Procedures for rating seat/head restraints

Head restraint geometry explained: The necessary first attribute of an effective head restraint is good geometry. If a head restraint isn't behind and close to the back of an occupant's head, it can't prevent a "whiplash" injury in a rear-end collision. Institute researchers regularly evaluate the geometry of head restraints in passenger vehicles based on the height and backset relative to an average-size male. A restraint should be at least as high as the head's center of gravity, or about 9 centimeters (3.5 inches) below the top of the head. The backset, or distance behind the head, should be as small as possible. Backsets of more than 10 centimeters (about 4 inches) have been associated with increased symptoms of neck injury in crashes.

Geometric ratings: Geometric ratings are good predictors of how well people will be protected in rear-end crashes — drivers with restraints rated good are less likely than those with poor restraints to claim neck injuries. Head restraint geometric ratings for hundreds of passenger vehicles are listed by vehicle make and series. Various head/seat combinations are rated (not every available seat option in every series has been measured). The restraints are measured with the angle of the torso at about 25 degrees, a typical seatback angle. Each restraint is classified according to its height and backset into one of four geometric zones — good, acceptable, marginal or poor.

Head restraint measuring device with ratings chart

Since 1995, the Institute has been publishing model-by-model ratings of head restraint geometry, based on a procedure for taking geometric measurements. The rating for a fixed head restraint is straightforward — the zone into which its height and backset place it also defines its rating. The rating for a head restraint that adjusts in height and/or backset depends on whether it locks in the adjusted position. If it doesn't lock, its rating is defined by its height and backset in the down and/or rear position. If it does lock, height and backset are measured twice — in the down position and in the most favorable adjusted and locked position. The final rating is the better of the two, except that if the rating as adjusted is used, it's downgraded one category because so few motorists adjust their restraints. Many vehicle models have more than one seat option — if seat differences affect the head restraint rating, more than one rating is shown. This procedure is used to rate the head restraints in 1995-99 models.

A modification of the above procedure has become an international standard available from the Research Council for Automobile Repairs (RCAR). Ratings for fixed head restraints and adjustable restraints that don't lock is unchanged under the international RCAR protocol. For adjustable restraints that lock in position when adjusted, the rating is based on the midpoint of the best (highest and closest) and worst (lowest and farthest) positions in relation to an average-size male. Active head restraints that are designed to move closer to the backs of occupants' heads in rear-end crashes are not rated for geometry at this time. The Institute rates head restraints in 2000 and later models according to the RCAR procedure.

Dynamic ratings: Seat/head restraints with geometry rated good or acceptable (current and recent model cars) are tested in a simulated rear impact conducted on a sled to assess how well the seats support the torso, neck and head of a BioRID dummy. The test simulates a rear-end crash with a velocity change of 10 mph, approximately equivalent to a stationary vehicle being struck at 20 mph by a vehicle of the same weight.

A seat/head restraint's dynamic rating depends on performance in the sled test. There are two sets of criteria for evaluating performance. The first criteria are the two seat design parameters, time to head restraint contact (must be ≤70 ms to pass) and torso acceleration (must be ≤9.5 g to pass). The second set of evaluation criteria are the maximum neck shear force and maximum neck tension measured on BioRID during the test. These neck forces (classified low, moderate or high) indicate how well or how poorly an occupant's head and neck would be supported in a rear impact at low to moderate speed. A seat that passes at least one of the seat design parameters and has low neck forces earns a dynamic rating of good.

Overall ratings: Then the geometric rating and the dynamic rating are combined to produce a seat/head restraint combination's overall evaluation. Seats with head restraints rated marginal or poor for geometry aren't tested dynamically. They're assigned overall ratings of poor because of inadequate geometry.

Dummy and sled used in dynamic tests of seat/head restraints: Dynamic testing of seat/head restraints requires a dummy with a realistic spine and neck. Until the development of BioRID, or biofidelic rear impact dummy, existing dummies had rigid spines and necks that didn't interact with vehicle seats the way human spines and necks do. BioRID was developed for rear testing by a consortium of Chalmers University, restraint maker Autoliv, Saab and Volvo. This dummy, representing an average-size man, has a spine composed of 24 articulated vertebra-like pieces. The spine interacts with vehicle seats during tests in much the same way as a human spine. Plus BioRID's segmented neck can produce the motion observed by human necks in real-world crashes in which vehicles are struck from behind.

The device on which dynamic tests of seat/head restraints are conducted is a steel flatbed sled that runs on fixed rails. The sled is moved to simulate vehicle crash accelerations, recreating the forces on occupants inside vehicles during real-world crashes. The changing acceleration or deceleration over the time duration of a crash is referred to as a crash pulse, and the key aspect of a sled is that it can be programmed to produce specific crash pulses. To evaluate seat/head restraints, vehicle seats and their attached restraints are fixed to the sled, which is accelerated to simulate a stationary vehicle that's rear-ended by another vehicle of the same weight going 20 mph. To accomplish this, compressed air is pumped into a special cylinder, thrusting a ram forward in a pre-programmed pattern of acceleration (crash pulse). Peak acceleration in the sled test is 10 g (5 g mean acceleration), and the duration is 91 ms.