890602
`
`rash Protection in Near=Side Impact -
`Advantages of a Supplemental Inflatable
`Restraint
`Charles Y. Warner, Charles E. Strother, Michael B. James,
`Donald E. Struble, and Timothy P. Egbert
`Collision Safety Engineering
`Orem, UT
`
`ABSTRACT
`
`Collision Safety Engineering, Inc. (CSE),
`has developed a test prototype system to
`protect occupants during lateral impacts.
`It
`is an inflatable system that offers the
`potential of improved protection from thoracic,
`abdominal and pelvic injury by moving an impact
`pad into the occupant early in the crash.
`Further, it shows promise for head and neck
`protection by deployment of a headbag that
`covers the major target areas of B-pillar,
`window space, and roofrail before head impact.
`Preliminary static and full-scale crash tests
`suggest the possibility of injury reduction in
`many real-world crashes, although much
`development work remains before the production
`viability of this concept can be established.
`A description of the system and its preliminary
`testing is preceded by an overview of side
`impact injury and comments on the recent NHTSA
`Rule Making notices dealing with side-impact
`injury.
`
`PROBLEM DEFINITION
`
`Side impacts, according to the National
`Highway Traffic Safety Administration (NHTSA),
`account for 30% of all fatalities and 34% of
`all serious injuries to passenger-car occupants
`(1,2)*. The problem of improving side-impact
`protection has received much attention in
`recent years, leading to NHTSA's
`
`issuance of two notices which propose changes
`to the present Federal Motor Vehicle Safety
`Standard ( FMVS S) 214:
`its Notice of Proposed
`Rulemaking (NPRH Jan. 27, 1988) and Advance
`Notice of Proposed Rulemaking {ANPRM - Aug. 19,
`1988). The January NPRM addresses torso and
`p~lvic injuries, while the August ANPRH
`
`<'•Numbers in parentheses indicate references are
`at end of paper.
`
`addresses the issues of head and neck injuries
`and ejection (1, 2).
`
`INJURY STATISTICS
`
`Many statistical studies of side-impact
`accidents are reported in the literature
`However, significant variations in data
`collection and analysis procedures make it
`difficult to directly compare the results of
`these studies,
`INJURIES - Rouhana and Foster ( 3) made an
`excellent compilation of some of the most
`important side-impact statistical studies_
`They analyzed the National Crash Severity Study
`(NCSS) side-impact data and then compared their
`results (insofar as possible) to other
`published studies of NCSS and other data
`Some
`of their findings include:
`(1) Approximately 40% of all accidents are
`side impacts.
`(2) With regard to occupant seating
`position, near-side occupants
`experience three times the incidence
`of serious or immediately-fatal
`injuries as do far-side occupants.
`(3) Serious injuries are three to ten
`times more likely if the passenger
`compartment sustains intrusion
`( 4) Wh i 1 e
`tho r a c i c in j u r i e s are mos t
`prevalent among "serious injuries,"
`head and neck injuries are most
`prevalent among "immediately-fatal
`injul"ies" (3)
`Although there has been no detailed, in(cid:173)
`dep th analysis of the National Accident
`Severity Study (NASS) data with regard to side
`impact collisiOTis, Hackney, et al. (4) did
`compare the available NASS data to the NCSS
`analysis made by Partyka and Rezabek (5). They
`found that "upper torso/side surface" injuries
`were the major injury category in both NASS and
`NCSS as analyzed by Partyka and Rezabek.
`Using the selection criteria of Rouhana
`and Foster, we looked at the NASS data to see
`
`3
`
`Page 1 of 17
`
`KSS 1017
`
`

`
`whether they would also show that head and neck
`injuries are the most prevalent among the
`immediately-fatal injuries. We found that head
`and neck injuries made up slightly more than
`o~e half of the immediately-fatal injuries,
`with chest injuries making up the remaining
`half,
`In both the NCSS and NASS data a
`significant number of head injuries in side
`impacts have unknown contact points_ While
`this lack of data makes it difficult to know
`the precise mechanisms of these injuries, it is
`reasonable to expect that some (and maybe even
`a subst:antial number) of the side-impact head
`injuries are attributable to the head passing
`through the side window opening and contacting
`either the lower window frame or parts of t:he
`oncoming vehicle.
`ACCIDENT SEVERITY - It is essential to
`establish some basis for categorizing injuries
`according to accident severity, if accident
`statistics are to be used effectively in
`designing for injury reduction .. Both NASS and
`NGSS use the GRASH3 program to calculate
`vehicle center-of-gravity-change in velocity
`{delta-V), which is used as a measure of
`accident severity.
`Research by CSE and others now in progress
`on the NASS side impact data indicates there
`may be important reasons to question the
`validity of the delta-V's found in both NCSS
`and NASS, particularly for side impact
`1)
`collisions. Five areas of concern are:
`missing data, 2) CRASH3 stiffness coefficients
`3) the effect of principal direction of fore~
`(POOF) on crush energy computation, 4) NASS
`field procedures for measuring vehicle side
`crush, and 5) the "missing vehicle" algorithm.
`Each of these five areas are briefly discussed
`below.
`The missing data problem has been
`acknowledged (6), but its ramifications have
`never been seriously analyzed. For side-impact
`collisions the NCSS study has delta·V
`information for only
`55% of the reported
`cases, while NASS has delta-V values for only
`45% of the reported cases.. Drawing conclusions
`on the basis of these minorities of cases is
`equivalent to assuming that the cases with
`delta-V data represent a random sample of all
`cases_ The correctness of that assumption
`needs to be explored.
`It seems likely that the
`cases with delta-V information are not randomly
`distributed through the data base but rather
`are grouped in some way that skews the overall
`picture ..
`The CRASH3 program uses pre-programmed
`stiffness coefficients to compute the crush
`energy from vehicle deformation measurements.
`These stiffness coefficients are selected
`according to the vehicle's wheelbase and the
`location of deformation (side, frontal rear).
`Recent analysis of crash-test data shows that
`the stiffness coefficients used by CRASH3
`significantly overestimate vehicle deformation
`energy associated with relatively small values
`of frontal crush (striking vehicle) and
`
`4
`
`underestimate energy associated with side crush
`(target vehicle)
`Since it is the total
`deformation energy that is used in CRASH3
`to
`calculate the delta-V,s in vehicle-to-vehicle
`collisions (damage algorithm), these errors !!@Y
`t~nd to compensate for each other. However, in
`single vehicle side-impact collisions, the
`GRASH3 program may consistently under predict
`the delta-V.
`Another confounding effect in the way
`CRASH3 computes crush energy from vehicle
`d:for~ation is the effect of the principal
`direc t1on of force ( 7)
`The program multiplies
`the computed crush energy by a so-called
`"correction factor" (of up to 2) which is a
`function of the angle between POOF and a
`perpendicular to the deformed surface of the
`vehicle,
`In many vehicle-vehicle side impacts
`the struck vehicle has significant velocity and
`thus the POOF's in many of these collisions
`differ significantly from the perpendicular.
`~·he "corr:ction factor" is therefore very large
`in many instances
`Compounding this is the
`~eality_ that in these instances the POOF angle
`is :yp1cally very difficult for even expert
`accident reconstructionists to estimate.
`Further, there has been no adequate
`justification given for the particular
`formulation for this factor, which seems to
`assume (contrary to experience) that vehicle
`structures are stiffer rather than more
`compliant when loaded angularly.
`The field procedures used by the NASS
`teams to measure vehicle deformation are also
`in need of careful re-evaluation. For side
`impacts, the crush depth is measured at the.
`maximum crush unless there is also sill crush
`in which case the maximum crush and sill crush
`are numerically averaged. Therefore, for two
`identical vehicles with identical maximum
`crush, the one which has sill crush in addition
`to the maximum crush (and hence which has
`logically absorbed more energy) will actually
`be computed as having absorbed less energy.
`Finally there is the problem of the
`"missing vehicle" algorithm
`This "missing
`vehicle" technique is presumably the result of
`NHTSA's attempts to reduce the missing data
`problem (8). A study of the NCSS file indicates
`that a substantial number of cases do not have
`a calculated delta-V because one of the
`vehicles was not available for inspection by
`the investigating team, An algorithm was
`developed to estimate the energy absorbed by
`the missing car by calculating the apparent:
`inter-vehicle forces from
`t:he crush and
`stiffness coefficients for the known car.
`Oelta-V computations made by this method are
`thus subject to greater errors, particularly in
`view of the problems with the frontal and side
`stiffness coefficients noted earlier.
`In conclusion, studies that attempt to
`address the relationships between injury and
`accident severity using NGSS and NASS data must
`be viewed with some skepticism in light of the
`many-faceted problems involved in the
`estimation of delta-V's, as delineated above.
`
`Page 2 of 17
`
`KSS 1017
`
`

`
`SIDE VERSUS FRONTAL COLLISIONS
`
`The vehicle safety community has made
`significant progress in the past 30 to 40 years
`in crashworthiness design improvements, Host
`of that effort, for appropriate reasons, has
`been concentrated on frontal collisions.
`Frontal and side collisions, however, are
`dramatically different for several traditional
`reasons:
`1) the amount of crush space and
`structure present on the side of a vehicle are
`substantially less than that on the front of a
`vehicle, 2) as a result, in frontal impacts,
`intrusion is not a factor in producing injury
`in all but the most severe collisions, whereas
`in side impacts there is almost always
`intrusions into occupant seating areas, 3)
`ingress and egress of vehicles is through the
`side, and 4) sides have windows (also referred
`to as"glazing") that open and close.. Since the
`automobile must operate within a worldwide
`system of streets, highways, garages, and
`parking facilities, it is unlikely to see these
`constraints altered.
`LIMITED SIDE STRUCTURE AND CRUSH SPACE -
`The general approach to reducing injury
`exposure is to reduce the deceleration
`experienced as the occupant changes speed to
`match that of the vehicle during the collision
`This is achieved by increasing the distance
`over which the occupant undergoes the
`speed
`change.
`In most frontal collisions, this is
`accomplished by linking the occupant to the
`occupant compartment by means of a restraint
`system..
`If the occupant is so restrained, both
`the substantial frontal crush and the space
`forward of the occupant inside the compartment
`("rattlespace") can be effectively used to
`diminish occupant loadings. Obviously, this
`approach cannot offer as much benefit in side
`impacts.
`In a side impact, the near-side occupant
`is seated very close to the collision object,
`separated only by a structure which, for
`practical reasons, cannot be sufficiently
`strong and stiff to keep the inner panel from
`moving inward in any but the lightest impacts,
`Figure 1. The problem is further compounded by
`architectural incompatibilities (e g_ in car(cid:173)
`te-car crashes by the height mismatch between
`the bumpers of striking vehicles and the sill
`and floor structures of struck vehicles, and by
`the broad space between pillars in fixed-object
`collisions).. Thus, in this type of collision
`situation, intrusion is almost always an issue_
`The near-side occupant is usually contacted by
`an accelerating inner door panel, resulting in
`a higher speed change for the occupant than for
`the center of gravity of the vehicle,
`In this
`situation, as shown in Figure 2, the presence
`of rattlespace can increase injury potential by
`allowing the interior surface to get a "running
`start" at the occupant (9).
`Intrusion above
`the struck vehicle beltline and into the window
`opening can also be an issue in side impacts,
`as structure from the striking vehicle may
`approach or penetrate the window opening,
`
`presenting a hard, generally blunt impact
`surface for the occupant's head.
`On the other hand, intrusion is rarely a
`problem in frontal crashes because of the
`extensive structure in front of the passenger
`compartment and the relatively large amount of
`space available forward of the occupant.
`Frontal intrusion is seen mainly in severe
`high-speed crashes, unusual underride and
`override or narrow object situations, or
`oblique sideswipes in which structure may
`actually be peeled away from the passenger
`compartment,
`INGRESS AND EGRESS - Since vehicles must
`have doors on the sides, the design challenge
`is significantly more complicated for side as
`opposed to frontal structures.
`Instead of
`continuous structural members running down the
`length of the side, there must be separate and
`distinct structures tied together at hinge and
`lock locations to accommodate the door,
`The
`force concentrations at these connections can
`be very great in collisions. Design options
`are limited by these connection points and
`their associated structure,
`GLAZING - The upper portion of the side
`structure of almost all passenger cars is
`limited in terms of occupant protection. Side
`window safety glass, even if in the "up"
`position at impact, will generally disintegrate
`early in a side impact collision, usually
`before any occupant contact, The resulting
`openings are certainly a factor in partial and
`full ejections and therefore, head and neck
`injuries. There has been some effort in
`developing retractable side windows with an
`anchored inner plastic layer (10, 11)
`At the
`present time, applications of fixed side or
`retained membrane side glazing have yet to be
`developed to a marketable stage.
`PADDING - Most efforts at improving side(cid:173)
`impact occupant protection have focused on some
`combination of padding and stiffened side
`structure. Well-designed padding can reduce
`injury exposure in two ways:
`l)by increasing
`somewhat the effective acceleration distance of
`the occupant, thereby reducing contact loads
`and 2) by distributing forces over larger
`areas, thereby reducing localized occupant
`loadings, The effectiveness of padding in side
`impacts is limited, however, because of the
`limited space available and the apparent
`negative reaction from consumers to a reduction
`in "elbow room".
`The potential for reducing intrusion
`velocity by increasing door stiffness is also
`severely limited. The forces involved in
`moderate to severe side impacts are simply too
`great to allow practical side structures to
`prevent intrusion in most instances; intrusion
`velocity will always be of concern for near(cid:173)
`side occupants in the crush zone. Consistent
`with this view is NHTSA's evaluation of the
`present FHVSS 214 (which essentially requires a
`door beam), which evaluation shows that the
`standard is effective only when the impact
`
`5
`
`Page 3 of 17
`
`KSS 1017
`
`

`
`5
`
`4
`FIGURE 1
`KEY POINTS OF INTEREST IN A FIXED-OBJECT SIDE IMPACT
`
`3
`
`OCCUPANT CONTACT
`
`-::x:
`
`0..
`::E:
`
`>-
`t-
`1-1
`u
`0
`-I
`w.J
`>
`
`15
`
`10
`
`5
`
`/ POLE-1
`
`• 0
`
`COMPARTMENT
`INTRUSION 17"
`
`CRUSH 23"
`
`0
`
`10
`
`20
`
`30
`
`40
`
`50
`
`60
`
`70 80 90 100
`
`TIME (MSEC)
`
`FIGURE 2
`VELOCITY-Ill-IE GRAPHS FOR A SUBCOMPACT VEHICLE IN A
`20 mph FIXED POLE SIDE IMPACT
`
`6
`
`Page 4 of 17
`
`KSS 1017
`
`

`
`forces are primarily frontal, rear or non(cid:173)
`horizontal (12)
`Very high collision forces combine with
`physical constraints to greatly reduce the
`potential for significantly reducing intrusion
`velocities with reasonable structural
`reinforcements. Repeated research attempts to
`achieve significant benefit from stiffer
`structures have led several researchers to the
`conclusion that once proper attention is paid
`to door attachments and direct load paths,
`additional resources are better spent in
`padding (9). The most effective side structure
`may turn out to be the one that furnishes the
`best backup for the padding it supports, thus
`helping the padding to contact the occupant
`early, spread the intrusion contact and
`restraint forces somewhat, and provide the
`gentlest possible acceleration or "ride-up" to
`the intruder's velocity, while moving the
`occupant as carefully as practicable away from
`the space consumed by the intruder.
`
`NHISA' S PROPOSAL FOR THORACIC AND PEI.VIC IN.JURY
`
`On January 27, 1988, the NHISA issued its
`NPRM to revise the existing Federal Standard
`(FMVSS 214) on side-impact protection (1).
`This proposal involved the substitution of a
`full-scale dynamic-vehicle crash test to
`replace the current static side structure
`strength and stiffness requirements.
`In this
`proposed test, the striking vehicle is to be
`the newly-developed NHISA Moving Deformable
`Barrier (MDB), originally intended to be a
`representation of an intermediate-sized vehicle
`(13). Compliance with the proposed revised
`standard would be on the basis of the Thoracic
`Trauma Index (TII), an acceleration-based
`injury criterion using thoracic accelerometer
`data provided in a specialized Side-Impact
`Dummy (SID)
`The NHISA proposal is an interesting
`attempt at progress toward better side-impact
`protection_ On the positive side, most safety
`researchers would agree that the proposed
`dynamic test can be more realistic than the
`present static crush test required by FMVSS
`214, given appropriate dummy performance and
`injury criteria. Dummy-injury measures seem to
`be at least a potentially more rational method
`of judging a side- impact design as opposed to
`structural strength and door exterior
`deflection measures. On the other hand, NHTSA
`has proposed a relatively complex compliance
`test. The design tasks necessary to ensure
`compliance with the proposed standard will also
`be complex, especially since serious questions
`remain about the benefits of NHTSA's proposal,
`given the spectrum of real-world side impacts.
`Such benefits will depend on the nature of the
`final rule and the efficacy of the resulting
`designs in mitigating injury.
`It is hoped the NHISA notices will have
`the effect of rallying societal effort to
`identify and pursue rational objectives for
`
`evolving improvement in side impact. For this
`to happen, the research and development
`community must achieve agreement on what
`constitutes rational objectives. We believe
`there is basis for re-examination of some
`facets of the NHTSA proposal.
`- Although the
`HONEYCOMB BARRIER FACE
`concept of a standardized crushing surface is
`conceptually appealing as a means of simulating
`the deformation of most side-impact partners,
`the NHISA proposed aluminum honeycomb face for
`the MDB falls short of its reasonable
`performance standardization goals on several
`counts.
`First, the honeycomb face itself does not
`demonstrate standardized or repeatable
`performance.
`Its manufacturing specifications
`do not properly regulate its crush
`characteristics, nor is it entirely reasonable
`to expect standardized crush performance from
`the honeycomb in the oblique buckling mode
`introduced by the crabbed-barrier
`configuration. An energy-absorbing material
`With reduced directional crush sensitivity is
`probably necessary if test variability is to be
`minimized_ Second,
`the current honeycomb
`specification is admittedly too stiff to
`represent the frontal crush of virtually all
`passenger cars, though it is thought by some to
`represent light trucks reasonably well (13),
`Third, the honeycomb material is costly, and in
`short supply, introducing significant
`logistical and financial burdens to testing and
`research programs
`In summary,
`the non-standard, too-stiff,
`too - expensive, aluminum honeycomb barrier face
`does not add realism or effectiveness to the
`test, but does add cost, not only in material
`and logistical senses, but also in invalid test
`results, wasted time, and decreased test
`repeatability.
`It should be eliminated from
`the test requirement.
`In its place, a
`contoured rigid moving barrier should be used
`at an appropriately reduced test speed.
`If
`this approach should need refinement
`a
`subsequent NPRH could be issued to upgrade
`the
`performance test to include an improved
`deformable barrier face when a device with
`appropriate performance, cost, availability,
`and repeatability has been developed and
`proven.
`It is clear that the
`-
`THE SID DUMMY
`anthropometric test device (AID} chest requires
`special treatment for human biofidelity in
`lateral impact, and that existing ATD thoraxes
`designed for frontal biofidelity have proven
`inadequate for the task (14). The proposed SID
`dummy simulates upper-arm inertia, introducing
`what may be an artifact on padding designs.
`It
`is not clear that its use will reduce real(cid:173)
`world injuries unless it can be shown that (a)
`the majority of seriously-injured side-impact
`occupants are loaded through the upper arm in
`its anatomical unextended position and (b) that
`appropriate thoracic and abdominal dynamics are
`represented by the cadaver test data used to
`develop the SID. Recent studies call these two
`
`1
`
`7
`
`Page 5 of 17
`
`KSS 1017
`
`

`
`points into question (15).
`THORACIC TRAUMA
`INDEX. (TTI) INJURY
`CRITERION - The NPRM assumes a reduction of
`side-impact injury as a result of reductions in
`the TTI, based on a body of cadaveric-colerance
`data. While the statistical correlation may be
`a good representation of the cadaver test data,
`it is difficult to apply to the design process
`until a reliable and economical computer
`simulation is available. The lumped-mass model
`presented in Reference 13 !!§Y be a good start
`toward the evolution of such a model. As it
`now stands, a designer is faced with performing
`multiple tests to evaluate padding-design
`changes, a costly and time-consuming procedure
`A simpler injury index, more reflective of the
`physics of the injury process, would be
`preferred by the authors and by the vast
`majority of safety researchers with whom this
`topic has been discussed.
`Given the broad spectrum of masses,
`stiffnesses, shapes and angles of trees, poles,
`posts, rails, car and truck fronts and corners
`and motorcycle components which may try to
`penetrate a car door, the task of improving
`occupant protection begins to look formidable
`indeed. But while a simple solution capable of
`resolving all side-impact issues may not be
`found at once, no progress can be made unless
`and until the basic vehicle-occupant kinematics
`in side impact are first understood.
`
`NEAR-SIDE OCCUPANT/VEHICLE KINE~tATICS
`
`illustrates a typical fixed(cid:173)
`Figure l
`object impact situation with a near-side
`occupant and identifies key points in the
`object/vehicle/occupant system (9). Figure 2
`is a velocity- versus-time plot of the motion
`of these points in a 20 mph lateral test of a
`baseline 1972 Ford Pinto into a 14" diameter
`rigid pole (16). ·rhe velocity curves in Figure
`2 are approximations of the veloci_ties of the
`key points as identified in Figure 1. Since
`acceleration data for these points were not
`available, accelerations were estimated by
`using deformation measurements together with
`high-speed films.
`'Ihe outer door sul:'face (point 2) adjacent
`to the impact location comes immediately to
`rest upon impact,.
`In contrast, the occupant
`compartment (point 5, represented by a point in
`the car side opposite the impact) comes to rest
`more gradually, in this case over a period of
`about 100 msec. The hatched area between these
`two curves (essentially the area under the
`curve for point 5) represents the vehicle crush
`in the plane of the collision, about 23 inches.
`The door -inner panel (point 3) comes to rest
`much more quickly (in about 27 msec)
`than does
`the occupant compartment.
`the shaded area
`between the door inner and outer panel velocity
`curves (points 2 and 3) represents the door
`crush, about 6 inches, and the area between the
`door inner panel (point 3) and the occupant
`compartment (point 5) represents the intrusion
`into the compartment in the plane of the
`
`8
`
`collision, about 17 inches ..
`The motion of the outboard surface of a
`near-side occupant in the plane of the
`collision is illustrated by the curve for point
`4 in Figure 2, starting about 4 inches away
`from the door inner panel Under this baseline
`condition, occupant contact with the door inner
`panel (hip or torso) is estimated to occur at
`about 30 msec. By this time,
`the door inner
`panel (point 3) is at rest so that no ride-down
`benefit is realized, Additional padding of the
`door interior and inner panel could cause
`occupant contact to occur earlier, allowing
`some ridedown and peak-Bhaving benefits.
`Padding could also reduce the level of occupant
`deceleration by contributing a percentage of
`the additional padding distance as "stopping
`distance " An inflatable system in the door
`might also cause occupant deceleration to begin
`earlier and create a mechanism for increasing
`stopping distance. For any benefit to be
`realized, such a system would have to begin
`imparting significant occupant deceleration
`within the first 20-25 msec of the colli~ion
`event.
`Figure 3 identifies the key points of
`interest in an intersection-type car-to-car
`collision. Figure 4 is a time plot of the
`velocity of these points in a test collision
`involving full-size Fords (17) _
`In this test,
`the struck vehicle was stationary and the
`striking vehicle moved at 40 mph. Figure 4
`plots the velocities of the striking car's
`firewall (point 0) and bumper (point 1), the
`struck-side door inner panel (point 3), the
`far-side occupant compartment (point 5), and a
`dummy occupant: (point 4) positioned on the
`struck side adjacent to the intruding door.. As
`seen in Figure 4, the onset of change in
`occupant velocity is delayed until about 25-30
`msec while the occupant "waits" until door
`intrusion advances through the rattlespace. As
`in the case of the fixed-object collision,
`padding and inflatable systems offer a
`potential of producing earlier occupant
`acceleration and increased acceleration
`distance. To be beneficial in this type of
`collision, an inflatable restraint would have
`to begin imparting occupant acceleration within
`20 msec after initial vehicular contact.
`Figure 5 is a velocity- time ploc of a test
`impact in which a prototype of the recently
`proposed NHTSA moving deformable barrier (HDB)
`struck the side of a Chevrolet Citation at an
`angle of 60 degrees just behind the A-pillar.
`In accordance with the proposed test procedure,
`the MDB was "crabbed" at an angle and struck
`the stationary Citation at 33 mph to simulate a
`collision in which a striking vehicle travels
`at 30 mph and a struck vehicle at 15 mph (18).
`The severity of this configuration resulced in
`the door inner panel velocity actually
`exceeding the struck vehicle final velocity of
`about 22 feet per second (9) Occupant contact
`in this case was initiated at about 30-35 msec,
`the worst possible time, since the door inner
`panel was at or near its peak velocity of
`
`Page 6 of 17
`
`KSS 1017
`
`

`
`5
`
`3
`
`4
`
`FIGURE 3
`KEY POINTS OF INTEREST IN A
`CAR-TO-CAR LATERAL COLLISION
`
`STRIKING VEHICLE FIREWALL-0
`
`-~~fii@fff. STRIKING VEHICLE...A·.
`'~'f~!f BUMPER & DOORSKI!l-1, 2 ... ,
`
`· • · .. ·
`
`-------!~-.t..-.. --- ----
`___ .::.~ .. .-r-
`
`PADDING AND DUMMY
`DEFLECTION RECOV/_..---
`
`------
`
`,...-~ COMPARTMENT-5
`
`DOOR PADDING PENETRATION
`AND CHEST COMPLIANCE
`
`/
`
`-
`
`!X" ·
`. ......__ Gp=:46
`
`40
`
`35
`
`30
`
`25
`
`20
`
`15
`
`5
`
`-::c
`
`a.
`::s
`........
`>-......
`
`1--1
`(...)
`0
`,_I
`w
`>
`
`0
`
`10
`
`20
`
`30
`
`40
`
`50
`TIME
`
`60
`
`70
`
`80
`
`90
`
`100
`
`(MSEC)
`
`FIGURE 4
`VEHICLE AND OCCUPANT KINEMATICS IN A
`40 mph CAR-TO-CAR LATERAL IMPACT
`(FIGURE REPRODUCED FROM REFERENCE 17)
`
`9
`
`Page 7 of 17
`
`KSS 1017
`
`

`
`In this case,
`nearly 40 feet per second.
`padding or an inflatable system could have
`similar benefits to the fixed-object case
`above, assuming occupant accelerations could be
`initiated before 20 msec.
`
`The prototype
`the outer door skin
`configuration also employs an upward-deploying
`head bag which is interposed very quickly
`between the occupant's head and surfaces likely
`to cause head injuries.
`
`H muca or t1r.u1:~
`
`S(OE OttllPMIT rnvct.V.Clmn"
`
`40
`
`30
`
`>-
`!::;
`'.3 10
`w
`>
`
`~u:sw oit 11t:Ali(cid:173)
`su1t ruo 11u•o!1S€
`
`-lO'---...---+--<--+-r~i--t--+--1---+--+--+--<~i---+----+-4--+--+--'
`200
`60
`80
`100
`120
`140
`160
`180
`40
`20
`0
`
`TIH£ (MSEcl
`
`FIGURE 5
`VEHICLE KINEMATICS IN A SIMULATED
`MOVING-MOVING CAR-10-CAR LATERAL IMPACT
`(FIGURE REPRODUCED FROM REFERENCE 18)
`
`In summary, at the theoretical level, an
`inflatable cushion in the door panel appears
`promising. It could potentially increase the
`effective occupant stopping distance by
`expanding into the "rattlespace" and could
`effectively distribute loads over large surface
`areas of the occupant. The major theoretical
`concern is whether the bag can inflate fast
`enough to provide a benefit without posing a
`significant deployment threat.
`
`THE DOORBAG CONCEPT
`
`Collision Safety Engineering has developed
`a prototype of a deployab:e door-mounted
`inflatable air cushion and pad system that
`offers·the potential for significantly
`improving the side-impact injury-reduction
`capability of the vehicle interior, as compared
`to the performance of baseline vehicles.
`Conceptually, the doorbag system
`incorporates mechanical performance features
`that address the significant parameters of the
`side-impact crash-protection problem,
`In the
`present configuration, padding is employed to
`provide load distribution and limit occupant
`accelerations. The padding is propelled toward
`the occupant by the deploying doorbag,
`utilizing available interior rattlespace and
`providing early occupant acceleration away from
`the intruding surfaces. Much of the normally
`empty space in the door interior is filled with
`foam to help provide a better load path for
`eacly load application. The doorbag is
`designed to be triggered very early in the
`event by a positive contact switch just inside
`
`10
`
`DEVELOPMENT OF A TEST PROTOTYPE
`
`The sensor, inflator, air cushion
`envelope, interior padding, and polyethelene
`foam were all integrated into a production(cid:173)
`configured 1980 Chevrolet Citation driver's
`door in such a way that the window and door
`mechanisms could be operated normally. The
`door interior was provided with a two-inch
`thick layer of 20 psi polyethelene foam
`padding, enclosing the air cushion system. The
`system was designed so as to maximize the
`probability that an eventual production(cid:173)
`engineered system could endure anticipated
`preimpact storage and function over the
`anticipated car life, without conflict with the
`normal door operation. The different features
`of the prototype system are briefly described
`as follows (Figure 6).
`
`FIGURE 6
`EXPLODED VIEW OF PROTOTYPE DOORBAG SYSTEM
`
`INFLATOR - The inflater was made by
`Thiokol and was similar to those used in the
`Mercedes driver-system airbags, except for
`upscaled gas flow and pressure output, The
`inflater was located near the upper rear
`shoulder of the driver to minimize inflation
`time. Other locations were potentially slower
`and more complicated, requiring ductwork and
`diffusers to preserve window functions.
`AIRBAG ENVELOPE - A single-chamber bag
`incorporating accordian folding was developed.
`A bag/inflator support structure was fabricated
`to support and to facilitate assembly of the
`system into the cutaway baseline door inner
`panel. High-density (30 psi) polystyrene foam
`was hand cut to fit into and occupy the door
`voids, without interference with window- and
`door-operating hardware, while providing a
`secure mounting for the prototype stripswitch
`sensor. Bag volume for the prototype was
`determined by comparing anticipated needs with
`
`Page 8 of 17
`
`KSS 1017
`
`

`
`volumes of successful driver systems. A 60
`liter (2,l cubic foot) bag volume was chosen.
`INTERIOR P~DDING - The interior upholstery
`was moved inward approximately 2 inches by
`installing a layer of 20 psi polyethelene foam
`on the interior door panel
`Polyethelene foam
`was chosen for this research application
`because of its extremely good stiffness-to(cid:173)
`we ight ratio, low cost, and ease of
`