COMPARATIVE PERFORMANCE TESTING OF PASSENGER ...

COMPARATIVE PERFORMANCE TESTING OF PASSENGER CARS RELATIVE TO FMVSS 214 AND THE EU 96/EC/27 SIDE IMPACT REGULATIONS: PHASE I Randa Radwan Samaha Louis N. Molino Matthew R Maltese National Highway Traffic SafetyAdministration PaperNumber 98-S8-O-08 hasrecentlypublisheda final rule institutionalizing the process [39]. The final rule setsforth the processthat the agencywill usein comparingU.S. andforeign vehicle safetystandardsand in determining appropriaterulemaking response,if any. The rule realIirms NBTSA’s policy of actively identifying and adopting those foreign vehicle safety standardsthat require significantly higher levels of safety performance than the counterpart U. S. standards. The rule also outlines the agency’s policy in the casewhere the comparisonindicates that the foreign standard’ssafety benefits are approximately equal to thoseof a counterpartU. S. standard.

ABSTRACT

Basedon a long recognizedneed,the National Highway Traffic Safety Administration (NHTSA) has begun to reexaminethe potential for international harmonization of sideimpact requirements.To this endNHTSA, as directedby the U. S. Congress,has recently submitted a report to the Congresson the agencyplans for achieving harmonizationof the U. S. and European side impact regulations. The first phaseof this plan involves crashtesting vehicles compliant to FMVSS 214 to the EuropeanUnion side impact directive To begin gathering the data necessaryto make the 96/27/EC. This paper presents the results to date of this NHTSA initiated a research research. The level of safety performanceof the vehicles functionalequivalenceassessment, basedon the injury measuresof the European and U.S. side program by testing eight U.S. production FMVSS 214 compliant vehiclesto the EU Directive 96/27/ECrequirement. impact regulationsis assessed. This paperfocuseson the results of the testing in terms of the level of safety performanceof the vehicles for both the U.S. and EU regulations. INTRODUCTION The National Highway TraIKc Safety Administration (NHTSA) has long recognized the need for international harmonization of side impact requirementsand the potential of addedsafetybenefits resulting from such harmonization. Although the U.S. and EU side impact regulations ideally addressthe samesafetyproblem,they differ in test procedures, barriers, dummies, and injury criteria. Recently, the U.S. Congressdirected NHTSA to study the differencesbetween the U.S. and proposedEuropeanside impact regulationsand to develop a plan for achieving harmonization of these regulations. Also, manufacturers believe that these differences lead to different vehicle designs, thus posing unduefinancial burdensin termsof dual development,testing, manufacturinganddistribution of vehiclesinvarious markets. NHTSA submitteda side impact harmonizationplan to the U.S. Congressin April of 1997 [l]. The first phase of the plan is an attempt at assessing whether the safety performanceof vehicles is functionally equivalentrelative to the European regulation (EU Directive 96/27/E(J) and the FederalMotor Vehicle Safety Standard(FMVSS) 214. The Functional Equivalence AssessmentProcess(Appendix A) was developedby the U. S. and Australia in coordinationwith foreign governments,industry andconsumergroups. NHTSA

1727

Current U.S. and European Side Impact Standards The U.S. regulation on side impact is FMVSS 214; m Impact Protection [2] addressingthoracic and pelvic fatalities and injuries in vehicle-to-vehicle crashes. The dynamic requirement,or crash test portion of this standardwas added in October of 1990. It was phased-inbeginning with 1994 model year (MY) cars such that all cars by the 1997MY had to meetthe requirements. Starting with the 1999MY, trucks, buses, and multipurpose passengervehicles under 2,721 kg (6000 lbs) must meet the dynamic part of this standard[3]. The EuropeanUnion (EU) side impact regulation, EU Directive 96/27/EC was approved in October of 1996. It applies to new and redesignedMl and Nl vehicle types beginning with the 1999 MY. Ml vehicles are those with a capacityof nine or lessoccupantsand would include passenger cars, multipurposepassengervehicles, and mini buses. Nl vehiclesarethosewith the capacityof carrying up to 3.5 metric tons, e.g. vans and chassis cabs. Vehicles with R-point of lowest seat>700 mm are excluded. All Ml and Nl vehicles starting in the 2004 MY must meetthis regulation.

The test proceduresof both regulationsare similar in that a stationary test vehicle is struck with a moving deformable barrier (MDB). Thesedynamic test proceduresfocus on the measurementof anthropomorphictest dummy responsesto compute injury criteria. However, the two regulations use different test procedures, barriers, dummies, and injury criteria. Figures 1 and 2 show a schematicof the test setupfor the U.S. and EU regulations. Table 1 comparesthe relevant crashtest parameterssuchasimpact direction, impactvelocity and barrier face dimensions. 4 1

WHEELBASE ON) 194Ommi

0.5w

targetvehicle. The lateral striking position is aligned with the occupant seatingposition rather than the vehicle wheelbase. The EU MDB is centered about the R-point or seating referencepoint defined as the H-pt for lowest and rearmost driving seatposition.

d 1

HICLE A

IMPAC POINT

1i

i

i ) V=50 km/h

VEHICLE B CG

Figure 2. EU 96/27/EC Side Impact Test Configuration. Table 1. Crash Test Parameter Comparison Figure 1. FIWVSS214 Side Impact Test Configuration.

The FMVS S 2 14 dynamic test simulatesthe 90 degree impact of a striking vehicle traveling 48.3 km/h into a target vehicle traveling 24.2 km/h. This is achievedby a moving deformablebarrier with all wheels rotated 27 degrees(crab angle)from the longitudinal axis, impacting a stationarytest vehicle with a 54 km/h closing speed.For a typical passenger car, the left edge of the FMVSS 214 MDB (214MDB) is 940 mm forward of the mid point of the struck vehicle wheel base. In the EU 96/27/EC dynamic test, the EuropeanMDB (EUMDB) impacts the target vehicle at 50 km/h and 90 degreeswith no crab angle. This differs from FMVSS 214 in that no attempt is made at simulating the movement of the

1728

1

Ir

EU 96/27/EC

II

300 mm Barrier Face Ground Height I

IIFaceWidth

I 1500mm

Performance Barrier Defined Material *same as seatingreferencepoint

1

FMVSS 214

280 mm Bumper 330 mm II I I 1676mm Aluminum Honeycomb

II

FMVSS 214 and EU 96/27/EC Movable Barriers - The dimensionsand material characteristicsof the 214MDB face are shown in Figure 3. The aluminum honeycombof

the barrier faceis specifiedby design. The bottomedgeof the MDB is 280 mm from the ground. The protruding portion of the barrier simulating a bumper is 330 mm from the ground. The 214MDB has a total massof 1367kg initially derived from the weights of passengercars and lights trucks in the U.S. fleet with a adjustmentmade assuminga downwardtrend in vehicle massdue to fuel economyneeds[4, pg IIIAd]. The dimensionsof the EUMDB face are given in Figure 4. The Europeanbarrier face is segmentedinto six blocks with force deflection performancecharacteristicsspecifiedin the EU regulation. The lower blocks are stifler than the top blocks and the centerblocks are stiffer than the outboardelements. The ELJMDB face is about 20% smaller than the 214MDB in terms of face area. It is also much softer than the 2 14MDB face on the blocks closestto the sides. The bottom edgeis the most forward part of the EuropeanMDB and is 300 mm from the ground. The Europeanbarrier has a massof 950 kg, 40% less then the massfor the U.S. barrier.

q~q50Jo,q I T-

r

300 i

It

A

/-A

Y GROUND LEVEL

SECTION A-A 800

-1

GROUND LEVEL

I INOTE: Length dimensions in millimeters

1

Figure 4. EU 96/27/EC Side Impact Deformable Barrier Face.

1676

J

I

HONEYCOMB / 838

1

BUMPER

(( HONEYCOMB

533

(

LEFT SIDE VIEW

FMYSS 214 and EU 96/27/EC Dummies and Iniury Criteria - In both regulations,successfultest performanceis determinedby dummy injury criteria. However, the regulation differ in both the test dummy and injury criteria. Figure 5 is a schematicof the two side impact dummies,the U.S. side impact dummy (SID) usedin FMVSS 214 and the EU dummy EUROSID-1 used in Directive 96/27iEC.

I 203 I I

559

NOTE: Length dimensions in millimeters

Figure 3. FMVSS 214 Side Impact Deformable Barrier Face.

1729

Although both dummies ideally representa 50th percentileside impact anthropomorphicdevice, they are basedon different designsand have different measurement capabilities. In particular, Eurosid-1 has an articulating half arm, while the responseof the arm is folded into the design of the thorax in SID. FMVSS 214 requiresthat a SID be placed in both the front and rear seatsof the test vehicle. The EU Directive requiresthat only one EUROSID- 1 be placed in the front seat. The injury criteria for each regulation, given in Table 2, relate to the measurement capabilities of the dummy used.

protection criteria (HPC) is derived from head acceleration over a head contact time duration and must remain below 1000. A rib deflectioncriterion @DC) allows a maximum of 42 mm of deflection in the thorax. A soft tissue viscous criterion (V*C), computedfrom combinedrib deflectionand velocity, is to be reportedwith a proposedlimit of 1 m/s. It is worth noting that for the first two yearsin which EU 96/27/EC becomeseffective,V*C valuesare to reportedbut not usedas a pass/fail criterion . A review of the EU directive is planned in the year 2000 during which the statusof V*C as a required injury criteria will be decided. The abdominal peak force (APF) is limited to 2.5 KN. Finally, the pubic symphysispeak force (PSPF),which is in the pelvic region, must be lessthan 6kN.

EUROSID-1

SID

VEHICLE MATRIX NOTE:

Dimensions

in millimeters

(mm)

Figure 5 Schematic of Side Impact Dummies of FMVSS 214 and EU 96/27/Ec Table 2. Test Dummy Injury Criteria I

EUROSID-1

I

Table 3 lists the U.S. productionvehiclesthat were tested to the EU 96/27/ECrequirements.The vehicleswereidentical in design to vehicles tested to FMVSS 214 in the NHTSA compliancetestprogram.The matrix includedfour 4-doorand four 2-doorpassengercars. Table 3. FMVSS 214/ EU 96/27/EC Test Matrix Vehicle

U.S. SID

IIHIC - 1000 I(RibDeflection 2 42 mm

1 TTI < 85 or 90 G

IIv*c 5 1m/s

I

II

AbdominalForce < 2.5 KN

IIPubic SymphysisForce

1996Ford Taurus*

4-Dr

1995Volvo 850

I I Pelvic Accel. I 130 G

SID was designedto measureonly the accelerationof the ribs, spine and pelvis to computethoracic and pelvic injury criteria [20]. The rib and spine accelerationsare combined into a single metric called the Thoracic Trauma Index (TTI(d)) which hasan 85g limit for 4-doorvehiclesand a 90g limit for 2-doorvehicles. The pelvic accelerationhas a 130g limit.

FMvss 214 test

Side NCAP

Production

1996

Yes

539K

1995

No

1 63K

1997Nissan Sentra

/ 4-Dr 1

1996

Yes

/ 72K

1997Hyundai Sonata

/ 4-Dr 1

1996

Yes

15K

1997Ford Mustang

1 2-Dr /

1996

/

No

j 170K

1997Lexus SC300

1 2-Dr 1 1995

/

No

/ 13K

No

58K

1995GeoMetro

2-Dr

1996

2-Dr 1996 1llK 1997Mitsubishi No Eclipse I I I EUROSID-1hasadditionalmeasurement capabilitiesthan *The EU test for the 1996Tauruswas performedby Ford SID, including force and displacementas well as acceleration Motor Company. basedreadings[5]. The EU regulation placeslimits on five dummy criteria to determinevehicle performance.The head

1730

The selectioncriteria for the vehicles in order of importance occupant responsesand vehicle intrusion profiles [7,8]. As were the following: such, the recently developedhoneycombface from Plascore waschosento ensurethe bestcurrently availablefit to EUMDB 1. FMVSS 214 test results were usedto provide a range of performancerequirements. performancefrom marginal to goodperformerswithin the In both the 2-Dr Mustang and Eclispe, there was no room set of 4-Dr vehicles and correspondinglywithin the set of to fit a rear Eurosid-1 dummy in the EU tests. In the FMVSS 2-Dr vehicles. 2. Vehicles to be tested in the side impact New Car 2 14 compliancetests,there was no room to fit a rear SID only AssessmentProgram (NCAP) were included as much as in the Eclipse. It is worth noting that the Eurosid-1 dummy possible in order to provide an additional comparative has a slightly higher seatedheight specified at 904 f 7 mm data set at a higher performancelevel for the current test versus 899 *lo mm for the SID. On the other hand the SID program and for possible future ECE 95 testing at a has a wider hip width specifiedat 373 &18 mm versus355 &5 mm for the Eurosid-1. higher performancelevel. 3. Vehicles built or sold by all U. S. manufacturersor their subsidiarieswould be represented,and similarly, to the Eurosid-1 Calibration Issues- Problemsin certification of extent possible, for those built or sold by foreign Eurosid-1 lumbar spine parts were encounteredin the set up manufacturers. for the EU vehicle testing. As such, a round robin calibration 4. The highest production vehicles would be represented. of three new lumbar spineswas performedat three U.S. sites. The results are listed in Appendix A. The baseangle output requirementswere only met about 50% of the time for two of EU 96/27/EC TESTS SETUP the parts although the lumbar pendulum pulse requirements were typically met. The differencesin baseangle outputswere With the exception of placing a Eurosid-1 dummy in the small and consistentsuggestingthat the new partswere similar rear outboard position, the proceduresof the EU 96/27/EC in constructionbut the calibration corridors maybetoo narrow. Directive were followed in performing the European side The resultswerepresentedto TN0 and ISO/TC22/SC12/WG5 impact tests of the U.S. production vehicles. In addition, working group, Anthropomorphic Test Devices, in June of experts from TNO, Netherlands, provided training on the 1997. TN0 has initiated a round robin researchactivity to latest dummy seatingpractices. They also providedguidance addressthis lumbar spine calibration issue. on common EU 96/2/7EC test set up practices,especially in areaswhere the EU directive is not specific. Although the seattrack position for the front dummy is not specifiedin the COMPARISON OF OCCUPANT RESPONSES EU Directive, the Eurosid-1 in the driver position was seated in the mid-track position to provide the best comparisonwith The first level of comparisonof results was basedon the FMVSS 214. In addition, comparison checks of the test normalized injury criteria of eachregulation. Tables4. and 5. vehicle options,test weight and attitude, and dummy H-points list the computedinjury criteria for the FMVSS 214 and EU and lateral clearancesbetweenthe FMVSS 214 and EU test 96/27/ECtestsof the eight vehiclesfor both the driver and rear setupswere performed. This was done to ensure minimum dummies. In general,the basisof the comparisonsmadebelow differencesin the vehicles,dummy positioning, and testsetups was to normalize the computedinjury values by the limit of for the comparisontesting. the criteria specified by each regulation. For example, the TTI(d) was normalized by 85 for the 4-Dr vehicles and by 90 The Plascorelayeredhoneycombconstructionbarrier face, for the 2-Dr vehicles, and RDC was normalized by 42 mm. was usedfor the EU Movable DeformableBarrier (EUMDB) Overall, the results indicate a higher severity for the driver elements. The choice of barrier face was basedon a recent dummy in the EU tests for the RDC thoracic criterion when evaluation of barrier faces performed through the comparedto the TTI(d) in the FMVSS 214 tests. No trend is International StandardsOrganizations (ISO) working group seenfor V*C versusTTI(d). The results also indicate possibly on Car Collision Test Procedures [6]. The evaluation a higher severityfor the driver dummy in the 4-Dr vehiclesfor comparedthe characteristicsof the Cellbond/TRL, Plascore, PelvicG in the FMVSS 214 testswhen comparedto the PSPF and AFLAJTAC EUMBD faces and indicated that the pelvic criterion in the EU tests. This apparent trend is Plascore face best fits the force performance corridors reversedfor the driver dummy in the 2-Dr vehicles where a specifiedby EU 96/27/EC. It hasbeenestablishedby various lower severity is indicated for PelvicG when compared to researchersthat different EUMDB face designs lead to PSPF. significantly different vehicle performanceresults,for both the

1731

1997 Mitsubishi Eclipse 2-Dr 1995 Volvo 850 SW 1996 .--_. Ford _.- Taurus .--.-_

4-Dr-.

1997 Nissan Sentra 4-Dr 1997 Hyundai Sona’-rar-vr’ m m

/ I

82 49 50

I

ii

I

ii

I

49.0

1

;ei

I

ii2

~

!

-n IU

1’

InIUL

1’

enT LJ.I

i’

1res

1

0.60

~

1 I

86 58 61

/ 1 I

49.6 29.9 40.0

1 I

yes yes “es

/ 1 I

1.04 0.38 094

/

4.097 1.686 2196 4.531 3.490

/ j 1 1 1

1.429 0.719 1131 1.029 1.369

1 / I I j

nc nc 67 231.9 nc

~ 1 1 116.1 I 145.65 1 1 42.2 / 48 1 1

94.9 23.5 90.8 389 177.6

1

1

* Numbers in bold are in excessof the criterion. nc= no contact nd= not determined

Table 5. FMVSS 214 and EU 96/27/EC Test Results (Rear Passenger)

The resultsalso indicate a much lower severityin the EU tests for the rear passengerdummy when both EU thoracic criteria are comparedto TTI(d) in the FMVSS 214 tests. No trendis apparentwhenPSPFwascomparedagainstPelvicGfor the rearpassengerdummy.

of the vehiclesand higher by 27.5% for the remainingfour vehicles. Therewere no apparenttrendsin thesedifferences for either the 2-Dr or 4-Dr setsof vehicles.

Thoracic Injury Criteria

214 vs EU Driver Thoracic Injury Criteria iom

With the caveatthat the Eurosid-1rib deflectionswhich form the basisfor computingRDC and V*C are questionable (Refer to section “Flat-Top”Anomalies in Eurosid-1 Rib Deflection Responsesbelow),the following observationsare made. For the 4-Dr vehicles,the NissanSentradriver dummy, exceeded the RDC and V*C criteria (SeeFigure6). For the 2Dr vehiclesdriver dummy,the GeoMetro exceededRDC and the Mitsubishi Eclipse exceededboth RDC and V*C. With the exceptionof the Sonataandthe LexusSC300,the normalizedTTI(d) was on the average26.8%lower than RDC for the driver dummy. For the Sonataand the Lexus,TTI(d) Figure 6, was 12% and 3% higher than RDC. As to V*C, the results were more of a mismatch, with normalizedTTI(d) on the average27.1% lower than V*C for the driver dummyfor four

1732

cdl !3MaxRibDefl

CiVmVC~

PelvicG was on the averagegreaterthan PSPFby 16% for three of the four vehicles. The exceptionwas the Sentra,in which PelvicG was lessthan PSPFby 3%. In contrast,for the 2-Dr vehicles, PelvicG was on the average12% lower than PSPFfor threeof the four vehiclesfor the driver dummy. The exceptionwas the SC300, in which PelvicGwas larger than PSPFby 19%.

214 vs EU Rear Occupant Thoracic Injury Criteria q nl(dJ

q MaxRib Defl DMaxVCI

0.8

214 vs EU Driver Pelvic Injury Criteria E Pelvic Gs EC¶ Pubic Syrrphysls

Metro

Taurus

Volvo

850

Sonata

Peak Force

Sentra

Figure 7.

The average normalized thoracic criteria for the 4-Dr vehiclesand the 2-Dr vehiclesare listed in Table 6. For the driver dummy, the averagenormalizedTTI(d) and RDC are higher for the 2-Dr vehicles. In contrast the average normalizedV*C is lower for the 2-Dr vehicles.

SC300

Mustang

Eclipse

Memo

TaUrUS

V&O 850

sonata

sentla

Figure 8. For the rearpassengerdummy,the normalizedPelvicGwas on the average38% higherthan PSPFfor threeof the vehicles and lower by 14% for the remaining three. There was no apparenttrend in thesedifferencesfor either the 2-Dr or 4-Dr setsof vehicles.

Table 6. Average Normalized Thoracic Criteria 4-Dr Vehicle Set 2-Dr Vehicle Set I I I I driver rear driver rear dummy dummy dummy dummy*

1

TTI(d)

69%

64%

78%

60%

RDC

88%

35%

96%

52%

214 vs EU Rear Occupant Pelvic Criteria q

Pelvic Gs ‘a Peak Symphysis Pubic Forcx~

0.9

v*c

0.8

66% 14% 81% 7% *averageresultsfrom only 2 vehicle tests

: 3 o.7 .Z 0.6 5 g 0.5 ; 0.4 N y 0.3 E 6 0.2 2 0.1

Pelvic Injury Criteria

For all the vehicles tested and for both front and rear dummy, none of vehicles exceededthe criteria for either 0.0 regulation(SeeFigures8. and 9.). For the driver dummy, the resultswere more of a m ismatchwhen comparingthe results for the two regulations. The normalized PelvicG in the Figure 9. FMVSS 2 14 testswas on the average8% higher than PSPFin the EU testsfor four of the vehiclesand lower by 10%for the other four. When looking at the 4-Dr vehicles separately,

1733

IVktro

Taurus

Votvo

850

Sonata

Sentra

Abdominal Injury Criterion

EU Rear Occupant Abdominal Criteria

For the driver dummy in the EU tests, the Mustang normalized APF was 92% of the limit specified in the regulation and the Volvo APF was 29%. The normalized APF for the remaining six vehicles was clustered closer with an averageof 53% of the limit.

Peak Force

030 i 5

0.25

2 p

0.20

6 ; 0.15 .z

EU Driver Abdominal Injury Criteria 0 Abdominal

a Atdominal 035

I 0.10 b z 0.05

Peak Force

1.0 , 0.9 j

0.8

9

0.7

0.02 Metro

2 .s 06 6 0.5 z.i

&we

votvo

Sonata

650

Sl?llt~

11.

0.4

EU Driver Head Injury Criteria

g 0.3 2

Taurus

0.2 q Head Protection Criteria B Reference

0.1

HlC[

0.45

00 SC303

Mustang

Eclipse

M&O

TaUrllS

V&O 850

.salata

sentra J

Figure 10.

f

0.40

z

0.35

.z 0.30 iii g 0.25

- 0.20 w 9 0.15 $ 0.10

Overall, the average APF for the driver dummy in 2-Dr vehicleswas 67% of the limit. The averageAPF for the driver z 0.05 dummy for the 4-Dr vehicles was lower at 42 %. In contrast, 0.00 for the rear dummy, the averageAPF was only 63% of the SC300 limit for the 2-Dr vehicles and 22.3% for the 4-Dr vehicles with no value exceeding30% of the limit. The APF resultsare Figure 12. presentedin Figures 10. and 11.

Mustang

Eclipse

Metro

Volvo 650

Taurus

Sonata

Senba

EU Rear Occupant Head Injury Criteria q Head Protection Criteria I8 Reference

Head Injury Criterion For the driver dummy in the EU tests, headcontactoccurred for four of the eightsvehicles,with an averagenormalizedHPC of only 8.9% of the limit specifiedin the regulation. For the rear dummy, contactoccurredfor three of the six vehicleswith an averageHPC of 27.8% of the limit. The normalized I-PC and lXC36 values from the EU tests are presentedin Figures 12. and 13. HPC is plotted only if headcontact occurred.

HIC

3 0.4 $? 0.3 .g c t D .B 1 !i P

0.3 0.2 0.2

0.1 0.1 0.0 SC300

Figure 13.

1734

Metro

Taurus

Volvo

850

Sonata

Sentra

“Flat-Top”Anomalies Responses

in

Eurosid-1

Rib

Deflection 1995 Volvo 850 SW Left Front Dummy

Figures 15. through 28. presentan overlayof the raw (Class 1650)rib displacementsfor both the front and rear dummy for the vehicles tested. As shown in the figures, sustainedpeaks (plateausof flat-tops) as long as 15 ms in at leastone of the rib displacementcurvesare presentfor the driver Eurosid-1for all the vehicles tested. These flat-tops are present for the rear dummy in three of the six vehicles.The displacementlevels of theseplateausrangeanywherefrom 1.5to 50 mm andarebelow the full range of the rib potentiometersof the dummy. Lau reportedon this phenomenonin a seriesof Ford LTD crash tests [ 111. Using a pneumatic impactor to impact the Eurosid-I, the sustainedpeaks in the rib displacementwere producedbut they could not be createdwith pendulumimpact. (Integration of the rib and spine accelerationsin the impactor impacts indicated that they were moving away from the impactor at similar speeds.) Hensonet al. reporteda similar Figure 16. rib deflection problem when testing Pontiac 6000’s using the FMVSS 214 procedurewith the Eurosid-1 1121. This was believedto occur when impacts were more rearward than the lateral center of the ribs. The American Automobile Manufacturers Association (AAMA) has highlighted this anomalousrib behavior in their list of mechanical concerns relative to the Eurosid-1 dummy [ 131. The AAMA has attributedthis behaviorto binding in the rib damping modules due to off-axis loading. Transport Canada has recently reproducedthis flat-top behavior with the Eurosid-1 ribs with pendulum impacts at -15 degreesfrom the coronal plane, i.e. posterior or rearward of the center of the ribs [14]. In those tests, the pendulum face contactedthe projecting back plate causing an alternative load path through the spine box, howeverthe rib deflectionswere near their maximum. 0.000

1995 Volvo 850 SW Left Rear Dummy

0.025

1997 Ford Mustang 2-Dr. Left Front Dummy

Figure 17. Review of the high speedfilm from the seriesof eight test presentedin this paperindicatedthat dummy rotation was visible in most of the tests. Contact with the intruding door was mademore rearwardthan the lateral centerof the ribs on some of the tests, while contact was forward of the center of the ribs in some of the tests suggesting off-axis loading conditions. This flat-top rib behavior of the Eurosid-1 was reproducedand further investigatedwith bumper pendulum impact tests outside the full vehicle test environment as outlined below. Figure 15.

1735

1997LexusSC300 2-Dr.

1996 Ford Taurus 4-Dr. Left Rear Dummy

LefI Rear Dummy

igure 20.

igure 18.

1997 Mitsubishi Eclipse Left Front Dummy

1997 Lexus SC300 2-Dr. LefI Front Dummy 40 30 -

20

E g

IO

E p

o

:: s

-10 -20 -30 O.WO

0.025

0.050

0.075

0.100

Time (SC)

Figure 21.

Figure 19.

1736

2-Dr.

1997 Hyundai Sonata 4-Dr. Left Rear Dummy

1996 Ford Taurus 4-Dr. Left Front Dummy

‘igure 22.

F igure 24. 1997 Hyundai Sonata 4-Dr. Left Front Dummy

1997 Nissan Sentra 4-Dr. Left Front Dummy 60

40

P :: a

E E Em E a z

0

I -20 o.wo

o

j -20

0.025

0.050 lime

0.075

0.100

o.ow

0.125

(set)

0.025

0.050 lime

Figure 23.

Figure 25.

1737

(SeC)

0075

0.100

1995 Geo Metro 2-Dr. Left Front Dummy

1997 Nissan Sentra 4-B. Left Rear Dummy

o.ow

0.025

0.050 lilne

0.075

O.lW

I i

I I I I I i I I I I I

-20

37 __ 0.030

0.025

0050

0.075

0.100

0 125

Time (set)

(WC) L

Figure 28.

Figure 26.

1995 Geo Metro 2-Dr. Left Rear Dummy

t 0.030

0.025

0.060 lime

0.075

0.100

0.125

(SC)

Figure 29. Eurosid-1 Bumper Pendulum Tests Setup.

Figure 27. Eurosid-1 Bumper Pendulum Tests - To further investigate the flat-top behavior in the rib displacement responses, a seriesof bumperpendulumimpact testswith the Eurosid-1wereperformed.The generaltest setupis shownin Figure29. The part 581bumperpendulumwhich hasa mass around 907.4 kg was used in all the tests. The bumper pendulumwas rotatedto a sufficient height and also given an initial velocity via springsto get a closing speedsimilar to the door contactspeedsencounteredin the seriesof EU full scale side impact tests. The test conditions are describedin the following:

1738

The impact speedswere around18.2km/lx. The pendulumimpact face was a steelplate with a 3/4” plywoodcover,6 10 mm wide measuredfrom the seating surfaceat a height of 239 mm correspondingto the top of the moldedpelvis of the Eurosid-I. Basedon the reviewof the full scalecrashtest films, two impactfaceheightswereused:239-503mm (Low Face)& 239-558 mm (High Face)with the seatingsurfaceat z = 0. The low face extendedfrom the pelvic/abdomen junction to approximatelyl/3 down the arm (arm at 40 degrees) from the shoulderpivot bolt. The high faceextendedfrom the pelvic abdomenjunction to the centerof the shoulder attachmentbolt. The dummywasplacedon the horizontalflat steelseating surfacewith legsextended.

5. Tests included +/-15 degrees, and 0 degreesleft dummy side impacts (e.g. + 15 degrees is an anterior oblique impact and requires rotating the dummy by +15 degrees aboutthe z-axis using a right-hand coordinatesystemwith x positive in the posterior-anteriordirection, andy positive lateral to the left.) The 18.2 km/hr impact was on the low end of the range of 25-40 km/hr door contact speedsencounteredin the EU full scale tests but, as can seen from the following figures, reasonablerib displacements ranges of 10 to 40 mm were achievedas compared to the values obtainedin the full scale tests.

ESPO23: +15 Degrees - High Face 35 30 f i! G f

-

Upper Rib Defe13m

25 al 15 10 5 0 -5

3000 6

2000 rom 0 ,c

30

As shown in Figures 30. through 34., for both impactor face heights significant flat-top rib deflections responseswere present in the +15 degreeimpact condition, while 0 degree impactsproduceda lesserflat-top. The flat-topbehaviordid not occurfor the testsat - 15 degrees,neither with the Low Facenor with the High Face.

I

I

I

0.02

0.04

0.06

25

0-m

I/\

---

T1 y Acceleation TlZyAmleration

0.08

0.10

0.12

Figure 31.

ESPOO7: +15 Degrees - Low Face

ESKlO6: -15 Degrees - Low Face

3om e

2003

8 2

10m 0 I 50 40 30 20 10 0 -10 0.m

I

/

,

/

I :z

4rn en u, OI

: al 0

:g 0.02

0.04

0.m

0.08

0.10

0.m

0.12

Milliseconds

0.02

0.04

0.06

Milliseconds

Figure 30.

Figure 32.

1739

0.08

0.10

0.12

1) Binding of the rib damper module: Henson et al. suggested that momentis transferredacrossthe rib damper in the Eurosid-1thorax during obliqueimpact [ 121. Such a moment could causeexcessivefriction between the piston and cylinder wall of the damper,and causeit to bind.

ESPOZ2: -15 Degrees - High Face

6

2000

z3

low 0

m 0

100 80 m 40 20 0 -20 -40 O.CU

0.02

0.W

0.M

008

0.10

0.12

Milkcon&

2) Load Bridging: In this test condition, each of the elementsof the torso (the shouldercomplex,eachof the thoracicribs, abdomen,and possiblythe backplate)share the total force transferredto the spinebox to accelerate Eurosid-1through the changeof speedof the test. Load bridging can occurwhen an element’sadjacentneighbors experiencean increasein stiffness(dueto a constitutive relationshipin a materialor structure,and/ora mechanical binding) and thus the element’sdeflectionis governedby its neighbors. For example,shouldthe upperand lower rib of the Eurosid-1thorax bind due to somemechanical defectand thus increasetheir stiffness,the deflectionof the middle rib should not exceedthat of the upper and lower ribs. Load bridging may be presentbetweena binding shoulderand the abdomen,or the shoulderand pelvis, or direct contact with the back plate, thereby limiting the deflectionof the rib modules.

Figure 33.

EWOOS: 0 Degrees - Low Face

30 c +z

20 10

=

0

I

-10

I

I

I

mm

= mm 3 2 lam

---

-

Front Abdminal Force Middle Abdominal Fcrce

- RearAbdOrni”d Face

0

/

c

I

0 -10 4

o.m

I

0.02

I

004

I

O.C6

/

008

I

0.10

I

0.12

Milliseconds

Figure 34. Basedupon this work and that of other researchers, the authorshavedevelopedthe following hypotheses for the cause of rib flat-tops in the Eurosid-1: 1740

A seriesof tests was conductedat +15 and 0 degreesto determine the load sharing between the thorax and the abdomen.In theseteststhe load wall contactedthe abdomen and thorax only, while the arm was rotated 180 degreesto point straight up, such that it was not contactedby the impactor. Figures 35. and 36. show the resultsfrom these tests, and flat-tops are observedin theseconditions. These resultsshow that after the flat-top rib responses,abdominal loadsdrop off significantly, and then rise again as the thorax beginsits expansion.Newton’sfirst law appliedto deformable bodiesloadedin parallel requiresthat a) the total load applied equalthe sumof the loadscarriedby eachof the bodies,andb) the load onebodycarriesis directlyproportionalto its stiffness andmass(stressfollows elasticmodulusand density). Based on this law, we can then concludefrom the datathat the drop in force on the abdomenindicatesan increasein force on the thorax. Thuswe would concludefrom this testthat the flat-top is a result of binding within rib modulesor quite possiblyof the dampermodulesthemselves.However,this may notbe the only mechanismof rib binding in other test environments.

platesof the shoulderassembly,and useof plastic spacersin the lumbar spine and neck similar to thoseusedin the SID. The modificationsin this upgradekit areminor in natureand would not seemtot addressthe major issuessuch as alleged binding of the damperin the rib cage,the influence of the kinematicsof the shoulderstructureon the rib cagedeflection, and the deformabilityof the pelvis bone.

ESPO28: +I 5 Degrees - Face #I - Arm up

-

Upper Rib Defledicm Middle Rib Dde%m !

mm

-2u 4 0.m

I

I \ I

I

I/ I

-

I

’

Tdal Abdominal Face Front Abdminal Fme --Middle Abdominal Face - - Rear AbdOiT,“al Face

I

I

I

I

I

0.02

o.c-4

0.03

0.08

0.10

ESFQ29 - 0 Degrees - Face #I - Arm up

I 0.12

Milliseconds

Figure 35.

3000 $

2om

3z

tom

Front Abdominal Face --MiddkAbdainal Fwce - - Rear Abdminal Fwce

0

Another seriesof testswith the Eurosid-1was performed with the Eurosid-1jacket removedat + 15 degrees,the results of which are shown in Figure 37. While these tests were originally designedto permit viewing of the thorax and 0.03 shoulderduring impact, the results indicate that there is a significant changein the dummybehavior betweenthejacket on andjacket off tests. In Figure 37., the resultsfrom three 1 repeattestsat + 15 degrees,with the Eurosid-1jacket removed Figure 36. in two of the tests,are presented. The resultsshow that the flat-top behavioroccurredin all of the threetests. Theseresults indicatethat the flat-top behavioris not only reproduciblebut is also repeatableunderthe given test conditions. O ther researchershave suggestedthat the impactor/door surfacemay comein contactwith the backplateof the dummy, therebyoff-loading the rib structureand causinga flat-top. Inspectionof films from theseseriesof bumperpendulumtests indicate that no impactor-to-backplate contact occurred. Moreover,contact with the back plate is only likely with a combination of oblique posterior loading and excessiverib deflection. Eurosid-1 BummerPendulum Tests with Uugrade Kit 30 TN0 has recently developeda researchkit tool upgradeto $20 addresssomeof the widely acceptedmechanicaldummyissues ~E 10 % 0 with the Eurosid-1 (Referto sectionEurosid-1 Mechanical Deficiencies) . The researchupgradekit for the Eurosid-1 addresses a numberof minor issues.Theseincludesmoothing sharpedgeson the projectingtorsobackplate,use of bumper washersto minimize impacts betweenthe femur shaft and pubic load cell mounting hardware,beveling sharpedgeson Figure 37. the clavicle link to preventbinding with the aluminum guide 1741

/ 0.02

I 0.04

0.06

/ 0.08

0.10

Milliseconds

Three Repeat Tests: +I5 Degrees - Low Face

I 0.12

In April 1998,the Eurosid-1 researchupgrade kit was made available to NHTSA for evaluation. To date, a preliminary evaluationwas performedthrough a seriesof bumperpendulumtestsunder test conditions similar to thosedescribedin the previous section. Tests were performedwith the Low Faceand were run at + 15 and 0 degrees.The purposeof performing repeattestswith the upgradekit installed in the Eurosid-1 was to investigateif modifications in the kit addressthe flat-top anomaliesin the rib potentiometerresponses.

ESP036: 15 Degrees - Low Face - Upgrade Kii

Figures 38. through 40. show the results from two repeattests at + 15 degreesand one test at 0 degrees.As can be seenfrom the figures, the flat-tops are still observedin theseconditions.

ESFOSS: 15 Degrees - Low Face - Upgrade Kit

I-

Figure 4.

ESP036: 0 Degrees - Low Face - Upgrade Kit

Conclusions Regarding Eurosid-1 Rib ResDonses “Flat-Top” Anomalies Irrespectiveof the mechanismcausingthe anomalous behavior, the flat-top behavior in the rib potentiometer responsesindicate that what should be the true peak rib deflectionsmay not be occurring and thus the resulting rib deflectionsare in doubt. The V*C computationwhich is basedon the rib deflection would also be suspect. In light of this, the RDC and V*C valuesfor the EU 96/27/EC vehicle tests presentedin this paper are questionable.

1742

g

.,” 40 *n 4” 20. 10 0 -. -10 7 0.00

I !

_-

/i, I :/ /

I

/

.__ 1, /\ \ \ I\

I Tl-yAccelmtw TV-y Accderabm

I

I

I

0.06

0.08

-y-c-

1 0.02

---

0.04

I

Milliseconds

Figure 5.

’

---/-==--010

0.12

COMPARISON OF VEHICLE RESPONSES MDB Ewapement - The EUMDB is 176 mm or 10.5 % narrower than the 214MDB and in the EU procedureit is centered on the driver seating reference point while the 214MDB is positionedmore forward and is positionedrelative to the center of the wheelbase. This resulted in no MDB to A-pillar engagementin the EU tests while the A-pillar was engagedin all of the FMVSS 214 tests. Also, the EUMDB right edgeengagedthe vehicle rearwardof the 214MDB right edgefor all of the vehicles tested. (SeeTable 7.). The vehicle contact areasfor the EUMDB and 214MDB, drawn to scale, are shown for the Lexus SC300 and Volvo 850 in Figures 41. and Figures 42. as examples. Table 7. FMYSS 214 vs EU 96/27/EC MDB’s to Vehicle Contact I I rt Vehicle Left edge Bight edge difference* difference* II

=::,t5’_’

Side Intru Vehicle

Nissan Sentra

I

274

I

98

II

Hyundai Sonata

I

236

I

60

II

Ford Mustang

I

311

I

135

II

Lexus SC300

I

433

I

257

II

Geo Metro

337

358 Mitsubishi Eclinse * 2 14MDB-EUMDB; positive is forward

Figure 42. Volvo 850: EUMDB and 214MDB Contact Area. Table 8. n at Door Si Driver H-pt. i Mid Door Door Sill Level max and average :rush (mm)

Driver Hpt level max and average crush (mm)

161 182

f c-4 z 5 Ik

E s 2 &

Ford Taurus*

N/A

N/A

N/A

N/A

N/A

N/A

Volvo 850 SW

110

150

70

69

284 220

264 178

280 227

270 189

Nissan Sentra

166 125

217 120

310 270

372 287

280 172

377 217

Kyundai Sonata

147 91

281 164

388 326

443 291

394 332

435 305

Ford Mustang

132 99

266 153

254 220

333 211

234 171

335 197

Lexus

157 109

130

320 272

330 216

351 239

304 193

160

112

112

70

239 227

249 179

226 161

262 141

SC300

GeoMetro

Figure 41. Lexus SC300: EUMDB and 214MDB Contact Areas. 1743

Mid Door level max and average crush (mm)

100

Mitsubishi 178 196 304 333 296 333 Eclipse 157 107 287 265 258 248 Crush profile data was not availablefor the TaurusEU test

Geo Metro Crush Profile - Door Sill

Side Crush Profile Comparison - In order to facilitate I 400 comparisonwith the intrusionprofile in the FMVSS214 tests, II 350 pre and post test side crush measurements were collectedfor the EU testsasspecifiedin the FMVSS2 14testprocedure[ 151. The maximumside crush at the door sill and m id door levels for the EU and F’MVSS214 testsare presentedin Table 8. It is worth noting that relative magnitudefor the maximum intrusionat theselevelsdid not correlatewith the thoracicand pelvic criteria valuesfor the different vehiclesneitherfor the 0 FMVSS 214 nor for the EU tests. With the exceptionof the Volvo 850, the maximum static intrusion at the driver H-pt levelfor the EU testswas on the average4 1 m m largerthan for I the FMVSS 214 tests. At the m id door level, with the Figure 43. exceptionof the Volvo X50and Lexus SC300, the maximum static intrusion at the driver H-pt level for the EU testswas on the average62 m m larger than for the FMVSS 214 tests. 450

At the sill level, with exceptionof the area around the 0 B-pillar, intrusion was larger for the FMVSS 214 tests of the Metro, SC300,Sentra,and Eclispe. In contrast,the intrusion was significantly larger at the sill level for the EU testsof the Sonataand Mustang. Figure 44.

The Lexus SC300was the only vehicle which had more intrusion at both the sill and m id door levelsfor the FMVSS 214 test. For the Volvo 850, which is designedto meetboth regulations,intrusionat both levelswas comparablefor the two regulations.It is worth noting that both the 850 andthe SC300 were the best performersin the 4-Dr and 2-Dr vehicles sets relativeto the requirementsof both regulations,

1000 Location

Referenced

2000

:+214*EU;

500 Longitudinal

,000 Location

Referenced

1500

2030

2500

to 214 impact Point (mm)

1+214*EUi

350 i

0

500 Longitudinal

,000 Location

Referenced i-1

Figure 45. 1744

1500

ta 2f4 impact Point (mm!

Geo Metro Crush Profile - Mid Door

The static crushprofiles at the door sill and m id door levels are presentedin Figures43. through56. In general,in the EU tests,the crush profile is more roundedwith larger intrusion aroundthe B-pillar and the rear sectionof the front door. In the FMVSS 2 14 tests,the crush profile is more rectangularin shapewith the intrusionmoreevenlydistributedalongthe area of MDB-to-vehicle engagement.This is attributed to the characteristics of the EUMDB and 214MDB and their positioningas describedearlier.

At the m id door level, also with the exceptionof the area aroundthe B-pillar, intrusion was larger for the EMVSS 214 tests of the Metro, Sonata, and Eclispe. In contrast, the intrusion was significantly larger at the m id door level for the EU testsof the Sentraand Mustang,specificallyaroundthe Bpillar and rear sectionof front door areas. In fact, the B-pillar was split in half in the EU test of the Mustang.

500 Longitudinal

1MO

2000

to 214 Impact Point (mm]

2500

Lexus SC300 Crush Profile - Door Sill

Mustang Crush Profile - Mid Door

100

100

50

50

0

0 0

500 Longitudinal

1000 Location

1500

Referenced *214

2000

25w

0

1500

Referenced

2000

2500

to 214 Impact Point (mm)

/GGGy

Figure 49.

Mitsubishi Eclipse Crush Profile - Door Sill 450,

Lexus SC300 Crush Profile - Mid Door I

350

350

300

-

250 200

E 250 s 2fA 2w

150

u

.C

!j L v

IWO Location

*EU

Figure 46.

P g

5w Longitudinal

to 214 Impact Point (mm]

3w

150 I(M

100 50 I

0-I 0

500 Longitudinal

1OW Location

15w

Referenced +214

2ow

”

25w

0

to 214 Impact Point (mm)

x0 Longitudinal

IWO Location

*EU

*214

Figure 47.

1500

Referenced

2wo

2500

to 214 Impact Point (mm)

*EU

Figure 50

Mitsubishi Eclipse Crush Profile - Mid Door 450 /

Nissan Sentra Crush Profile - Door Sill I

300 P g 250 .c g 200 L v 150

c g 200 L o 150 100

50

/

\

50

0

0 0

500 Longktudinal

1000 Location

15w

Referenced !+214

2ow

2500

0

to 214 Impact Point [mm)

500 Longitudinal

+EUI

1WO Location

Referenced ‘214*EU

Figure 48.

Figure 51. 1745

1500

2030

to 214 Impact Point [mm)

2500

Nissan Sentra Crush Profile - Mid Door 450

j

Volvo 850 Crush Profile - Door Sill I

300

P g r 2L

P g 250 .c fn 2 200 o 150

o

3w 250 2cc 150 icm

‘1” 0

50 0 1OW

500 Longitudinal

Location

15w

Referenced I+214

2ow

25w

0

to 214 Impact Point (mm)

500 Longitudinal

1WO Location

1MO

Referenced

+EU/

2000

2500

to 214 Impact Point (mm)

FGzG

Figure 52.

F igure 55. Volvo 850 Ctush Profile - Mid Door

Hyundai Sonata Crush Profile - Door Sill

100 50 0 0

5w Longitudinal

IWO Location

1500

Referenced

2wo

2500

to 214 Impact Point (mm)

!GG=G 0

500 Longitudinal

1000 Location

Referenced

,503

2000

2500

to 214 impact Point (mm)

Figure 56.

l=zGGY

Figure 53. COMPARISON RESEARCH

300 250 200 150 100 50 0 0

500 Longbtudinal

low Location

Referenced l=icq

Figure 54.

,500

2o(M

to 214 Impact Point (mm)

RELEVANT

PREVIOUS

Becauseof the fluid natureof the Europeantest procedure, the databaseof full scalevehicle crash tests which can be directly comparedto testing performed with the current procedure is lim ited. The data are further lim ited if comparisonsare requiredbetweenthe samevehiclestestedto both the U.S. andEU regulations.Satakeet al. reportedon 24 full vehicletestson five Japanese vehiclesusing the U.S. and (BCE/R.95)procedures[9]. TheECE/R.95procedurematches that of the EU Directive, howeverthe barrier height was 260 m m ratherthen 300 m m . Sometestswere run at 300 m m for comparison. The barrier face used was made by UTAC of aluminumwith a triangular pyramid-shapeddesign.

Hyundai Sonata Crush Profile - Mid Door

F g c 2L o

WITH

2500

In the baselinetest, both 4-Dr vehiclesexceededthe rib deflectioncriteria whentestedto the ECE. Both 4-Dr vehicles were below the TTI lim it in the U.S. test although one was closeto it. Comparingthe 2-Dr vehiclesto the 4-Dr, in the 21746 Dr vehicle, the abdominaland pelvic loads increasedfor the

ECE test, and the rib deflection was lower. For one of these vehicles, the abdominal force exceededthe limit in the ECE test, but for the samevehicle testedto the U.S. procedure,all injury criteria were below their limits. For the other 2-Dr vehicle, the U.S. procedurewas more severe,with a very high TTI and pelvic acceleration. The ECE procedureresulted in rib deflection, abdominalloads and pelvic loads at or slightly abovethe limits. These results were obtainedwith the barrier height at 260 mm. A 4-Dr vehicle and 2-Dr vehicles were testedwith both a 260 and a 300 mm barrier height. All injury criteria were greaterfor the 300 mm barrier exceptabdominal force which had about the sameresult.

than for 2-Dr vehicles. This is consistent with the results reportedin [lo] and [27]. However,using a studentst-test it is shown that this differenceis not significant r&=0.59). The averagednormalizedRDC was 7.9% greaterfor 2-Dr vehicles. This is in contrastto [lo] and [27]. This differenceis also not significant @=0.66). For the U.S. procedure, the average normalizedTTI was 12% greaterfor 2-Dr vehicles. This was consistentwith resultsreport in [9]. Howeverthe differenceis not significant @=0.40). So for the chest injury criteria no consistent or significant difference is evident for either regulation relative to the number of vehicle doors.

For the EU procedurethe normalizedPSPFwas greaterfor the The aboveresults may indicate that, for 4-Dr vehicles, the 2-Dr vehicles by 3 1% @=0.31) which is consistentwith [9], currentEU Directive is moredifficult to passthan FMVSS 214. [lo] and [27]. While for the U.S. procedurethe normalized PelvicG is nearly the samefor both 2-Dr and 4-Dr vehicles, Bergmann et al also performedtests using the ECE/R.95 which is not consistent with [9]. Finally, the normalized procedure[27]. Testing was doneat barrier heights of 260 and averageAPF is 57% greaterfor 2-Dr vehicles than for 4-Dr 300 mm with barrier facesof Kemnont, Fritzmeier, Hexcel and vehicles. This differenceis consideredsignificant at p=O.O53. AFL elements. Average results across all tests were An increasein APF for 2-Dr vehicles was also reportedin [9], determined. Both V*C and RDC for the 4-Dr vehicles were [lo] and [27]. much higher than for the 2-Dr vehicles. V*C was in the vicinity of the criterion limit. However,for APF and PSPFthe 2-Dr vehicles had higher values than the 4-Dr. Table 9. Vehicle Type With Greater Average Normalized Injury Beusenberget al. found a similar dependenceon the number of doors a vehicle has when tested to the (EEVC) procedure [lo]. Seventeen 4-Dr tests and five two-door tests were analyzed. It is not clear what barrier face construction was usedfor thesetestsnor what barrier height was employed.The averageand maximum V*C and rib deflection,in general,were below the injury criteria for 2-Dr vehicles. For 4-Dr vehicles the averageV*C was above the criteria and the averagerib displacementwas equal to the criteria. For 2-Dr vehicles the abdominalforce was abovethe criteria. For 4-Dr vehicles the abdominal forces were below the criteria. Pubic loads were 0.59 0.66 0.40 0.97 0.31 0.053 P higher for 2-Dr vehicles than 4-Dr, but both were well below the criteria. In the current set of test, it is not clear if the relative severity of each regulation is influenced by the number of vehicle doors. Table 9. gives the vehicle type (2-Dr or 4-Dr) which has the larger normalized injury criteria and the percentageby which it is greater. Also given, is the result of a student’s t-test to determine if the difference in the injury criteria is statistically significant. Significance will be determined at ~10.10. However, the determination of significance is certainly influencedby the small samplesize of 4 vehicles in each category. The following discussionis limited to the front seatdummy. Some results seem consistent with previous work whereas others do not. This is mainly attributed to barrier height differencesand the lack of consistentperformanceamongstthe various European barrier faces. For the EU procedure,the averagenormalized V*C was 23% greater for 4-Dr vehicles 1747

INITIAL ASSESSMENT EQUIVALENCE

OF

FUNCTIONAL

From NHTSA’s perspective,in basic terms, a foreign vehicle safetystandardis consideredfunctionally equivalentto a counterpart U.S. standardwhen the two standardsaddress the samesafetyneedand provide similar safetybenefit in the U.S. crash environment. Relative to the European and U.S. sideimpact regulations,FMVSS 214 has only recentlybeenin full effect for passengercars and will apply to LTV’s by the end of this year, and EU 97/26/EC is not yet in effect. As such, there is currently insufficient real world safety data to assessthe effectivenessof either regulations whether in the U.S. or Europeanreal world environments.

Data from compliancetesting, such as the seriespresented in this paper, can be used as a surrogate. Injury risk curves would be usedto assessoccupantrisk in the real world from the computedinjury criteria obtainedvia crashtesting. Currently, injury risk curvesare not available for the abdominal, pelvic, and headEU injury criteria. In addition, due to the volume and quality of the earlier injury and impact data, the EU injury risk functions originally developedfor the thoracic region need to be improved [16]. Moreover, those thoracic injury risk functions were based on the responses of the Production Prototypeversions of the Eurosid dummy and would needto be updatedfor the production Eurosid-1. In addition, the aspectof how well the test conditions and movabledeformablebarrier of the EU regulationrepresentthe real world U.S. crashenvironmentcannotbe overlookedwhen assessingthe relative safety benefit of the two standards. A dynamic crash test requirementin a safety regulation should simulate the crash environment to the extent possible. More importantly, the dynamic requirement should provide for realistic injury causingmechanisms.The representativeness of the EUMDB as the striking vehicle mustbe consideredbecause a large portion of the U.S. side impact casualtiesare the result of impacts with light trucks, vans and sportsutility vehicles. Issuesin severalareasthat needto be addressedbeforeany conclusivedeterminationof the functional equivalenceof the two side impact regulationsare outlined below. Vehicle Issues In terms of the overall vehicle performance, similar comparative testing of vehicles designed to European requirements would be needed to assessif such vehicles perform well relative to FMVSS 214. The testing presentedin this paper is only one part of a general matrix to assessthe respective comparative performance. Testing of vehicles designedto both standardsand testing of additional vehicles equipped with side air bag systems, which are becoming prevalent in the U.S. fleet, would be a part of this general matrix. The repeatabilityand reproducibility of testing to both regulationswould also needto be addressed. Vehicle Compliance and Rankings - Basedon this series of comparativetesting, FMVSS 214 and EU 96/27/EC did not provide similar vehicle performance rankings nor pass/fail resultsbasedon the respectivethoracic and pelvic criteria (see Tables lo., 1land 12.). Overall, the vehicles testedhad been chosenbasedon complianceto FMVSS 214. However’three out of theseeight vehicles failed one or two of the EU criteria for the driver.

1748

Table 10. FMVSS 214 vs EU 96/27/EC Criteria: Pass/Fail

IIFord Taurus IA4 IINissan Sentra I‘dPI

IJ IPI

F

P

IFI

F

IIFord Mustang 121PI

P

[PT-F:

121PI

P

IPI

II

Hyundai Sonata

IILexus SC300

P

jg2J.O : ‘, : : : * 80 % = Exceed80% of Criteria

For statistical confidence to be achieved, certain manufacturesrequirethat, asa vehicle designbasis,regulation requirementsmustbe met by a margin considerablybelow the actual limits specified. As such, pass/fail results based on SO%of the criteria were also investigated for this series of comparativetesting. Three of the eight vehicles exceeded80 % of at leastoneof the FMVSS 214 requirementswhile five of the vehicles exceeded80 % of at least one of the EU requirements.The Metro and Eclipse exceeded80% of oneor more of the requirementsfor both regulations. The Taurus, Sentra,and Mustang exceeded80% of the requirementsof the EU regulation only while the Sonata exceeded80% the requirementsof FMVSS 214 only. Considering the vehicle rankings for the 4-Dr vehicles basedon the driver dummy criteria, FMVSS 2 14 TTI(d) rated the Hyundai Sonataas fourth, while the EU RDC rated the Sonataasfirst and the V*C rated it as second.Rankingsbased on the pelvic criteria for the 4-Dr vehicles were a much better match with only the third and fourth position switched.

Table 11. FMVSS 214 vs EU 96/27/EC 4-Dr Vehicle Rankings: Driver Vehicle

TTI rank

RDC rank

V*C rank

PelvG rank

PSPF rank

Volvo 850

1

2

1

1

1

Hyundai

Since the relative rankings of the vehicles testeddid not look promising, linear regression analysis was applied to evaluatethe degreeof correlation between the thoracic and pelvic criteria of the two regulations. p2,the regressionoutput that indicateshow well onevariable can be predictedthrough a linear transformation of anothervariable, is presentedin Table 13 . Table 13. RegressionAnalysis (~3 of FMVSS 214/EU 92/27/EC Criteria for the 4-Dr/2-Dr Vehicles

I 4 I 1 I 2 I 4 I 3

Table 12. FMVSS 214 vs EU 96/27/EC 2-Dr Vehicle Rankings: Driver Driver Vehicle

TTI rank

RDC rank

V*C rank

PelvG rank

PSPF rank

Ford Mustang

1

2

3

1

4

As to vehicle rankings for the 2-Dr vehicles basedon the driver dummy criteria, there was a good match for the thoracic criteria with only the first and second position switched. Rankings based on the pelvic criteria were a poor match. PelvicG rated the Ford Mustang as first while PSPF rated the Mustang as fourth. PelvicG rated the Mitsubishi Eclipse as fourth, while PSPFrated the Eclipse as second. It is worth noting that the Volvo 850 which ranked first amongstthe 4-Dr vehiclesfor the all the injury criteria of both regulations,with the exception of ranking a close secondfor RDC, was the only vehicle in the matrix tested which was designedto meet both regulation. In addition, it has a side mountedair bag system. As to the Lexus SC300,which ranked first or secondamongstthe 2-Dr vehiclesfor both regulations, it was actually designed to meet FMVSS 214. Its- good performancerelative to the EU requirementsmay be attributed to its inherent design, with a sporty wider track and considerablecrush spacebetweenthe occupantand inner door, and betweenthe inner and outer door. 1749

* Regressionwas not performedfor 2-Dr rear occupantsince there was only two data points Overall, the results indicate mediocre or no correlation betweenthe thoracic and pelvic criteria of the two regulations for the eight vehiclestested. In particular, the correlation for the driver dummy thoracic criteria for both the 4-Dr and 2-Dr vehiclesis poor. The correlationfor the rear occupantthoracic criteria for the 4-Dr vehicle is mediocre. Finally, the correlation for the driver dummy pelvic criteria is relatively good, p2=0.77, for the 4-Dr vehiclesbut very poor for the 2Dr vehicles. Real World Vehicle Issues - In Figure 6., the FMVSS 2 14andEU 96/27/ECthoracic injmy criteria valuesare sorted by vehicle weight from left to right for the 2-Dr and 4-Dr vehicles, The vehicle weights are presentedin Table 14. below. In the FMVSS 214 tests, TTI(d) exhibited a trend of better performance,i.e. lower valuesfor the heavier vehicles, for both the 2-Dr and 4-Dr vehicles. This is consistentwith the real world performancein the U.S. crash environment as indicated by a recent study by Farmer et al of vehicle-tovehicle side impact crash study basedon 1988-1992National Accident Sampling System CrashworthinessData System (NASSKDS) [17]. The study, which excluded crashes involving rollovers or ejections, indicated that occupantsof heaviervehicleswere lesslikely to be seriouslyinjured (AIS z 3) in a sideimpact than occupantsof lightweight vehicles. For every extra 45.4 kg in the weight of the subjectvehicle, there was a corresponding7-13% decreasein the odds of serious injury. In contrast,in the EU 96/27/RCtests, the lighter 4-Dr Sonataperformedbetter than the heavier Taurus, while the heavier 2-Dr vehicles performed better overall for the EU thoracic criteria.

Also, in earlier comparativefull scaletesting by Dalmotas et al., using 1988 U. S. production vehicles, the small Chevrolet Sprint performed much better than the large Chevrolet Caprice and Pontiac Bonville when tested to the Europeanprocedure[ 181.The matrix of sevenvehiclesusedin the comparativetesting by Dalmotas et al. exhibited better performancefor the larger vehicleswhen testedto the FMVSS 214 procedure. This trend was relative to TTI(d), computed from the SID and a production prototype Eurosid which was usedas the basis of comparativeperformancefor thesetestsat Transport Canada. Additional full vehicle testing would be neededto further investigate this possible anomaly in the performanceof large vehicles relative to the EU requirements. Table 14. 1

Volvo 850

1666

2670

Hynndai Sonata

1540

2700

Nissan Sentra

1307

2535

Lexus SC300

1819

2685

Ford Mustang

1617

2574

Mitsubishi Eclipse

1459

2515

Geo Metro

1039

2365

II

in the U.S. Full vehicle testing of LTV’s would then also be neededto assessthe relative benefits of the U.S. and EU regulations as applied to LTV’S and in particular assessthe adequacyof the EUMDB in such tests. Movable Deformable Barrier and Test Conditions Issues MDB Issues- As indicated previously, a dynamic crash test requirement in a safety regulation should simulate the crash environment to the extent possible and should also provide for realistic injury causing mechanisms. As shown in Figure 57., over 43% of the fatalities and 37% of the serious injuries (MAIS 1 3) in U.S. light vehicle side impact crashes are in side impacts where an LTV is the striking or bullet vehicle. This is basedon a yearly averagefrom the current U.S. crash environment (19881996 NASWDS and FARS). As shown in Figure 58., when the trend of fatalities in struck vehicles is reviewedfrom 1980through 1996FARS, fatalities in car to car side crashesare decreasingwhile fatalities in LTV to car side crasheshave more than doubled. In fact, a recent study by Gabler and Hollowell indicated that basedon 1996 FARS, side impacts in which an LTV was the bullet vehicle resultedin 56.9% of the all the fatalities in side struck light vehicles [25]. This initial assessment,combinedwith the fact that the LTV population is growing in the U.S. fleet, suggests the following: The MDB in the dynamic test procedurefor a side impact regulation in the U.S. should provide for injury causing mechanisms similar to those caused by the LTV vehicle class in order to provide a good representationof the current and future U.S. side crash environment. The MDB weight, stiffness, and geometrycharacteristicswould needto be evaluatedon this basis.

Awlication of the Standards - FMVSS 214 becomes100 percent effective for light trucks, vans, and multiple purpose vehicles (LTV’s), a growing proportion of the U.S. fleet, in the Fatalities/Injuries in VTV Crashes by 1999 MY. In the EU regulation, vehicles with R-point of 198893 NASS’CDS No RIO lowest seat >700 mm are excluded. The H-points of large !a %Ccc kz % Serious Injury 0 pickups, sportsutility vehicles,largevans,someof the compact pickups, and the majority of minivans are typically larger than 700 mm. As such, the EU regulation does not apply to the majority of LTV’s. The currentU.S. crashenvironment(19881996NASSKDS and Fatality Automotive Reporting System, (FARS)), whenviewed as a yearly average,indicatesthat LTV occupants,are relatively safe when involved in side crashes, accountingfor 14 % of involvement and resulting in 10 % of the severeinjuries and the 11 % of the fatalities. Never the less, LTV’s are currently 34% of vehicle registrations’, and Car Srrrall L-IV their proportion of the U.S. fleet is growing asseenby thetrend Bullet Vehicle in market share,with LTV’s making up 43% of vehicle sales in 1996. As such, there is a need for the EU regulation to Figure 57. addressthe LTV vehicle class if it were to becomeapplicable

’ SourceR. L. Polk Co, 1996

1750

Bullet Vehicle %Killed

Standard LTV

2500

2000 0 1500 p a if! 1000

500

II

on resultsfrom the New Car AssessmentProgram(NCAP) frontal tests. The quotientof the total barrier force and the corresponding displacement of theoccupantcompartmentwas

Fatalities Trend in Light Vehicles FAR.9 01980 ml984 q l988 01992 ml996

usedasa cursorymeasure of vehicle“linear”stiffnessfromthe NCAP frontal test results. This cursorystudy indicatesthat overall,LTVs haveabouttwice the frontal linear stiffnessof cars. It is worth noting that the standarddeviationsin the averagelinearstiffnessfor passenger carsandeachof theLTV vehiclecategoriesis large. This indicatesthat thereis a wide rangeof linear stiffness values within each of the vehicle categories.These initial results indicate that the Plascore EUMDB has a frontal stiffnessrepresentative of a passenger car and is lessstfl thanthe 214MDB. Strengthcomparisons and force versusdeflectioncomparisonsof the EUMDB and 214MDB are alsopresentedin F igure61. and 62.

c

c

0

Car Into Truck Struck

Truck Into Car

Truck into Truck

Vehicle

F igure 58. Average Linear Stiffness (NCAP Frontal Test 1970-1997)

As a vehicle class,LTVs are heavieron the averageas a vehicleclass. A studyby Kahaneof 19851993passenger cars and light trucks indicated that LTVs were on the average heavierby 358 kg than carswith a slowly growingweight m ismatchbetweenthe two classes[26]. In 1993,the sales averagedmassof LTV at 1770kg was422kg heavierthanthat of passengercars at 1348kg. F igure 59. presentsthe test weight of the EUMDB and 214MDB alongwith the average test massof cars,m u ltipurposevehicles(MPV) or sportutility vehicles,pickups,andvansin NCAP frontal testsconducted by NHTSA.

MPV

Average Test Weight (NCAP Frontal Test 1970-1997)

Will

214MDB

EUMDB

F igure 60.

I

2072

I

Pickup

2ooo p E $ s 5; : ; 4 -c

1500

IWO ,I h.

I

I’ “I

SW

0 PC

MW

Pickup

Van

214MDB

EUMDB

F igure 59. iTVs alsotypically havea stiff frame-raildesignversusthe softer car unibodydesigns. F igure 60. presentsthe average stiffnessof the PlascoreEUMDB calculatedfrom the force deflectionperformancecorridors,and the averagestiffnessof the 214MDB derivedfrom the force deflectionresponsein a 40 km/h rigid barrierimpact. Thefigurealsopresentsaverages of the linear stiffnessfor carsandLTV vehiclecategories based 175

F igure 61. EUMDB and 214MDB Force Deflection ResponseComparisons. As a final note,the geometryof the striking bullet and,as such, a representativeside impact MDB should also be addressed. LTV pickupandsportsutility vehicleshavehigher hoodheight than passenger cars. Also, LTVs typically ride higher than cars. As indicatedby the studyby Gabler and

Hollowell, the sport utility vehicle categoryof the LTV class agreementwith vehicle to vehicle damagepatternsthan those hasthe highestride height with an averagerocker panelheight producedby the EUMDB [lS]. of 390 mm. In 1996, Bergmann et al. reported that in a series of vehicle tests performedby Volkswagen AG, the deformation patterns in the vehicle to vehicle impacts can be compared MDB Comparisons neither with thosein the ECER.95 testsnor with thosein the MDB for FMVSS NO114 FMVSS 214 tests [27]. Nevertheless,the data presentedby BergmaM et al. did indicate that the deformationpatternsin the FMVSS 214 tests were a closer match than those of the ECE R.95 tests. Bergmanet al. also reportedthat their vehicle to vehicle testsshowedsevereloading of the struckvehicles in the lower side region, and that the penetrationresistancemust be increased(safetycatch, increasedsill overlap area,etc.) for real accidents.They statedthat suchvehicle design,however, leadsto increasedthoracic loading in the ECE/R.95 test. They EUROMDB with PLASCOBE FACE also statedthat in the developmentphaseof new vehicles, a vehicle can be well abovethe injury limits in FMVSS 214 but exhibit very low occupant loadings in ECE/R.95. As mentionedpreviously, the ECE/R.95 procedurematchesthat of the currentEU Directive, except the barrier height was 260 mm rather then 300 mm. VTV Longitudinal Delta-V Side Impacts 198896

Note: Strength’s(s) are in tits of psi and stifhesses (K) in Ibflin

NASSKDS,

No R/O

B %Ucc w % Sericus Injury 0 o/d