Today, two agencies conduct crash testing in the USA: the NHTSA and the IIHS. Both have a rating system, and both of their rating systems come with upsides and downsides. In this post, I look at frontal crash tests; the NHTSA full-frontal and the IIHS offset tests (moderate and small overlap).
The NHTSA has used a common 35-mph full-frontal crash test since 1979, with one major change occurring for the 2011 model year. In total, there have been four "generations" of grading for this test; three for the 1979-2010 version, and one for the 2011-present version.
The NHTSA has used a common 35-mph full-frontal crash test since 1979, with one major change occurring for the 2011 model year. In total, there have been four "generations" of grading for this test; three for the 1979-2010 version, and one for the 2011-present version.
The first two generations were based on a "pass/fail" system; if a vehicle had a HIC of less than 1000 and less than 60 Chest G's, it was considered to "pass". The system used from 1979-1981 was just that; "pass" or "fail". This system was obviously flawed, there was no way of knowing whether a vehicle "nearly passed" or failed miserably. That meant that consumers saw, for instance, the 1980 Ford Thunderbird and 1979 Peugeot 504 both as "fails", even though the Thunderbird failed because its driver HIC was 1,001, still easily survivable - or the 504, which had a driver HIC of 4,611, a definite fatality.
Sometimes, a vehicle that "passed" could have a higher injury risk than one that "failed"; for instance, the 1979 Oldsmobile Cutlass Supreme, with a 33% risk of severe injury to the driver, "passed" because its HIC of 998 and 58 Chest G's both narrowly cleared the thresholds, whereas the 1981 Ford Escort "failed", even though its passenger dummy's risk was only 23%, because it's passenger HIC was 1,010.
Sometimes, a vehicle that "passed" could have a higher injury risk than one that "failed"; for instance, the 1979 Oldsmobile Cutlass Supreme, with a 33% risk of severe injury to the driver, "passed" because its HIC of 998 and 58 Chest G's both narrowly cleared the thresholds, whereas the 1981 Ford Escort "failed", even though its passenger dummy's risk was only 23%, because it's passenger HIC was 1,010.
The next generation, from 1981-1993, was much improved as it actually gave the numbers. Now there was an incentive to get forces as low as possible; consumers would see the difference between a vehicle that "squeaked by" vs. one that had very low forces, and similarly, they could see the difference between a mediocre vehicle that just barely failed vs. a true death trap. However, the numbers were difficult to make sense of; how much HIC and Chest G's were required to cause X% injury risk? At what force level was fatal, and what force level would you walk away from?
The third generation, used from 1994-2010, used a "star" system. 5 stars meant <10% risk of severe injury, 4 star 11-20%, 3 star 21-35%, 2 star 36-45%, and 1 star 46-100%. This was a simple and effective method of communicating injury risk, though a "0 star" rating, perhaps anything above 80-90% injury risk, could have differentiated the death traps from the poor performers. Luckily, by 1994, there were very few vehicles that would have fallen in the 80%+ injury risk category for either occupant.
All three of these generations of NHTSA grading shared some common flaws. For one, it took into account only head and chest injury, completely ignoring the lower body. Second, it ignored structure and dummy kinematics. A vehicle could be structurally deficient and/or have a high risk of serious leg injury and still get a high rating. Although high femur loads (above 2,250 lbs) were noted by NHTSA, they weren't always mentioned in all rating releases and didn't affect the star rating.
The 2011 update to the NHTSA testing methodology takes into account head, neck, chest, and leg injury, combining it into an injury risk and a star band. The threshold has been lowered from "severe" to "serious" injury, and now 5 stars means <10% risk, 4 stars 10-15%, 3 stars 15-20%, 2 stars 20-40%, and 1 star 40% or more.
The IIHS uses a different rating methodology; in each test, they give a rating for structure, dummy kinematics, and four different categories of injury: head/neck, chest, and left and right legs (for small overlap tests, they use hip/thigh and lower leg/foot categories instead of the left and right leg categories). For each category, several different forces are measured: for instance, for the head/neck category, HIC, Nij, neck tension, and neck compression are measured. The worst value from these forces is used for the rating; for instance, to get a Good head/neck rating, the HIC, Nij, neck tension, and neck compression forces all have to fall in the Good range. If, for instance, the HIC was in the Marginal range but the other forces were all Good, the head/neck rating would drop to Marginal.
All six sub-ratings are then combined via this demerit system for moderate overlap and this one for small overlap (see page 16). Different things are weighted differently; for instance, a poor rating to a leg will only limit the overall rating to Acceptable, but if the head/neck or chest forces are Marginal or Poor, the rating can be no better than Marginal or Poor, respectively.
The IIHS rating system seems to be perfect - except for one thing: it doesn't take the overall risk of injury into account, like NHTSA's rating system does. Even the Good rating bands carry some risk of injury. For instance, 50 mm of chest compression is the maximum for a Good chest rating (although most modern vehicles are below 35) yet a chest compression of 50 mm carries a serious injury risk of over 25%, enough by itself to drop a vehicle into the 2 star range in the NHTSA system.
However, a vehicle could get an Acceptable chest rating and a Good overall rating (having used up only two of the maximum 3 demerits for a Good rating), barely meeting the Acceptable chest compression cut-off of 60 mm, which would have a serious injury risk of over 45% (which, for an older occupant, could be over 90%). Theoretically, a dummy in a vehicle could have all of its force measurements near the bad-end of the Good rating, and while each individual injury measurement may not seem that bad, the injury risk would really add up.
Luckily, in the real world, this tends not to happen; the highest injury risk I was able to calculate among any Good-rated vehicle was 32%, and almost all are under 20%, with most newer vehicles being under 13%.
Both rating systems do well for what they're intended for, but both could learn something from the other.
All three of these generations of NHTSA grading shared some common flaws. For one, it took into account only head and chest injury, completely ignoring the lower body. Second, it ignored structure and dummy kinematics. A vehicle could be structurally deficient and/or have a high risk of serious leg injury and still get a high rating. Although high femur loads (above 2,250 lbs) were noted by NHTSA, they weren't always mentioned in all rating releases and didn't affect the star rating.
The 2011 update to the NHTSA testing methodology takes into account head, neck, chest, and leg injury, combining it into an injury risk and a star band. The threshold has been lowered from "severe" to "serious" injury, and now 5 stars means <10% risk, 4 stars 10-15%, 3 stars 15-20%, 2 stars 20-40%, and 1 star 40% or more.
The IIHS uses a different rating methodology; in each test, they give a rating for structure, dummy kinematics, and four different categories of injury: head/neck, chest, and left and right legs (for small overlap tests, they use hip/thigh and lower leg/foot categories instead of the left and right leg categories). For each category, several different forces are measured: for instance, for the head/neck category, HIC, Nij, neck tension, and neck compression are measured. The worst value from these forces is used for the rating; for instance, to get a Good head/neck rating, the HIC, Nij, neck tension, and neck compression forces all have to fall in the Good range. If, for instance, the HIC was in the Marginal range but the other forces were all Good, the head/neck rating would drop to Marginal.
All six sub-ratings are then combined via this demerit system for moderate overlap and this one for small overlap (see page 16). Different things are weighted differently; for instance, a poor rating to a leg will only limit the overall rating to Acceptable, but if the head/neck or chest forces are Marginal or Poor, the rating can be no better than Marginal or Poor, respectively.
The IIHS rating system seems to be perfect - except for one thing: it doesn't take the overall risk of injury into account, like NHTSA's rating system does. Even the Good rating bands carry some risk of injury. For instance, 50 mm of chest compression is the maximum for a Good chest rating (although most modern vehicles are below 35) yet a chest compression of 50 mm carries a serious injury risk of over 25%, enough by itself to drop a vehicle into the 2 star range in the NHTSA system.
However, a vehicle could get an Acceptable chest rating and a Good overall rating (having used up only two of the maximum 3 demerits for a Good rating), barely meeting the Acceptable chest compression cut-off of 60 mm, which would have a serious injury risk of over 45% (which, for an older occupant, could be over 90%). Theoretically, a dummy in a vehicle could have all of its force measurements near the bad-end of the Good rating, and while each individual injury measurement may not seem that bad, the injury risk would really add up.
Luckily, in the real world, this tends not to happen; the highest injury risk I was able to calculate among any Good-rated vehicle was 32%, and almost all are under 20%, with most newer vehicles being under 13%.
Both rating systems do well for what they're intended for, but both could learn something from the other.
Comments
Post a Comment