RECONSTRUCTION AND ANALYSIS OF LATERAL COLLISIONS

Peter Niederer Ph . D . Institute for Biomedical Engineering Swi ss Federal Institute of Technology and Univer sity of Zurich

Felix Wal z M . D . , Ulrich Zollinger M . D . Institute of Forensic Medicine University of Zurich

ABSTRACT

In 1 9 7 6 a one year study on severely and fatally inj ured users of seat belts has been per formed in Switzerland . During this program 304 accidents involving 4 1 0 persons ( 2 5 7 OAIS 2 - 6 , 1 5 3 OAIS 6+ ) were investigated . In this report certain aspects of lateral col­li sions in connection with the data obtained in the field study are analysed . The average ISS for near s ide and offside occupants is shown to increase significantly with increasi�g 6v, whereas the inf luence of the mass ratio of the involved vehicles is less pro­nounced . A mathematical simulation analysis of the influence of the intrusion and the lateral distance of the occupant from the impacted door on the attained shoulder force levels is demon­strated .

INTRODUCTION

In 1 9 7 6 the Swi ss Federal Police Administration initiated a one year study on severely and fatally inj ured users of seat belts in Switzer land in connection with a mandatory seat belt law ef fective

January 1 , 1 9 7 6 . Dur ing this program 304 acc idents invo lving 4 10

inj ured persons ( 2 5 7 OAIS 2 - 6 , 1 5 3 OAI S 6 + ) were analysed . I t

was e stimated { l } , that the samp l e contai ned 1 0 % o f a l l seat belt

user s which were i n j ured with a seve r i ty o f OAI S 2 , 2 5 % of the

OAIS 3 ca s e s , and 3 7 % of the ones with OAI S 4 and 5 . With regard

to the fata l i t i e s the sample is complete .

As b e l t e f fec tivene s s has s u f f ic iently been proven in f i e ld in­

ve s t igations in the past no comparative performance study was in­

tended . I nstead , spec i f ic problems were to be analysed including

inj ury patterns of seat belt wearer s , frequency and s t a t i s t i c a l

sign i f icance o f c a s e s with adve r s e b e l t e f fects and b e l t failure s .

The app l ied methodology and the general results are pub l i shed by

Wal z et a l . { l } and Niederer et a l . { 2 } . The re levant inj ury pat­

terns are repor ted by Wal z et a l . { 3 , 4 , 5 } , whi le in { 2 } spec i a l

attention i s g iven t o be lt fai lures and adverse bel t e f f e c t s a s

we l l a s t o the correlat ion between inj ury seve r i ty and the r e l e ­

vant c o l l i s ion parameter s , e spec i a l l y 6v1 ) and ma s s ratio2 ) .

In this report certain aspec t s with regard to lateral co l l i s ions

( 3 and 9 o ' c lock impact direction) are trea ted . Fir s t , the influ­

ence o f &v and the mass ratio o f the involved vehi c l e s i s e s ­

tabl i shed . S econd , an occupant dynamics s tudy with the a i d o f com­

puter s imulation demons trates the dependence of the calcu la ted

peak shoulder force on the intrus ion and o� the latera l d i s tance

o f the occupant from the impacted door .

1 ) � v i s d e f ined a s the v e l o c i ty change during the c o l l i s ion phase . This parame ter i s ei ther e stimated from the ava i l able documen­tation or calcu lated with the aid of SMAC ( S imulation Model o f Automobi l e Col l i sion s , McHenry { 6 , 7 } ) un l e s s a mean ingful value cannot be a s s igned in a case ( e . g . complex c o l l i s ion s ) .

2 ) The mass ratio is d e f ined here a s the ratio between the mass of the vehicle under consideration and its c o l l i sion partner . With increa s ing mas s o f the c o l l i s ion partner , the mass ratio there­fore decrea s e s .

INFLUENCE OF 6V AND MASS RATI O ON THE INJURY SEVERITY IN LATERAL

COLL I S IONS

For the whol e samp l e a corr e l a t ion ana l y s i s { e . g . 8 } reve a l s { 1 , 2 } that both 6 v and the ma s s ratio exhib i t a stati s t ic a l l y s igni f i ­

cant inf luence o n the inj ury seve r i ty measured a s average ISS { 9 } . I f the same procedure i s appl ied to the c la s s e s o f near s ide and

o f f s ide occupants in c o l l i s ions with a 3 and 9 o ' c lock impact d i ­

rection the fol lowing r e s u l t s are obtained :

6 v : near s id e occupant s , 3 3 ca s e s :

o f f s ide occupant s , 1 6 c a s e s :

mass rat i o : near s id e occupant s :

o f f s ide occupant s :

r = . 8 1 ( * * * ) , r = . 5 3 ( * ) ,

r = . 3 8 ( * ) , r = . 1 1 n . s . ,

f iqure

table

f igure

table

1

1

2 2

The inf l uence o f 6 v is seen to be s t a t i s t ica l ly s i g n i f icant for

both near s ide and o f f s ide occupant s . In contra s t to the frontal

c o l l i s ions { 2 } , the ma s s ratio appear s to be of a l e s ser impor­

tance with the s ide co l l i s ions of the sample �

ev. ISS

80

40

8 4 0 II ni:of c ....

o- 111- 31- 411- 111 - 81- U lt.V(km/h) 1& 30 4& eo eo 100

F i g . 1 : Average I S S v s . 6 v , near s ide occupants

ev. ISS

nr. of c• ...

2 9 • 2 10 11 > 1.2 .8- .8- . e- < . II 0 . m 1.2 •• .6

Fig . 2 : Average I S S v s . ma s s ratio (m) , nearside occupants

Av av. ISS 1 • nr . o! km/h i caaea

- - ··- 1 ---1

15 0 - 4 - l 1 1 1 6 - 30 l 9 . 5 8 . 9 4

31 - 45 1 6 . 3 S . 2 6

4 6 - 60 37 . 5 4 6 2

6 1 - 80 - - 0

81 - 100 1 54 - l unknown '

1 3 9 . 9 2

_J_ -- ..

Table 1 : Average I S S vs . 6v , o f f s ide occupants ( s : std . dev . )

1 1

Maas ratio av . ISS • nr. o! C&M8

l . 2 20 - l

. 8 - l . 2 1 3 . 6 6 . 9 5

. 6 - . 8 70 - l

. 5 - . 6 8 - l . 5 22 . 9 6 . 4 7

fixed obj . 1 9 - l i

Table 2 : Aver age I S S vs . mass ratio , o f f s ide occu� pant s

MATHEMATICAL S IMULATION OF LATERAL COLL I S I ONS

The avai lable documentation on l ateral de formation charact e r i s t i c s

o f vehic l e s i s even l e s s s u f f ic ient than the one i n a fronta l d i ­

rection . The application o f SMAC a s we l l a s the " hand " e s tima t i on

o f öv i s based to a con s iderable extent on the knowledge of d e ­

forma t ion character i st ic s . However , in many c a s e s o f l ateral c o l ­

l i s ions one o f the involved vehic les experiences a frontal impact

such that the r e lat ive extent of crush can s erve as an indicat ion

of lateral deformab i l i ty . As an examp l e , in f igure 3 a c a se o f

lateral c o l l i s ion i s documented which wa s reconstruc ted with the

a id of SMAC . The be l ted r ight front pa s senger in the vehicle V l s u s t a i ned fatal head and thorax i n j ur ie s . The intrus ion o f thi s

vehi c l e i s found to be 5 5 cm , 6 v was determined a s approx . 4 6 km/h

and the c a lcula ted deceleration curve i s shown in f i gure 4 .

U t i l i z ing the cal c u lated decel eration pu l s e o f f igure 4 a parame­

tric study wa s performed by adapting the two d imen s ional occupant

dynam i c s program PSOS ( Program for the S imulation and Optim i z a tion

o f S a fety Be l t s , Niederer { 1 0 } ) to a lateral motion . Two dimen­

s ional s imu lation o f side c o l l i s ions has previous l y been made by

Bowman et a l . { 1 1 } u s ing the MVMA 2D mode l ( Robbins et al . { 1 2 } ) .

�-/ :--.....

I I

-

Sehwyr, dlft 16. Aprll 1976

F i g . 3 : P o l i c e documentation and SMAC reconstruction o f a l ateral col l i s ion betwee n vehic l e 5 (Vl , see photograph at end o f paper ) and vehic l e 7 ( V 2 ) . The impact o f vehic le 1 0 wa s o f minor severity . The t ire tracks 3 were caused by Vl ( po s i t ion 4 i s an e a r l ier phase ) the tracks 9 by vehic le 10 .

F i g . 4 : Decelera tion pul s e a s determined by SMAC f o r vehicle V l .

dec:el. (g) 20�

10 f-

. L 40 L - --.o-·---'-'--1-'-20- t (IMK)

The 8 l inks , 1 0 degree s o f freedom model PSOS doe s not a l l ow for a

s imu l a t ion of a l l the l imbs o f a v ictim ( f igure 5 ) . I t i s believed

however that the inf luence of the m i s s inq l imbs as we l l as of the

mo tion out of plane are minimal dur inq the f i r s t ma j or impact ( up

to � 1 0 0 msec ) . To substantiate thi s hypothe s i s a compar ison with

the C a l span 3D Crash Victim S imu lation Proqram ( F l eck et al . { 1 3 } ,

f igure 6 ) was made which showed comparable resul t s .

I n the f o l l owing , the i n f l uence of the intr u s ion I and the d i s tance

D of the head of the v i ctim from the impacted door on the peak load

exerted on the shoulder is analysed . Thi s peak l oad i s found as the

maximal sum of the hor i zontal force components acting on the upper

tor so and upper arm e l l ip s e . The a s s umed arrangement of the veh i c l e

pane l s a s we l l a s the ir d e formation charact e r i s t i c s a r e purely

hypothetica l . The latter are comparable to s imilar pane l s d e f ini ng

the MODROS base line cond i tions ( Danforth et a l . { 1 4 } ) . The c a l ­

cul ated absolute magni tude of the force leve l s ha s therefore a

l imited signi ficance and should ra ther be viewed a s reference

value . The intrus ion was s imulated by a l inear lateral mot ion

with re spect to the vehicle of the two pan e l s def ining the door

and the window during the f i r s t 8 0 msec up to a prescr ibed d i s ­

plac emen t .

0.0 HSEC

0

.. ,.„ . ..... . • . • „ .... :u:. .„_ ·-· .\. i\• „. ..... X ICHI

eo.o "51:t

. . 1 . .

'°-0 HSEC

F i g . 5 : Mot ion sequence o f the two d imensional model PSOS in a s ide impact with intruding door and window pane l s . The side impact is d e f i ned by the dece leration pul s e shown in figure 4 . The in­trusion it mod e l l ed' by a r e lative motion o f the door and window to the vehicle .

F i g . 6 : Examp l e o f a s ide impact mod e l l ed with the C a l span 3 D Crash Victim S imulation Program .

The s imulation r e s u l t s ( f igure 7 ) exhibit e ssent i a l ly two f e a ­

ture s :

At sma l ler d i s tanc e s D , in general , lower peak force leve l s are

obta ined ( un l e s s the occupant has no contact with the impacted

door ) , whereas with increas ing intrus ion the peak force increa­

s e s a l s o . This latter e f fect i s in agreement with observed in­

j ury seve r i ty (Walz et al . { 4 } , Hartemann { 1 $ } , Jones { 1 6 } ) .

- With increasing d i stance D the min imal peak force i s no longer

a s sociated with z ero intrusion . I n spite o f the fact that the

intrus ion proc e s s is not yet compl eted at the time of max imal

load , relative impact veloc ities o f the victim on the door are

the sma l l e s t when an appropr iate amount o f intrus ion occur s .

The higher the d i stance D , the higher this appropriate amount

is found to be .

Head S I values vary from � 1 5 0 up to �800 . There e x i s t s no obvious

corr e lation between the peak shoulder load and the head S I .

CONCLUSIONS

- In the samp l e of 3 3 nears ide and 16 o f f s ide bel ted occupants

who sustained i n j ur i e s of OAIS � 2 in c r a shes with an impact

d irection of 3 or 9 o ' c lock , a stat i s t i c a l ly s igni f icant d e ­

pendence o f t h e inj ury seve r i ty o n 6v could b e proven . I n con­

trast to tha t , the inf luence o f the ma s s ratio is l e s s pro­

nounced ; with the o f f s ide victims no correlation is found with

the in j ury severity .

- With increasing s i tting d i s tance from the impacted door ( a t

l e a s t up to a value at which the occupant s t i l l contacts the

door ) and with i nc r e a s ing intrusion peak forces exerted on the

shoulder genera l ly increase a l s o . Howeve r , at higher d i s tances

from the impacted door at the beginning o f the crash the mi n i -

20

; 5 1--��-t--r-"'--1��--i���t--����

;

0 ;

/ /

/

/ /

20 40 eo l (cm)

F i g . 7 : Dependence o f the peak shoulder force F on the intrusion m I and the d i s t ance D between head and door . At D = 2 0 cm , the

upper arm e l l ip s e is approx . 1 cm apart from the door pane l .

ma l peak force need not be a s soc iated with no intrusion . With

an appropr iate amount o f intrus ion impact veloc it i e s o f the body

on the door can be reduced .

ACKNOWLEDGEMENTS

The Computer Departmen t o f the Swi s s Federal I n s t itute of Tech­

nology is acknowledged for the computer time granted .

194

REFERENCES

{ 1 } F . Wal z , U . Zollinger , A . Renfer , R . Wegmann , M . Meier , P . Niederer , H . Rud i n . Unfal luntersuchung Sicherheitsgurten . E idg . Justiz- und Poli z eidepartement , Bern , Mai 1 9 7 7 .

{ 2 } P . Niederer , F . Wal z , U . Zollinger . Adverse Belt Effects and Causes of Belt Failures in Severe Car Acc idents in Switzer­land 19 7 6 . To appear in Proc . of the 2 l st Stapp Car Crash Conference , 1 9 77 .

{ 3 } F . Wal z , P . Niederer , U . Zollinger , A . Renfer . Analysis of 1 1 5 Killed and 2 0 5 Severely Inj ured ( OAIS�2 ) Seat Belt Users . Proc . of the 6th IAATM Conference , 1 9 7 7 .

{ 4 } F . Wal z , P . Niederer , u . Zollinger . Belted Occupants in S ide and Oblique Impacts . Same Volume .

{ 5 } F . Walz . Verletzungen der Brust- und Lendenwirbelsäule be j Gurtenträqern . To appear in Hefte zur Unfallhei lkunde .

{ 6 } R . McHenry , D . Segal , J . Lynch , P . Henderson . Mathematical Reconstruction of Hiqhway Accidents . CAL Report ZM- 5 0 4 6 -V- l , 1 9 7 3 .

{ 7 } R . McHenry . Computer Aids for Acc ident Investigation . SAE SP-4 1 2 , 1 9 7 6 , p . 8 5 .

{ 8 } P . Armitage . Statistical Methods in Medical Research . Blais­del l , London , 1 9 7 4 .

{ 9 } s . Baker et al . The I n j ury Severity Score . Proc . of the AAAM 1 9 7 4 , p . 5 9 .

{ 1 0 } P . Niederer . Mathematische Optimierung von Sicherheitsgurten . Automobiltechn . Z e itschr . 2 , 1 9 7 7 .

{ l l. } B . Bowman , L . Schneider , D . Foust . S imulated Occupant Res­ponse to S ide-Impact Collisions . Proc � of the 1 9th Stapp Car Crash Conference , 1 9 7 5 .

{ 1 2 } H . Robbins , R . Bennett , B . Bowman . MVMA 2D Crash Victim S i ­mulation Vers ion I . HSRI , Uni v . o f Mich . , UM-HSRI -BI -7 3 - 3 .

{ 1 3 } J . Flec k , F . Butler , S . �rogel . An Improved Three Dimensional Computer S imulation of Crash Victims . DOT-HS- 8 0 1 5 0 7 / 5 1 0 .

{ 1 4 } J . Danforth , C . Randal l . Modi f ied ROS Occupant Dynamics Si­mulation User Manual . GM Res . Labs . , Res . Publ . 1 2 5 4 , 1 9 7 3 .

{ 1 5 } F . Hartemann . Occupant Protection in Lateral Impacts . Proc . of the 2 0th Stapp Car Crash Conf . , 1 9 7 6 .

{ 1 6 } I . Jones . Benefits of Restraint Use with Particular Reference to Acc idents Involving C0mpartment Intrus ion . Proc . of the 5th Conf . of the IAATM , 19 7 5 .

„

III il

Damage extent of vehicle Vl ( see figure 3 ) .