Draft list of criteria as initially prepared by JASIC and modified during GTR 7-Phase II informal group discussions in Gothenburg.
5.a. Candidate Whiplash Injury Criteria
The sub-Group reviewed the Gothenburg candidate criteria list (WCWID-01-02).
KM noted that IVrot and NDCx were deleted in Gothenburg, due to the difficulty of instrumenting the dummy. The sub-Group noted that the NDCx has to be done from film analysis. Most labs can do this, but the problems are calibrating the camera system so that assessments at all labs are comparable. AI noted that SAE have tried to address this with many experts and have been unable to ensure reproducibility.
JD noted that the NDCrot value (12°) is well within the physiological range of most people, so not injurious. There must be a time component that says that this range happened too quickly.
BL reiterated that the goal has been for a long time to give the public protection from long-term impairments, loss of work, costs etc – i.e. the consequences of an injury or disorder.
The sub-Group discussed whether IV-NIC = 1.1, which gives a WAD2+ risk of 82.9% and AIS risk of 50%, should be used as a basis for deciding a threshold. It was noted that NHTSA are still looking at the appropriate statistical method, so the number could change.
KM noted that using both NDCrot and NIC captures components similar to V*C. It was noted that it is important to have criteria that address extension injury and flexion injury.
It was agreed that NDCx should be eliminated, due to the difficulty of making the measurement at all laboratories.
It was agreed that NIC should be used.
BL asked if NDCrot can be measured directly with angular rate sensors (ARS) with sufficient reliability. KM noted that new ARS are much more reliable than older designs and are suitable. NHTSA have done a lot of work on these for BrIC and there is a test procedure in the NPRM for the Q3S that uses ARS and the NPRM includes a specification for the ARS.
There are at least two manufacturers.
It was agreed that NDCrot should be used, using appropriately specified angular rate sensors.
JD presented some updates and clarifications from yesterday’s presentation:
Permanent Medical Impairment (PMI) classification requires multiple medical assessments over a period of 1-3 years. People with PMI classification typically do not return to their previous work.
Action JD to provide more details on the PMI classification.
KM noted that the data is from older tests, where the flexion bumpers were not as well controlled as they are today. Need to do some fleet assessment tests with the current
dummy. JD noted that the EEVC results should be considered preliminary, due to the reliance on data from previous versions of the dummy. However, based on the data the NDCrot indicates that no seats will fail, other than WIL. KM noted that JNCAP have found a much wider range, with 50% failing the proposed threshold.
JD showed EEVC data for NIC, showing that the majority of the seats evaluated would not pass. However, the results may be somewhat different for the current version of the dummy and some of the seats that would fail have already been removed from the market. JD noted that at Gothenburg, EEVC recommendation on a NIC threshold was pragmatic – based on eliminating older seat designs with poor real-world performance.
KM asked what criteria would be recommended by the EEVC work. JD noted the following:
Lex van Rooij (LvR) presented all the 2010-2011 NIC results from Euro NCAP, for all three pulses. Visually, about 80% of the seats were below NIC 15 and 95% below NIC 20.
Action LvR to provide the Euro NCAP NIC graphs for 2010-2011.
KO presented the 2009-2011 NDCrot and NIC JNCAP data. It was noted that NDCrot relates well to the JNCAP rating. BL cautioned that there still needs to be a link between the rating and the real-world performance.
BL noted that NIC in Euro NCAP is only assessed up to head restraint contact. YSK noted that NIC peaked before head restraint contact in the VRTC tests. KB and NP agreed with the use of NIC. AK noted that the correlation was not high in the VRTC tests. Can it be judged that the NIC can discriminate between good and bad real-world performance? It seems okay at the extremes (good and poor), but maybe not in the intermediate range. BL: This is a problem for the consumer testing, but not for regulation. We need to be sure that we have a positive effect in the field – i.e. provide people with protection to a minimum level decided at the political level, and don’t guide seat design in the wrong direction. The latter is very important for regulation, because it can take a long time to remove something from legislation. PP noted that it is important not to take excessive budget from other safety features that relate to more serious injuries, which can be achieved by making the requirements reasonable for low-cost cars.
BL noted that the GTR-7 still has the dynamic test as an option – if a seat passes the static geometry requirements then a dynamic test is not necessary. This is still the case in FMVSS 202a. Something different is conceivable for adoption in a UN Regulation.
JD noted that there is only one, no longer produced seat that has good real-world performance and a high NIC. KM identified that it would be important to review the seat and the test data and see what is the reason for this and whether it would still apply if it was tested today.
Action JD to check this seat, and also the high NDCrot with one modern anti-whiplash seat design. Data may be available from other laboratories for these seats which may help clarify the results.
It was agreed that upper and lower neck Fz should be deleted.
Upper and lower neck My(flexion/extension) were discussed. BL asked whether the NDCrot and upper neck My(flexion/extension) were well correlated, such that it would not be necessary to have both. NDCrot well correlated (R2=0.80) with upper neck My(flexion), for pre-contact and contact phases. Poor correlation for upper neck My(extension) (R2=0.42).
For the lower neck, there was no correlation (R2=0.01) for My(flexion) and poor correlation for My extension (R2=0.2). However, the NDCrot value was for flexion only, and it does not make sense for the NDCrot(flexion) to correlate with My(flexion).
BL: Do we need both upper and lower neck loads? TF: At the upper neck, the correlation is good with NDCrot, so duplicate parameter, and injuries tend to be identified in the lower neck.
There was no information on the correlation between NDCrot(extension) and upper or lower neck My(extension), and probably won’t get this from JNCAP data because most seats don’t
allow large extension any more. KM noted that this data will be available from the VRTC fleet analysis, so it is possible that this could be updated soon.
There was no information to indicate that upper neck Fx would be duplicated by NDCrot or any other criteria already selected. KM noted that there isn’t an obvious reason why they would be correlated.
AK asked if the neck shears were related to injury. KO presented correlations between upper and lower neck Fx and symptoms, and KM noted that the Fx was well correlated (approx. R2 0.75) with injury, where Fx was estimated from inverse dynamics only up until head restraint contact.
AK asked if the neck shears were correlated with NIC, such that Fx would be a duplicate requirement. LvR presented 2010-2011 Euro NCAP data and there was no correlation. BL noted that the Euro NCAP lower performance limit is 30 N for upper neck Fx, which is less than the force on the due to gravity when you get out of bed –with such low forces it is maybe no surprise that there is no correlation. Additionally, the limits were derived pragmatically from tests with 30 seats at one laboratory, and the head restraints then were not as high as they are now, so some of the heads would wrap over the head restraint – giving low Fx and high Fz compared to many modern seats.
KO presented JNCAP data showing no correlation between NDCr(flexion or extension) and upper neck Fx(flexion or extension. and no correlation between NDCrot(extension) and upper neck My(extension). In most tests, NDCrot(extension) was zero. There was no correlation between NDCrot(extension) and lower neck My(extension), so no reason to delete lower neck My(extension).
KM noted that the recommendation for My and Fx comes from Japan. KO noted that it comes from the volunteer and accident reconstruction simulation studies, because it correlates with the strains that are directly related to the injury. NDCrot doesn’t capture injuries that are due to high strains during the s-shaped phase. Upper neck and lower neck My more related to global motion, so duplicated by NDCrot, but the shear forces are necessary. KM noted that if IVrot were used, Fx may not be required, but we have excluded IVrot so we need Fx to control loading during the s-shape. KO: Yes.
BL asked if legislation would misdirect head restraint design if upper neck My was deleted? It was agreed that it would not. It was agreed that upper neck My should be deleted provided that NDCrot has requirements for both flexion and extension.
BL asked the same question for lower neck moment. It could be possible to review the data with new knowledge in some years’ time. It was agreed that lower neck My should be
deleted provided that NDCrot has requirements for both flexion and extension.
BL asked the same question for upper and lower neck Fx. KM noted that both are probably required. KO considered that these are required in order to fully assess loading during the s-shaped phase. DH noted that he would be uncomfortable with a proposal that did not adequately assess loading during the s-shaped phase, because in some cases the strains have been shown to be high in this phase, even with modest head rotation relative to T1.
Deleting the neck forces and moments entirely could leave a gap in the assessment and therefore the safety. NP noted that it is not clear that the shear forces can be deleted safely.
KM noted that if the intervertebral kinematics were used, the shear would not be required, but given that NDCrot is a global assessment the Fx is still required. BL reminded that the Fx were important in the Japan volunteer and accident reconstruction simulations. PP asked whether both the upper and lower neck Fx were required. KM noted that the lower neck Fx would be expected to be more important in the pre-contact phase, and the upper neck Fx more important in the head restraint contact phase. PP noted that he understood the discomfort with removing the Fx, but it could be very sensitive to the initial positioning of the dummy. There should be some information on this from the upcoming VRTC fleet study.
It was reiterated that this is consideration of the measurement as an injury criterion, not a consideration of whether the dummy measures this reliably.
AK asked whether there was a correlation between upper neck Fx and lower neck Fx in the JARI modelling.
Action KO to check whether the upper and lower neck shear forces are correlated in the JARI accident reconstruction simulations.
It was agreed that upper and lower neck Fx should be used.