad mc Tyre carcass mass (front) kg 9.5
κ Longitudinal slip coefficient - Tyre carcass mass (rear) kg 11.5
ct Heat capacity of tread kJ/kg K 2.4
Table A.2. Thermodynamic variables cc Heat capacity of carcass kJ/kg K 1.6
htt Track-tyre heat transfer coef- kW/m2 K 12
Parameter Description Units ficient
Qf rict Heat generated by tyre friction W cwf Contact patch width (front) m 0.20
Qdef l Heat generated by tyre deflection W cwr Contact patch width (rear) m 0.22
Convective heat transfer between acp Contact patch length con- m/kN 0.056
Qconv.T A W
tread and ambient air stant
αc Reference sliding/non-sliding ◦ 8
Conductive heat transfer be-
Qcond.T T W
tween tread and track slip angle
Conductive heat transfer be- cs1 Sliding/non-sliding reference - 0.3
Qcond.T C W
tween tread and carcass 1
Acp Area of the contact patch m2 cs2 Sliding/non-sliding reference - 0.8
Atot Total area of tyre tread m2 2
D Tyre diameter m 0.66
W Tyre width (front) m 0.355
Tyre width (rear) m 0.380