Embed Size (px)
DESCRIPTION
zxdfvsdf
Citation preview
SIMULATION FOR AIRBAG SENSOR CALIBRATION
Jérôme LAGRUT
2nd European HyperWorks Technology Conference
2
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PAB VOLVO GROUP – Business Area
Volvo Trucks Renault Trucks Mack Trucks Nissan DieselConstruction
Equipment
Buses Volvo Penta Volvo Aero Financial Services
3
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PAB VOLVO GROUP – Corporate Values
QUALITYQUALITY
SAFETYSAFETY
ENVIRONMENTENVIRONMENT
4
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PAB VOLVO GROUP – Corporate Values
QUALITYQUALITY
SAFETYSAFETY
ENVIRONMENTENVIRONMENT
“Safety is – and must always be – the fundamental principle
of our design work-”
Assar Gabrielsson and Gustaf Larson,founders of Volvo
5
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PAIRBAG TEST
FMFM
30 km/h30 km/h
00°°TRAILER BACKTRAILER BACK
6
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PAIRBAG CONTROL
CRASHCRASH
SITUATIONSITUATION
NON FIRE
FIRE
ECUElectronic Control Unit
algorithm
ECU
accelerations
7
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3P
ECU
NON FIRE
FIRE
algorithm
accelerations
TODAY =TODAY =
TESTSTESTS
FUTURE =FUTURE =
SIMULATIONSIMULATION
accelerations
HOW TO CALIBRATE AN AIRBAG ECU ?
8
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3P
LOW SPEEDLOW SPEED
TEST 3TEST 3
LOW SPEEDLOW SPEED
TEST 2TEST 2
LOW SPEEDLOW SPEED
TEST 1TEST 1
HIGH SPEEDHIGH SPEED
TEST 3TEST 3
HIGH SPEEDHIGH SPEED
TEST 2TEST 2
HIGH SPEEDHIGH SPEED
TEST 1TEST 1
DEVELOPMENT
TESTSSensorCALIBRATION
LOW SPEEDLOW SPEED
TEST 3TEST 3
LOW SPEEDLOW SPEED
TEST 2TEST 2
LOW SPEEDLOW SPEED
TEST 1TEST 1
HIGH SPEEDHIGH SPEED
TEST 3TEST 3
HIGH SPEEDHIGH SPEED
TEST 2TEST 2
HIGH SPEEDHIGH SPEED
TEST 1TEST 1
VERIFICATION
TESTS
algorithm
Acceleration signals
SENSOR
SENSOR
SUPPLIER
SUPPLIER
HOW TO CALIBRATE AN AIRBAG ECU ?
An airbag sensor has to be calibrated based on acceleration signals.
9
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PRESEARCH AND DEVELOPMENT PROJECT
A two years research and development project has been done to define the simulation
methodology:
• MESH GUIDELINES: recommendations for the mesh quality, mesh size (mesh class) and element formulation
• TEST SENSOR ACCELERATION: frequency analysis to find the frequency spectrum that needs to be caught
• FILTERING: cut off frequency and sampling rate have been set up to filter numerical noise and catch physical accelerations
• SENSOR MODELING: recommendations to represent the sensors
10
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PPROJECT APPLICATION: FH trucks
A new airbag sensor will be introduced in FH trucks and in the mean time, some parts will be modified:
20072007--2008: 1st application 2008: 1st application
of airbag sensor of airbag sensor
calibration with simulation calibration with simulation
help in the VOLVO Grouphelp in the VOLVO Group
11
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PPROJECT APPLICATION: FH trucks
Airbag sensor calibration matrix
LOW SPEEDLOW SPEED
TEST 3TEST 3
LOW SPEEDLOW SPEED
TEST 2TEST 2
LOW SPEEDLOW SPEED
TEST 1TEST 1
HIGH SPEEDHIGH SPEED
TEST 3TEST 3
HIGH SPEEDHIGH SPEED
TEST 2TEST 2
HIGH SPEEDHIGH SPEED
TEST 1TEST 1
DEVELOPMENT
TESTSSensorCALIBRATION
LOW SPEEDLOW SPEED
TEST 3TEST 3
LOW SPEEDLOW SPEED
TEST 2TEST 2
LOW SPEEDLOW SPEED
TEST 1TEST 1
HIGH SPEEDHIGH SPEED
TEST 3TEST 3
HIGH SPEEDHIGH SPEED
TEST 2TEST 2
HIGH SPEEDHIGH SPEED
TEST 1TEST 1
VERIFICATION
TESTS
algorithm
Acceleration signals
SENSOR
SENSOR
SUPPLIER
SUPPLIER
SIMULATION
SIMULATION
SIMULATION
12
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PPROJECT APPLICATION: FH trucks
Complete chassis + Engine + FUPS +
Power train
Complete BIW + DIW + Dummies +
Dashboard + Seats + Steering column + …
Finite Element model
1.5 million elements
RADIOSS 5.1 solver
13
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PPROJECT APPLICATION: FH trucks
Test versus simulation correlation: Simulation / Physical test
BA
RR
IER
FO
RC
E
AIR
BA
G S
ENSO
R
AC
CELER
ATIO
N
AIR
BA
G S
ENSO
R
VELOC
ITY
14
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PPROJECT APPLICATION: FH trucks
Test versus simulation correlation: validation of the model with the airbag sensor supplier
SENSOR
ALGORITHM
Same Time To Fire
� Model is validated
Different Time To Fire
� Model is not validated
accelerations
PHYSICAL TEST(from the development test phase)
SIMULATION
accelerations
15
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PPROJECT APPLICATION: FH trucks
Sensitivity analysis: done if the model is validated to improve the robustness of the algorithm
• influence of new or modified parts
• influence of cab variants
• influence of chassis variants
• influence of truck mass
• influence of test speed
• influence of barrier position (height, angle…)
Influence on the Influence on the
airbag airbag Time To FireTime To Fire
�������� Algorithm Algorithm
adjustmentsadjustments
16
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PPROJECT APPLICATION: FH trucks
Airbag sensor calibration matrix
LOW SPEEDLOW SPEED
TEST 3TEST 3
LOW SPEEDLOW SPEED
TEST 2TEST 2
LOW SPEEDLOW SPEED
TEST 1TEST 1
HIGH SPEEDHIGH SPEED
TEST 3TEST 3
HIGH SPEEDHIGH SPEED
TEST 2TEST 2
HIGH SPEEDHIGH SPEED
TEST 1TEST 1
DEVELOPMENT
TESTSSensorCALIBRATION
LOW SPEEDLOW SPEED
TEST 3TEST 3
LOW SPEEDLOW SPEED
TEST 2TEST 2
LOW SPEEDLOW SPEED
TEST 1TEST 1
HIGH SPEEDHIGH SPEED
TEST 3TEST 3
HIGH SPEEDHIGH SPEED
TEST 2TEST 2
HIGH SPEEDHIGH SPEED
TEST 1TEST 1
VERIFICATION
TESTS
algorithm
Acceleration signals
SENSOR
SENSOR
SUPPLIER
SUPPLIER
SIMULATION
SIMULATION
SAVING:
~ 200 k€
SIMULATION
17
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PCONCLUSIONS: DIFFICULTIES ENCOUNTERED
� Level of details needed in the models (pipes, tanks, brackets…)
� Material characteristics for crash application
� Rupture, tearing (welds, glue, bolts, sheet metal…)
� Model size: 64 bits workstations and software needed
� Model size : CPU time
� Test robustness
18
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PCONCLUSIONS: CAE BENEFITS
� Many variants can be simulated (cab size, engine...)
� Details can be seen in simulation and not in test (transparent vehicle)
� The worst case scenario can be found by computation and influence the
verification test configuration
� New test conditions can be tested (speed, angle...)
� Simulations will bring higher robustness to the algorithm
� Simulations can show how new introduced parts (eg Front lid in steel
instead of plastic) will effect the sensor signal and the algorithm
19
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PTHANK YOU FOR YOUR ATTENTION
20
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PBACKUP SLIDE: sampling rate
If we need 10 sampling points per period and if we want to catchaccelerations up to 1500 Hz we need a sampling frequency :
1 10.06ms
10 15/ mssampT
ν= = =
21
Volvo 3P
EHTC 2008 Sept. 30 - Oct. 1
Jerome LAGRUT
Volvo 3PBACKUP SLIDE: mesh size
• We need 10 elements (10.lc) per wavelength (c.T = c/ν) for the highest
frequency and the slowest (plastic) stress wave :
• For thicknesses of 2.mm or less this is close to the mesh convergence
requirement
50mm/ms3.333mm
10 15/ msc
cl
ν≤ = =
![Techniken der Projektentwicklung - Klassendiagrammeiluetkeb/2005/session06.pdf · [Sichtbarkeit] attribut_n:Typ[=Initialwert] {Zusicherung} p_i stehr für den i-ten Parameter der](https://img.pdfslide.net/doc/110x75/5d4eb4d888c9934b5e8bce18/techniken-der-projektentwicklung-klassendiagramme-iluetkeb2005session06pdf.jpg)
