analysis and simulation of lap time: innovative tools for design and structural behavior evaluation...
TRANSCRIPT
ANALYSIS AND SIMULATION OF LAP TIME: INNOVATIVE TOOLS
FOR DESIGN AND STRUCTURAL BEHAVIOR EVALUATION OF A
COMPETITION GO-KART CHASSIS
University of Rome “Tor Vergata”
PH.D in Mechanical Systems Design
XXIV Cycle
Ing. Marco Urbinati
Tutor : Prof. Marco E. Biancolini
Coordinator: Prof. Carlo Brutti
University of Rome“Tor Vergata”
Go-Kart... In
tro
du
cti
on
University of Rome“Tor Vergata”
Go-Kart... In
tro
du
cti
on
University of Rome“Tor Vergata”
Go-Kart... In
tro
du
cti
on
University of Rome“Tor Vergata”
Go-Kart... In
tro
du
cti
on
University of Rome“Tor Vergata”
Go-Kart... In
tro
du
cti
on
2
1
Desig
n T
ech
niq
ues
Desig
n T
ech
niq
ue
s
Design processes state of art In
tro
du
cti
on
University of Rome“Tor Vergata”
Telemetry experimental
data
Complex kinematic model
Simplified model
Wo
rk f
low
ch
art
University of Rome“Tor Vergata”
Chassis characterization
Te
lem
etr
y s
iste
m
University of Rome“Tor Vergata”
From the vehicle to PC
Moving
vehicle Sensors Electrical
signal
Anlog to digital
converter
Data stored
into ECU
1
Data stored
into ECU
Data stored
into PC
Base 2 to real
data conversion Graphical representation
and analysis
2
Te
lem
etr
y s
iste
m
University of Rome“Tor Vergata”
From the vehicle to PC
Moving
vehicle Sensors Electrical
signal
Anlog to digital
converter
Data stored
into ECU
1
Data stored
into ECU
Data stored
into PC
Base 2 to real
data conversion Graphical representation
and analysis
2
Te
lem
etr
y s
iste
m
University of Rome“Tor Vergata”
From the vehicle to PC
Moving
vehicle Sensors Electrical
signal
Anlog to digital
converter
Data stored
into ECU
1
Data stored
into ECU
Data stored
into PC
Base 2 to real
data conversion Graphical representation
and analysis
2
Te
lem
etr
y s
iste
m
University of Rome“Tor Vergata”
Go-Kart telemetry system
Dashboard
GPS
Brake& Accelerator
potentiometer
Expansion with
lateral&longitudinal
acceleration sensor
Water, exhoust and tire
temperature sensoors
Exhoust valve sensor
Simplification of the real model to a
material point constrained on a
trajectory.
GPS data: objective
information of the
vehicle position along
the track Sim
pli
fied
mo
de
l
University of Rome“Tor Vergata”
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l INPUT
(GPS data)
Trajectory sampling
Acceleration calculation
Speed profile determination during braking
and acceleration
OUTPUT:
-velocità
-accelerazioni
-laptime
Lap time definition
Comparison with experimental data
Curvature determination along the
trajectory
Chassis data input:
• sproket
• Weight
• tire
Dynamic data input:
• Lateral&longitudinal
acceleration
• Frontal area
• Cd
Engine data input:
• Power
• Torque
• ηtr
Ste
p 1
D
ata
in
S
tep
2
Ev
alu
ati
on
Trajectory sampling and curvature determination
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Trajectory sampling and curvature determination
Coordinate transformation
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Minimum speed point ≡ max lateral acceleration
Trajectory sampling and curvature determination
Coordinate transformation
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Trajectory sampling and curvature determination
Coordinate transformation
Sa
mp
ling
tim
e b
ase
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Trajectory sampling and curvature determination
Coordinate transformation
Sa
mp
ling
tim
e b
ase
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Trajectory sampling and curvature determination
Coordinate transformation
Sa
mp
ling
sp
ace
ba
se
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Trajectory sampling and curvature determination
Coordinate transformation
Sa
mp
ling
sp
ace
ba
se
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Four “models” for curvature determination
are tested:
1. Resampling constant model
2. Interpolation of experimental data
model
3. Curvature direct formula model
4. Osculating circle model
Trajectory sampling and curvature determination
Coordinate transformation
Sa
mp
ling
sp
ace
ba
se
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Four “models” for curvature determination
are tested:
1. Resampling constant model
2. Interpolation of experimental data
model
3. Curvature direct formula model
4. Osculating circle model
Trajectory sampling and curvature determination
Coordinate transformation
Sa
mp
ling
sp
ace
ba
se
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Trajectory sampling and curvature determination
Coordinate transformation
Sa
mp
ling
sp
ace
ba
se
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
[m]
Curvature
Input data…
Acceleration diagram
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Power & Torque
of the engine
5 103
1 104
1.5 104
2 104
0
10
20
30
40
0
10
20
30
Motore_rawid_m otore 2 Motore_rawid_m otore 3
Motore_rawid_m otore 1
Cx & Frontal
Area
Tire data
Defining point of minimum speed ad the speed
profile for braking and acceleration
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Defining point of minimum speed ad the speed
profile for braking and acceleration
Envelope of in and out speed for each corner
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Acceleration Braking
Defining point of minimum speed ad the speed
profile for braking and acceleration
Envelope of in and out speed for each corner
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Acceleration Braking
Tuning of the lumped model
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Min
speed
fitting
Brake
fitting
Exit
fitting
Tuning of the lumped model
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l
Simulated Lap time: 47.172 sec VS Real data lap time: 47.110
Speed
Longitudinal acceleration Lateral acceleration
University of Rome“Tor Vergata”
Sim
pli
fied
mo
de
l Sensitivity Test
rFactor: what and why…
1. Advanced car racing simulation
2. Ability to play any kind of four-wheeled vehicle
3. Includes complex simulation models of the wheels and aerodynamics and a
physics engine with 15 degrees of freedom
4. Open architecture for simulator implementations by external users (cars,
circuits, championship, general add-on)
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
rFactor: what and why…
1. Advanced car racing simulation
2. Ability to play any kind of four-wheeled vehicle
3. Includes complex simulation models of the wheels and aerodynamics and a
physics engine with 15 degrees of freedom
4. Open architecture for simulator implementations by external users (cars,
circuits, championship, general add-on)
5. Can be used with most complex systems of driving simulators
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
rFactor…inside
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
rFactor…inside
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
rFactor…inside
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
rFactor…inside
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
rFactor…inside
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
rFactor…inside
Constraint: a suspension
system for the front and rear
had to exist
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
1 Engine & sproket
KZ Category
KF Category
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
1 Engine & sproket
KZ Category
KF Category
// KF1 25.0 N a 11.500
// Engine data generated by PhysicsEditor
RPMTorque=( 0.0, -0.7, 0.0)
RPMTorque=( 500.0, -1.1, 1.71)
RPMTorque=( 1000.0, -1.6, 3.40)
RPMTorque=( 1500.0, -2.2, 5.09)
RPMTorque=( 2000.0, -2.8, 6.74)
RPMTorque=( 2500.0, -3.4, 8.37)
RPMTorque=( 3000.0, -4.0, 9.96)
RPMTorque=( 3500.0, -4.7, 11.50)
RPMTorque=( 4000.0, -5.3, 12.99)
RPMTorque=( 4500.0, -6.0, 14.42)
RPMTorque=( 5000.0, -6.7, 15.78)
RPMTorque=( 5500.0, -7.4, 17.06)
RPMTorque=( 6000.0, -8.1, 18.27)
RPMTorque=( 6500.0, -8.8, 19.39)
RPMTorque=( 7000.0, -9.5, 20.42)
RPMTorque=( 7500.0, -10.1, 21.36)
RPMTorque=( 8000.0, -10.8, 22.20)
RPMTorque=( 8500.0, -11.4, 22.93)
RPMTorque=( 9000.0, -12.0, 23.56)
RPMTorque=( 9500.0, -12.5, 24.07)
RPMTorque=( 10000.0, -13.0, 24.48)
RPMTorque=( 10500.0, -13.5, 24.77)
RPMTorque=( 11000.0, -13.9, 24.94)
RPMTorque=( 11500.0, -14.2, 25.00)
RPMTorque=( 12000.0, -14.5, 24.75)
RPMTorque=( 12500.0, -14.8, 23.99)
[…]
Code implementation
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
2 Tires
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
2 Tires
//
[SLIPCURVE]
Name="Default"
Step=0.009000 // Slip step
DropoffFunction=0.00 // See explanation above
Data: 0.000000 0.174836 0.349483 0.518060 0.668882 0.790665 0.878928 0.936783 0.971287 0.989751
0.997978 1.000000 0.999865 0.999478 0.998839 0.997952 0.996820 0.995447 0.993838 0.992000
0.989937 0.987659 0.985172 0.982486 0.979609 0.976552 0.973322 0.969932 0.966391 0.962709
0.958897 0.954965 0.950924 0.946785 0.942557 0.938251 0.933876 0.929442 0.924958 0.920432
0.915874 0.911292 0.906693 0.902084 0.897473 0.892866 0.888269 0.883688 0.879128 0.874595
0.870092 0.865624 0.861195 0.856808 0.852466 0.848173 0.843931 0.839741 0.835607 0.831529
0.827510 0.823550 0.819651 0.815813 0.812037 0.808324 0.804674 0.801087 0.797563 0.794103
0.790705 0.787371 0.784098 0.780888 0.777739 0.774651 0.771624 0.768656 0.765747 0.762895
0.760102 0.757364 0.754682 0.752055 0.749482 0.746961 0.744492 0.742074 0.739707 0.737388
0.735117 0.732894 0.730717 0.728585 0.726497 0.724453 0.722452 0.720492 0.718572 0.716693
0.714852 0.713050 0.711285 0.709556 0.707863 0.706205 0.704581 0.702990 0.701431 0.699904
0.698409 0.696943 0.695508 0.694101 0.692722 0.691371 0.690047 0.688750 0.687478 0.686231
0.685009 0.683811 0.682636 0.681484 0.680355 0.679247 0.678161 0.677095 0.676050 0.675025
0.674020 0.673033 0.672065 0.671115 0.670183 0.669268 0.668370 0.667489 0.666624 0.665775
0.664941 0.664123 0.663319 0.662530 0.661754 0.660993 0.660245 0.659511 0.658789 0.658080
0.657383 0.656699 0.656026 0.655365 0.654716 0.654077 0.653449 0.652832 0.652226 0.651629
0.651043 0.650466 0.649899 0.649341 0.648793 0.648253 0.647722 0.647200 0.646686 0.646181
0.645683 0.645194 0.644712 0.644238 0.643771 0.643312 0.642860 0.642415 0.641977 0.641545
0.641120 0.640702 0.640290 0.639884 0.639484 0.639091 0.638703 0.638321 0.637945 0.637574
0.637209 0.636849 0.636494 0.636145 0.635800 0.635461 0.635126 0.634796 0.634471 0.634151
0.633835 0.633523 0.633216 0.632913 0.632614 0.632320 0.632029 0.631742 0.631460 0.631181
0.630906 0.630635 0.630367 0.630103 0.629842 0.629585 0.629332 0.629081 0.628834 0.628590
0.628350 0.628112 0.627878 0.627646 0.627418 0.627192 0.626970 0.626750 0.626533 0.626319
0.626107 0.625898 0.625692 0.625488 0.625286 0.625088 0.625000
Code implementation
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
3 Chassis
Hummer
Stock car
Touring car
Street car
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
3 Chassis
Hummer
Stock car
Touring car
Street car
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
3 Chassis
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
3 Chassis
Structural data
CarFactory software: for rfactor vehicle suspension editing KPChassis software: for
chassis structural analysis
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
3 Chassis
Driveline definition for Artifical Inteligence
GPS data line
AI path line
Fitting the line manualy
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
3 Chassis
Driveline definition for Artifical Inteligence
Kart model implementation…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
3 Chassis
The choice…
Detailed cornering analysis
Best A
I la
p t
ime c
om
pare
d
Data comparison between the chosen model (red) and real data (blue)
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
Model optimization…
Data comparison between the chosen model (red) and real data (blue)
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
Model optimization…
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
Model optimization…
1 Tire model
New set values:
SpringBase=10000
SpringkPa=850
Damper=500
DryLatLong=(2.6, 2.85)
WetLatLong=(1.85, 2.15)
Data comparison between the chosen model (red) and real data (blue)
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
Model optimization…
1 Tire model
New set values:
SpringBase=10000
SpringkPa=850
Damper=500
DryLatLong=(2.6, 2.85)
WetLatLong=(1.85, 2.15)
2 Chassis inertia
Software CarFactory:
Used for inertia conditions
optimization for vehicle
models implemented in
rFactor
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
Model optimization…
Final model results
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
Model optimization…
Final model results
University of Rome“Tor Vergata”
Des
ign
wit
h lap
tim
e
Standard design process
New chassis
design Production
Track test and
race
Historical
experience
Evaluation tests and
driver feedback
University of Rome“Tor Vergata”
Des
ign
wit
h lap
tim
e
Standard design process
New chassis
design Production
Track test and
race
Historical
experience
Evaluation tests and
driver feedback
Proposed design process
New Chassis
design Production
Track test and
race
Historical
experience
Evaluation tests and
driver feedback
Simulation
Compare and
upgrade data
University of Rome“Tor Vergata”
Des
ign
wit
h lap
tim
e
Proposed design process
KP-Chassis implementation
Updating rFactor code with ChassisMod
Simulation and testing in rFactor
Develop of
the geometry
of the frame
Optimized
shape of
the frame
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
Sensitivity Test
Chassis1 Road Rebel:
chrome molybdenum steel tubing
32 mm diameter tubes
Chassis2 Black Star:
chrome molybdenum steel tubing
30 mm diameter tubes for rails
32 mm diameter tubes for crossbars
ChassisMod: software developed for quikly implement chassis data
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
For each type of frame were compared to the actual data with those obtained during the
simulation, in order to identify the differences and verify that the recorded response had
remained sufficiently reliable
TEST A
Real (blue) VS
Virtual (red)
Chassis 1
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
For each type of frame were compared to the actual data with those obtained during the
simulation, in order to identify the differences and verify that the recorded response had
remained sufficiently reliable
TEST A
Real (blue) VS
Virtual (red)
Chassis 1
Real (blue) VS
Virtual (red)
Chassis 2
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
For each type of frame were compared to the actual data with those obtained during the
simulation, in order to identify the differences and verify that the recorded response had
remained sufficiently reliable
TEST A
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
TEST B
The real data of the test of the two frames were compared between them by determining
what were the differences due to the two different geometries, then the same analysis
was carried out on simulated data to verify that the sensitivity of the simulator was able to
actually declare the same result.
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
TEST B
The real data of the test of the two frames were compared between them by determining
what were the differences due to the two different geometries, then the same analysis
was carried out on simulated data to verify that the sensitivity of the simulator was able to
actually declare the same result.
Real data of
Chassis 1(blue) VS
Chassis2 (red)
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
TEST B
The real data of the test of the two frames were compared between them by determining
what were the differences due to the two different geometries, then the same analysis
was carried out on simulated data to verify that the sensitivity of the simulator was able to
actually declare the same result.
Real data of
Chassis 1(blue) VS
Chassis2 (red)
Virtual data of
Chassis 1(blue) VS
Chassis2 (red)
University of Rome“Tor Vergata”
rFa
cto
r m
od
el
TEST B
The real data of the test of the two frames were compared between them by determining
what were the differences due to the two different geometries, then the same analysis
was carried out on simulated data to verify that the sensitivity of the simulator was able to
actually declare the same result.
1. The simplified model shows a good sensitivity to the variations of
the input parameters, allowing the use to a level of analysis mainly
limited to them.
2. The adoption of a more complete and complex model allowed us to
identify the kinematic configuration that provides the best
correspondence between the behavior of the frame present in the
simulator and the real one.
1. It was created a simplified and quick to use go-kart simulation
model;
2. The representative model of a go kart frame was implemented into
rFactor, approximating a three-body model with stiffness obtained
from structural analysis of the same.
3. It was presented and trained a innovative design process and a
wide margin of development.
University of Rome“Tor Vergata”
Co
nclu
sio
ns
an
d f
utu
re d
evelo
pm
en
ts
1. Extension of the work implementing all the tracks of a
championship to broaden the spectrum of response of the model.
2. Promote the design process presented for the design of go kart
frames in which the dynamic simulation can become an essential
tool as much as track testing.
3. Extend the possibilities of using the model associating it with a
physical simulator (under construction) for driver training and
chassis optimization on individual tracks.
University of Rome“Tor Vergata”
Co
nclu
sio
ns
an
d f
utu
re d
evelo
pm
en
ts
Thanks for your attention ...
University of Rome“Tor Vergata”