MODELING AND DESIGN OPTIMIZATION OF A FORMULA

STUDENT RACE CAR

K. Arvanitopoulos-Darginis
1
, A. Pouchias
1
, K. Zisis
1
, C. Mantelakis
1
, K.-E.
Karyotakis
1
,
L. Ntziachristos
1,2
, N. Michailidis
1,3
*

1
Aristotle Racing Team, Aristotle University of Thessaloniki, GR-54124 Thessaloniki,
Greece
2
Laboratory of Applied Thermodynamics, Mechanical Engineering Department,
Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece
3
Physical Metallurgy Laboratory, Mechanical Engineering Department, Aristotle
University of Thessaloniki, GR-54124 Thessaloniki, Greece

KEYWORDS
Formula Student, FEM Modeling, Components design, Optimization

ABSTRACT
Formula Student Series is a competition where teams from universities around the
world design and construct a single-seater race car to participate in related
international events. Teams have to conform to a set of rules imposing restrictions at
the general specifications of the car, such as the dimensions, the engine capacity
and the safety. The paper presents a series of case studies from the use of Finite
Element Method (FEM) modeling and design optimization of various components
from the current race car of the Aristotle Racing Team (ART), investigated with the
aid of ANSA and software packages. More specifically, various structural parts
were modeled and analyzed to reduce weight and deformation, given the restrictions
imposed by materials mechanical and physical properties. Moreover, extensive
Computational Fluid Dynamics (CFD) analysis of the intake manifold was performed
for optimizing the air flow to obtain a better engine performance. Additionally, multiple
crash analyses of the front part of the car were carried out for reducing the number of
the experiments, thus minimizing cost and development time of the car impact
attenuator. The use of optimization tools in several components, led to a significant
reduction of weight without compromising in structural stiffness.

1. INTRODUCTION

Aristotle Racing Team ART was created in 2006 at the Mechanical Engineering
Department of the Aristotle University of Thessaloniki. Since then, ART has
successfully developed two race-cars, which have competed in several European
FSAE events. Based on the teams experience from previous years, ART started the
design of its third vehicle in September 2011 (see Figure 1). The
Figu

aim
of 8
201
The

The
ART
pac

2. C

In th
syst
the
dec
than
Figu
ure 1 ART
was to rem
8 months. Th
2, contribut
e basic desi
Lower
previou
Improve
Increas
e present pa
T13 race
ckages of BE
Carbon Driv
he concept
tem (see F
drive-shaft
celeration pe
n 400 gr fro
ure 2 - 3D v
T13 single-s
main compe
he feedbac
ted to a furt
gn objective
the centre
us car and re
e vehicle er
se engine pe
aper describ
car, with r
ETA-CAE S
ve-shaft
of weight
igure 2) fro
ft reduces
erformance
m each driv
view of drive
seater race
titive while
k from the j
ther evolutio
es of ART1
of gravity
educe weig
rgonomics a
erformance
bes the des
reference t
Systems.
reduction, A
om steel tub
rotating m
. The use o
ve-shaft.
e-train syste
car, in the f
completing
udges in 2
on of the ve
3 were to:
(CG) by at
ght.
and driveab
e and efficie
sign aspect
to modellin
ART conve
be to a carb
mass which
of carbon fi
em
frame of Fo
the project
Formula St
ehicle.
t least 25%
bility.
ncy.
s of severa
ng solution
rted the dri
bon fibre on
h is benefi
bre leads t
ormula Stud
t within a tig
udent partic
% compared
l parts and
s facilitate
ve-shaft for
ne. Removi
cial for ac
o a weight
dent.
ght time-fram
cipations in
d to the tea
componen
ed by softw
r the drive-
ing weight f
cceleration
saving of m


me

ams
ts of
ware
train
from
and
more

Design and Technical requirements

Maximum torque was calculated to 700 Nm by considering 1.3 g longitudinal
acceleration to the system and by applying a safety factor of 2. The length of each
tube is 308 mm long, with an inner diameter of 45 mm and wall thickness of 2.95
mm. Under these design constraints it was tried to model carbon fibre tubes using for
main material orientation 45
0
.

F.E. Analysis and Model Comparison

The type of analysis applied was static structural and the objectives were to increase
the torsional stiffness and reduce weight. Shell elements were used for the carbon
fibre tube and a transversely orthotropic material was defined with the use of material
matrix MAT8.
First, an approximation model with the use of laminate tool was made, trying to
achieve a filament winding model. The thickness of the layers was in the range of 0.3
mm to 0.5 mm. The first layer has a thickness of 0.5 mm and orientation 5
0
as
described from Brazier effect. The rest layers have a thickness in the range of 0.3
mm to 0.4 mm and orientation 45
0
. The mechanical properties used in the model
are given in Table 1.

Table 1 Mechanical properties of the laminated material.
Symbols E
1
E
2
v
12
G
12
RHO
Units GPa GPa - GPa g/cc
Value 135 10 0.3 5 1.60

The maximum rotational displacement of the carbon fibre tube is 0.047 rad,
distributed as presented in Figure 3.

Figu


3. P

One
syst
thus
con
wou
acco
Des
Mak
and
ped
that
brak
case
con
fram

F.E

The
sim
to d
ure 3 - Rota
Pedal Box a
e of the up
tem that wo
s lap times.
cept, to a t
uld be to h
ommodate
sign and Te
king an adju
d weight, so
dal box is e
t it should b
ke pedal m
e a maxim
strains app
me.
. Analysis
e first objec
ulation that
design a stiff
ational displ
and Brake
pgrades tha
ould allow t
In this con
two-pedal o
have an ad
drivers of d
echnical re
ustable ped
o the minim
essential to
be able to
ust be able
mum displa
ply to the pe
and Optim
ctive was t
t would yield
ffer and ligh
acement (ra
Pedal Anal
at the team
he driver to
text, the pe
one, used ju
djustable pe
different heig
equirement
dal box wou
mization of t
conform to
withstand a
e to withstan
acement of
edal box, w
ization Pro
the proper
d reliable re
ter
ad)
lysis
m decided to
o change ge
edal box had
ust for throt
edal box, a
ght.
ts
ld require g
the volume
o the FSAE
a force from
nd this forc
f 1 mm w
which distrib
ocess
set-up of
esults. Base
o add in th
ears faster,
d to be rede
ttling and b
allowing a
greater widt
e of the ped
E technical
m the driver
ce at a mini
was deeme
butes the fo
a good qu
ed on that, t
he car was
improving
esigned from
raking. A fu
freedom of
h and thus
dal box was
requiremen
r up to 200
mum defor
ed necessa
rces from t
uality mesh
the ultimate
s a hand cl
driveability
m a three-p
urther evolu
f positionin
greater vol
s necessar
nts, which im
00 N. Thus,
rmation, for
ary. The s
the pedal to
and a co
e goal woul
lutch
and
pedal
ution
g to
ume
ry. A
mply
, the
that
ame
o the
rrect
d be

Figu
(bot

ped
was
adju
eno
para
was
ther
Afte
weig
prog
men
TOS
proc
and
The
afte
man
Res
The
non
ure 4 De
ttom right).
dal box. The
s imposed
ust stiffness
ough, or red
ameters. Th
s also follow
re was less
er several o
ght that wa
gram in the
ntioned rest
SCA Struct
cessor platf
d configurati
e optimized
er few more
nufacturing.
sults
e final desig
n adjustable
esign optim
e process c
to the des
s, checking
duce volum
his loop (se
wed on the
space for im
ptimization
s initially se
design pha
trictions, in
ture topolog
form. The
ions an opti
solution ha
e calculatio
.
n was at ha
e pedal box,
ization of t
consisted o
scribed ana
g whether t
me if the de
e Figure 4)
brake and
mprovemen
loops, the n
et. Thus it w
ase so that
the least a
gy optimize
model was
imal result w
ad to be red
n runs, the
alf the weigh
that was u
the pedal b
of an initial
alysis. Sub
that would
eformation
) was repea
throttle ped
nt.
new design
was decide
it would del
amount of ti
er, due to
set-up and
was finally r
designed to
e final desig
ht of the ini
used in the p
box, manua
design of t
sequent ve
add volum
and strains
ated severa
dals but with
ns were not
d to integra
liver the bes
me. The so
its compati
d after adju
reached.
o ensure an
gn of the c
tial one and
previous ca
ally and sof
the new pe
ersions wer
e if the pa
s were insi
l times. The
h fewer iter
converging
ate a topolo
st results w
oftware pac
bility with t
usting the k
n easy man
omponents
d was even
ar. The over
ftware assi
edal box, w
re designe
art was not
ide the def
e same proc
rations beca
g to the targ
ogy optimiza
within the ab
ckage used
the ANSA
key parame
nufacturing,
s was ready
lighter than
rall result w
isted
which
d to
stiff
fined
cess
ause
eted
ation
ove-
was
pre-
eters
and
y for
n the
was a
light
(see

Figu
opti

4. C

Com
mot
aero
hold
com
both
inta

ter, equally
e Figure 5).
ure 5 - von
mum pedal
CFD in Form
mputational
tor sports
odynamics,
ds true for
mpared to o
h around th
ke manifold
y stiff and a

Mises stres
l box design
mula Stude
Fluid Dyn
engineering
as well a
r Formula
other motor
e vehicle an
d is given in
adjustable p
ss distributio
n.
ent
amics play
g both in
s in that o
Student r
r sport class
nd inside th
Figure 6.
pedal box t
on in the ad
ys an increa
the design
of several v
race-cars,
ses, benefi
he engine. A
that conform
djustable pe
asingly imp
n and ana
vehicles su
which, tho
t greatly fro
An example
ms with FS
edal box, as
portant role
alysis of e
ub-compone
ough relativ
om optimiz
of ART13 n
SAE regulat

s defined fo
e in the fiel
external veh
ents. This
vely low-sp
ing the air
newly desig
tions
r the
ld of
hicle
also
peed
flow
gned
Figu

Des

One
aero
grea
con
stud
regu
phy
rest
and

Des

The
Forc
pres
of th
by i
Ano
avo
20%
thes
deta


ure 6 - 3D v
sign and Te
e of the ma
odynamics,
atly contrib
sists of the
dent car is
ulations, wh
ysical limit
trictor, D) th
d E) the runn
sign goals
e FSAE int
cing all the
ssure losse
he restricto
mproving th
other part th
id further a
% of the int
se areas o
achment.
view of the n
echnical re
ain applicati
is the impr
buted to the
following p
a butterfly
hich limits t
to its powe
he plenum,
ners, which
take system
e air to pas
es. To acco
r. Hence, o
he diffuser p
hat required
erodynamic
takes perfo
of recirculat
new intake
equirement
ons of CFD
rovement o
e engines
parts: A) the
y valve, B)
the amount
er, C) the
the volume
are the tub
ms are limi
ss through
unt for thes
our first des
profile throu
attention w
c losses du
ormance. T
ing flow as
manifold.
ts
D in Formul
of the air flo
performan
e throttle sy
the 19
t of air ente
diffuser, to
e where air
bes that con
ited by reg
such a sm
se losses, a
sign goal wa
ugh CFD an
was the inta
ue to second
Thus, the s
s much as
a Student r
ow through
ce. A typic
ystem, whic
mm air re
ering the en
o regain ai
coming fro
nnect the ple
gulations w
mall diamet
a diffuser is
as to minim
nalysis.
ake plenum
dary flow, a
second des
possible a

racing, apar
the intake m
cal FSAE i
h in the cas
estrictor, ac
ngine and t
ir pressure
m the diffus
enum to the
with a 19
ter causes
always use
mize these p
where it wa
as these ac
ign goal w
and avoid
rt from exte
manifold, w
intake man
se of a Form
ccording to
thus provid
e loss from
ser is stabil
e engine.
mm restri
massive s
ed downstr
pressure los
as necessa
ccount for a
was to minim
boundary l
ernal
which
nifold
mula
o the
es a
the
lized
ctor.
static
ream
sses
ry to
bout
mize
ayer
Res

Afte
prev
led
min
(Fig

a)
Figu

As
sym
arra
runn
des
rout
and
(Fig
sults
er analysing
vious intake
us to a new
imal separa
gure 7).

ure 7 (a) A
for the pl
mmetrical p
angement.
ner resultin
igned to gu
te. Finally, d
d were sum
gure 8).
g the flow
e, flow sepa
w diffuser p
ation result

ART9 intake
enum desi
pattern aro
This leads
g in a smo
uide the air
during the d
mmarily corr
through th
aration was
profile desig
ting to high

e diffuser, (
ign, the in
ound the
to the min
oother engin
flow to eac
design proc
rected by te
he restricto
detected a
gned to op
her pressure


b) ART13 in
nitial conce
diffuser in
nimum dive
ne operatio
ch runner w
cess severa
esting each
or- diffuser
at high air ve
erate at ve
e recovery
b)

ntake diffus
ept was to
stead of
ergence of
on. Moreove
without sepa
al areas of r
h design ve
assembly
elocities (M
locities up
at the end
ser.
place the
the more
the flow p
er, the plen
aration occu
recirculation
ersion using
of the tea
Mach>0.5) w
to Mach 1
d of the diff
e runners
typical in
profile for e
num walls w
urring along
n were dete
g CFD ana
ams
which
with
fuser
in a
n-line
each
were
g the
ected
lysis
Figu

5. C

A F
min

Des

Imp
is m
of th
aim
abs
a st
be s
dec
intru
defl
the

F.E

Exp
cras
perf
in th
ure 8 Visu
Crash analy
Finite Elem
imizing the
sign and Te
pact Attenua
much more m
he front par
to the safe
orb at least
tationary ob
such as pro
celeration of
usion plate
ect by more
material ch
. Analysis
perimental t
sh analysis
formance a
he experime
ualised resu
ysis Impa
ment analys
number of
echnical Re
ators length
manoeuvra
rt of the veh
ety of the d
t 7350 J. Ba
bject with a
oviding a ma
f less than
(which in o
e than 25.4
hosen for thi
and Model
tests with t
. Therefore
s the exper
ental tests
ult of CFD a
act attenua
sis of the
required dy
equiremen
h reduction
ble, given t
hicle. Also a
river, it mu
asically , this
velocity of
aximum ave
or equal t
our case is
mm (1 inch
is applicatio
s
the materia
e, a a mod
rimental res
has to be re
analysis of t
ator
car's impa
ynamic tests
nts
was the ma
hat the leng
according to
st be an en
s converts t
impact of 7
erage dece
to 40g. On
s a 2 mm s
h) after the t
on.
al had been
del had to b
sults. Basica
reproduced
he intake p
act attenuat
s to reach t
ain goal in o
gth of this s
o the compe
nergy-absor
to a total m
7.0 m/s. Als
eleration of
the other
solid steel
test. Rohac
n performe
be built tha
ally this mea
by the mod
lenum.
tor is nece
he desired g
order to hav
structure def
etitions rule
rbing struct
mass of 300
so, the defo
less than 2
hand, no p
plate) shou
cell 110 IG (
ed prior to
at would lea
ans that the
del. Also the
essary tow
goal.
ve a vehicle
efines the le
es, which ma
ture that sh
kg running
ormation sh
0g, with a p
part of the
uld permane
(PMI) foam
conducting
ad to the s
e strain reac
e model sh

wards
that
ngth
ainly
ould
into
hould
peak
anti-
ently
was
g the
ame
ched
hould
pred
For
be a
as t
So a
the


Figu
com

A m
dam

6. C

A se
vari
inve
part
rest
dict the acc
the purpos
added that
the contacts
a model wa
strain throu
ure 9 Im
mparison wit
major probl
mping of the
To use
recomm
To reac
Conclusion
eries of cas
ous compo
estigated w
ts were mo
trictions imp
celeration fu
se of such a
this solver
s the elem
as develope
ughout the l
pact attenu
th the teste
em though
e Anti-Intrus
e a PID fo
mended.
ch a more p
ns
se studies
onents of t
ith the aid
odelled and
posed by
unction of ti
analysis, Fi
requires m
ment type
ed as show
oading is si
uator dynam
d accelerat
h with this
sion Plate. S
r the mate
precise strai
with the us
the current
of ANSA a
d analyzed
the mecha
ime and the
inite Eleme
much care w
the materia
wn below (F
imilar to the
mic simulat
ion.
model is t
Some impro
erial that w
n output in
se of FEM
t race car
and s
to reduce
anical and
e damping
nt code Ls-
with many a
al type/data
igure 9). Al
e experimen
tion model
that it cann
ovements fo
would perm
the model.
modelling a
of the Aris
software pa
e weight an
physical p
of the Anti-
-Dyna was
aspects in th
base etc.
so it should
nts.
and accele
not predict
or this mode
it Failure I
and design
stotle Racin
ckages. Va
nd deforma
roperties o
-Intrusion P
used. It sh
he model, s
d be added
eration grap
accurately
el would be:
Index is hi
optimizatio
ng Team w
arious struc
ation, given
of the mate
Plate.
hould
such
that
ph
y the
:
ighly
on of
were
ctural
the
erials

used. Moreover, extensive CFD analysis of the intake manifold was performed for
optimizing the air flow to obtain a better engine performance. Multiple crash
simulations of the front part of the car were carried out for reducing the number of the
experiments, thus minimizing cost and development time of the car impact
attenuator. Also, the use of optimization tools in the design of several components,
led to a significant reduction of weight without compromising in structural stiffness.

7. Acknowledgements

The authors on behalf of the ART would like to thank all the sponsors that supported
the team in the task of creating our current single-seater race car and especially
BETA CAE Systems for financial and technical support

8. References

1. Formula SAE Rules, SAE International, 2013.