energies

Review
Aerodynamic and Thermal Modelling of Disc
Brakes—Challenges and Limitations
Alexey Vdovin 1, * and Gaël Le Gigan 2
1 Department of Mechanics and Maritime Sciences, Chalmers University, 412 96 Göteborg, Sweden
2 Volvo Car Corporation, 418 78 Göteborg, Sweden; [email protected]
* Correspondence: [email protected]; Tel.: +46-31-772-3626




Received: 28 November 2019; Accepted: 25 December 2019; Published: 1 January 2020


Abstract: The brake system is a critical component for any passenger vehicle as its task is to
convert the kinetic and potential energy of the vehicle into heat, allowing the vehicle to stop. Heat
energy generated must be dissipated into the surroundings in order to prevent brake overheating.
Traditionally, a lot of experimental testing is performed to ensure correct brake operation under all
possible load scenarios. However, with the development of simulation techniques, many vehicle
manufacturers today are looking into partially or completely replacing physical experiments by
virtual testing. Such a transition has several substantial benefits, but simultaneously a lot of challenges
and limitations need to be addressed and understood for reliable and accurate simulation results. This
paper summarizes many of such challenges, discusses the effects that can and cannot be captured,
and gives a broader picture of the issues faced when conducting numerical brake cooling simulations.

Keywords: brake cooling; computer-aided engineering (CAE) simulations; test replication;

computational fluid dynamics (CFD); thermal simulations

1. Introduction
Friction brakes are used to convert the kinetic, and sometimes potential energy of the vehicle,
into thermal energy, hence, allowing the vehicle to decelerate or stop when required. Since brakes can
generate a substantial amount of heat energy in a short instance, overheating of the brake discs is a
possible scenario to be avoided. Overheating can cause brake fading, brake judder, increased wear, as
well as thermal cracking and even brake fluid vaporization [1,2], which constitutes a serious safety
issue. Furthermore, other parts of wheel suspension, wheel cover or tire, can also be affected by high
temperatures, especially during the thermal soak process [3,4].
In order to avoid brake system overheating, it is a necessity that the brake assembly is appropriately
dimensioned and designed as to store and dissipate energy without reaching critical temperatures, even
in the most extreme braking scenarios. Traditionally, the brake systems have been tested experimentally
using different test benches, such as full-scale brake dynamometer, as well as through complete vehicle
testing. However, nowadays, some tests can be replicated by CAE simulations, which are cheaper
and can also provide information difficult to obtain in real life tests. Moreover, the simulations can be
performed at much earlier stages, such as concept phase, hence, avoiding possible later problems in
vehicle development, reducing both prototyping costs and lead time for parts and by that reducing the
total length of vehicle development.
The complexity of the computer-aided design (CAD) models for numerical simulations can vary
significantly. Simple studies looked into modelling of a single channel inside the brake disc [5], a sector
of the brake disc [6,7] or just the brake disc with some surrounding geometries [8–10]. More complex
studies would include more geometrical parts [11,12], which have been shown to be more representative
of the on-road conditions [11]. At the same time, from the physics perspective, these simulations

Energies 2020, 13, 203; doi:10.3390/en13010203 www.mdpi.com/journal/energies

Energies 2020, 13, x FOR PEER REVIEW 2 of 12

Energies 2020, 13, 203 2 of 12

representative of the on-road conditions [11]. At the same time, from the physics perspective, these
simulations would often focus either on computing convection heat transfer coefficients from a
computational
would fluid
often focus dynamics
either (CFD) convection
on computing point of viewheatortransfer
on conduction modelling
coefficients through the solid
from a computational fluid
parts. Simulations
dynamics (CFD) pointthatofinclude
view or both effects are more
on conduction challenging
modelling through and
thedemanding
solid parts. in terms of setup
Simulations that
since they require coupling between fluid and thermal solvers to account for changes
include both effects are more challenging and demanding in terms of setup since they require coupling in both flow
and temperature
between fluid andfields
thermal[12–14].
solvers to account for changes in both flow and temperature fields [12–14].
Complete replacement
Complete replacement of of the
the brake
brake cooling
cooling experimental
experimental tests
tests by
by numerical
numerical simulations
simulations isis aa
demanding task as it requires accurate modelling of all three heat transfer mechanisms:
demanding task as it requires accurate modelling of all three heat transfer mechanisms: convection convection
(through airflow
(through airflow passing
passing inside
inside and
and around
around the
the disc),
disc), conduction
conduction (through
(through the
the support
support and
and contact
contact
areas) and, at last, radiation. This paper addresses the complexity of this modelling
areas) and, at last, radiation. This paper addresses the complexity of this modelling and discusses and discusses
simplifications and
simplifications and limitations
limitations that
that need
need to
to be
be accounted
accounted for.for.

2. Complications of Modelling Aero-Thermal Problems

different tests
Several different tests are
are usually
usually performed
performed to ensure
ensure the
the correct
correct brakes
brakes system
system cooling
performance, some of them
performance, them lastlast several
several minutes
minutes andand some
some can
can last
last much
much longer.
longer. Moreover, the the
dominant heat
dominant heat transfer
transfer mechanism
mechanism would would differ
differ depending
dependingon on the
thetest
testscenario.
scenario.
Convection isisusually
Convection usuallythe themost
mostimportant
important mechanism
mechanism forfor cooling.
cooling. Radiation
Radiation is important
is important onlyonly
for
for high
high temperatures,
temperatures, such such as a glowing
as a glowing brake brake disc.
disc. The The contribution
contribution of conduction
of conduction through through
the support the
support
is is less important
less important since it is since
a slowit process
is a slow process
[10]; [10];ithowever,
however, it will beif dominant
will be dominant if the test
the test includes the
includesphase
soaking the soaking
when thephase whenisthe
vehicle vehicleand
stopped is stopped and forcedisconvection
forced convection replaced by is replaced by natural
natural convection.
convection. Nevertheless,
Nevertheless, all these mechanisms
all these mechanisms need to for
need to be modelled be modelled
every test for every iftest
scenario thescenario if the
test results aretest
to
results
be are to accurately.
replicated be replicated accurately.
Historically, CFD
Historically, CFD codes were were used
used toto compute
compute convection heat transfer coefficients (HTCs),
while separate
while separate thermal
thermal solvers
solvers were
were used
used toto model
model conduction
conduction and and radiation
radiation [12,13,15].
[12,13,15]. Specifics of
aerodynamic and thermal models are discussed in later chapters; here, the focus is on on the
the coupling
coupling
process and
process and data
data exchange
exchange between different solvers. solvers. It should be noted that there are studies that
combine fluid and and solid
solid domains
domains within
within the same finitefinite volume
volume solver
solver [16];
[16]; however,
however, theythey still
still
experience most
experience most ofof the
the issues
issues presented
presented in in this
this paper.
paper.

2.1.
2.1. Coupling
Coupling Approaches
Approaches
The
The data
data exchange
exchange process
process between
between twotwo models,
models, also
also known
known as as coupling,
coupling, usually
usually involves
involves
transferring HTCs and reference fluid temperatures from aerodynamic solver to
transferring HTCs and reference fluid temperatures from aerodynamic solver to the thermal solver.the thermal solver.
The thermal solver uses these data to estimate the convection heat fluxes for part surfaces
The thermal solver uses these data to estimate the convection heat fluxes for part surfaces and and combines
it with computed
combines it with conduction and radiation
computed conduction andfluxes to estimate
radiation fluxesthe
to changes
estimatein components
the changes intemperatures
components
during a specified time period. New surface temperatures are then being transferred to
temperatures during a specified time period. New surface temperatures are then being transferred to the aerodynamic
solver for recalculating
the aerodynamic solverthe
for airflow around
recalculating the
the simulated
airflow aroundcomponents, see Figure
the simulated 1.
components, see Figure 1.

Figure 1.1. Simplified coupling

Figure coupling process.
process. Twall is the
wall is the temperature
temperature of the surfaces
surfaces while
while TTfluid
fluid is the
the
reference air temperature next to the surfaces.
reference air temperature next to the surfaces.

The
The communication
communicationbetween
betweentwo twoseparate solvers
separate cancan
solvers be set
be up
set in
upseveral different
in several ways.ways.
different The most
The
accurate method
most accurate would would
method be to use
beatofully-transient approachapproach
use a fully-transient when changes
when in the airflow
changes and
in the HTCsand
airflow are
Energies 2020, 13, 203 3 of 12

continuously modelled for the whole duration of the test. However, simulating several minutes of the
transient airflow around the vehicle can be extremely expensive in terms of computational resources;
therefore, other simulation approaches have been developed.
One of the popular approaches is to consider the flow around the rotating brake disc to be
independent of the surface temperatures. This assumption allows performing only a few aerodynamic
simulations with different vehicle velocities. Interpolation is then used to obtain surface HTC
dependency curves for all vehicle velocities. These curves are then used in the thermal model without
any additional flow recalculation loops required. Such simplification works relatively well for high
speed scenarios [13]; however, if the simulated vehicle velocities are slow, the buoyancy effects become
more pronounced, and this approach thereby becomes invalid.
More accurate simulation approaches, usually referred to as “semi-transient”, use two-way
coupling between thermal and aerodynamic solvers, i.e., the airflow and HTCs are being updated
several times during one simulation. In this case, surface HTCs are considered dependent on both the
vehicle velocity and component surface temperatures.
An important parameter for semi-transient simulations is the coupling frequency. Some approaches
would assume a step curve for HTCs of the parts and require frequent recalculations of them, while
others would assume linear behaviour of the surface HTCs and, hence, allow much longer time
intervals between coupling [13,14,17]. Both methods have pros and cons, depending on the test
scenario being simulated.

2.2. Data Mapping

Addressing the complications of a coupling process, it is important to remember the data mapping
between codes. Even though having a conformal mesh on the two sides of the aero-solid interface would
be advantageous for matching heat fluxes between solid and fluid cells, quite often non-conformal
interfaces are being used. This occurs because the mesh requirements for two solvers are quite different
and having coarser mesh for thermal simulations reduces the computational effort. At the same time,
non-conformal interfaces mean that a single cell from the thermal simulation would need to be mapped
to several cells from the CFD side.
HTCs on the aerodynamic side can be quite sensitive to the CFD mesh quality; hence, extra
attention is needed when mapping them. Furthermore, it is important to make a good selection for the
reference temperatures to be mapped in parallel with the HTCs as some of the parts are going to be
heated up and some cooled down by the air.

3. Aerodynamic Modelling
Simulating wheelhouse flows is a topic of many scientific publications as it is a very complex
problem by itself. Moreover, standard aerodynamic simulations do not include the energy equation
or any temperature related effects. To obtain the convection heat transfer coefficients required for
coupling process, the aerodynamic simulations need to be enhanced with energy and gravity models
to account for temperature gradients in the air and their influence on the airflow.
The following chapter summarizes some of the main considerations and limitations for such type
of simulations.

3.1. Level of Model Details

The level of details in the virtual model would depend on the type of braking test being simulated,
for example, emergency brake test and drag braking. Unfortunately, to simulate an actual on-road test
scenario, a fully detailed vehicle model is required. The airflow in the wheelhouses depends not only
on external air entering from underneath the vehicle and from the sides but also on the air coming
from the engine bay. Therefore, due to asymmetrical packaging and heat distribution inside the engine
compartment, the velocity and temperature fields inside the wheelhouses would be significantly
Energies 2020, 13, 203 4 of 12

different
Energies [18].
2020, 13, xConsequently, the
FOR PEER REVIEW cooling performance of brakes on the left and right sides would 4 of be
12
different as well.
3.2. Airflow Simulation Approaches
3.2. Airflow Simulation Approaches
In standard aerodynamic simulations, brake discs are often simulated with moving wall
In standard
boundaries, whichaerodynamic simulations,
can significantly affect brake discs are
the amount of often
mass simulated withthe
flow through moving
discswall
and,boundaries,
hence, the
which can significantly affect the amount of mass flow through the discs
calculated cooling performance. There are two more advanced approaches to simulate and, hence, the brake
calculated
disc
cooling performance. There are two more advanced approaches to simulate brake disc rotation:
rotation:
• approach, when
Multiple Reference Frame (MRF) approach, when the
the additional
additional forces
forces imposed
imposed by
by rotation
rotation are
are
introduced to the volumetric region inside the disc;
• Rigid
Rigid Body Motion
Motion (RBM)
(RBM) approach,
approach, when
when the
the mesh
mesh of
of the vanes (cooling channels
channels inside the
the
disc) is physically rotated in space.
Both
Both approaches
approaches can
can be
be applicable,
applicable, but
but the
the latter,
latter, though
though being
being more
more accurate,
accurate, requires
requires using
using
unsteady
unsteadymethods
methodsforforsimulating the
simulating theairflow, hence,
airflow, it isitmuch
hence, more
is much computationally
more expensive.
computationally The
expensive.
MRF method, on the other hand, is beneficial since it allows to use steady-state Reynolds
The MRF method, on the other hand, is beneficial since it allows to use steady-state Reynolds Averaged Averaged
Navier
Navier Stokes
Stokes simulations,
simulations, which
which are
are significantly
significantly faster
faster and
and atat the
the same
same time
time provide
provide good
good enough
enough
accuracy for the
accuracy for the mass-flow
mass-flowthrough
throughthethevanes
vanesandandsurface
surface HTCs
HTCs [19].
[19]. AA typical
typical volume
volume of rotation
of rotation for
for any of these methods would be limited to the vanes of the disc, as presented
any of these methods would be limited to the vanes of the disc, as presented in Figure 2. in Figure 2.

Figure 2.
Figure Rotating region
2. Rotating region inside
inside the
the brake
brake disc.
disc.

It should
It should be
be mentioned
mentioned that
that this
this rotating
rotating region
region is
is only
only an
an addition
addition that
that is
is required
required for
for correct
correct
airflow simulation around the brake discs. Other rotating parts like rims, tires and even
airflow simulation around the brake discs. Other rotating parts like rims, tires and even the cooling the cooling
fans inside
fans inside the
the engine
engine bay
bay still
still need
needtotobe
betreated
treatedappropriately
appropriatelyasasthey
theyshould
shouldbebe inin
any other
any type
other of
type
aerodynamic simulation. Moreover, both the airflow inside and around brake discs
of aerodynamic simulation. Moreover, both the airflow inside and around brake discs and the surfaceand the surface
heat transfer
heat transfer coefficients
coefficients are
are quite
quite sensitive
sensitive to
to the
the mesh
mesh settings.
settings.
3.3. HTC Averaging
3.3. HTC Averaging
Independently of using steady-state or unsteady methods for the airflow simulations, it is
Independently of using steady-state or unsteady methods for the airflow simulations, it is
important to remember that coupling between aerodynamic and thermal codes usually happen only a
important to remember that coupling between aerodynamic and thermal codes usually happen only
certain amount of times during a simulation of one test cycle. Moreover, the outer disc surfaces are still
a certain amount of times during a simulation of one test cycle. Moreover, the outer disc surfaces are
typically modelled as moving wall boundaries. This means that it may be necessary to average the
still typically modelled as moving wall boundaries. This means that it may be necessary to average
heat transfer coefficients of the rotating parts not to introduce any unrealistic temperature gradients on
the heat transfer coefficients of the rotating parts not to introduce any unrealistic temperature
the disc surfaces. As the airflow for the inboard and outboard side of the disc is usually significantly
gradients on the disc surfaces. As the airflow for the inboard and outboard side of the disc is usually
different, a recommendation is to split the brake disc surfaces to at least three regions: two outer
significantly different, a recommendation is to split the brake disc surfaces to at least three regions:
surfaces and the vanes.
two outer surfaces and the vanes.
3.4. Simulating Tests with Soaking Phases
3.4. Simulating Tests with Soaking Phases
Some of the test scenarios, for example, mountain descent test, include the soaking phase when
Some of the test scenarios, for example, mountain descent test, include the soaking phase when
the vehicle is standing still and the thermal energy from the disc is dissipated into the rest of the
the vehicle is standing still and the thermal energy from the disc is dissipated into the rest of the
braking system and the surrounding air. For the cases of pure natural convection, the velocity field
braking system and the surrounding air. For the cases of pure natural convection, the velocity field
inside the wheelhouse is directly affected by the temperatures of different parts; hence, it is necessary
inside the wheelhouse is directly affected by the temperatures of different parts; hence, it is necessary
to recompute the flow field during this phase.
to recompute the flow field during this phase.
Moreover, extra attention is needed for the transition between driving and soaking phases. First,
it is important to have some data about the fan after-run curves and temperature distribution inside
the engine compartment, as the fan is going to push warm air from the engine bay to the wheel arches.
Energies 2020, 13, 203 5 of 12

Moreover, extra attention is needed for the transition between driving and soaking phases. First,
it
Energies 2020, 13, xto
is important have
FOR some
PEER data about the fan after-run curves and temperature distribution inside
REVIEW 5 of 12
the engine compartment, as the fan is going to push warm air from the engine bay to the wheel arches.
It should
It should bebe noted
noted that
that such an additional
additional air circulation
circulation was
was shown
shown to to increase
increase the
the rate
rate of the soaking
soaking
process for
process for the
the brake
brake system
system in
in comparison
comparison toto a case without fan after-run
after-run [18].
[18]. Secondly, one should
Secondly, one should
keep in
keep inmind
mindthatthatthe
theswitch
switchtotonatural
natural convection
convection does
does notnot happen
happen instantaneously,
instantaneously, andand hence,
hence, it canit
canadvantageous
be be advantageous to simulate
to simulate this transition
this transition periodperiod
usingusing transient
transient approaches.
approaches.

4.
4. Thermal Modelling
A
A typical
typicalthermal
thermalmodel
modelusually
usuallyincludes
includes all all
parts thatthat
parts are in
areimmediate proximity
in immediate to theto
proximity hotthe
brake
hot
disc and participate in the heat dissipation from it: brake disc, brake pads with shims
brake disc and participate in the heat dissipation from it: brake disc, brake pads with shims and and backplates,
calliper withcalliper
backplates, all internal
withparts, wheel parts,
all internal rim and tire, rim
wheel dustand
shield,
tire,wheel hub and
dust shield, knuckle.
wheel A typical
hub and mesh
knuckle. A
for the model will contain 3 to 5 million cells. An example is given in Figure 3.
typical mesh for the model will contain 3 to 5 million cells. An example is given in Figure 3.

Figure 3. Typical thermal

thermal model
model of
of aa brake
brake assembly
assembly including
including wheel
wheel rim
rim and
and tire.
tire.

As
As mentioned
mentioned previously,
previously, the
the main
main task
task of
of the
the thermal
thermal solver
solver is
is to
to calculate
calculate conduction
conduction and
and
radiation
radiation heat fluxes, combine them with the convective data obtained from the aerodynamic solver
heat fluxes, combine them with the convective data obtained from the aerodynamic solver
and
and advance
advance in
in time,
time, calculating
calculating new
new temperature
temperaturedistributions
distributionsinside
insidethe
theparts
partsmodelled.
modelled.

4.1.
4.1. Conduction
Conduction Modelling
Modelling
Depending
Dependingon onthethe
brake test test
brake that is being
that is modelled and the vehicle
being modelled and thesetup, the brake
vehicle setup,discthetemperature
brake disc
can reach up to 500–700 ◦ C or higher. Other parts, even though reaching lower temperatures, are still
temperature can reach up to 500–700 °C or higher. Other parts, even though reaching lower
subjected
temperatures,to significant spatial and
are still subjected temporal temperature
to significant gradients,
spatial and temporal and as a result,
temperature their material
gradients, and as a
properties
result, their material properties are going to be affected. Hence, to model thermal conductionofwithin
are going to be affected. Hence, to model thermal conduction within parts, it is high
importance to obtain temperature dependent properties, mainly specific heat
parts, it is of high importance to obtain temperature dependent properties, mainly specific heat and and thermal conductivity,
for all parts
thermal participating
conductivity, for in
allthe heat
parts dissipation,in
participating orthe
at least
heat for the brakeor
dissipation, disc
at and
leastbrake
for thepads.
brakeThese
disc
properties can also often be anisotropic, especially for the pads.
and brake pads. These properties can also often be anisotropic, especially for the pads.
The
The conduction
conduction between
between partsparts in in typically
typically handled
handled by by thermal
thermal interfaces
interfaces oror thermal
thermal links.
links. Two
Two
surfaces in contact always have a certain amount of thermal contact resistance.
surfaces in contact always have a certain amount of thermal contact resistance. This resistance This resistance depends
on many parameters,
depends including material
on many parameters, including properties,
material surface roughness,
properties, surfaceand the contact
roughness, and pressure [20],
the contact
which can change substantially even depending on the position of tightening
pressure [20], which can change substantially even depending on the position of tightening bolts [10]. bolts [10]. In the
simulation, the contact
In the simulation, resistance
the contact can be can
resistance added to thermal
be added links orlinks
to thermal modelled by adding
or modelled virtualvirtual
by adding layers
of material between components. It can also be used to replace thin parts
layers of material between components. It can also be used to replace thin parts that would requirethat would require too fine
meshing otherwise, for example, brake pad shims. However, the assumed
too fine meshing otherwise, for example, brake pad shims. However, the assumed values to be used values to be used for contact
resistance
for contactareresistance
quite difficult to estimate
are quite as there
difficult are many
to estimate asvariables
there are that mustvariables
many be consideredthat [21].
must be
One area
considered [21]. that requires particular attention is the contact between the piston and calliper body
whereOnetheyarea
are that
separated
requiresby particular
a thin brake fluid film.
attention is the contact between the piston and calliper body
where they are separated by a thin brake fluid film.
4.2. Radiation Modelling
One of theModelling
4.2. Radiation easiest approaches to model radiation between parts is to use surface-to-surface radiation
model. This model requires emissivity coefficients being assigned for every surface participating in the
One of the easiest approaches to model radiation between parts is to use surface-to-surface
radiation model. This model requires emissivity coefficients being assigned for every surface
participating in the heat transfer process. These coefficients can be measured (e.g., using reflectometer)
or estimated most of the times, but for some of the surfaces, it can be a challenging task.
Energies 2020, 13, 203 6 of 12

heat transfer
Energies 2020, 13,process.
x FOR PEERThese coefficients can be measured (e.g., using reflectometer) or estimated 6most
REVIEW of 12
of the times, but for some of the surfaces, it can be a challenging task.
The brake
The brake disc
disc is
is aapart
partthat
thatisismostly
mostlyaffected
affectedby bythe
theradiation
radiationheat transfer
heat transfer asasit is
it subjected
is subjected to
the highest temperatures. Therefore, the authors have measured emissivity of
to the highest temperatures. Therefore, the authors have measured emissivity of a large variety of a large variety of brake
discs (all
brake discsmade of similar
(all made cast cast
of similar ironiron
alloy material),
alloy material),newnewandand used,
used,and
andfound
foundthatthatthethe emissivity
emissivity
coefficients can vary between 0.15 to 0.9 depending on the disc surface conditions.
coefficients can vary between 0.15 to 0.9 depending on the disc surface conditions. This finding is This finding is also
also
supported by
supported other investigations
by other investigations available
available inin literature
literature [10,22].
[10,22]. Furthermore,
Furthermore, operating
operating brake disc
brake disc
would have
would have different
different emissivity
emissivity values
values compared
compared to to one
one that
that is
is standing
standing still.
still. This
This is
is due
due toto the
the fact
fact
that the emissivity values are temperature dependent and also because there
that the emissivity values are temperature dependent and also because there is usually an oxidation is usually an oxidation
layer at
layer at the
the surfaces
surfacesof ofthe
thebrake
brakethat
thatisisbeing
beingcleaned
cleaned away
away bybythethe brake
brake pads
pads as soon
as soon as the
as the brakes
brakes are
are applied while the vehicle is
applied while the vehicle is moving [23]. moving [23].
As the
As themesh
meshofofthe the brake
brake disc
disc is often
is often modelled
modelled stationary
stationary in a thermal
in a thermal model model
even duringeven driving
during
driving scenarios,
scenarios, an extraan extra
step step is required
is required to account
to account for the for the rotation
rotation and local
and avoid avoidhot localandhotcold
andspots
cold
spots formation on the disc surfaces. The radiation patches on each of the two sides
formation on the disc surfaces. The radiation patches on each of the two sides of the brake disc should of the brake disc
should
be be combined
combined to workto aswork as asource/receiver
a single single source/receiver of radiation.
of radiation.

4.3. Brake Fluid Modelling

There are several approaches to model the brake fluid and the selectionselection between them would
depend on on the
theimportance
importanceofofthethe
brake fluid
brake simulation
fluid for afor
simulation particular test scenario
a particular that isthat
test scenario being
is
simulated.
being simulated.
The first and the simplest method is to use a single concentrated mass node representing brake
fluid inside the calliper. In this case, thermal links
links need
need to be set up in between this node and the
corresponding surfaces of the piston and calliper. This This method
method implies that the entire brake fluid
temperature.
volume has the same temperature.
The second approach is to have the brake fluid volume wrapped and meshed so that the brake
simulated as aa solid
fluid can be simulated solid body
body with
with fluid
fluid material
material properties.
properties. This method
method is much more
accurate, but as the mixing of the fluid due to convection is not allowed, it can result in rather high
temperature gradients inside the fluid; see Figure 4. Consequently, one should pay attention to the
measurement sensor
position of the measurement sensor during
during the
the experiment
experiment and
and its
its equivalent
equivalent in
in the
thesimulation.
simulation.

Figure 4. Example of temperature gradients inside the brake fluid volume.

Figure 4. Example of temperature gradients inside the brake fluid volume.
Another option is to try simulating brake fluid as fluid together with the brake lines, but that
would Another
increaseoption is to try simulating
the computational cost of brake fluid as fluid
such simulations together
even more. with the brake lines, but that
would increase the computational cost of such simulations even more.
4.4. Energy Input and Disc-to-Pads Contact
4.4. Energy Input and Disc-to-Pads Contact
Measuring the amount of energy that goes into the brakes and its distribution between inboard
Measuringsides
and outboard the amount of energy
of the disc that hard
are rather goes issues
into thethat
brakes and itstodistribution
are going between
be discussed in the inboard
Testing
and outboard
section. sides
From the of the disc
simulation areitrather
side, hard to
is possible issues thatthis
address are problem
going to using
be discussed in the method
finite element Testing
section.approach
(FEM) From the[24,25].
simulation side, itnot
However, is possible to address
to increase this problem
the complexity of theusing finite element
simulations, method
the amount of
(FEM) approach [24,25]. However, not to increase the complexity of the simulations, the amount of
thermal energy added to the system can be estimated. Yet, even the application of this energy inside
the thermal model may be a challenging task.
Energies 2020, 13, 203 7 of 12

thermal energy
Energies 2020, added
13, x FOR PEERtoREVIEW
the system can be estimated. Yet, even the application of this energy inside 7 of 12
the thermal model may be a challenging task.
One of the approaches to add energy to the brake system is to apply volumetric heat flux to the
circular area of the disc that passes underunder the brake pads. At the same time, in real life, the energy is
generated by friction in the contact pair between the brake disc and and brake
brake pads.
pads. The energy is then
distributed between the contacting surfaces. Therefore, Therefore, the
the volumetric
volumetric flux approach may lead to
overprediction of
overprediction of the the disc temperatures
temperatures and incorrect energy
incorrect energy flows for the contacts between the brake
disc and brake pads.
An alternative approach is to use virtual thermal nodes in between the brake disc and the the brake
pad surfaces that go in in contact when the brakes are applied.
are applied. Figure 5 shows an example of how this
type of surfaces can be defined. The energy is then “added” directly to the virtual node, allowing to
have heat fluxes, that go into the brake pads and and the brake disc,
disc, changing over time based on the
material properties
properties and temperatures of the corresponding parts as
and temperatures of the corresponding parts as they
they would
would in
in reality.
reality.

(a) (b)
Figure 5. Extra
Extrasurface
surfacepart
partIDs
IDs(PIDs)
(PIDs)assigned
assignedforfor
brake disc
brake in in
disc green (a),(a),
green andand
thethe
brake padpad
brake in purple (b).
in purple
(b).
If the simulated test involves stages of driving without braking or soaking phases, the thermal
If the simulated
links between the disctest
andinvolves
the padsstages
mustof
bedriving without
adjusted braking or soaking phases, the thermal
accordingly.
links between the disc and the pads must be adjusted accordingly.
5. Testing
5. Testing
The numerical models and methods developed for replicating any physical test need to be
The numerical models and methods developed for replicating any physical test need to be
validated against real life experimental data. Therefore, it is important to know and understand the
validated against real life experimental data. Therefore, it is important to know and understand the
limitations and possible issues with the experimental testing. This chapter covers some of the vital
limitations and possible issues with the experimental testing. This chapter covers some of the vital
testing issues that are difficult to predict or incorporate into numerical models.
testing issues that are difficult to predict or incorporate into numerical models.
5.1. Typical Test Scenarios and On-Road Testing
5.1. Typical Test Scenarios and On-Road Testing
One of the easiest ways to test brake cooling performance is to do it on a component level, for
One of the easiest ways to test brake cooling performance is to do it on a component level, for
example, using brake dynamometer for reduced brake assembly. Unfortunately, component tests,
example, using brake dynamometer for reduced brake assembly. Unfortunately, component tests, even
even with partial vehicle geometry [11], do not represent the on-road driving as there are many
with partial vehicle geometry [11], do not represent the on-road driving as there are many uncertainties
uncertainties that must be accounted for one way or the other.
that must be accounted for one way or the other.
Two typical full-scale on-road vehicle tests scenarios validating brake cooling performance are:
Two typical full-scale on-road vehicle tests scenarios validating brake cooling performance are:
mountain descent (also known as alpine descent) and the “Auto Motor und Sport” (AMS) test.
mountain descent (also known as alpine descent) and the “Auto Motor und Sport” (AMS) test.
Mountain descent test, as the name implies, is conducted on mountains and includes two phases:
Mountain descent test, as the name implies, is conducted on mountains and includes two phases:
1. Heat-up
1. phase: low
Heat-up phase: lowspeed
speeddown-hill
down-hill driving
driving withwith constantly
constantly engaged
engaged brakesbrakes on aofslope
on a slope of
around
around
10%; 10%;
2. Thermal soaking phase: the vehicle stands still, and the heat from the brakes is transferred into
2. Thermal soaking phase: the vehicle stands still, and the heat from the brakes is transferred into
the brake system and the environment.
the brake system and the environment.
AMS test consists of 10 repeated cycles of accelerating up to 130 km/h and braking from 130 to
0 km/h on a flat surface that are sometimes followed by the cool-down phase with constant speed
driving or standing still.
Energies 2020, 13, 203 8 of 12

AMS test consists of 10 repeated cycles of accelerating up to 130 km/h and braking from 130 to
0 km/h on a flat surface that are sometimes followed by the cool-down phase with constant speed
driving or standing still.
Any of the on-road tests will inevitably suffer from the lack of repeatability due to unpredictable
and uncontrollable environment. The main considerations are the following:

• Air temperature and humidity variations: outside air conditions are changing every day; moreover,
considering mountain descent, these conditions can be significantly different at the top and the
bottom of the mountain, which in some cases could be compensated by proper monitoring of the
driving conditions and calibration of potential numerical factors;
• Wind conditions: wind speed and direction change all the time. Considering the cases of low
speed driving and especially soaking phase when the vehicle is standing still, the presence of
wind can significantly affect the cooling performance of the brake system;
• Initial part temperature variations: quite often, several tests are conducted one after another to
test various brake and wheel specifications or counteract repeatability issues. Unfortunately, it is
almost impossible to get the vehicle into the same exact initial state as it was in before the first run;
• Road variations and driver behaviours: none of the roads or drivers are the same, and even if
driving the same road with the same driver there will be variations in how the vehicle behaves
and what forces it is subjected to. The usage of a driving robot may remove driver behaviour
from the equation, but even there, the road will never be exactly the same and neither will the
weather conditions.

A lot of these variations can be removed if the tests are simplified and ran in a more controlled
environment, but it also means that some of the important variables can be lost. Another approach to
tackle repeatability issues is to use statistical analysis; however, testing is expensive and repeating a
test 100 times may not be an option.
Other important considerations for matching the simulations and on-road testing are the fan
control strategies and the vehicle body movement during various parts of the tests. The first has
already been discussed, as for the body movement, when driving downhill, the weight distribution
of the vehicle is shifted forwards affecting ground clearance and, hence, affecting the flow inside the
wheelhouses. Addition of the braking only increases this effect. Moreover, even on the flat road,
inertial radial expansion of tires and aerodynamic lift forces would significantly increase the ground
clearance of the vehicle [26,27]. Furthermore, the magnitude of this effect depends on the vehicle
velocity, which can be of high importance for the AMS tests.

5.2. Wind Tunnel Testing

One of the ways to reduce uncertainties is to conduct tests in a controlled environment. The best
option available is to do it in the wind tunnel. This does not only allow choosing the vehicle and
wind velocity precisely but also permits to have much better instrumentation as the vehicle is fixed in
its position.
Many modern wind tunnels are equipped with various moving ground systems that exist to
prevent the formation of the boundary layer and to simulate the ground movement under the vehicle.
Unfortunately, wheel drive units are often not powerful enough to withstand the high torque required
to replicate braking scenarios. Therefore, braking tests need to be conducted without ground simulation
with only the wheel rotating on the dynamometer, see Figure 6. In the example given, a dynamometer
part of the wind tunnel test section has a stationary floor, the rear wheels are standing still, and some
additional pipes are present to remove the exhaust gasses. All these modifications are providing extra
uncertainties as they alter the airflow compared to on-road conditions.
Energies 2020, 13, 203 9 of 12
Energies 2020, 13, x FOR PEER REVIEW 9 of 12

Figure 6. Test object during the wind tunnel braking investigation.

When
When imitating
imitating mountain descent test
mountain descent test inin the
the wind
wind tunnel, constant braking
tunnel, constant braking force
force has
has to
to be
be
maintained. Unfortunately, increasing temperatures of the brake pads and the
maintained. Unfortunately, increasing temperatures of the brake pads and the brake discs alter the brake discs alter
the friction
friction coefficient
coefficient between
between these these
parts; parts;
hence, hence, the pressure
the pressure in thesystem
in the braking brakingmustsystem must be
be constantly
constantly
adjusted. adjusted.
Most
Most ofof the
the full-scale
full-scale wind
wind tunnels
tunnels have
have aa closed
closed loop
loop setup,
setup, which
which means
means that
that the
the air continues
air continues
to move inside
to move in aa circular
inside in circular motion,
motion, andand it
it is
is impossible
impossible to
to bring
bring the
theair
airspeed
speedtotofull
fullstop
stopimmediately.
immediately.
Consequently,
Consequently, in order to imitate a vehicle stopping at the beginning of the soaking phase, aa wind
in order to imitate a vehicle stopping at the beginning of the soaking phase, wind
shield
shield can
can be
be used
used toto protect
protect the
the front
front of
of the
the vehicle.
vehicle.

5.3. Sensors and

5.3. Sensors and Measurements
Measurements
Thermocouples
Thermocouples are are widely
widelyused usedininindustry
industrytoto measure
measure local
local component
component temperatures.
temperatures. They They
are
are cheap, easy to mount, accurate, and reliable. There are two
cheap, easy to mount, accurate, and reliable. There are two types of thermocouples: embedded types of thermocouples: embedded
thermocouple
thermocouple and and rubbing
rubbing (also (also called
called sliding)
sliding) thermocouple.
thermocouple. In In general, thermocouples are
general, thermocouples are wired
wired
to measuring equipment, which records the voltage produced
to measuring equipment, which records the voltage produced and maps the results to calibrated and maps the results to calibrated
temperatures. Wireless thermocouples
temperatures. Wireless thermocouples can can also
also be be used,
used, butbut they
they areare less
less reliable
reliable duedue to to possible
possible loss
loss
of connection and not convenient because of the added weight and
of connection and not convenient because of the added weight and space taken by the wireless unit. space taken by the wireless unit.
However,
However, they they allow
allow measuring
measuring temperature
temperatureon on rotating
rotatingpartspartseasily.
easily.
Thermal
Thermal cameras
cameras can can bebe used
used in in laboratory,
laboratory, but but they
they have
have the the emissivity
emissivity issue:issue: the
the temperature
temperature
measurement
measurement needs to be compensated at high values due to the change of emissivity, which
needs to be compensated at high values due to the change of emissivity, which as as for
for
today,
today, cannot be dynamically modified while recording. Moreover, the thermal cameras cannot be
cannot be dynamically modified while recording. Moreover, the thermal cameras cannot be
used
used forforbrake
brakedisc disctemperature
temperature measurements
measurements as theas wheel rim blocks
the wheel the view.
rim blocks theFurthermore, thermal
view. Furthermore,
cameras are very expensive
thermal cameras (for the wide
are very expensive (for temperature
the wide temperatureranges observedrangesatobserved
braking),atbulky, and fragile.
braking), bulky,
The
and fragile.usual approach is to embed one thermocouple 2 mm below the outer brake pad friction
surface,
Thei.e., the approach
usual brake pad,iswhich to embed is closer
one to the wheel rim.
thermocouple 2 mm Thebelowthermocouple
the outerisbrake inserted padfrom the
friction
backplate, making sure that the piston or calliper finger will not
surface, i.e., the brake pad, which is closer to the wheel rim. The thermocouple is inserted from the put pressure on the wire, which
could damage
backplate, making it orsure
in the thatworst case, cut
the piston or it. Embedding
calliper the not
finger will thermocouple
put pressure at the
on thefriction
wire,surface
which
can
could damage it or in the worst case, cut it. Embedding the thermocouple at the friction Also,
destroy it since embedded thermocouples are not meant to be subjected to friction. surface when
can
inserting the thermocouple, the hole can go through or stop 2
destroy it since embedded thermocouples are not meant to be subjected to friction. Also, when mm before the friction surface, so that
the thermocouple
inserting is not exposed
the thermocouple, to frictional
the hole can go through debris. or stop 2 mm before the friction surface, so that
When brake disc temperatures
the thermocouple is not exposed to frictional debris.are recorded in some of the experiments, rubbing thermocouples
are often
When used
braketo disc
measure the average
temperatures are surface
recorded temperature
in some of the on experiments,
a certain radius at thethermocouples
rubbing contact patch.
Also, pyrometers can be used for the same result. Alternatively,
are often used to measure the average surface temperature on a certain radius at the contact the brake disc can also be drilledpatch.
with
thermocouples
Also, pyrometers embedded
can be used underforitsthesurface.
same result.This setup would require
Alternatively, the brakea slip ring
disc canconnector on the
also be drilled
outside of the wheel to transfer the signal from the rotating disc,
with thermocouples embedded under its surface. This setup would require a slip ring connector on similar to the one in Figure 6.
When the
the outside brake
of the fluidtotemperature
wheel transfer theneeds signaltofrom be measured,
the rotating a special thermocouple
disc, similar to the one embedded
in Figurein6.the
brakeWhen
calliperthebleeding
brake fluid screw is normallyneeds
temperature used;to seebeFigure 7. This
measured, is not where
a special the brakeembedded
thermocouple fluid reaches in
its
the brake calliper bleeding screw is normally used; see Figure 7. This is not where thetemperature
highest temperature, but this is the only location that is safe enough to measure the brake fluid
without
reaches affecting
its highest thetemperature,
strength of the butcalliper.
this is Brake
the only fluidlocation
temperature,that isassafe
provided
enough by to
themeasure
experiment,the
will be a single point measurement, while as it was mentioned earlier
temperature without affecting the strength of the calliper. Brake fluid temperature, as provided by in Section 4.3, CAE simulations
allow seeing the will
the experiment, brake befluid temperature
a single gradient. while as it was mentioned earlier in Section 4.3,
point measurement,
CAE simulations allow seeing the brake fluid temperature gradient.
Energies2020,
Energies 2020,13,
13,203
x FOR PEER REVIEW 10 of
10 of 12
12

Figure 7.
Figure 7. A modified
modified bleeding
bleeding screw
screw for
for measuring
measuring brake
brake fluid
fluid temperature.
temperature.

Capturing
Capturingspatial
spatialtemperature
temperaturegradients
gradientsof of
thethe
brake fluid
brake experimentally
fluid is hardly
experimentally possible,
is hardly and
possible,
even for the parts that are easier to access, there is always a limit of how many thermocouples
and even for the parts that are easier to access, there is always a limit of how many thermocouples can be
used.
can beAny
used.additional thermocouples
Any additional do notdo
thermocouples only
notincrease the time
only increase required
the time for test
required forpreparation but
test preparation
also alteralter
but also the airflow around
the airflow parts.
around parts.

5.4.
5.4. Disc
Disc Deformations
Deformations and
and Hot-Spotting
Hot-Spotting
Most
Most of of the
theparts
partswillwillexperience
experiencethermal
thermalexpansion
expansion whenwhen subjected
subjected to tohigh
hightemperatures.
temperatures. For For
the
the brake cooling tests, the most important deformations happen to the disc as it experiences highest
brake cooling tests, the most important deformations happen to the disc as it experiences highest
temperature
temperaturechanges.
changes.Unfortunately,
Unfortunately, thermal
thermalexpansion
expansionis not
is the
notonly
the deformation that occurs
only deformation that during
occurs
braking. Conning is the type of deformation that occurs when the brake
during braking. Conning is the type of deformation that occurs when the brake disc deforms disc deforms elastically and/or
plastically
elastically at the neck
and/or due to cyclic
plastically at the cooling
neck due andto heating. Another
cyclic cooling andimportant effect is hot-spotting.
heating. Another important effect
Hot-spotting
is hot-spotting. is related to a phenomenon called Thermo-Elastic Instabilities (TEI) [28]. When
braking, the pair of brake pads is pushed towards the brake disc, which,
Hot-spotting is related to a phenomenon called Thermo-Elastic Instabilities (TEI) [28]. in return, generates frictional
When
heat flux the
braking, at the
pairsliding
of brake interface, which is
pads is pushed proportional
towards the braketo the
disc,contact
which,pressure. Therefore,frictional
in return, generates if some
uneven
heat fluxpressure distribution
at the sliding arises
interface, at the
which is friction surface,
proportional the contact
to the areas where the pressure
pressure. Therefore,is ifhigher
some
experience a higher temperature increase. This, in turn, causes greater
uneven pressure distribution arises at the friction surface, the areas where the pressure local thermal expansion and,
is higher
thereby, leads
experience to further
a higher local pressure
temperature increase
increase. This,[29]. Hot-spot
in turn, causespositions change
greater local with every
thermal new load
expansion and,
cycle; therefore, measuring the disc temperatures with embedded thermocouples
thereby, leads to further local pressure increase [29]. Hot-spot positions change with every new load can have serious
repeatability issues
cycle; therefore, even when
measuring thetesting in the wind tunnel.
disc temperatures with embedded thermocouples can have serious
Any uneven pressure distribution at the
repeatability issues even when testing in the wind tunnel. contact between the brake disc and brake pads leads
to uneven deformations and, consequently,
Any uneven pressure distribution at the contact betweennon-uniform heatthe
fluxes at the
brake disccontact
and brake interface. These
pads leads to
effects
uneven aredeformations
typically unaccounted for in the brake
and, consequently, cooling simulations
non-uniform heat fluxesbut at should be kept
the contact in mind These
interface. when
comparing
effects are the resultsunaccounted
typically between realfor andinvirtual models.
the brake cooling simulations but should be kept in mind
when comparing the results between real and virtual models.
5.5. Energy Input Estimation
5.5. Energy Input Estimation
Temperatures in the brake system are directly affected by the amount of energy that goes into
contact between the brake disc brake and pads. However, estimating this amount for each pad-to-disc
Temperatures in the brake system are directly affected by the amount of energy that goes into
contact is not a trivial task.
contact between the brake disc brake and pads. However, estimating this amount for each pad-to-
disc When
contactthe vehicle brakes, one can have a good estimation of changes in kinetic and potential
is not a trivial task.
energies. Unfortunately, not all of this energy goes into the braking system. Some of it is spent on
When the vehicle brakes, one can have a good estimation of changes in kinetic and potential
overcoming rolling and aerodynamic resistance and some ends up in drivetrain and other losses.
energies. Unfortunately, not all of this energy goes into the braking system. Some of it is spent on
Moreover, depending on the vehicle weight distribution and brake system setup, front and rear brakes
overcoming rolling and aerodynamic resistance and some ends up in drivetrain and other losses.
would have an uneven split between them. Typically, front brakes take most of the load due to higher
Moreover, depending on the vehicle weight distribution and brake system setup, front and rear
vertical force on front wheels and vehicle stability reasons. Additionally, since the vehicles are not
brakes would have an uneven split between them. Typically, front brakes take most of the load due
symmetrical, brakes on the left and right side of the vehicle may have slightly different temperatures,
to higher vertical force on front wheels and vehicle stability reasons. Additionally, since the vehicles
which would lead to changing friction and different energy input for them. The same principle applies
are not symmetrical, brakes on the left and right side of the vehicle may have slightly different
to inboard and outboard sides of the brake disc where the energy input would differ. Furthermore,
temperatures, which would lead to changing friction and different energy input for them. The same
the ratios applies
principle in thesetosplits would
inboard andchange during
outboard sidesthe braking
of the braketest
disccycle,
where making it almost
the energy inputimpossible to
would differ.
estimate them exactly at any given moment.
Furthermore, the ratios in these splits would change during the braking test cycle, making it almost
impossible to estimate them exactly at any given moment.
Energies 2020, 13, 203 11 of 12

During the tests in the wind tunnel, the amount of energy added to the braking system is easier to
estimate as the braking force is recorded by the dynamometer. Nevertheless, the amount of losses in
various system parts and the distribution between two contact surfaces of the disc are still unclear.

6. Concluding Remarks
Brake cooling performance evaluation is important since it allows safe vehicle operation in even
the most extreme circumstances. Currently, most of this evaluation is performed using expensive
experimental testing methods; however, there is extensive ongoing work to replicate such testing using
CAE simulations. In this paper, some important effects, findings, and issues needed to be considered
during this process were summarized.
Brake cooling CAE simulations are complex since a large amount of material data and other
properties are missing or even impossible to estimate. Additionally, there are several simulation
approaches, all having different advantages and disadvantages, and a common denominator that they
all require considerable simplifications.
Nevertheless, it has been shown that CAE simulations give satisfactory results, especially when
comparing different design choices. The simulations are not only cheaper but can also be performed
at earlier stages of the vehicle development process, even when there are no prototypes available.
Furthermore, as the simulations have perfect control over used boundary conditions, they provide
perfect repeatability, which is something that the experimental testing is certainly struggling with.
In addition, virtual models give important physical insights into the complicated heat dissipation
process. The CAE simulations provide not only the temperature gradients within parts, but also the
heat flux distributions for convection, conduction, and radiation for every part surface or part contact
at any given time. For instance, experimentally determined brake fluid temperature is measured in a
single point, while CAE simulations offer spatial temperature gradients inside the calliper.
At this time, it is not yet possible to replace the experimental testing with simulations, but two
approaches can be used together to complement each other. Moreover, as the computational resources
become more affordable, the amount of virtual testing will increase. However, much more work is still
needed to improve the existing methods and develop new ones before they can completely replace the
physical testing.

Author Contributions: Most of the writing was performed by A.V., G.L.G. contributed with a substantial part of
the “Testing” chapter as well as provided feedback, editing and proof-reading for the other parts of the article. All
authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Acknowledgments: The authors would like to thank engineers from Brakes, Thermodynamics and Aerodynamics
groups at Volvo Car Corporation for their time and support that made this work possible.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Stephens, A. Aerodynamic Cooling of Automotive Disc Brakes. Master’s Thesis, RMIT University, Melbourne,
Australia, 2006.
2. Emery, A.F.; Kumar, P.; Firey, J.C. Experimental Study of Automotive Brake System Temperatures; Washington
State Transportation Center, University of Washington: Washington, DC, USA, 1997.
3. Michelin Truck Tire Service Manual; Michelin Tire Company: Clermont-Ferrand, France, 2004.
4. Ratamero, L.; Ferreira, R. Wheel Cover Affected by the Heat Generated in the Brake System; SAE Technical Paper
2010-36-0380; SAE: Warrendale, PA, USA, 2010. [CrossRef]
5. Palmer, E.; Mishra, R.; Fieldhouse, J.; Layfield, J. Analysis of Air Flow and Heat Dissipation from a High Performance
GT Car Front Brake; SAE Technical Paper 2008-01-0820; SAE: Warrendale, PA, USA, 2008. [CrossRef]
6. Barigozzi, G.; Perdichizzi, A.; Donati, M. Combined Experimental and CFD Investigation of Brake Discs
Aero-thermal Performances. SAE Int. J. Passeng. Cars-Mech. Syst. 2009, 1, 1194–1201. [CrossRef]
Energies 2020, 13, 203 12 of 12

7. Reddy, S.; Mallikarjuna, J.; Ganesan, V. Flow and Heat Transfer Analysis of a Ventilated Disc Brake Rotor Using
CFD; SAE Technical Paper 2008-01-0822; SAE: Warrendale, PA, USA, 2008. [CrossRef]
8. Yigit, S.; Penther, P.; Wuchatsch, J.; Werner, F. A Monolithic Approach to Simulate the Cooling Behavior of
Disk Brakes. SAE Int. J. Passeng. Cars-Mech. Syst. 2013, 6, 1430–1437. [CrossRef]
9. Choi, B.; Park, J.; Kim, M. Development of the Virtual Test Technology for Evaluating Thermal Performance
of Disc Brake. Trans. Korean Soc. Automot. Eng. 2009, 17, 40–47. [CrossRef]
10. Voller, G.; Tirović, M.; Morris, R.; Gibbens, P. Analysis of automotive disc brake cooling characteristics. Proc.
Inst. Mech. Eng. Part J. Automob. Eng. 2003, 217, 657–666. [CrossRef]
11. Stephens, A.; Watkins, S.; Dixon, C. Aerodynamic Testing of a Vented Disc Brake; SAE Technical Paper
2003-01-0932; SAE: Warrendale, PA, USA, 2003. [CrossRef]
12. Schuetz, T. Cooling Analysis of a Passenger Car Disk Brake; SAE Technical Paper 2009-01-3049; SAE: Warrendale,
PA, USA, 2009. [CrossRef]
13. Palmer, E.; Jansen, W. Development of a High Fidelity CAE Model for Predicting Brake System Temperatures; SAE
Technical Paper 2017-01-0145; SAE: Warrendale, PA, USA, 2017. [CrossRef]
14. Vdovin, A.; Gustafsson, M.; Sebben, S. A coupled approach for vehicle brake cooling performance simulations.
Int. J. Therm. Sci. 2018, 132, 257–266. [CrossRef]
15. Bhambare, K.; Haffey, M.; Jelic, S. Brake Duty Cycle Simulation for Thermal Design of Vehicle Braking System;
SAE Technical Paper 2013-36-0015; SAE: Warrendale, PA, USA, 2013. [CrossRef]
16. Yigit, S. A CFD Method to Simulate Brake Liquid and Disc Temperature Monolithically. In Proceedings of
the 12th FKFS Conference, Stuttgart, Germany, 1–2 October 2019.
17. Morello, G. Simulation of Transient Thermal Situation in Hill Driving Using CFD. Master’s Thesis, Chalmers
University, Gothenburg, Sweden, 2018.
18. Vdovin, A.; Löfdahl, L.; Sebben, S. Numerical and Experimental Investigations of the Brake Cooling for
Passenger Cars. ATA Ing. Dell Auto. 2015, 68, 11–21.
19. Dekker, T.; Krishnan, V. Numerical Investigations of Brake Cooling Performance. Master’s Thesis, Chalmers
University, Gothenburg, Sweden, 2018.
20. Yovanovich, M. Four Decades of Research on Thermal Contact, Gap, and Joint Resistance in Microelectronics.
IEEE Trans. Compon. Packag. Technol. 2005, 28, 182–206. [CrossRef]
21. Babu, K.N. Thermal Contact Resistance: Experiments and Simulation. Master’s Thesis, Chalmers University,
Gothenburg, Sweden, 2015. Available online: https://hdl.handle.net/20.500.12380/223157 (accessed on 30
December 2019).
22. Eisengräber, R.; Grochowicz, J.; Schuster, M.; Augsburg, K.; Koch, L. Comparison of Different Methods for
the Determination of the Friction Temperature of Disc Brakes. SAE Trans. 1999, 498–504. [CrossRef]
23. Dragomir, G.; Pancu, R.; Bungau, C.; Beles, H.; Georgescu, L. Studies About Emissivity Variation Depending
on The Temperature for Car Brake Disc. Ann. Oradea Univ. Fascicle Manag. Technol. Eng. 2014, 2014, 253–256.
[CrossRef]
24. Wawrzonek, L.; Bialecki, R.A. Temperaturein a disk brake, simulation and experimental verification. Int. J.
Numer. Methods Heat Fluid Flow 2008, 18, 387–400. [CrossRef]
25. Adamowicz, A.; Grzes, P. Convective cooling on a disc brake during single braking. Acta Mech. Autom. 2012,
6, 5–10.
26. Schnepf, B.; Tesch, G.; Indinger, T. On the Influence of Ride Height Changes on the Aerodynamic Performance
of Wheel Designs. Int. J. Automot. Eng. 2015, 6, 23–29.
27. Vdovin, A.; Löfdahl, L.; Sebben, S.; Walker, T. Investigation of vehicle ride height and wheel position influence
on the aerodynamic forces of ground vehicles. In Proceedings of the International Vehicle Aerodynamics
Conference, Loughborough, UK, 14–15 October 2014; pp. 81–90.
28. Barber, J.R. Thermoelastic instabilities in the sliding of conforming solids. Proc. R. Soc. Lond. Math. Phys. Sci.
1969, 312, 381–394. [CrossRef]
29. Le Gigan, G.; Vernersson, T.; Lundén, R.; Skoglund, P. Disc brakes for heavy vehicles: An experimental study
of temperatures and cracks. Proc. Inst. Mech. Eng. Part J. Automob. Eng. 2015, 229, 684–707. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).