copyright
`
`76
`1
`
`society of Automotive
`Engineers,
`Inc.
`
`840403
`
`Injury and Intrusion
`in Side Impacts and Rollovers
`
`Charles E. Strother, Gregory C. Smith,
`Michael B. James and Charles Y. Warner
`Collision Safety Engineering
`Orem. UT
`
`AEHRME
`
`relationship between occupant crash
`The
`injury and occupant compartment
`intrusion is
`seen in the perspectives of the velocity-time
`analysis and the NCSS statistical data for two
`important accident
`injury modes.
`lateral and
`rollover
`collisions. Restraint
`system use.
`interior impacts. and vehicle design features
`are considered.
`Side
`impact
`intrusion is
`analyzed from physical principles and further
`demonstrated by reference to staged collisions
`and NCSS data.
`Recent publications regarding
`findings of
`the N68 data for
`rollovers. as
`well as the NCSS data itself. are reviewed as a
`background
`for kinematic
`findings
`regarding
`occupant
`injury in rollovers with roof crush.
`The findings in both modes are consistent with
`the proposition that occupant injury is related
`to occupant-interior velocity history.
`rather
`than intrusion Er g, for all but
`the most
`violent crashes. and that any association seen
`between injury and intrusion is likely a mutual
`manifestation of crash severity. rather than a
`cause-effect relationship between intrusion and
`injury.
`Indeed.
`intrusion may have beneficial.
`detrimental. or neutral effects upon injury.
`depending upon seating position and restraint.
`Attempts. to stiffen vehicle structures as
`a means of controlling intrusion will succeed
`as
`injury prevention measures only if they
`accomplish
`a meaningful moderation of
`the
`velocity history of
`the
`contacted surface
`relative to the occupant - otherwise they may
`be counterproductive.
`
`INIRUSION may be defined as the reduction in
`occupant
`compartment
`dimensions
`due
`to
`collision deformation.
`‘The degree of intrusion
`is typically calculated by measuring post-
`impact residual deformation.
`
`the
`is
`paper
`this
`of
`objective
`The
`the possible
`relationship
`consideration of
`between intrusion and injury.
`Intrusion is the
`consequence of the f.i;§_j-_ , i.e.. vehicle
`contact with some collision partner.
`Injury.
`on the other hand. is generally the consequence
`of the second i.mpagt: occupant contact with a
`part of
`the occupied vehicle.
`the collision
`partner. or some other object.
`It is important
`to note that because the severity of both the
`1st and 2nd impacts is related to the overall
`accident severity. one would exm_c_t to see some
`correlations between 1st collision effects,
`such as intrusion. and 2nd collision severity.
`as reflected in the resulting injury. But the
`existence of such correlation or association
`does
`not
`establish
`a
`cause
`and
`effect
`relationship between
`intrusion and
`injury.
`Rather. both are seen to be different effects
`
`of one common cause: the accident severity.
`Intrusion can occur in all accident modes.
`but each collision possesses unique features.
`and requires its own analysis.
`For example.
`the relationship between intrusion and injury
`in
`a
`lateral
`collision
`is
`substantially
`different
`than in a rear collision.
`In the
`same regard,
`the relationships are different
`for various vehicle types.
`The analysis of
`injury
`causation
`in
`frontal
`collisions
`involving
`forward
`control
`vehicles
`is
`understandably different
`than
`for
`similar
`frontal crashes in full sized passenger cars.
`As will be shown. depending on the specific
`circumstances of the collision.
`intrusion can
`be detrimental. it can be of no consequence. or
`it can actually be beneficial.
`Intrusion effects in lateral collisions
`and rollovers are discussed in some detail
`below.
`‘me physical approach employed may be
`applied
`to
`other
`intrusion
`situations.
`Statistical data from the NCSS files and crash
`test data have been employed, where available.
`to
`demonstrate
`the
`applicability of
`the
`analysis.
`»
`
`317
`
`IPR2016-01790
`IPR2016—0l790
`AVS
`AVS
`Exhibit 2007
`Exhibit 2007
`
`
`

`
`HISIDRICAL PERSPECFIVE
`
`The early literature on occupant crash
`safety did
`not
`recognize
`the
`distinction
`between first and second collisions.
`Instead,
`perhaps due to the overwhelming visual
`impact
`of twisted metal and broken glass. many early
`works deplored any deformation which reduced.
`the "occupant survival space."
`Intrusion was
`thought
`to
`be
`the _ of
`injuries.
`Developnent of
`the velocity-time vehicle and
`occupant analysis in the late fifties led to a
`better understanding of occupant dynamics and
`injury causation within
`the vehicle.
`but
`despite its 25 year history this technique has
`not yet attained sufficiently wide application
`within the engineering community.
`Even in the
`current
`literature.
`we
`find
`attempts
`to
`preserve an undeformed "nonencroachment“ volume
`within the vehicle. often assuming that
`the
`volume itself*would provide a superior level of
`safety.
`(1.2)
`Safety Vehicle
`Experimental
`The U.S.
`program.
`initiated by
`the National Highway
`Traffic Safety Administration (NHTSA)
`in the
`late
`sixties.
`incorporated
`a
`"strongbox"
`concept which. while yielding some useful
`scientific demonstrations, also resulted in the
`development of impractical. tank-like prototype
`vehicles which were too heavy.
`too large on the
`outside. and generally too small on the inside
`to be
`suitable
`for
`the marketplace.
`and
`generated sane negative public reaction to
`safety.
`The NHTSA subsequently decided, as
`part of a program reevaluation.
`that the ESV
`goals regarding intrusion were too strict. and
`could
`not
`be practically achieved.
`The
`successor NHTSA program.
`the Research safety
`Vehicle (RSV) . made only grudging recognition
`that
`intrusion preventionn. mt &,
`is not a
`sufficient design criterion.
`Intrusion limits
`persisted, although they were somewhat relaxed
`compared to prior ESV specifications.
`(3)
`An
`improved understanding regarding the
`actual
`role of
`intrusion in injury causation
`has gradually become more widely recognized
`(4). although the "strongbox" philosophy still
`appears in the crashworthiness literature.
`
`INTRUSION IN LATERAL IMPACT
`
`impacts were
`side
`on
`first works
`The
`concerned mostly with vehicle trajectory and
`the collection of limited vehicle acceleration
`data.
`As
`the work of
`reducing injury in
`frontal collisions progressed_ in .the_ early
`1976's. more serious interest in side impacts
`developed. Actual accident case histories were
`studied. mathematical models were developed and
`accident data were collected for statistical
`analysis.
`It became clear that the side impact
`problem would not
`lend itself to the same
`solutions being applied to frontal collisions.
`
`in parenthesis refer to references
`Numbers
`listed at the end of the paper.
`
`318
`
`The most obvious difference between the
`typical side impact and the typical frontal
`collision.
`aside from the
`shorter vehicle-
`exterior-to—occupant distance. was
`the degree
`of
`intrusion into the occupant compartment.
`studies
`spoke
`in terms of occupants being
`impinged
`upon
`by
`the
`intruding structure.
`Early
`statistical
`studies
`showing
`a
`" “ between
`injury
`severity and
`degree of
`intrusion were employed, suggesting
`that intrusion was the injury mechanism in side
`impacts.
`Based on the available information.
`the
`"obvious"
`solution to the
`side
`impact
`problem appeared to be the elimination. or at
`least moderation of.
`intrusion.
`The primary
`goal was
`to maintain the integrity of
`the
`"survival space.“
`The ESV program set a goal of 3-4 inches
`of lateral intrusion in fixed pole (15 mph) and
`car-to-car
`(33-40
`mph
`closing
`velocity)
`impacts.‘
`Once
`this
`goal
`had
`been
`(4)
`formulated.
`reinforced side
`structures were
`developed
`and
`sophisticated
`tests
`were
`conducted in an attempt
`to achieve it.
`The
`most
`promising method
`appeafed
`to‘
`be
`the
`strengthening
`(or
`stiffening)
`door
`structures.
`although it was
`recognized that
`other
`options were
`available.
`including
`softening the
`front
`end
`structure of
`the
`impacting car. Researchers discovered that the
`data did not always match the theory; i.e. the
`problem was not as simple as first supposed.
`stiffer doors reduced intrusion. but sometimes
`increased test dummy injury measures.
`_
`Using the information generated by these
`pioneer efforts (5-32).
`injury mechanisms
`in
`side impacts were re-evaluated, ‘leading to a
`more comprehensive set of injury criteria.
`(35)
`The
`following presents
`an overview of
`the_
`present understanding of side impact phenomena.
`(22.23.29,33)
`.
`Occupants involved in lateral collisions
`can be
`injured by one
`(or more)
`of
`five
`principal mechanisms:
`
`the
`(1) Contacting
`undeformed)
`side structure of
`vehicle.
`
`cr_
`(deformed
`the occupied
`
`(2) Directly
`object_ or vehicle.
`
`contacting
`'
`
`the
`
`striking
`
`(e.g.
`contacted by objects
`(3) Being
`passengers) which were formerly on the opposite
`side of the occupied vehicle.
`
`(4) Being compressed between the impacted
`side structure and other parts of the occupant
`compartment.
`
`totally
`(5) Being either partially or
`ejected from the subject vehicle (which may
`subsequently
`roll).
`resulting in occupant/
`terrain contact
`or
`terrain/occupant/vehicle
`compressive loading.
`
`

`
`is generally not
`Ejection (Mechanism 5)
`thought to be intrusion—related and will not be
`elaborated upon in this paper. although dmr
`omning.
`intrusion. and injury level are all
`manifestations of accident severity.
`It
`is
`interesting to note that ejections in side
`impacts are generally the result of either
`rollover
`or
`door
`openings.
`and
`that door
`openings are usually the result of excessive
`longitudinal
`forces
`on
`the
`door
`retention
`system (hinges.
`latch. etc.).
`These forces
`would still exist. and may even be greater. if
`the side structure were incapable of
`lateral
`crush and intrusion.
`True compressive-type injuries (Mechanism
`4) are not of great statistical concern because
`the accident severity required to produce the
`dimensional
`changes
`necessary
`for
`such
`compression .
`in most
`cases .
`produce
`lethal
`contact—type
`injuries
`prior
`to
`occupant
`Attention in this section is thus
`compression.
`the
`first
`three
`focused on
`(contact-type )
`V
`V
`injury mechanisms cited above.
`Contact
`injuries to the thorax, abdomen.
`pelvis
`can
`be
`serious
`indeed,
`but
`and
`definitive
`criteria
`relating
`to
`specific
`injuries are not in a form useful for vehicle
`design.(34)
`Further. although much has been
`accomplished. anthropometric test dummies are
`still in the development stage so far as side
`impact injuries are concerned.(35.36.37.38)
`It
`is
`clear
`that
`a
`simple
`accelerometer
`measurement’ on a
`dummy will give only an
`indication
`of
`injury
`potential
`for
`some
`mechanisms of
`injury.
`The BLUR criterion has
`been ‘proposed by Robbins et.al. as an injury
`correlator..(39)
`The AFIR technique developed
`by Eppinger. et.al. (40) also shows some promise
`of
`injury correlation. but neither BLUR nor
`AFIR have yet generated accepted specific
`design guidelines.
`side impact
`The
`following analyses of
`injury causation focus on the occupant—docr
`Contact velocity as a physically identifiable.
`albeit
`incomplete parameter of
`side impact
`injury causation.
`
`LATERAL COLLISIONS WITH FIXED OBJECTS-
`Fixed-object
`and car—to—car
`lateral
`impacts.
`although they share many commonalities. are
`significantly different and deserve separate
`attention.
`since the fixed object collision is
`easier to understand. it is discussed first.
`Figure 1 identifies several key points in
`the object/vehicle/occupant
`system.
`It
`is
`important to understand the motion of the these
`points during the collision. not only in order
`to describe the phenomena. but
`to assess the
`role that intrusion plays in injury causation.
`Figure 2 is a pair of velocity-ti.me graphs of
`the motion of
`these points
`in two 2% mph
`lateral pole tests discussed in Reference 41.
`Figure 2a describes an unmodified vehicle (1971
`Pinto; Test Dl) while Figure 2b represents a
`vehicle
`with
`extensive
`side—structure
`modifications to reduce intrusion (1972 Pinto;
`
`Since acceleration traces for the
`Test D2) .
`key points in Figure 1 were not available from
`these tests. approximations of the velocities
`of these points were made by the authors using
`post—itmpact
`deformation
`measurements.
`Velocity-time histories are extremely useful
`for
`understanding
`and
`explaining
`crash
`phenomena.
`since
`all
`three.
`kinematic
`parameters .
`velocity .
`acceleration .
`and
`displacement. appear on the same graph. either
`directly. as slopes (acceleration) . or as areas
`(relative displacements).
`the
`Directing attention to Figure 2a.
`outer door surface (point 2) of the unmodified
`vehicle comes immediately to rest upon impact.
`The occupant compartment (point 5) on the other
`hand. comes to rest more gradually — in this
`case over a time interval of approximately 1%
`msec. The area between these two curves (shown.
`hatched) is the vehicle crush in the plane of
`the collision (about
`23
`inches).
`In this
`impact the door inner panel (point 3) comes to
`rest in about 25-30 msec.
`The area under the
`door
`inner panel velocity. curve is the door
`exterior crush (about 6 inches) and the area
`between this curve and the compartment velocity
`curve
`(shown shaded)
`is the intrusion in the
`plane of the collision (measured: 17 inches).
`Now consider
`the motion of occupants at
`various positions and conditions of restraint
`in this unmodified vehicle impact.
`The motion
`of a belted or unbelted near—side occupant
`(about 4 inches away from the door inner panel)
`in the plane of the collision is illustrated by
`the solid line in Figure 2a. Occupant contact
`with the door
`inner panel
`(hip.
`torso) or
`struck object
`(head)
`is estimated to occur at
`about 35 milliseconds. All potential contact
`surfaces are at rest and will remain at rest at
`this time butQnl¥amu£lA39f.1.'.11es1ent_ual
`intrusion has taken plagx
`the
`In Figure 2b. a similar graph for
`stiffened vehicle.
`the occupant is observed to
`again strike objects at rest.
`Thus in both
`cases.
`the nearside occupant contacts the door
`panel/fixed object at a velocity equal to the
`closing velocity prior to impact. The occupant
`in both cases is brought violently to rest by
`contact with the door panel/fixed object. but
`intrusion is szbserxed
`1:9
`continue
`after
`occupant
`impact as the vehicle deforms around
`the struck object. This further intrusion has
`no affect on the injury potential
`for
`the
`occupant (barring a compression—type injury. or
`ejection).
`Studies
`have
`indicated
`that
`increased intrusion may have
`the effect of
`creating a more hostile environment when the
`occupant makes
`contact with
`the
`vehicle
`interior (i.e. sharp poinm or exposed edges).
`(ll.l6.l7)
`This certainly is possible. but
`observe in the example that the occupant has
`made violent contact with the door panel at a
`point where
`intrusion is
`relatively small
`(about
`5-6
`inches:
`less
`than
`the
`total
`intrusion for the modified vehicle).
`In any
`event.
`the measure of
`post—impact
`residual
`
`319
`
`

`
`FIGIRE I-KEY POINTS ll-'
`
`INTEREST IN A FIXED-QJECT SIE IQFACT
`
`mmms
`
`2.3>:8.m>
`
`DCCUPANT (4)
`
`UCCUPANT CONTACT
`
`INTRUSION <~17"> ‘
`A
`
`Z causu <~23">
`
`E5_E8.n._>
`
`
`
`V\ .:\\%\\§m»>..“..
`§\mm.
`
`88
`
`mE,....
`
`DCCUPANT (4)
`
`WW
`
`6(~\mm
`
`0.
`
`CDMPARTMENT <5;
`
`
`
`FIGRE 2 VELOCITY-TIME GRAPHS FOR A BASELINE AND
`MODIFIED SUBCDNPACT VEHICLE" IN A 20 mph
`FIXED POLE (14-inch DIAMETER)
`IMPACT
`
`320
`
`
`
`
`
`

`
`intrusion is seen to be a completely unreliable
`indication of the injury potential for a near-
`side occupant in the plane of the fixed—object
`collision (where maximum intrusion is likely to
`occur) .
`V
`as
`kinematics.
`of vehicle
`study
`The
`depicted in Figure 2. will disclose that near-
`side belted or unbelted occupants who are
`positioned away from the region of
`intruded
`side structure in fixed—object collisions may
`very well
`txanefij;
`from intrusion.
`These
`occupants will
`tend to encounter a vehicle
`interior that
`is moving with a velocity more
`characteristic of
`the
`compartment.
`It
`is
`obvious
`that higher compartment acceleration
`levels will be associated with stiffer modified
`structure.
`This
`increase
`in
`comparunent
`acceleration can prevent
`the off-impact near-
`side
`occupant
`from benefitting
`from the
`"ridedown"
`characteristic
`that
`intrusion
`-
`provides.
`consideration of Figure 2 will
`Again .
`indicate that far—side belted occupants. linked
`by
`their
`restraint
`system to the vehicle
`compartment.
`likewise can benefit from themore
`gradual
`compartment
`arrest
`accompanying
`intrusion.
`Of
`course
`this benefit can be
`somewhat negated if intrusion becomes so large
`that
`occupants make violent
`contact with
`arrested near—side surfaces.
`'
`A
`Figure 2 suggests that
`injuries to far-
`side unbelted occupants
`in the plane of
`the
`collision will generally be unaffected by
`intrusion
`levels.
`‘me
`relatively
`large
`distance the far-side occupant must
`travel
`before contact is made will usually assure that
`the intruded interior panel. whether stiffened
`or not. has come to rest before contact.
`On
`the other hand. much more deformation will have
`occurred prior
`to far-side occupant
`impact.
`The potential
`is thus greater
`for
`increased
`hostility
`of
`contacted
`vehicle
`interior
`surfaces. Consideration of vehicle kinematics
`
`that unbelted far-side occupants "away
`shows
`from the
`intrusion zone will probably not
`benefit
`from the
`enhanced
`ridedown
`that
`
`intrusion provides their belted counterparts.
`due to the remoteness of
`the vehicle contact
`
`‘
`'
`_
`surfaces.
`can
`should be noted that vehicles
`It
`laterally impact
`"fixed" objects
`that may
`consequently be moved. Occupants sensitive to
`the compartment velocity (usually all occupants
`except
`those near-side occupants adjacent
`the
`impact) will
`generally
`benefit
`from the
`reduction in compartment velocity change that
`accompanies
`displacement
`of
`the
`collision
`partner.
`Near-side.
`adjacent
`occupants.
`however. may not realize such benefit if the
`object displacement
`is delayed significantly.
`In Reference 42. for example. a 1976 Rabbit was
`subjected to a 30 mph broadside collision with
`a surrrogate breakaway luminaire support.
`The
`breakaway feature reduced the vehicle velocity
`change
`to
`8-9
`mph
`but
`the
`near-side
`anthropometric dummy
`seated adjacent
`to the
`impact zone sustained a 3!?! mph impact against
`
`321
`
`his vehicle's side structure irrespective of
`intrusion.
`fme
`reason
`for
`the
`lack of
`effectiveness of the breakaway pole for this
`occupant is that the inertia of the pole and
`its required attachment strength resulted in a
`2E-36 millisecond delay in its displacement
`away from the impact (see Figure 3) ._ Near-side
`occupant contact and arrest occured during this
`timeframe. so that this occupant struck a side
`interior door that had been brought to rest by
`the "breakaway" pole before vehicle follow-
`through
`forced
`pole
`displacement without
`substantial structural intrusion.
`‘
`-
`In summary.
`the study of vehicle and
`occupant kinematics indicates_ that very litfle
`detriment
`is
`seen
`to be
`associated with
`intrusion in fixed object lateral impacts and
`some benefits are potentially present for near-
`side occupants outside the crush zone and far-
`side belted occupants.
`Post-impact intrusion.
`per se,
`is a poor and unreliable measure of
`crashworthiness
`in
`fixed-object
`lateral
`impacts.
`
`fixed—object
`COLLISIODG-In
`CAR-‘IO-(‘AR
`the vehicle
`collisions. occupants moving at
`stationary
`impact
`velocity
`often
`contact
`surfaces at a velocity equal
`to the vehicle's
`pre-impact lateral velocity.
`In most car-to-
`car
`side collisions.
`however." neither
`the
`struck vehicle nor
`its occupants have
`any
`appreciable lateral velocity before impact.
`Lateral velocity is imparted to unbelted and
`nearside belted occupants by virtue. of contact
`with the vehicle side structure whidz can. and
`often ‘ does.
`intrude
`into
`the
`occupant
`compartment.
`An analysis of a typical V—t
`curve
`shows
`that depending upon the injury
`criteria being employed. the depth of intrusion
`may or may not be as important as the velocity
`associated with this intrusion during its early
`stages.
`‘mat is.
`the important criterion may
`be the interior panel relative-velocity—impulse
`experienced
`upon
`occupant
`_y contact.
`.
`.-
`_
`(22.23.29,33)
`_
`,
`’
`'
`»Figure 4 identifies the key points of
`interest in a car-to-car collision.
`Figure 5
`is a graph of the velocity of these points in a
`"El-bone" collision involving full size -Fords.
`(43)
`In this test
`the struck vehicle was
`stationary andthe striking vehicle was moving
`at
`40 mph.
`Plotted on Figure 4 are the
`velocities of the striking car'~s firewall and
`bumper.
`the struck-side door
`inner panel.
`the
`far—side occupant
`compartment.
`and a
`dummy
`positioned on the struck side adjacent
`the
`intruding door.
`Intrusion is represented on
`this graph by the area between the door. inner
`panel
`and the compartment velocities.
`The
`near-side occupant contacts the door before it
`attains its final velocity (Vf) , He reaches a
`common velocity with the door at around 4
`milliseconds.
`and stays in contact with the
`door until about 65 milliseconds.
`Intrusion.
`on the other hand. continues for a period well
`beyond 65 milliseconds.
`The intrusion that
`
`'
`
`

`
`Figure 3 . Test Sequence for the Initial Collision Event
`of Full Scale Crash Test Number 1469—4A82
`(Reproduced frcn1Reference 42)
`-
`
`t = 0.150 sec
`
`322
`
`

`
`
`
`rxcuae 4-KEY PUINTS or INTEREST IN A CAR-T0-CAR LATERAL COLLISION
`
`STRIKIPI3 VEHIQ.E FIREVALL G)
`CAFFHIX.)
`
`-l>- S
`
`U.) U1
`
`',
`
`smxxm: vauus uzusa
`
`
`
`U.)G
`
`f\) U}
`
`.
`_._.._..._.._‘.1..r_.
`
`
`
`
`
`VELOCITY(mph)
`
`f\.) 8
`
`ha U1
`
`TO EIIIY
`1% INITIAL E?
`DI
`'3
`
`
`
`‘ Mmmznouuon
`
`,
`
`(---J"
`
`. \\\\\\\\\\\\L__--_.__"
`\\\x\\\\
`
`
`
`IE
`
`2%
`
`38
`
`40
`
`5!?!
`TIME
`
`Z
`ac)
`
`70
`
`8%
`
`9121
`
`121
`
`FIGURE 3 VEHICLE AND UCCUPANT KINEMATICS IN
`CAR LA
`L IMPA
`A 4U mph CA_R‘TU‘
`TE
`(FIGURE REPRUUUCEU FROM RE
`
`ENCE
`
`323
`
`

`
`takes place after the near-side occupant has
`experienced his velocity change is of little or
`no injury consequence-
`If injury is caused by
`a mechanism sensitive only to occupant contact
`velocity the intrusion that occurs subsequent
`to door contact (at about 25 msec) will be of
`little or no importance. On the other hand. if
`injury is caused by a mechanism more sensitive
`to body motion (e.g. chest compression or organ
`displacement)
`then only the
`intrusion that
`occurs during the ti.me required to bring the
`occupant
`up
`to
`door
`velocity would
`be
`important .
`In either
`case .
`the
`eventual
`intrusion would be irrevelant. Therefore.
`to
`improve the side structure of a vehicle. one
`must
`either reduce the intrlisien 1els&it¥
`during the first 25 Insets (e-9- BLUR- AFIR) or
`otherwise reduce the impulse imparted to the
`occupant. Reducing measured residual intrusion
`will not necessarily effect
`a decrease
`in
`expected occupant injury levels.
`Figure 6
`(Ref. 44)
`is another velocity-
`time graph for an impact in which a "typical"
`intermediate
`vehicle.
`represented
`by
`a
`crushable moving barrier. travelling at 3% mph.
`struck the side of a Chevrolet Citation which
`was traveling forward at 15 mph.
`The impact
`angle was 60 degrees. and the first contact was
`just
`behind
`the A-pillar.
`Due
`to this
`orientation, only a small portion of the side
`structure of the struck vehicle (Citation) was
`involved in the crash. which resulted in the
`door
`inner panel velocity actually exceeding
`the struck vehicle final velocity (about 22
`feet per second) during the period of near-side
`occupant contact. Thus.
`the near—side adjacent
`dummy occupant was
`subjected to an
`impact
`severity (as measured by his velocity change or
`delta-v)
`about 15..ear®nt greater than that
`indicated by the vehicle change in velocity.
`Again.
`intrusion continued well
`beyond
`the
`period of
`dummy contact.
`Vehicle kinematic
`consideration thus
`indicates that a stiffer
`side structure which did not either effect the
`door
`inner panel velocity during early stages
`or reduce the impulse imparted to the occupant
`would be of
`little or no benefit
`in this
`situation.
`Further consideration of vehicle and
`occupant kinematics will disclose the
`same
`observations
`regarding
`(1)
`the
`possible
`benefits of
`intrusion to near—side occupants
`remote from the impact zone and to belted far-
`side
`occupants.
`and
`(2)
`the
`probable
`irrelevance of
`intrusion levels for unbelted
`far-side occupants.
`apply to the car-to—car
`case as well as to the fixed-object collision.
`Using the reduction of residual intrusion
`as a design goal
`in crashworthiness research.
`without consideration of occupant kinematics
`and
`injury
`criteria
`is
`a
`tempting,
`but
`dangerous .
`approach .
`Unfortunately .
`it
`continues
`to be employed.
`Certainly post-
`imrpact
`intrusion levels are easily measured,
`but today's level of instrumentation technology
`is such
`that data more
`relevant
`to early
`
`324
`
`readily
`be made
`can
`velocity
`intrusion
`It should be clear from the above
`available.
`that residual
`intrusion should only correlate
`with injury to the extent
`that it correlates
`with accident severity; that is. to the extent
`that
`it correlates with the more probable
`injury producer in the majority of cases:
`door
`velocity during occupant impact.
`‘
`The latest example known to the authors of
`dangers
`inherent
`in
`using
`residual
`the
`intrusion as a design criterion in lateral
`impacts
`arose
`in the NHTSA's
`"Lightweight
`Subcompact Vehicle Side Structure Program". (36)
`In this program. a VW Rabbit was the subject
`vehicle
`for
`a
`series
`of
`structural
`modifications aimai at
`intrusion in
`car-to—car side impacts. The modified vehicles
`were
`subjected to dynamic tests duplicating
`earlier tests with baseline cars.
`Padding was
`added to the door
`interiors in later tests to
`
`evaluate a "complete door protection system".
`Three degrees of structural modifications were
`developed:
`lightweight. middleweight.
`and
`heavyweight.
`It was envisioned that they would
`show increasing effectiveness with increasing
`weight penalty.
`In the final lightweight and
`middleweight designs. which added about 27 and
`66 pounds respectively to the vehicle. a number
`of
`features were
`combined, all
`aimed at
`increasing the bending strength of the door and
`its perimeter attachment
`to the door opening
`(strengthened door
`beams,
`sill-to-door—ties.
`strengthened door hinges). Although intrusion
`levels were reduced,
`the reductions resulted
`from increases
`in
`stiffness which were
`manifested only relatively late in the accident
`sequence.
`As seen in Figure 7
`(taken from
`Reference
`313) .
`the
`increased Rabbit
`side
`structure stiffness was not evident during the
`first 19 inches or so of side crush.
`the regime
`during which near-side occupant contact with
`the door
`inner panel would most likely occur.
`Thus.
`although
`the
`immediate objective
`of
`reducing intrusion was achieved, it was done
`without significantly changing the door
`inner
`panel velocity during the occupant contact
`period-
`i-e.
`the occupant
`iniury notential
`PI.Q12a.l2l¥ was not changed.
`Ta.ble l. extracted
`from Reference 45.
`summarizes the crash test
`results which illustrate this.
`
`above.
`seen
`PADDIM-As
`IMPACT
`SIDE
`structural modifications that effect an early
`decrease in door
`interior velocity offer the
`potential
`for
`reducing
`near—side
`adjacent
`occupant
`injuries
`in
`car-toecar
`lateral-
`collisions. Another obvious countermeasure for
`consideration is the addition of padding to the
`door
`interior panel.
`Padding will not affect
`occupant change—in—velocity. of course. but can
`decrease acceleration levels and improve load
`distribution during occupant/vehicle contact.
`Padding futhermore has a lesser weight penalty
`and can be effective in both fixed-object and
`car-to—car impacts.
`Space is limited, however.
`so that 3-4 inches of padding is about
`the
`
`

`
`
`
`VELOCITY(Ft/coo)
`
`-18
`
`AREA EBUALS INTRUSI0N522"
`(MEASURED STATIC INTRUSION-'-15")
`
`«t BASED ON NEAR-SIDE
`RIB RESPONSE
`
`23
`
`4E
`
`153 125 145 153 133 253
`83
`53
`TIME (mace)
`
`FIQRE 5 VEHICLE KIXHATICS IN A SIKLATED KJVIM3-KJVIM;
`CAR-T0-CAR LATERAL IIPACT CFIGIRE REPRCIIIZED
`FROM REFERENCE 44)
`
`Middleweight/—
`éghtweight
`
`I
`
`I
`
`-._._
`
`
`
`Force(Kips)
`
`/
`
`f__,Baseline
`
`30
`25
`2
`Displacement (in)
`
`Figure 7 .
`
`Ccmparison of the Static
`Crush Strengths of the
`Side Structures of Baseline
`and Modified vw Rabbits.
`(Figures reproduced from
`Reference 30)
`
`325
`
`

`
`TABLE
`
`1
`
`CRASH '1'E,s'1‘ SUMMRY (Table reproduced from Reference 45)
`
`
`
`
`
`
`
`Contact
`Dynamic Intrusion
`Peak Acceleration
`Velocitz
`@ 120 msec
`
`Chest
`Pelvis
`
`Resultant
`
`Lower Door
`Chest
`Pelvis
`
`
`(inches)
`
`
`
`Upper Door
`(inches)
`
`No
`
`Side Structure
`
`
`
`
`
`
`
`
`
`1°
`
`
`
`
`
`
`
`TABLE 2 — The Proportion of Near—Side Occupants in Lateral Collisions
`(8-10, 2-4 o'clock, door damage) Sustaining Severe, Serious,
`or Critical Injury as a Function of Vehicle
`43V and In-
`trusion Levels (NCSS Data, Phase 2)
`
`
`
`
`(10.82)
`
`.122
`
`Intrusion
`Level
`(in.)
`
`Vehicle
`
`
`
`(1/1) 1.000
`
`0-10
`
`11-20
`
`(3/31)
`
`.097
`
`(6/50)
`
`.120
`
`(2/23)
`
`.130
`
`
`
`6-10
`
`(0/34)
`
`.000
`
`(13/49).265
`
`
`
`(16/106)
`
`.151
`
`(10/32)
`
`.313
`
`
`
`.474
`
`
` (10/d6).278
`IIIIHEEIEEHNIIHHRIIII
`
`
`(9/19)
`
`iIl%i%HIIIHHHIIN
`|INHHHHl||HH%|I
`
`(0/4)
`
`.000
`
`
`
`
`
`INIIENHHIIIIIN
`
`326
`
`

`
`factors
`human
`with
`consistent
`maximum
`considerations and marketability. Also. adding
`padding to the upper door structure can prevent
`the upper door sheet metal from deforming (and
`functioning as padding) as has been observed to
`occur
`(45).
`It seems clear at this point
`in
`side impact research that some combinations of
`well—conceived
`structural modifications
`and
`carefully
`constituted
`and
`configured
`side
`
`padding will evolve as viable steps forward in
`occupant protection.
`'I'he evolution of
`such
`systems is dependent on the completion of side
`impact dummy development and injury criteria
`selection.
`
`of
`number
`S'fl&'1'IS'l‘IC‘S-A
`IMPACT
`SIDE
`studies have been conducted of
`statistical
`lateral
`collisions
`using treasured accident
`data. (46)
`Both NCSS and the National Accident
`Severity Study (NASS) record intrusion in terms
`of
`the Collision Deformation Classification
`(CDC) as defined by SAE J224.
`'Ihe CDC code is
`a very rough categorization of exterior crush
`(rather than interior intrusion) with the first
`of the five crush extent categories believed to
`correspond to minimal or no intrusion.
`NCSS
`has
`also undertaken
`some
`limited studies
`involving
`actual
`intrusion
`dimension
`measurements.
`A
`significant
`problem in
`attempting to use these data is that a post
`impact examination of
`the extent of intrusion
`into the occupant space does not define the
`velocity relationship between the door interior
`and the occupant during contact.
`A more
`extensive case vehicle examination of the the
`type detailed in Reference 29. or perhaps a
`computerized reconstruction analysis is needed.
`Reference
`47
`is an effort
`to compare
`injury prediction in laboratory (sled)
`tests
`with accident data from the NCSS data base.
`E‘ifty—one selected side impact accident cases
`were
`studied
`in detail
`by
`reviewing
`the
`accident reports used to code the individual
`NCSS cases.
`‘me basis for selection of a case
`for clinical review was
`(a)
`the existance of a
`near—side occupant and (b) a 3 or 9 o'clock
`principal direction of
`force
`(PDF).
`It was
`found that direct occupant compartment
`impact
`yielded injuries consistent with laboratory
`results. Furthermore.
`impact not directly into
`the occupant compartment (i.e.
`impacts on the
`occupant
`side
`but
`forward or aft of
`the
`occupant) produced observably lower
`injuries
`for
`the same vehicle change in velocity. as
`reconstructed using the CRASH 2 algorithm.
`In
`a broader set of side collisions from the NCSS
`data. selected for near—side occupant crashes
`at 2-4 or 8-10 o'clock PDF impacts.
`the results
`supported the same conclusions.
`Based upon these findings, a relationship
`between injury and direct
`intrusion into the
`occupant
`compartment was postulated. noting
`that vehicle change in velocity is insufficient
`to correlate
`injury.
`These
`findings are
`consistent with the velocity analysis above and
`the hypothesis that injury is closely related
`
`327
`
`to the door velocity during occupant contact.
`As was evident in this velocity analysis, near-
`side adjacent occupants will contact vehicle
`side structure (or external objects) at a
`velocity near or.
`in some car-to-car cases.
`exceeding the vehicle velocity change.
`This
`situation thus
`approximates
`the
`laboratory
`situation in which cadavers are impacted into
`stationary side structure at a velocity equal
`to the
`vehicle
`change-in-velocity.
`‘Best
`occupants away from the intrusion zone. on the
`other hand. contacted the side structure moving
`at
`a velocity more characteristic of
`the
`vehicle compartment and hence benefitted from
`crush/intrusion—created ridedown.
`me NCSS intrusion data was examined in an
`attempt to gain insight into the correlation of
`measured maximum residual
`intrusion with
`observed injury cases which involves near-side
`front
`seat
`occupants
`(drivers
`and
`front
`passengers)
`involved in 8-lfl and 2-4 o'clock
`PDF impacts involving door deformation. Table
`2
`is a summary of
`the results. giving the
`proportion of such occupants receiving severe.
`serious. or critical injuries (OAIS levels 3 to
`6)
`(48) as a function of vehicle change-in-
`velocity (delta-V) and maximum residual
`(door.
`B-pillar)
`intrusion.
`The data of Table 2 are
`limited (239 cases)
`and portray a somewhat
`confusing picture of the role of intrusion and
`vehic