Marco Anghileri Dipartimento di Ingegneria Aerospaziale Politecnico di Milano Italy. NCHRP 22_24 Interim report Meeting. Washington 2008-01-11 Robust:

Marco AnghileriDipartimento di Ingegneria Aerospaziale

Politecnico di MilanoItaly.

NCHRP 22_24 Interim report Meeting.Washington 2008-01-11

Robust: “Road Upgrade of Standards”

NCHRP 22_24 Interim Report meeting. Washington 2008-01-11

Robust

. G

RD

1-2

002-7

002

1

2

Robust

• EU FP5 sponsored three year project.

• Scope: create knowledge to improve EN-1317.

• Strong integration with CEN/TC226/Wg1/Tg1.

• The project covered 3 main items (referred to EN1317):

– 1 Study of the current situation.

– 2 Experimental procedures.

– 3 Computational mechanics.


Robust

. G

RD

1-2

002-7

002

1

4

Robust activities. 1 Study of the current situation

• To verify the relevance of 1317 on roadside safety.

– Statistical study of real life accidents and injuries to verify the contribution of 1317.

– Verification of the correspondence between testing procedures and real life accidents.


Robust

. G

RD

1-2

002-7

002

1

5

Robust activities.2 Experimental procedures.

• Testing procedures.• Vehicle influence

• Impact tolerances (velocity/angle)

• Soil influence

• Instrumentation mounting

• Data acquisition.

• Data analysis.• Influence of software.

• Severity indices.• Moving average

• PHD definition


Robust

. G

RD

1-2

002-7

002

1

6

Robust activities. 3 Computational mechanics.

• To improve the use of CM inside the certification process.

– Vehicle models improvement.

– Simulation procedures.

– Parametric studies.

– Validation procedures.


Robust

. G

RD

1-2

002-7

002

1

7

Item 1: Current situation.

• Statistical analysis of current level of safety around Europe.– Collection of data from real life accidents.

– Impacts only against EN1317 safety barriers.

– To investigate the application of EN1317 standards.

– To verify that EN1317 testing procedures is representative of real life.

• Incident data from:– Single vehicle incidents.

– Incidents an ALL roads

– Incidents which a VRS has been hit

– Only incidents from 1° Jan 1990


Robust

. G

RD

1-2

002-7

002

1

8

Data collection table.

• General incident information (date, road type, speed limit).

• Details of the vehicle (make, model, weight).

• Movement and damage to the vehicle.

• Severity and location of injuries to the occupant(s).

• Type of VRS.

• Impact speed and angle.

• Data source(s).


Robust

. G

RD

1-2

002-7

002

1

9

Current situation.Results and recommendations

• For all impact testing with cars, an impact speed of 110 km/h is recommended for all testing;

• The impact angle for cars remains at twenty degrees.

• The number of registered motorcycles be reordered, and the occurrence of such incidents analyzed in future years.

• It is recommended that the weight specified for the small car test (currently referred to as TB11) should be changed so as to be 950 ± 50 kg;


Robust

. G

RD

1-2

002-7

002

1

10

Current situation.Results and recommendations

• It is recommended that the vehicles used for the testing of VRS are no more than 5 years old;

• The data have shown that one dummy (either instrumented or not instrumented) should be seated in the vehicle during all testing between cars and VRS;

• It is further recommended that the dummy should be placed on the driver’s side of the vehicle and this should be the impact side. The dummy should be restrained by a seat belt;


Robust

. G

RD

1-2

002-7

002

1

11

2 Experimental procedures.

• Testing procedures.• Vehicle influence

• Impact tolerances (velocity/angle)

• Soil influence

• Instrumentation mounting

• Data acquisition.

• Data analysis.• Influence of software.

• Severity indices.• Moving average

• PHD definition


Robust

. G

RD

1-2

002-7

002

1

12

2 Testing procedures.

• Scope:

– Reduce scatter between labs.

– Improve testing procedures.

– Obtain benchmarks for computational mechanics activity.

• Activities:

– Round robin activity with rigid barrier

– Round robin activity with deformable barrier

– Round robin with heavy vehicles.

– Same test repeated in the same test house

– Collection of procedures used during testing inside the different test houses.

– Statistical analysis of already performed test (TB11 - 900 kg 100 km/h 20°) to investigate procedures for severity indices evaluation and possible correlation between existing severity indices.


Robust

. G

RD

1-2

002-7

002

1

13

Round robin with rigid barrier

• Rigid barrier:

– Same rigid barrier

– 900 kg 100 km/h 20°

– 6 tests with same new car

– 4 tests with test house standard cars

– 2 additional tests with the same new car and improved procedures

– In total 12 tests.


Robust

. G

RD

1-2

002-7

002

1

14

Round robin with rigid barrier

• Analysis of results showed and analyzed scatter from:– Transducer mounting procedures.

– Data acquisition (example: zeroing)

– Severity indices evaluation:• Software• 1317 procedures

– Small vehicle differences:• Tires-ground interaction• 2 – 4 doors

– Sensitivity to vibration of some severity indices (Phd,ASI).

Med max % min %Asi 1.86 0.05 2.83 -0.03 -1.48THIV 32.89 1.31 3.98 -0.49 -1.5PHD 14.15 3.55 25.1 -2.75 -19.4


Robust

. G

RD

1-2

002-7

002

1

15

Round Robin on deformable barrier.

• One TB11 test in each of the test houses on the same N2 steel barrier with the same car


Robust

. G

RD

1-2

002-7

002

1

16



Robust

. G

RD

1-2

002-7

002

1

17


• Scatter for these series of tests is bigger than the scatter observed for Round Robin with rigid barrier.

• The only difference was the use of a soft barrier and different soils conditions.


Robust

. G

RD

1-2

002-7

002

1

18

Round Robin with heavy vehicles.

• Tests with two different heavy vehicles, TB42 and TB51, one performed identically three times and one performed twice, to evaluate experimental result scatter


Robust

. G

RD

1-2

002-7

002

1

19

Round Robin with heavy vehicles.


Robust

. G

RD

1-2

002-7

002

1

20

Statistical analysis of existing data

• Example: Thiv-ASI correlation

Corrélation THIV - ASI (All data)

y = 13.748x + 11.409

R2 = 0.6359

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

50.00

0 0.5 1 1.5 2 2.5 3

ASI

THIV

THIV/ASI-All

VAL limites

Lineare (THIV/ASI-All)

Corrélation THIV/ASI

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

0.00 0.50 1.00 1.50 2.00 2.50 3.00

ASI

THIV

Autostrada

LIER

..

TRL


Robust

. G

RD

1-2

002-7

002

1

21

Statistical analysis of existing data

• ASI-dynamic deflection correlation:

y = 1.461e-0.688x

R2 = 0.589

0.00

0.40

0.80

1.20

1.60

2.00

0.00 0.50 1.00 1.50 2.00

Dynamic deflexion

AS

I ASI-DD

Espo. (ASI-DD)


Robust

. G

RD

1-2

002-7

002

1

22

Composite mounting fixing.• Installation of accelerometers can

affect measures and severity indices.

• The weight of the mounting block is a crucial problem


Robust

. G

RD

1-2

002-7

002

1

23

Deceleration tests.


Robust

. G

RD

1-2

002-7

002

1

24

Deceleration tests. Results.

• Frequency response of impact pulses:

– Strong differences between different mounting blocks.

– Differences at frequencies where severity indices work.

0 20 40 60 80 1000

1000

2000

3000

4000

5000

6000

7000

8000spectrum power density. y direction. 70° impact

AluminiumComposite

2 4 6 8 10 12 140

1000

2000

3000

4000

5000

6000

7000

spectrum power density. y direction. 70° impact

AluminiumComposite


Robust

. G

RD

1-2

002-7

002

1

25

ASI. Acceleration Severity Index

– “The index ASI is intended to give a measure of the severity of the vehicle motion for a person seated in the proximity of point P (CG) during an impact.”

– Steps:

• Measure three acceleration components of vehicle CG according with CFC180.

• Apply a 50 ms moving average on these acceleration.

• Evaluate Asi as:

2 22

lim lim lim

( ) y zx

x y z

a aaASI t

a a a


Robust

. G

RD

1-2

002-7

002

1

27

Moving average

• Questions:

– Can moving average distort signals?

– Can moving average give wrong information?


Robust

. G

RD

1-2

002-7

002

1

30

50 ms moving average – standard filtering gain.

• Gain=output/input

0 20 40 60 80 100

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

freq [Hz]

Gai

n

Butterworth filter


Robust

. G

RD

1-2

002-7

002

1

31

50 ms moving average – standard filtering attenuation.

• Attenuation.

100

101

102

103

104

-350

-300

-250

-200

-150

-100

-50

0

freq [Hz]

Am

plitu

de [D

b]

Moving averageFilter

0 20 40 60 80 100-350

-300

-250

-200

-150

-100

-50

0

50

100

freq [Hz]

Am

plitu

de [D

b]

Moving average


Robust

. G

RD

1-2

002-7

002

1

32

Moving average does not preserve energy.

• Velocity evaluation with:– Original signal

– Filtered signal (Butterworth)

– Moving Average

0 0.2 0.4 0.6 0.8 1-2

-1.5

-1

-0.5

0

0.5

1Velocity evaluation

time

velo

city

Moving average Original signalFiltered signal

0 0.2 0.4 0.6 0.8 1 1.2 1.4-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6Velocity evaluation

time

velo

city

Moving average Original signalFiltered signal


Robust

. G

RD

1-2

002-7

002

1

35

Is this a real problem?

• To verify the presence of this problem:

– Test cases obtained from some standard crash tests.

• Acceleration time-history.

– Original

– Filtered with moving average.

– Filtered with a “correct” filter.


Robust

. G

RD

1-2

002-7

002

1

36

Example.

0 0.2 0.4 0.6 0.8 1-50

-40

-30

-20

-10

0

10

20

30TEST 3: accelerazione X

time [s]

g

0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

5

10

15

20TEST 3: accelerazione Y

time [s]

g 0 0.2 0.4 0.6 0.8 1-15

-10

-5

0

5

10

15TEST 3: accelerazione Z

time [s]g


Robust

. G

RD

1-2

002-7

002

1

37

0 0.2 0.4 0.6 0.8 1-16

-14

-12

-10

-8

-6

-4

-2

0

2

time [s]

gTEST 3:Confronto tra media mobile e filtraggio. Componente X

Media mobileFiltro

0 0.2 0.4 0.6 0.8 1-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

time [s]

g

TEST 3:Confronto tra media mobile e filtraggio. Componente Y

Media mobileFiltro

0 0.2 0.4 0.6 0.8 1-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

3

time [s]

g

TEST 3:Confronto tra media mobile e filtraggio. Componente Z

Media mobileFiltro


Robust

. G

RD

1-2

002-7

002

1

38

Moving average

• The modification of original signals driven by the moving average has been demonstrated but:

• Is this strange behavior of moving average desired by the original designer of ASI?

• Or was simply not observed?


Robust

. G

RD

1-2

002-7

002

1

39

History of ASI.– I.Laker: “ A short summary of three vehicle Impact Severity

Measure- ASI THIV/PHD NCHRP 230” 1991.

• First step: the Shoemaker table


Robust

. G

RD

1-2

002-7

002

1

42

History

• 1955 Stapp tests.

• 1969 Limits in 3 direction from Cornell Aeronautical Laboratories

• 1971 Moving average.

• 1972 Ellipsoidal Envelope from US Air force documents


Robust

. G

RD

1-2

002-7

002

1

43

Acceleration time histories.

– Aeronautical deceleration:

• Source: Us Army “Aircraft Crash Survival Design Guide”.

• Single peak: from 15 to 30 g

• Time duration from 0.1 to 0.15 s

– Safety barrier deceleration.

0 0.2 0.4 0.6 0.8 1-15

-10

-5

0

5

10TEST 1: accelerazione X

time [s]

g


Robust

. G

RD

1-2

002-7

002

1

44

ASI History

• ASI was based on research on the injury assessment of vehicle and aircraft occupants in phenomena such as re-entry space capsule impacts and combat airplane maneuvers.

• These phenomena have limited or no oscillations throughout the event.

• For this reason, computing an average over a period of 50 ms was used to obtain an average value to be compared with the tolerable limits.

• Impacts against road restraint systems generally have a duration much greater than 50 milliseconds, and show strong oscillations at different frequencies.

• The 50 ms moving average when applied over such longer acceleration pulses becomes a low-pass filter, but it does not behave like filters used conventionally for crash analysis distorting randomly the original signals.


Robust

. G

RD

1-2

002-7

002

1

45

Modification of ASI formulation.

• 126 tests analyzed

• Evaluation of ASI using filtering instead of moving average.

• 2 poles Butterworth forward – backward filter (to avoid time shift). 4 total poles.

• Cut off frequencies tested:

10 – 12 – 14 – 16 – 18 – 20 hz

• The final cut off frequency has been identified as the one with the better correlation with the standard ASI formula. The idea is to avoid, if possible, modification of the current limits for the ASI formula.


Robust

. G

RD

1-2

002-7

002

1

46

Filtered ASI.

• Best correlation (not in all the domain):

– 12 hz cutoff frequency.

– 2 pole forward-backward Butterworth filter. (4 resultant poles)

0.60.70.80.91.01.11.21.31.41.51.61.71.81.92.02.12.2

0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2

ASI standard

AS

I filt

ered

asi

12 hz

Lineare (asi)

Lineare (12 hz)


Robust

. G

RD

1-2

002-7

002

1

47

Experimental activities output

• Proposal of PHD deletion. (Phd has been deleted)

• Study on accelerometer mounting block (in the revision of 1317 mounting block will be deleted and accelerometers will be placed directly on the tunnel).

• Study on Asi evaluation procedure. (ASI formula has been modified deleting the moving average).

• Soil must be better defined and controlled during tests (TG1 is discussing how to measure soil variation before tests).


Robust

. G

RD

1-2

002-7

002

1

48

3 Computational mechanics

• Create a vehicle data base.

• Produce information for CME group.

• Identify procedures to be used during computational mechanics activity for certification.

– Vehicle modeling.

– Soil modeling.

– Materials.

– Numerical data acquisition and severity indices evaluation.

• Improve the level of confidence inside computational mechanics

• Study validation procedures.


Robust

. G

RD

1-2

002-7

002

1

49

Vehicle data base. Robust.


Robust

. G

RD

1-2

002-7

002

1

50

Geo metro models.

• Situation:

• R0: NPRA modification of NCAC

• R1: Polimi modification R0 (suspension, meshing).

• R2: R1+seats and gebod

• R3: R1+ steering

• R4: R3+gebod+seats


Robust

. G

RD

1-2

002-7

002

1

51

First modifications. Wheel/suspension problems.


Robust

. G

RD

1-2

002-7

002

1

52

Front suspensions and steering system

• The actual Geo-Metro front suspensions are McPherson type.


Robust

. G

RD

1-2

002-7

002

1

53


• Original Geo Metro model has been strongly remeshed and than modified adding suspension and steering.

• In the old version of the Geo-Metro FE model the two front wheels were fixed to the respective side of the vehicle’s frame by a deformable beam

• Each wheel cannot steer unless the beam doesn’t fail.

• These two beams increased the stiffness of the suspensions assembly.

• Furthermore, in order to coordinate the steering movement of front wheels, a rigid link between the front rims was defined.

• Even if this link could synchronize the wheels during the steering action, it was not quite correct because it made the steering and suspension motion dependent: the suspension compression or rebound made the wheels steer, too!


Robust

. G

RD

1-2

002-7

002

1

54


• Front suspensions of the model have been enhanced by the redefinition of old joints and the creation of new ones which now allow a steering capability of the vehicle.

• Simple simulation have been run to check the system

Lower A-arm

Spindle


Robust

. G

RD

1-2

002-7

002

1

55

Steering system

• Steering of front wheels under the application of a force to the right wheel only and the vehicle raised from the ground.


Robust

. G

RD

1-2

002-7

002

1

56

Handling capabilities.


Robust

. G

RD

1-2

002-7

002

1

57

Steering force removed after 0.5 sec

Steering force applied for all the time


Robust

. G

RD

1-2

002-7

002

1

58

Measure of severity indices and time histories

• The measure of severity indices and time histories requires two important verification:

– Numerical data acquisition behavior

– Numerical definition of the transducer

• The numerical data acquisition must be able to acquire data that can reconstruct properly the physics of the phenomenon.

• The definition of the transducer must be comparable to the behavior of a typical transducer used during crash tests.


Robust

. G

RD

1-2

002-7

002

1

62

Output frequency influence.

• Round Robin scenario. Small vehicle 100 km/h 20° rigid barrier.

• Ls-dyna 970 solver 5434a version.

• To verify the behaviour of the numerical data acquisition system, accelerations have been sampled at different frequencies.

• Three output frequencies were considered: • 854 kHz (sampling time equal to the integration timestep),

• 100 kHz

• 10 kHz.

• The data output in the NODOUT ASCII file were used to compute the occupant risk factors, as prescribed in the EN1317 specifications, by means of the Test Risk Assessment Program (TRAP).

• The output data were initially filtered with a standard CFC180 filter and then processed by the software.


Robust

. G

RD

1-2

002-7

002

1

63

Comparison.

• How can we define which is the proper acceleration and the wrong and why an acceleration sampled during a numerical simulation can be wrong?

• Besides the acceleration time history also the velocity and displacement time histories can be obtained from these nodes.

• To understand which is the right acceleration and which is the wrong, we must verify that:

– the velocity and the displacement obtained integrating the acceleration

– And– the velocity and displacement directly sampled.

– Must be equivalent


Robust

. G

RD

1-2

002-7

002

1

64

Lateral velocity comparison

• 854 kHz

• 100 Khz

• 10 Khz

0 0.025 0.05 0.075 0.1 0.125 0.15-8000

-7000

-6000

-5000

-4000

-3000

-2000

-1000

0

1000

Time [s]

Velo

city [

mm

/s]

Vy 854kHz SampledVy 854kHz Integrated

0 0.025 0.05 0.075 0.1 0.125 0.15-8000

-7000

-6000

-5000

-4000

-3000

-2000

-1000

0

1000

Time [s]

Velo

city [m

m/s]


0 0.025 0.05 0.075 0.1 0.125 0.15-14000

-12000

-10000

-8000

-6000

-4000

-2000

0

2000

Time [s]

Velo

city [m

m/s]



Robust

. G

RD

1-2

002-7

002

1

65

Vertical velocity comparison

• 854 kHz

• 100 Khz

• 10 Khz

0 0.025 0.05 0.075 0.1 0.125 0.15-1000

-500

0

500

1000

1500

2000

2500

3000

3500

4000

Time [s]

Velo

city [m

m/s]

Vz 854kHz SampledVz 854kHz Integrated

0 0.025 0.05 0.075 0.1 0.125 0.15-1000

-500

0

500

1000

1500

2000

2500

3000

3500

4000

Time [s]

Velo

city [

mm

/s]


0 0.025 0.05 0.075 0.1 0.125 0.15-2000

0

2000

4000

6000

8000

10000

Time [s]

Velo

city [m

m/s]



Robust

. G

RD

1-2

002-7

002

1

66

Comparison results.

• Acceleration sampled at 854 Khz and 100Khz are able to reconstruct correctly the velocity and the displacement of the vehicle.

• Acceleration sampled at 10 Khz (standard sampling rate used for experimental testing) is not able to reconstruct the motion of the vehicle.

• Signals sampled at 10 Khz have problems similar to aliasing.


Robust

. G

RD

1-2

002-7

002

1

67

Data acquisition conclusion.

• This problem showed that numerical data acquisition has the same typical problems of the experimental data acquisition.

• Care must be taken for the definition of the sampling rate. High sampling rate means huge amount of data.

• This problem is mesh sensitive and code sensitive (Pam crash has pre-sampling filtering).

• The requirement we included in our procedure is that, to prove the proper data acquisition, the reconstruction of the motion must be demonstrated starting from acceleration time histories.


Robust

. G

RD

1-2

002-7

002

1

68

Numerical accelerometer definition.

• Now that we have demonstrated the capability of our numerical data acquisition the problem is shifted to the transducer itself.

• This numerical transducer must be compared to a standard real transducer. If we want to compare these two output these transducers (numerical and experimental) must be equivalent.


Robust

. G

RD

1-2

002-7

002

1

69

Numerical-experimental transducers.

• Numerical accelerometers are not damped. They can produce frequencies up to the natural frequency of the element where they are attached.

• We have seen how to acquire these signals but now we must make them equivalent to the experimental ones

• Typical real accelerometer frequency response.


Robust

. G

RD

1-2

002-7

002

1

70

Numerical accelerometer processing

– We have seen that sampling frequency must be adequate to avoid numerical aliasing problems (generally much higher that the experimental one).

– A new filtering (before the same filtering used during experimental data acquisition process) must be used to adequate the numerical frequency response to the real frequency response.

– Which frequency?


Robust

. G

RD

1-2

002-7

002

1

71

Example applied to the Round Robin Activity

• Round robin.

• Experimental sampling rate :10 kHz

• Numerical sampling rate: 100 kHz

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

x 104

0

0.5

1

1.5

2

2.5

3

3.5x 10

13

frequency [hz]

ampl

itude

Spectrum

Numerical original signal Numerical CFC1000 Numerical CFC600 NumericalCFC180 Numerical CFC60 Experimental unfiltered signal

0 0.5 1 1.5 2 2.5

x 104

0

2

4

6

8

10

12

14

16

x 1012

frequency [hz]

ampl

itude

Spectrum


0 2000 4000 6000 8000 10000 120000

1

2

3

4

5

6

7

8

x 1012

frequency [hz]

ampl

itude

Spectrum


0 1000 2000 3000 4000 5000 60000

0.5

1

1.5

2

2.5

3

3.5

4

x 1012

frequency [hz]

ampl

itude

Spectrum


0 500 1000 1500 2000 2500 30000

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

x 1012

frequency [hz]

ampl

itude

Spectrum


0 500 1000 15000

0.5

1

1.5

2

2.5

x 1011

frequency [hz]

ampl

itude

Spectrum

Numerical CFC180 Numerical CFC60 Experimental unfiltered signal

Frequencies relevant for severety indices

evaluation


Robust

. G

RD

1-2

002-7

002

1

72

Numerical accelerometer.

• To correctly sample acceleration time histories:

– Demonstrate that you are able to properly reconstruct the motion (with Geo Metro R4 100 kKhz).

• To correctly compare the numerical accelerometer to the experimental one:

– Pre-filter data to have a numerical frequency response similar to the experimental one (with Geo Metro R4 CFC60).


Robust

. G

RD

1-2

002-7

002

1

73

Parametric studies.

• The following points have been studied to better define procedures for the use of CM during the standardization process.

– Friction• between wheel / wearing course• between barrier and vehicle

– Boundary conditions for the barrier• Ground condition• Anchoring of the barrier• Length of the barrier

– Vehicle• Impact point• Speed of the vehicle• Weight of the vehicle• Spinning wheels• Instrumentation• Mounting Block• Location of the accelerometer


Robust

. G

RD

1-2

002-7

002

1

74

Other parametric studies.

• Parametric study on material properties

• ASI, THIV and Dynamic deflection variation on

– Material properties:

• E-module,

• Yield stress (S235 steel has a variation of yield stress from 235 to 360 Mpa)

– Material thickness based on % change

• The material properties should have an upper and a lower limit


Robust

. G

RD

1-2

002-7

002

1

75

Parametric study result.


Robust

. G

RD

1-2

002-7

002

1

76

Parametric study result.


Robust

. G

RD

1-2

002-7

002

1

77

Validation

• Validation methods have been studied using the benchmark cases obtained during the Round Robin activities.

• Validation has been identified not only on severity indices value but also on time histories.

• The method to be used for validation must be objective and able to validate also experimental tests performed inside the tolerances defined by the standards.


Robust

. G

RD

1-2

002-7

002

1

78

Validation based on velocity comparison.

• Accelerometers comparison does not fit for objective validation method.

• Was decided to use a global reference frame velocity comparison.

– Rotations of accelerations (unfiltered).

– Evaluation of global reference frame velocities.-0.05 0 0.05 0.1 0.15 0.2 0.25

-10

0

10

20

30

40

Ay filtered at 180 hz

AutostradeHut Cidaut Lier TRL


Robust

. G

RD

1-2

002-7

002

1

79

All RR tests.

• Same rigid barrier.

• Different vehicles.

• 12 tests

-0.4 -0.2 0 0.2 0.4 0.6 0.8 1-2

0

2

4

6

8

10

12Vy global Unfiltered

-0.2 0 0.2 0.4 0.6 0.8 1 1.214

16

18

20

22

24

26

28Vx global Unfiltered

-0.4 -0.2 0 0.2 0.4 0.6 0.8 1-3

-2

-1

0

1

2

3

4Vz global Unfiltered


Robust

. G

RD

1-2

002-7

002

1

80

Validation based on velocity comparison.

• To take into account the different importance of components, validation based on resultant velocity.

• After the point where the difference between simulated and measured resultant velocity is greater than:– Xx% of current velocity (suggestion:± 5% of initial

current velocity)?

– Xx% of initial velocity (suggestion:± 5% of initial velocity)?

The model is no longer validated.


Robust

. G

RD

1-2

002-7

002

1

81

Comparison based on resultant velocity.

-0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.618

19

20

21

22

23

24

25

26

27

28reultant velocity


Robust

. G

RD

1-2

002-7

002

1

82

Validation. All requirements.

• Severity indices:

– ASI THIV

• Barrier behavior.• Deformation (WW, DD).

• Failures (number and location).

• Vehicle:• Trajectories.

• Failures ? (during RR different failures)

• Resultant velocity.

• Yaw rate


Robust

. G

RD

1-2

002-7

002

1

83

Computational mechanics.

• Techniques been developed inside Robust project have been applied to all experimental test case (each test repeated several times):

– Rigid barrier

– Deformable barrier

• Different numerical test houses.

• Results available at:

http://www4.vegvesen.no/ROBUSTDOC/


Robust

. G

RD

1-2

002-7

002

1

84

Conclusion

• Robust produced an huge amount of results that now can be used to revise standards.

• Regarding CM

– Benchmarks

– Procedures

– Experience

Are available for the integration of this method inside the certification process.


Robust

. G

RD

1-2

002-7

002

1

85

Questions?

Documents

Marco Anghileri Dipartimento di Ingegneria Aerospaziale Politecnico di Milano Italy. NCHRP 22_24 Interim report Meeting. Washington 2008-01-11 Robust: