CR 173: Children in Adult Seat Belts and Child Harnesses; Crash Laboratory Comparisons of Dummy Responses (1997)

Rationale for project

The effectiveness of adult belt systems when used by children has long been an issue of debate, because of the incompatibility of the size and shape of the typical child with the geometry of the typical seat-belt installation. Concerns centre on the crash protection offered by these systems and on the possibility of increased risk and severity of belt-induced injuries.

Many children are still restrained in adult belts alone, even though seat-belt restraint is not optimal for small occupants. Studies of the effects of adult belts on child injury reduction and injury patterns are rare. Until very recently there have not been available sufficiently biofidelic child dummies to attempt crash simulation studies. However, a 18-month-old CRABI dummy is now available, as is an early model of the new Hybrid III six-year-old dummy. Further, the developmental Hybrid III three-year-old dummy became available in Australia for a limited period in March 1996.

Accordingly, a test program was designed to supplement the field observations made during a recent Australian field study by investigating the responses of the above dummies when restrained in adult lap/sash, lap-only and child harness belt systems.

18-month CRABI dummy

Both the lap belt and the lap/sash belt allowed the 18-month CRABI dummy excessive excursion, and as might be expected the kinematics were far from satisfactory. Nevertheless, the CRABI 18-month dummy was restrained by the lap/sash adult belt surprisingly well, although the dummy still showed considerable forwards rotation despite the upper torso remaining constrained by the shoulder belt. Even the lap/sash belt allowed the dummys head to contact the lower part of its legs.

When restrained by a lap belt, the dummy rotated much further forwards, so that the belt moved down on to the upper surface of the thighs. This allowed considerable excursion and the dummy's head impacted the front of the seat, including the wooden frame supporting the 156 mm deep cushion. The result was a high resultant head acceleration and HIC value.

These head contacts complicated the analysis of head and neck responses. In one lap-belted run, the head contact produced very high outputs for head acceleration, HIC, and neck moment. It is therefore difficult to determine any general difference, in terms of head accelerations and neck loads, between the two configurations of seat belt restraint. It was the head contact that determined the overall outcome.

Significant differences in terms of dummy response were apparent in the chest acceleration, lumbar load and pelvic acceleration responses. Lap-only belts produced higher lumbar loads and pelvic accelerations. Lap/sash belts produced greater chest accelerations.

Further, lap belt loads were much higher in the lap-only configuration, as would be expected. The lap-only belt showed lap belt webbing loads that were about double the loads in the lap portion of the lap/sash belt.

3-year old Hybrid III dummy

For the 3-year old Hybrid III dummy, the lap belt was the only restraint observed to allow definite head contact. As a result, the observed high head acceleration and neck loads may not be typical of the real world crash forces affecting these regions in the absence of head contact.

Except for axial tensile loads, which were highest in the lap belt, the lap/harness system produced the highest outputs for all neck force and moment measurements. The harness system also produced the greatest chest acceleration and higher pelvic accelerations than the lap/sash belt. The lap-only belt produced pelvic accelerations that were comparable to the harness system. The lap/sash system resulted in the lowest outputs generally.

All three restraints held the 3-year old Hybrid III in place during the entire crash sequence. The lap belt allowed excursion of the torso, although not to the extent observed with the 18-month CRABI. There was also some evidence of the 3 year old Hybrid III submarining to some extent under the lap belt, as in one test the lap belt ripped the "skin" of the dummy in the abdominal region.

6-year old Hybrid III dummy

All three restraint systems held the 6-year old Hybrid III dummy in place during the entire crash sequence. However, the lap belt did allow a large amount of forward excursion of the upper torso and head.

In general, the child harness system produced greater head accelerations, neck loads (in particular forward shear and compressive loads) and chest accelerations than the other restraint systems. However this restraint system produced the lowest pelvic accelerations.

Head accelerations and HIC were higher with the lap belt, because of head strike. Neck rearward shear was higher with the lap belt, and forward neck shear higher with the lap/sash belt. The average axial tensile loads were slightly higher with the lap belt than the lap/sash system, but contrary to the values for the three-year-old dummy the difference in this case is small.

As with the other dummies, loads in the lap-only belt were nearly twice as high in the lap portion of a lap/sash belt.


The sled test data for the three dummies showed mixed results for neck shear, axial tension and bending moments. Except for axial tensile forces in the two larger dummies and neck moments in the 18-month CRABI, the tendency was for the lap/sash system to result in rather higher readings than the lap-only belt. However, the lap/sash system, as well as minimising dummy head and upper torso excursion, was effective in minimising head acceleration and pelvic accelerations.

Head accelerations, HIC, chest accelerations and lap belt loads were higher with the lap belt alone than with the lap/sash belt. The absence of upper torso restraint in the lap-only system allowed excessive excursion of the Hybrid III 3-year old and 6-year old, but it did minimise dummy neck and chest response.

There was a tendency for neck forces to be highest in runs with the lap/harness system. Even head accelerations and HIC were high in the lap/harness system, in the absence of the head contacts that affected the lap-belt runs. Probably because the shoulder straps load the centre of the lap belt in this configuration, lap belt loads were higher and submarining is more likely than in a lap/sash belt.

Generally, the values for force and bending moment in this sled test series were high in comparison with previous similar research. However, in the field data collected in the Australian field study there were 19 children aged two years to 14 who were restrained in lap/sash belts in generally frontal crashes at a calculated delta-V of 45 km/h or over. More than two-thirds of these children (13) were in frontal crashes of 65 km/h or over. It is probable, therefore, that all these children were exposed to forces of the same order of magnitude that we found in our series of sled runs, generally above the tolerance criteria suggested for guidance by other workers.

Yet, among these children in the real world there was only one neck injury that was not AIS 1 or 2. This AIS 6 (fatal) injury in a three-year-old directly resulted from a heavy head contact with the windshield. The other neck injuries were all soft tissue injuries commonly associated with bruising and abrasions from belt loading.

In European crash reconstructions, neck tensile (Fz) forces of over 2.5 kN have been recorded in laboratory reconstructions, yet no neck injuries had been sustained in the real crashes. This relationship - high loads measured in laboratory, yet little or no injury in a real crash of equivalent severity - is consistent with our observations.


In summary, accepting some inconsistencies in the results from dummy to dummy, the results are in accord with the field data: broadly, that in return for a greatly reduced risk of head and abdominal injury, a lap/sash belt may present a slightly higher risk than a lap belt of minor inertial neck injury, equivalent to AIS 1 or 2. However, there is nothing in this set of sled test results to indicate that adding a sash belt to a lap belt places a child at a higher risk of serious neck injury.

There are many more head injuries than neck injuries in the data from field studies. Lap-belt-induced injury of the abdominal organs and lumbar spine are also far more common than inertial injuries to the cervical spine. In the development of design or performance criteria, for the minimisation of the risk of cervical spine injury it is important not to unreasonably raise the risk of other serious injuries, such as those resulting from head and chest impact.

The results indicate that the simple addition of a harness system to a lap belt, although reducing excursion of the head and torso, may lead to neck forces that are much higher than those seen in the lap/sash belt configuration. There are therefore grounds for concern on the performance of the lap belt/child harness configuration. The comparatively high readings for head and neck forces and accelerations indicate the need for some attention to design. Unlike the configuration in a forward-facing child seat, the shoulder straps of the child harness when used with a lap belt are anchored directly to the vehicle structure. In a child seat, the harness is attached to the seat and the seat attachments are separate. The lap/harness configuration is therefore much stiffer in its reaction to crash forces, and this may be why the dummy neck loads were higher in these tests. This suggestion is supported by outputs for chest accelerations, which were also highest in the lap/harness configuration.

In addition, lap-belt loads were high in this configuration because the lap belt is loaded at its centre by the shoulder straps. This also raises the risk of "submarining" under the lap belt.

If a child is to use an adult seat belt, the best alternative is the lap/sash configuration. The case is very strong for encouraging, or compelling, the fitting of lap/sash seat belts in the centre seat positions of all passenger cars where practicable.

Type: Research and Analysis Report

Sub Type: Consultant Report

Author(s): M Henderson

Topics: Child, Occ protection, Seat belts

Publication Date: 01/10/97