Department of Transportation
Department of Transportation |
|
| Preliminary Regulatory Impact Analysis | |
FMVSS 216, UPGRADE
ROOF CRUSH RESISTANCE
Office of Regulatory Analysis and Evaluation
National Center for Statistics and Analysis
August 2005
TABLE OF CONTENTS
APPENDIX A
This Preliminary Regulatory Impact Analysis examines the potential impacts of new performance requirements for passenger car and light truck roof strength. The intent of this proposal is to improve occupant protection in rollover crashes that involve roof crush.
Test Requirements
The agency is proposing to modify the test procedures in FMVSS No. 216 to require that vehicles be tested with the application of a force loading device up to 2.5 times the vehicle’s weight without the roof crushing to a level where it touches the head of a seated 50th percentile male dummy. This represents a change from the current requirement, which specifies a test load of only 1.5 times the vehicle’s weight without the device moving more than 127 millimeters (5 inches). The agency also examined an alternative proposal that would require a 3.0 load requirement.
Countermeasures
The agency believes that manufacturers will meet this standard by strengthening reinforcements in roof pillars, by increasing the gauge of steel used in roofs, or by using higher strength materials. The agency estimates that about 35 percent of all current passenger car and light truck models will require changes to meet the 2.5 vehicle load proposal, and that 75 percent would require changes to meet the 3.0 vehicle load alternative.
Benefits
The agency estimates that the proposed changes in FMVSS No. 216 requiring a load resistance of 2.5 times the vehicle weight would prevent from 13-44 fatalities and from 498 to793 nonfatal injuries. The range in these estimates reflects different methodological approaches that were employed to examine the injury profile of occupants in rollover crashes where the roof collapsed into the occupant compartment. The agency estimates that the alternative 3.0 times the vehicle weight load requirement would prevent from 49-135 fatalities and from 1,540-2,151 nonfatal injuries.
Costs
The design changes made to comply with higher test load requirements will add both cost and weight to the vehicle. This will increase the initial purchase price and reduce fuel efficiency, which will increase lifetime fuel usage costs. The agency estimates that compliance with the proposed 2.5 times the vehicle weight load requirement will increase vehicle costs by $10.67 per affected vehicle. Added weight is estimated to increase the lifetime cost of fuel usage by $5.33 to $6.69. The range in fuel costs reflects different discount rate assumptions of 7% and 3% respectively. The agency estimates that compliance with the proposed 3.0 times the vehicle weight load requirement will increase vehicle costs by $51.27 per affected vehicle. Added weight is estimated to increase the lifetime cost of fuel usage by $29.41 to $36.70 for the 3.0 times the vehicle weight load requirement alternative. The total annual consumer cost for the 2.5 times the vehicle weight load requirement proposal is from $88 to $95 million. The total annual consumer cost for the 3.0 times the vehicle weight load requirement proposal is from $1.2 billion to $1.3 billion.
Cost Effectiveness and Net Benefits
Cost effectiveness is a measure of the economic investment that is required to prevent a fatality. The cost effectiveness of this proposal was estimated under both 3% and 7% discount rate assumptions for each proposal. Nonfatal injuries were translated into fatality equivalents based on comprehensive valuations that included both economic impacts and valuations of lost quality of life. To reflect the present value of benefits that would be experienced over the vehicle’s useful life, the resulting equivalent fatalities were discounted over the vehicle’s life based on annual exposure to crash involvement as measured by annual miles traveled. The results indicate that the 2.5 times the vehicle weight load requirement proposal would cost from $2.1 million to $3.4 million per equivalent life saved. The 3.0 times the vehicle weight load requirement proposal would cost from $9.0 million to $14.4 million per equivalent life saved. The higher cost of the 3.0 times the vehicle weight load requirement alternative reflects the larger portion of the fleet that would have to be changed to meet the standard, as well as the greater cost and weight of design changes that would be required to achieve this level of roof strength.
The equivalent fatalities prevented are 39-55 fatalities for the 2.5 load factor and 122-171 fatalities for the 3.0 load factor. At the 3% discount rate this translates into 32-46 fatalities prevented for the 2.5 load factor and 100-140 fatalities prevented for the 3.0 load factor. At the 7% discount rate this translates into 26-37 fatalities prevented for the 2.5 load factor and 80-113 fatalities prevented for the 3.0 load factor.
Net benefits represent the difference between total costs and the total monetary value of benefits. The monetary value of benefits was estimated by assigning a value of $3.5 million to each equivalent fatality prevented. The 2.5 times the vehicle weight load requirement alternative would produce net benefits of from $3.0 million to $64.1 million. The 3.0 times the vehicle weight load factor alternative would result in net costs of $761 million to $909 million.
A. Current Requirements
FMVSS 216, "Roof crush resistance," became effective on September 1, 1973. This standard established strength requirements for the roof structure over the front occupants of passenger cars with a Gross Vehicle Weight Rating (GVWR) of 6,000 pounds or less. The purpose of the standard is to reduce deaths and injuries due to crushing of the roof into the passenger compartment area in rollover crashes. Since its inception, the roof crush standard has been amended, extending its requirements to passenger cars, trucks, buses, and multipurpose passenger vehicles (MPVs) with a GVWR of 2,722 kilograms (6,000 pounds) or less (55 FR 15510, April 17, 1991). The standard was also amended to modify the test device placement procedure to accommodate vehicles with raised and highly sloped (aerodynamic) roof structures (64 FR 22567, April 27, 1999).
The test procedure currently used to evaluate compliance with the standard involves securing a vehicle on a rigid horizontal surface, placing a flat steel rectangular plate on the vehicle’s roof, and using the plate to apply 1.5 times the unloaded weight of the vehicle (up to a maximum of 22,240 N, or 5,000 pounds, for passenger cars) onto the roof structure. During the test, the plate is angled and positioned to simulate vehicle-to-ground contact on the roof over the front seat area. [1] To achieve this contact, the plate is tilted forward at a 5-degree angle, along its longitudinal axis, and tilted outward at a 25-degree angle, along it’s lateral axis, so that the plate’s outboard side is lower than its inboard side. The test plate’s edges are also positioned with respect to fixed locations on the vehicle’s roof, depending upon the roof slope, to ensure that the plate stresses the roof over the front seat area. Compliance with the standard is achieved if the vehicle’s roof prevents the test plate from moving downward more than 127 mm (5 inches).
B. Previous Agency Rulemaking
1. Rollovers
In 1991, Congress mandated NHTSA to assess rulemaking on rollover occupant protection as a part of the Intermodal Surface Transportation Efficiency Act (ISTEA). ISTEA required the agency to initiate rulemaking to address the problems of rollover crashes. In response to that mandate, NHTSA published an advance notice of proposed rulemaking (ANPRM) (57 FR 242, January 3, 1991) that summarized the statistics and research in rollover crashes, sought answers to several questions about vehicle stability and rollover crashes, and outlined possible regulatory and other approaches to reduce rollover fatalities. NHTSA also published a report to Congress that detailed the agency’s effort in these areas (NHTSA 1999-5572-35).
The agency released a document entitled, "Planning Document for Rollover Prevention and Injury Mitigation," at a Society of Automotive Engineers (SAE) meeting on rollover on September 23, 1992. The planning document gave an overview of the rollover problem and a list of alternative actions that NHTSA was examining to address the problem. Activities described in the document included: crash avoidance research on vehicle measures for rollover resistance, research on antilock brake effectiveness, rulemaking on upper interior padding to prevent head injury, research on improved roof crush resistance to prevent head and spinal injury, research on improved side window glazing and door latches to prevent occupant ejection, and consumer information to alert people to the severity of rollover crashes and the benefits of seat belt use in this type of crash. NHTSA published a notice announcing the availability of the planning document and requesting comments (57 FR 44721, September 29, 1992).
In May 1996, NHTSA issued the "Status Report for Rollover Prevention and Injury Mitigation" (NHTSA 1996-1811-2). This document updated the progress of the programs discussed in the planning document. Under section 12 of the Transportation Recall, Enhancement, Accountability and Documentation (TREAD) Act of November 2000, NHTSA was mandated to develop a dynamic rollover test for the purposes of consumer information, to carry out a program of conducting such tests, and to conduct rulemaking to determine how best to disseminate test results to the public, as these tests are being developed. On July 3, 2001, the agency published a Request For Comment notice (66 FR 35179) discussing potential advantages and disadvantages of a variety of dynamic rollover tests that were selected to be evaluated in the agency’s research program. After a subsequent comment period, NHTSA published a NPRM on October 7, 2002, discussing the results of NHTSA's evaluation of numerous driving maneuver tests for the dynamic rollover consumer information program. The NPRM also proposed several alternative methods for using the dynamic rollover test results in the agency's consumer information for vehicle rollover resistance. NHTSA is now using the dynamic rollover testing to supplement the existing static ratings in the 2004 model year.
2. Roof Crush
On April 17, 1991, NHTSA published a final rule amending FMVSS 216 to extend its requirements to MPVs, trucks, and buses with a GVWR of 2,722 kilograms (6,000 pounds) or less (56 FR 15510). NHTSA justified the extension to light trucks by the increased use of light trucks [2] as passenger vehicles and the need to ensure that those vehicles offer safety protection comparable to that offered to passenger cars. The final rule adopted the same test requirement and procedure as those for passenger cars, except for the absence of the 22,240 Newton limit on the applied force. This amendment became effective on September 1, 1994.
On May 6, 1996, the agency received a petition for rulemaking from R. Ben Hogan, Smith and Alspaugh, P.C. (Hogan). Hogan commented that the current static requirements in FMVSS 216 bear no relationship to real world rollover crash conditions and therefore should be replaced with a more realistic test such as the inverted vehicle drop test defined in the SAE Standard J996. This request coincided with agency research testing that was being conducted using the inverted drop test procedure. The petitioner also requested that NHTSA require "roll cages" to be standard in all cars as requested by some commenters responding to the January 3, 1992, ANPRM on rollover occupant protection. NHTSA granted this petition on January 8, 1997, believing that the inverted drop test had merit for further agency consideration.
On April 27, 1999, NHTSA published a final rule regarding the test procedure in FMVSS 216 (64 FR 22567). Prior to the amendments made by the final rule, the existing procedure resulted in certain vehicles with rounded roofs (e.g., the Ford Taurus) being tested with the test plate positioned too far rearward on the vehicle roof. In this position, the plate did not test the roof over the front occupants. In addition, this position created the potential for contact between the front edge of the test plate and the roof, allowing the plate to penetrate the roof along the leading edge of the plate. Similarly, in following this procedure for vehicles with raised, irregularly-shaped roofs (such as some vans with roof conversions), the initial contact point on the roof may not be above the front occupants, but on the raised rear portion of the roof, behind those occupants. In both of these cases, the positioning of the plate relative to the initial contact point on the roof, instead of relative to a fixed location on the roof, resulted in too much variability in the plate positioning and reduced test repeatability.
The April 27, 1999 final rule addressed the problem of rounded roofs by specifying a new primary test procedure for all vehicles except those with certain modified roof configurations. Under the new procedure, the test plate is positioned so that the front edge of the plate is 254 mm (10 inches) in front of the most forward point of the roof. Positioned in this way, the front edge of the plate is always projected slightly forward of the roof instead of contacting it. The rule addressed the problem for vehicles with raised or modified roofs by specifying a secondary test procedure if the initial point of contact is rearward of the front seat area. Under the secondary test procedure, the plate is moved forward such that its rearward edge is positioned at the rear of the roof over the front seat area.
In June 1999, RVIA and Ford submitted a petition for reconsideration regarding the rear edge plate contact on certain aerodynamically shaped roof vehicles with the secondary test procedure. To provide temporary relief, vehicle manufacturers, until October 25, 2000, had the option of using the standard's original test plate placement procedure (1973) for light trucks that have a raised or altered roof, in place of the primary or secondary procedures defined above (65 FR 4579, January 31, 2000). The original procedure positioned the plate with respect to its initial point of contact with the roof. The initial point of contact was established by angling and lowering the plate as required, following the primary test procedure until the plate contacts the roof. After establishing the initial contact point on the vehicle, the test plate was moved upwards, and the forward edge was positioned 254 mm (10 inches) forward of the initial point of contact with the vehicle. This position was allowed to make testing possible for raised roof vehicles that experience contact with the plate's rearward edge when testing to the secondary test procedure.
On October 22, 2001, NHTSA published a Request for Comment notice (66 FR 53376) to assist in the upgrade of FMVSS 216. In the Request for Comment, the agency discussed issues regarding the current standard and posed eight questions grouped according to the following areas: (1) current test procedures; (2) alternative dynamic tests; and (3) limiting headroom reduction. The agency has received numerous comments from the public.
In response to comments received in the October 2001 FMVSS 216 Request for Comments, NHTSA initiated a review of National Accident Sampling System – Crashworthiness Data System (NASS-CDS) investigated crashes The study was intended to evaluate if there were significant changes observed in the patterns of roof damage in the current fleet and to compare real-world roof damage to both compliance and extended FMVSS 216 tests. [3] This study examined rollover cases with greater than 15.2 cm (6 inches) of vertical intrusion contained in the most recent five years of NASS-CDS (1997 to 2001). The agency evaluated the damage to the A and B pillars, roof rails and roof plane of the vehicles. All the examined vehicles had considerably greater damage than FMVSS 216 compliance tests. The NASS-CDS case review revealed that lateral roof deformation that was limited to one side of the vehicle’s roof, happened frequently.
After the NASS-CDS case review, a finite element study was initiated to examine the effect of using alternative roll and pitch angles for the current FMVSS 216 test procedure. A finite element model of a 1997 Dodge Caravan was used to simulate extended FMVSS 216 tests for approximately five inches of plate motion using a variety of roll and pitch angles. The simulations indicated that the Caravan roof would attain similar amounts of deformation at a lower force level using 10-degree pitch and 45-degree (10-45) roll angles compared to the current 5-degree pitch and 25-degree (5-25) roll angles. A 1997 Chevrolet S10 pickup model was also studied, but the results were less conclusive.
The results of the finite element study were encouraging enough to conduct a series of modified FMVSS 216 tests. Two tests were conducted on the Dodge Caravan and Chevrolet S10 vehicles using both the current 5-25 roll angles as well as using modified 10-45 roll angles. A third vehicle, a 2002 Ford Explorer, was also tested using the same test configurations. Each test was conducted until 254 mm (10 inches) of load plate displacement was achieved.
The roof damage between the two test configurations was generally similar. The tests using 10-45 roll angles had some additional lateral damage, but the damage was localized near the roof side rail and did not extend laterally to the midline of the vehicle. The force distribution applied to the front and back of the load plate changed considerably between the two test configurations. The test configuration using the 10-45 roll angles applied almost all of the force to the forward ram located near the front of the load plate. The 5-25 configuration applied only two-thirds of the force to the front ram. Based on the similarity of the post-test damage patterns and general force levels, there was not sufficient evidence to justify changing the load plate configuration. Thus, the agency proposed to keep using the 5-25 roll angles in future FMVSS 216 upgrade testing.
| Vehicle | Ram Travel @150% |
Cos 25o | Test% Weight |
Ram Travel @250% |
Cos 25o |
Test % Weight |
Ram Travel @300% |
Cos 25o | |||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2002 Toyota Camry | 25 mm | 1.0 in | 0.9 in | 250 | 50 mm | 2.0 in | 1.8 in | 300 | 60 mm | 2.3 in | 2.1 in |
| 2002 Honda CRV | 33 mm | 1.3 in | 1.2 in | 250 | 112 mm | 4.4 in | 4.0 in | 300 | 140 mm | 5.5 in | 5.0 in |
| 1997 Dodge Grand Caravan | 38 mm | 1.5 in | 1.4 in | 250 | 85 mm | .3.3 in | 2.9 in | 265 | |||
| 2001 Chevy Tahoe | 40 mm | 1.6 in | 1.4 in | 250 | 65 mm | 2.6 in | 2.3 in | 290 | |||
| 1998 Chevy S-10 PU | 42 mm | 1.6 in | 1.5 in | 250 | 65 mm | 2.6 in | 2.3 in | 276 | |||
| 2002 Ford Mustang | 39 mm | 1.5in | 1.4 in | 250 | 115 mm | 4.5 in | 4.1 in | 265 | |||
| 2002 Ford Crown Vic | 49 mm | 1.9 in | 1.7 in | 250 | 190 mm | 7.5 in | 6.8 in | 300 | 245 mm | 9.6 in | 8.6 in |
| 2002 Ford Explorer | 43 mm | 1.6 in | 1.5 in | 250 | 215 mm | 8.5 in | 7.7 in | 300 | 255 mm | 10.0 in | 9.0 in |
| 2002 Dodge Ram 1500 PU | 45 mm | 1.7 in | 1.6 in | 249 - Fail | 100 mm | 3.9 in | 3.5 in | ||||
| 1999 Ford E150 Van | 50 mm | 2.0 in | 1.8 in | 188 - Fail | |||||||
| Indicates Pass at 300% | Indicates Failure at 300 % | Indicates Failure at 250% | |||||||||
A set of ten recent model vehicles, were tested using an extended FMVSS 216 procedure to gather further roof strength data. These vehicles were tested to 25.4 cm (10 inches) of load plate displacement. These vehicles included the following: 2002 Dodge Ram 1500 Pickup, 2002 Toyota Camry, 2002 Ford Mustang, 2002 Honda CRV, 2002 Ford Explorer, 2002 Ford Crown Victoria, 2001 Chevy Tahoe, 1999 Ford E-150, 1998 Chevy S10 Pickup, 1997 Dodge Grand Caravan. The test results are shown in Table II-1.
Additional instrumentation was included in all of the extended tests to measure the motion of the roof directly above the head of the 50th percentile dummy. Three string potentiometers were connected from the roof to known locations near the floor of the vehicle. These displacement measurements enabled the tracking of the motion of a single point on the roof during the test. These measurements were intended to provide some measure of the relationship between external plate displacement and available occupant headroom throughout the test. All of the extended FMVSS 216 tests used the seating procedure from the FMVSS 208 test procedure.
All ten vehicles in Table II-1 attained the FMVSS 216 required applied force of 1.5 times the unloaded vehicle weight within 5.1 cm (2 inches) of plate displacement. Nine of the ten vehicles exceeded an applied force of 2.0 times the unloaded vehicle weight within 7.6 cm (3 inches) of plate displacement. The Ford E-150 van never exceeded an applied force of 2.0 times the unloaded vehicle weight during the test. Six of the ten vehicles tested (1998 Chevy S10 Pickup, 2002 Toyota Camry, 2002 Honda CRV, 1997 Dodge Grand Caravan, 2002 Ford Mustang, and 2001 Chevy Tahoe) exceeded an applied force of 2.5 times the unloaded vehicle weight within the 127 mm (5 inches) of plate displacement allowed in the current standard.
The roof displacement measurements were combined with the initial occupant seating position to determine the initial occupant headroom. For the 10 vehicles tested in the research program, the initial headroom from the occupants head to the hard roof, ranged from 98 mm (3.9 inches) for the Ford Mustang to 254 mm (10 inches) for the Ford E-150. The depth of the roof liner varied considerably among the vehicles tested. The Ford Mustang had the smallest distance from the roof liner to the hard roof of 8 mm (0.3 inches), and the Ford E-150 van had the largest distance of 61 mm (2.4 inches).
String potentiometer measurements were used to track the 3-dimensional motion of the roof location directly above the occupant’s head. The post-test location of the roof attachment point was unlikely to be directly above the occupants head, however the downward displacement of this roof point is expected to provide an indication of the occupant headroom throughout the roof crush test. At an applied force of 1.5 times the vehicle weight, the occupant head-to-hard roof clearance ranged from 95 mm (3.7 inches) to 252 mm (9.9 inches). All of the test vehicles had greater than 100 mm of head-to-hard roof clearance except for the Ford Mustang. At 2.0 times the vehicle weight, the head-to-hard roof clearance ranged from 68 mm (2.7 inches) to 181 mm (7.1 inches) for the 9 vehicles that reached this force level. At 2.5 times the vehicle weight, the head-to-hard roof clearance ranged from 14 mm (0.5 inches) to 176 mm (6.9 inches) for the eight vehicles that reached this force level during the extended testing. If the liner depths did not change as the roof crushed, then at an applied force of 2.5 times the vehicle weight, six vehicles still had over 25 mm (1 inch) of head to roof liner clearance and 5 vehicles had over 75 mm (3 inches) of head-to-roof liner clearance. Even at 3.0 times the vehicle weight, two of the vehicles (Toyota Camry and Honda CRV) still had over 75 mm (3 inches) of head-to-roof linear clearance. While the string potentiometers were not an exact measurement of occupant headroom, they did provide useful insights into the performance tradeoffs that can exist between roof strength and occupant headroom.
After completing the initial research program, NHTSA conducted ten additional extended FMVSS 216 tests with a 50th percentile Hybrid III dummy to better assess occupant headroom as a function of applied force, by determining when the roof contacted the dummy’s head. The following vehicles were tested to 25.4 cm (10 inches) of load plate displacement: 2003 Ford Focus, 2003 Chevy Cavalier, 2003 Subaru Forester, 2002 Toyota Tacoma, 2001 Ford Taurus, 2003 Chevy Impala, 2002 Nissan Xterra, 2003 Ford F-150, 2003 Ford Expedition, and 2003 Chevy Express 15-passenger van. The test results are shown in Table II-2.
These ten additional tests used the FMVSS 208 seating procedure to place the 50th percentile Hybrid III dummy in the occupant seat. The dummy was seated throughout the duration of the 25.4 cm (10 inch) displacement of the load plate. Video cameras tracked the interior roof as it approached the dummy’s head. For each vehicle, the agency tracked the force level and the load plate displacement, and recorded the point where the roof contacted the dummy’s head.
| Vehicle | Ram Travel @150% |
Cos 25o | Test % Weight |
Ram Travel @250% |
Cos 25o | Test % Weight |
Ram Travel @300% |
Cos 25o | |||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2003 Subaru Forester | 21 mm | 0.8 in | 0.7 in | 250 | 40 mm | 1.6 in | 1.5 in | 300 | 45 mm | 1.8in | 1.6 in |
| 2003 Chevy Impala | 32 mm | 1.3 in | 1.1 in | 250 | 53 mm | 2.1 in | 1.9 in | 300 | 63 mm | 2.5 in | 2.2 in |
| 2002 Nissan Xterra | 33 mm | 1.3 in | 1.2 in | 250 | 45 mm | 1.8 in | 1.6 in | 300 | 57 mm | 2.2 in | 2.0 in |
| 2003 Ford Focus | 30 mm | 12 in | 1.1 in | 250 | 60 mm | 2.4 in | 2.2 in | 275 | |||
| 2003 Chevy Cavalier | 32 mm | 1.3 in | 1.1 in | 250 | 68 mm | 2.7 in | 2.4 in | 263 | |||
| 2002 Toyota Tacoma | 38 mm | 1.5in | 1.4 in | 250 | 66 mm | 2.6 in | 2.4 in | 266 | |||
| 2003 Ford F-150 | 37 mm | 1.5 in | 1.3 in | 250 | 77 mm | 3.0 in | 2.7 in | 285 | |||
| 2001 Ford Taurus | 48 mm | 1.9 in | 1.7 in | 250 – Fail Head Touch |
90 mm | 3.5 in | 3.2 in | ||||
| 2003 Ford Expedition | 56 mm | 2.4 in | 1.6 in | 235 - Fail | 115 mm | 4.5 in | 4.1 in | ||||
| 2003 Chevy Express Van | 68 mm | 2.7 in | 2.4 in | 205 - Fail | 110 mm | 4.3 in | 3.9 in | ||||
| Indicates Pass at 300% | Indicates Failure at 300 % | Indicates Failure at 250% | |||||||||
All ten vehicles attained the FMVSS 216 required applied force of 1.5 times the unloaded vehicle weight prior to roof contact with the dummy’s head as shown in Table II-2. Seven vehicles exceeded an applied force of 2.5 times the unloaded vehicle weight, prior to roof contact with the dummy’s head, and three of the vehicles exceeded 3.0 times the unloaded weight of the vehicle.
The Subaru Forester even exceeded an applied force of 4.0 times the unloaded vehicle weight without roof contact with the dummy’s head. Among the ten vehicles tested, the roof contacted the dummy’s head between 82 mm (3.2 inches) and 185 mm (7.3 inches) of plate displacement.
A. Alternatives
The agency is considering upgrading the present roof crush resistance requirement in FMVSS 216 of 1.5 times the vehicle’s weight to 2.5 times the vehicle weight. The agency also considered upgrading the standard to 2.0 and 3.0 times the vehicle’s weight. All three alternatives include the additional requirement that no roof component can touch the head of a 50th percentile Hybrid III dummy when seated in the driver or front passenger seat. All three alternatives are discussed below:
B. 2.0 Times Vehicle Weight Alternative
As observed in Tables II 1-2 there are very few vehicles that would not pass a roof crush resistance of 2.0 times the vehicle’s weight when tested using an extended FMVSS 216 procedure. In fact, the Ford E-150 Van was the only vehicle of the 20 vehicles tested in these two tables that could not pass the 2.0 times the vehicle’s weight roof crush resistance. By this observation, the agency believes only increasing the roof crush resistance to a factor of 2.0 times the vehicle’s weight would accomplish very little, and the corresponding benefits would be equally inconsequential. Thus, the agency is not pursuing a 2.0 times the vehicle’s weight requirement, but will consider the higher 2.5 and 3.0 alternatives.
C. 2.5 Times Vehicle Weight Alternative
In Tables II-1 and II-2 seven of the twenty vehicles (35%) tested would not pass a roof crush resistance of 2.5 times the vehicle’s weight when tested using an extended FMVSS 216 procedure. When adjusted for current sales levels, failed vehicles represent 32.4% of the new passenger vehicle fleet. The resulting benefits and costs that correspond to these test results relative to a roof crush resistance requirement of 2.5 times the vehicle’s weight are discussed in Chapters IV and V, respectively.
D. 3.0 Times Vehicle Weight Alternative
In Tables II-1 and II-3 fifteen of the twenty vehicles (75%) tested would not pass a roof crush resistance of 3.0 times the vehicle’s weight when tested using an extended FMVSS 216 procedure. When adjusted for current sales levels, failed vehicles represent 84.3% of the new passenger vehicle fleet. The resulting benefits and costs that correspond to these test results relative to a roof crush resistance requirement of 3.0 times the vehicle’s weight are discussed in Chapters IV and V respectively.
E. Rollover Near Side/Far Side Roof Crush
The agency received comments from Public Citizen and the Center for Injury Research regarding near and far side testing. [4] The comments stated that vehicle occupants on the far side of the rollover have a much greater risk of serious injury than the occupants on the near side. The commenters suggested that the agency require that both sides of the roof on the same vehicle withstand a crush force of 2.5 times the unloaded vehicle weight. That is after the force is applied to one side of the vehicle, the vehicle is then repositioned and the force load is applied on the opposite side of the roof over the front seat area. Public Citizen cited a recent paper by researchers at Delphi Automotive and Saab, which compared the injury risk depending on the seating position of an occupant relative to the direction of the rollover crash. [5] From this study, Public Citizen concluded that belted, non-ejected occupants on the far side suffer 12 times the risk of serious injuries compared to belted, non-ejected occupants on the near side of the rolling vehicle. On July 26, 2004, JP Research, Inc. submitted an evaluation of the Delphi Automotive and Saab research paper. JP Research discussed the paper with the principal authors and verified that the paper contained errors. When these errors were corrected, the ratio of risk for serious injuries comparing the far to near side roof changed from 12 times, to between 2.4 to 1.
NHTSA performed an independent analysis using NASS-CDS from 1997 to 2002. The agency believes that there is no significant increase in risk for far side belted, non-ejected occupants. NHTSA analyzed NASS-CDS (1997 to 2002) data to evaluate the Delphi research paper with respect to merits of testing both sides of the roof over the front seat area. The analysis included belted front outboard adults who were not totally ejected in a manner similar to the Delphi research paper, but it further restricted the analysis to vehicles that rolled only two to four quarter turns to the side. We estimate the risk of a serious injury, defined as a maximum AIS injury of 3 or greater, to be 29 seriously injured persons per 1000 "far side" occupants and 30 seriously injured persons per 1000 "near side" occupants for a ratio of about 1 to 1.
The agency is committed to further far side research and testing to assure adequate protection is provided to the occupants on both sides of the vehicle. The agency plans to further evaluate the safety need for near and far side testing of the roof over the front occupant area of the same vehicle, before proposing such a requirement.
A. Effectiveness of Increased Roof Strength
Although a significant level of roof crush frequently occurs in rollover crashes, previous efforts to link roof crush to occupant injury have yielded mixed results. Bahling [6] compared dummy neck loadings in production vehicles to rollcaged vehicles and found that neck loads resulted from "diving" impacts where the torso momentum compresses the neck against the vehicle interior. They also concluded that the reduction in roof deformation in the rollcaged vehicle had no effect in reducing neck loads in the area of ground impact. In an early study, Huelke [7] concluded that no statistical relationship exists between the Abbreviated Injury Scale (AIS) and roof crush for restrained occupants in rollovers. Likewise, MacKay [8] used field accident investigations to conclude that seat belts did not reduce head injuries in rollovers. By contrast, Huelke [9] concluded that fatality rates were lower for belted occupants, but primarily due to ejection mitigation. In another early study Partyka [10], found that data in the National Crash Severity Study (NCSS) were too sparse to draw conclusions. In a more recent study of NASS-CDS data, Partyka [11] was unable to ascertain the significance of roof intrusion on injuries in rollover crashes because of concerns that roof intrusion might be a surrogate for crash severity rather than a single cause of injury, and that data was not available for non-injury contacts. Rains and Kanianthra [12] analyzed NASS-CDS data and found that indications that head injury increases with reductions in head room, but also found trends indicating higher risk of head injury with headroom reduction.
These previous efforts compared injury rates to factors such as roof strength and pre-crash headroom. However, a recent NHTSA analysis, which examined the relationship between injury and post crash headroom (Austin et al) [13] found a statistically significant relationship between injury rates and roof crush based on roof contact with the occupant’s head. The injury pattern was noticeably different (less serious) in cases where roof intrusion did not exceed the pre-crash headroom of the occupant – in other words, when the deformed structure did not intrude below the occupant’s head. The initial Austin et al study examined all rollover non-ejected, belted cases of head injury regardless of other injuries that the occupant incurred, whether they were caused by intrusion or not. However, the changes required by this proposal are limited to specific test loads. Many crashes are essentially catastrophic in nature – they impart stress loads on the vehicle’s roof that would overwhelm even the increased strength required by an upgraded test procedure. In order to estimate benefits from revisions to the standard, an estimate must be made of the number of cases where changes made to comply with the standard will actually impact deaths and injuries by reducing the level of intrusion enough so that it no longer extends below the occupant’s head.
To accommodate this approach, the Austin analysis was re-run with a more specific set of restrictions – i.e., belted non-ejected [14] rollover fatalities or injuries with vertical intrusion over the injured occupant’s seat, where the injury was caused by roof contact and the injury is the single maximum level injury experienced by the occupant. [15] Unbelted occupants were not considered because in a rollover without restraints they will essentially become moving objects within the vehicle and tying their injuries to vehicle crush (as opposed to their own movement) is problematic. The restriction to vertical intrusion over the occupant’s seat was necessary to establish cause and effect. The restriction to sole maximum injury level was made because if an occupant has multiple injuries at the same maximum level, they would still be injured at that level even if the intrusion-caused injury were eliminated. However, NHTSA recognizes that this might still improve the injured occupant’s outcome, and thus this approach gives a conservative estimate of safety benefits.
The calculation of benefits will be based upon the predicted change in post-crash headroom. Post-crash headroom is defined as the pre-crash space over the occupant’s head minus the amount of vertical intrusion of a roof component. This calculation makes use of the headroom measures in Consumers Union’s Consumer Report vehicle specification file. The Consumers Union headroom measures involve a 5’9" person who adjusts the seat to a "comfortable" position. The driver’s headroom was used for both front outboard-seating positions. The Consumers Union measures begin with model year 1987. If the MSN Autos web site did not indicate a redesign of the vehicle’s body, then headroom was carried over from one model year to the next to complete years for which headroom was missing. [16] Post-crash headroom was then determined in two steps. The first step involved adjusting the Consumers Union headroom by the difference between the tester’s seated height and the NASS-CDS occupant’s seated height. In both cases, seated height was estimated by taking 48 percent of standing height, which is based upon the 50th percentile male dummy. The second step involved subtracting the magnitude of vertical intrusion. In a relatively small number of cases, the intrusion was reported as a range rather than by an exact number. In these instances the range was replaced by the median of the range. In another relatively small number of cases where the Consumers Union headroom was unknown, the average vehicle headroom for the vehicle’s body type was used. However, for a significant number of cases, post-crash headroom could not be determined because the occupant’s height was unknown. These cases were excluded from the calculation of effectiveness.
To determine effectiveness, the relationship between the maximum injury severity of a head, neck, or face injury due to an intruding roof component and post-crash headroom for all occupants in the target population was examined. A wide variety of functional forms were explored, including linear functions, logarithmic functions, step functions, and functions involving percent reduction in headroom, but the only statistically significant result could be obtained using a dichotomous cut-point between no remaining headroom (negative post-crash) and remaining headroom (positive post-crash).
The resulting analysis is summarized in Table IV-1. Note that the distribution of cases where headroom is remaining results in a less severe injury profile (shifted towards less serious injuries), compared with cases where the roof intruded below pre-crash headroom.
Based on the data in Table IV-1, the "No Headroom" cases were redistributed using the injury profile found in the "Headroom Remaining" cases. This produced a revised injury profile for those cases. By comparing the revised No Headroom profile to the initial No Headroom profile, a theoretical safety impact was calculated and used to estimate an effectiveness rate. This rate was calculated as follows:
E = (Na-Nr)/Na
Where:
E = effectiveness in injury reduction for each specific level.
Na = initial incidence of injuries when No Headroom
Nr = revised incidence of injuries based on Remaining Headroom distribution.
The results indicate that, for the restricted target population, 92 percent of fatalities would be eliminated. In addition, about 30 percent of all MAIS 3-5 injuries, 68 percent of MAIS 2 injuries, and 70 percent of MAIS 1 injuries would be eliminated. Uninjured counts would increase by about 11 percent. Although these effectiveness rates seem high, remember that they only apply to the very narrow target population they were designed to address. If effectiveness were computed based on a broader definition of the rollover injury problem, effectiveness rates would be much lower.
The effectiveness indicates the predicted percent of the injuries at each level that would be prevented by shifting occupants with no remaining headroom to remaining headroom. Table IV-1 also contains the cases with missing headroom for completeness, and the totals for each injury level correspond to the total target population figures noted in bold in Table IV-4. These numbers also highlight again that the benefits for any change to FMVSS 216 are likely to be small because of the small proportion of occupants with relevant injuries even when the vertical intrusion of a roof component is likely to be below the top of their head.
| No Headroom |
Headroom Remaining |
Effectiveness | Missing Headroom |
Total | |||
|---|---|---|---|---|---|---|---|
| No relevant injury | 47,319 | 85.61% | 54,372 | 95.42% | 7,602 | 109,294 | |
| MAIS 1 | 3,863 | 6.99% | 1,207 | 2.12% | 69.7% | 661 | 5,731 |
| MAIS 2 | 3,212 | 5.81% | 1,059 | 1.86% | 68.0% | 261 | 4,532 |
| MAIS 3-5 | 422 | 0.76% | 303 | 0.53% | 30.4% | 81 | 807 |
| Fatal | 457 | 0.83% | 40 | 0.07% | 91.5% | 98 | 596 |
B. Target Population
The stated purpose of FMVSS 216 is to "reduce deaths and injuries due to the crushing of the roof into the occupant compartment". This rationale, as well as the test procedure itself, suggest that the standard does not apply to all injured occupants of rollover crashes. Rather, it covers a more narrow set of injuries. Therefore, the target population, defined as occupants who are likely to benefit from a stronger roof due to an upgrade of FMVSS 216, is a subset of all occupants injured during a rollover. This section first explains the procedure for determining the target population. The section then describes the calculation of the population that may benefit from one of the two proposed upgrades.
The target population estimates are based on results from NASS-CDS from 1997 through 2002. The beginning year, 1997, is when NASS-CDS first recorded the exact magnitude of intrusion. At the time of this analysis, 2002 is the most recent year available. The vehicles considered for the benefit calculations are non-convertible light duty vehicles (NASS-CDS body type 2 through 49). Vehicle headroom is only available for vehicles of model year 1987 and later, and intrusion measures are only known for vehicles that were inspected by a NASS-CDS investigator. Therefore, older vehicles and vehicles that were not inspected were excluded from the analysis, but the sample weights were adjusted to reflect the estimated total number of NASS-CDS rollover vehicles. Finally, vehicles that rolled only one-quarter turn to the side or that experienced a collision with a fixed object (other than a bush, embankment, ditch, culvert, or the ground) to the top of the vehicle were excluded. Vehicles that rolled only one-quarter turn did not experience a roof-to-ground contact consistent with the test procedure. Likewise, vehicles where a tree or pole struck the roof experienced a more concentrated force than would occur solely with a ground contact.
The occupants counted for benefits calculations were not fully ejected, belted, aged 13 or older, and seated in one of the two front outboard seats of the vehicles described above. Based on data from 1997-2001 NASS-CDS we determined that an estimated 77 percent of the seriously and fatally injured with known serious injuries who were fully ejected in roof-involved rollovers received their most severe injury from outside of the vehicle and another estimated 3 percent received injuries of equal severity from both outside and inside the vehicle. Consequently, it appears that preventing ejection is the most important means for reducing injury to fully ejected rollover occupants. For occupants who were unbelted but not fully ejected, we could not establish a relationship between roof crush injuries and the magnitude of roof crush. The test itself measures roof strength over front outboard seats, for which NHTSA recommends that the occupants be aged 13 or older, and the potential benefits for other seating positions (if any) cannot be determined.
Occupants for whom injury severity was unknown and fatalities without injury information were excluded from the analysis, but the sample weights were adjusted to reflect the estimated total number of occupants at each injury level. Also, fatalities were adjusted to reflect the average number of non-convertible light vehicle rollover fatalities in Fatality Analysis Reporting System (FARS) from 1997 through 2002. Finally, the occupants in the target population had vertical intrusion of a roof component over their seating position, where a roof component includes the roof itself, roof side rails, front (windshield) and back (backlight) headers, A and B pillars, the sun visor, as well as any roof console, sunroof components, or roll-bar. This criterion ensures that affected occupants were exposed to roof crush. Finally, the occupant had to experience a relevant injury, which is defined as a head, face, or neck injury from a vertically intruding roof component into the occupant’s seating position. We also explored intrusion that was over the front middle seating position and lateral and longitudinal intrusion, but relaxing these assumptions did not change the results.
Table IV-2 demonstrates how each of the above restrictions reduces the injured population affected by roof crush relevant to FMVSS 216. The restrictions that matter the most appear to differ by the severity of injury. For serious injuries and fatalities, large drops occur when excluding fully ejected occupants. A second large drop occurs when excluding the unbelted who were not fully ejected. For minor and moderate injuries, large drops occur when excluding the unbelted. All four of the injury categories also experienced significant drops when excluding vehicles with no vertical intrusion of a roof component and when requiring that the occupant have a relevant injury. Finally, not all occupants with a relevant injury are applicable for benefits purposes. For survivors, only injured occupants where the relevant roof crush injury was the sole maximum severity (AIS) injury are included in the calculation of safety benefits lives saved. For fatalities, only those occupants for whom the relevant roof crush injury was either the sole maximum severity injury or one of the most severe injuries are counted for lives saved. This approach reflects the fact that eliminating one of two or more injuries of identical severity would not change the status of the occupant – they would still be injured at the same severity level. However, an analysis of head-injured fatality cases indicated that when 2 or more injuries of the same MAIS level occur to different body regions, the cause of death is overwhelmingly the head injury. Therefore, for fatalities, both sole MAIS injury cases and those cases with a head injury at MAIS that were not the sole MAIS were considered to be relevant injuries.
| Minor Injuries (AIS 1) |
Moderate Injuries (AIS 2) |
Serious Injuries (AIS 3-5) |
Fatalities | |
|---|---|---|---|---|
| Non-Convertible Light Vehicles in Rollovers | 212,340 | 39,379 | 23,793 | 9,942 |
| Roof-Involved Rollover | 173,428 | 33,930 | 21,005 | 8,585 |
| No Fixed Object Collision to Top | 165,730 | 30,616 | 19,454 | 7,426 |
| Not Totally Ejected | 161,893 | 26,404 | 12,833 | 3,559 |
| Using Safety Restraint | 118,098 | 14,738 | 9,592 | 2,026 |
| Front Outboard Seats | 103,932 | 14,119 | 9,003 | 1,780 |
| Not 12 Years Old or Younger | 101,654 | 14,064 | 8,974 | 1,764 |
| Roof Component Intrusion | 62,695 | 11,346 | 7,144 | 1,030 |
| Head, Neck, or Face Injury from Intruding Roof Component | 21,198 | 7,169 | 2,373 | 751 |
| Injury – Not MAIS | 0 | 2,262 | 1,354 | 155 |
| Injury at MAIS - Not Sole Injury | 15,467 | 375 | 213 | 371 |
| Sole MAIS Injury | 5,731 | 4,532 | 807 | 225 |
| Note: Occupants relevant for benefits calculations are in bold. | ||||
The target population relevant to FMVSS 216 (the bold numbers near the bottom of Table IV-2) is thus a relatively small subset of the occupants injured in rollovers. For fatalities, the estimated total for the target population is 6 percent of all non-convertible light vehicle rollover fatalities (596/9942). For minor and serious injuries, the estimated total is closer to 3 percent. There is an apparent jump in the moderate injury category where the target population is an estimated 12 percent of the total population, but this estimate is due in large part to two NASS-CDS cases with a combined annual weight of over 3,200. [17]
C. Approach and Methodology
Table IV-3 lists all relevant cases of fatal injury from the 1997-2001 NASS-CDS. These cases make up the weighted target population described in Table IV-2. In addition to case information, it lists occupant height, pre-crash headroom, intrusion levels, post-crash headroom, and the annual average weighted fatality counts that result from each case. Note that in 14 of the 32 fatality cases, intrusion levels exceed 10 inches, indicating a severe force level that caused significant crush to the roof. The impact that the proposed standards will have on safety will depend on the degree to which they reduce intrusion in these 32 cases.
There are two basic pass/fail criteria in the proposal. Vehicles will be tested using a standard plate mounted at an angle that simulates vehicle-to-ground contact over the front seat area. The plate will be pushed into the roof until a proposed force equal to 2.5 times the vehicle’s weight is attained. During this procedure, the roof cannot intrude below the head of a 50th percentile male dummy positioned in the drivers seat. Vehicles will fail the test if they are unable to withstand the force level, or if they withstand the force level but still allow the roof to intrude below the dummy’s head. In either case, the test would be terminated as a failure when contact is made with the dummy’s head.
| Year | CASE | OCC | CU Headroom (inches) | Intrusion (inches) | Occ Height (inches) | Post Crash Headroom (inches) | Post Crash Headroom 250% | Post Crash Headroom 300% | Est. Annual Average | Approach A: Target Pop 250% | Approach A: Target Pop 300% | Approach B: Target Pop 250% | Approach B: Target Pop 300% |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1997 | 48 | 1 | 3 | 12.6 | 70.9 | -10.5 | -7.8 | -7.4 | 4.48 | 0.0002 | 0.0006 | ||
| 1999 | 169 | 1 | 2.5 | 6.7 | 68.1 | -3.8 | -2.1 | -1.8 | 14.709 | 2.5318 | 3.4433 | ||
| 2001 | 20 | 1 | 3.5 | 13.4 | 70.9 | -10.8 | -8.1 | -7.7 | 9.081 | 0.0002 | 0.0005 | ||
| 2002 | 16 | 1 | 1.5 | 5.1 | 68.9 | -3.6 | -2.4 | -2.2 | 23.755 | 3.8165 | 4.824 | ||
| 2001 | 125 | 2 | 3 | 20.5 | 66.1 | -16.1 | -13.4 | -13 | 3.822 | 0 | 0 | ||
| 2002 | 73 | 1 | 4 | 11.8 | 68.9 | -7.8 | -5.1 | -4.7 | 20.637 | 0.0771 | 0.1704 | ||
| 1997 | 15 | 1 | 4 | 13.8 | 66.9 | -8.8 | -6.1 | -5.7 | 9.408 | 0.0002 | 0.0007 | ||
| 2002 | 59 | 1 | 4 | 5.9 | 74 | -4.3 | -2.8 | -2.6 | 13.41 | 5.3193 | 5.9173 | ||
| 2001 | 24 | 2 | 5 | 6.3 | 65 | 0.6 | 2.2 | 2.4 | 14.107 | 5.2648 | 5.671 | ||
| 1998 | 47 | 1 | 4 | 4.7 | 72 | -2.2 | -1.1 | -1 | 1.041 | 0.2348 | 0.2492 | ||
| 2001 | 131 | 2 | 5.5 | 3.9 | 74 | -0.8 | 0 | 0.1 | 9.08 | 9.0799 | 9.0799 | 0.1664 | 0.1757 |
| 2002 | 91 | 1 | 5 | 8.7 | 66.9 | -2.7 | -0.4 | -0.1 | 8.438 | 3.4905 | 4.1906 | ||
| 2002 | 189 | 1 | 3 | 2 | 79.1 | -1.5 | -1.2 | -1.2 | 13.41 | 0.2794 | 0.2794 | ||
| 2000 | 172 | 1 | 3.5 | 13.8 | 72 | -11.7 | -9 | -8.6 | 12.065 | 0.0001 | 0.0002 | ||
| 2000 | 2 | 1 | 2.5 | 5.5 | 70.1 | -3.5 | -2.2 | -2 | 21.821 | 5.7812 | 6.9538 | ||
| 1998 | 10 | 1 | 3 | 17.7 | 66.9 | -13.7 | -11 | -10.6 | 7.56 | 0 | 0 | ||
| 2000 | 68 | 1 | 4 | 18.1 | 66.1 | -12.7 | -10 | -9.6 | 29.935 | 0 | 0 | ||
| 2000 | 181 | 1 | 4.5 | 1.6 | 65 | 4.9 | 5.1 | 5.1 | 7.363 | 0.0053 | 0.0053 | ||
| 2001 | 143 | 2 | 4.5 | 18.9 | 70.1 | -14.9 | -12.2 | -11.8 | 3.822 | 0 | 0 | ||
| 2001 | 16 | 1 | 6 | 7.5 | 70.1 | -2 | 0 | 0.2 | 6.964 | 6.9644 | 6.9644 | 3.1832 | 3.3526 |
| 2000 | 167 | 1 | 4 | 11.4 | 64.2 | -5.1 | -2.4 | -2 | 12.065 | 0.0995 | 0.2053 | ||
| 2000 | 178 | 2 | 4 | 2.4 | 68.1 | 2.1 | 2.5 | 2.5 | 18.759 | 0.2237 | 0.2237 | ||
| 1998 | 130 | 1 | 2.5 | 11.4 | 16.075 | 0.0001 | 0.0004 | 0.0047 | 0.0126 | ||||
| 2002 | 182 | 2 | 4 | 9.4 | 63 | -2.6 | -0.1 | 0.2 | 11.232 | 11.232 | 1.3891 | 1.9311 | |
| 2001 | 28 | 1 | 4 | 2.8 | 74 | -1.2 | -0.7 | -0.6 | 6.795 | 0.1648 | 0.1863 | ||
| 2000 | 73 | 1 | 4.5 | 8.3 | 66.9 | -2.8 | -0.6 | -0.3 | 113.56 | 41.481 | 50.8107 | ||
| 1999 | 86 | 1 | 4.5 | 15.4 | 74 | -13.3 | -10.6 | -10.2 | 8.843 | 0 | 0 | ||
| 1997 | 180 | 1 | 5 | 12.2 | 68.1 | -6.8 | -4.1 | -3.7 | 14.241 | 0.1701 | 0.339 | ||
| 2000 | 76 | 1 | 6 | 6.7 | 77.2 | -4.6 | -2.9 | -2.6 | 27.609 | 8.0761 | 8.6514 | ||
| 1997 | 6 | 1 | 5.5 | 8.7 | 81.916 | 19.9148 | 25.7467 | 41.6019 | 48.0048 | ||||
| 2001 | 89 | 1 | 5 | 15 | 70.9 | -10.9 | -8.2 | -7.8 | 21.26 | 0.0003 | 0.0009 | ||
| Aggregate Impact | 36 | 53 | 123 | 146 | |||||||||
To estimate the impact that the proposed test requirements would have on roof intrusion, NHTSA examined the force deflection curves from a series of 20 quasi-static vehicle tests conducted to determine their performance under conditions that lead to roof crush. These tests were conducted over several minutes and produced about 900 separate data points to define the base deflection curves. The tests are further described in Chapter II, Agency Research and Test Results. An example of these curves (based on the Ford Taurus), which plot the displacement of the roof plate used in the test to the force applied (measured as a percent of vehicle weight), is shown (labeled baseline) in Figure A. The vehicles tested were designed to meet current testing standards, which specify a force of 1.5 times the vehicles’ weight. In order to estimate vehicle behavior at the higher levels being proposed, NHTSA derived an adjustment factor by comparing the peak vehicle load experienced prior to estimated head contact [18] with the load that would be required under the proposal. This was done for the 6 vehicles that failed to achieve the proposed 2.5 force level and the 1 vehicle that failed to achieve the 2.0 force level. For example, the Taurus experienced a peak load of 2 times its weight, which passes a 2.0 alternative, but fails both the 2.5 proposal and the 3.0 alternative. Under the 2.5 proposal, the adjustment factor would be 1.25 (2.5/2.0). This factor was used to scale the original force deceleration curves to levels consistent with the proposed alternative requirements of 2.0 and 2.5. Similarly, the factor for the 3.0 alternative would be 1.5.
Manufacturers normally build a level of safety into their vehicle designs so that vehicles can routinely pass required test requirements. Table IV-4 illustrates these margins for our test vehicles. Within the 20 vehicles tested, this safety margin averaged about 83 percent, and the minimum margin was 25 percent. Under more difficult test requirements, manufacturers may not be able or willing to maintain the high average safety margins they experience under the current standard. However, NHTSA believes it is likely that manufacturers will maintain a safety margin equal to at least a minimum margin of 20 percent. We therefore examined scenarios that assume a 20 percent safety margin as well as the basic requirements. Separate curves were thus developed for 2.5, and 3.0 scenarios with this 20 percent safety margin. The 2.5 and 3.0 curves in Figure IV-A were thus derived as a simple proportional relationship between the current standard and the proposed standard. For example, the peak force for the Taurus occurred at about 2.0 x its curb weight (Table IV-4) at about 40,000 N of force. Assuming manufacturers will maintain a compliance margin of 20 percent for whatever changes they make, for the 2.5 requirements, this implies they will actually strengthen it to 3.0 (2.5x1.2). We estimated that the peak force to failure at 3.0 would be proportional to the ratio of the strength requirements - in this case 1.5 (3.0/2.0). Therefore we assume that the peak force at 3.0 would be about 60,000 N. (1.5x 40,000 N). The same proportion is applied to each point on the curve to produce the estimated scaled force deflection curve for a 2.5 requirement. The resulting curves are shown together with the baseline curve for the Ford Taurus in Figure IV-A.
The benefits for the alternative load requirements will be evaluated using the crush measured in filed reported crash cases. For each crush measurement the change in static deflection between the baseline and alternative requirements, for equivalent energy absorption, will be used to estimate the change in vertical crush. Following development of these individual force deflection curves for the 7 vehicles that did not meet the 2.5 load requirement, each curve was integrated over its full test duration to estimate the total energy experienced at each point during the test. Using these data, energy was plotted as a function of deflection. The resulting curves are shown for the example vehicle in Figure IV-B. These curves were then analyzed to produce a set of reconfigured curves that plotted the change in deflection as a function of baseline deflection. These measurements were taken from Figure IV-B relative to the baseline curve, i.e., the deflection at each point on the baseline curve in Figure IV-B was compared to the deflection on each of the alternative load requirement curves to produce a difference that is reflected on the y-axis. The resulting curves are shown for the Ford Taurus in Figure IV-C. These curves describe the change that would be expected in intrusion levels for each level of base intrusion under the alternative roof strength levels. For example, for the Taurus, a rollover that would cause 200 mm of intrusion with its existing roof would cause about 40 mm less intrusion if the roof were strengthened to meet a 2.5 standard, and about 60 mm less intrusion in a roof that was strengthened to meet a 3.0 requirement.
| VEHICLE | ROOF STRENGTH %Vehicle Weight |
Compliance Margin |
|---|---|---|
| 1999 Ford E150 Van | 188 | 1.25 |
| 2003 Ford Taurus | 203 | 1.35 |
| 2003 Chevy Express Van | 205 | 1.37 |
| 2002 Ford Explorer | 235 | 1.57 |
| 2003 Ford Expedition | 238 | 1.59 |
| 2002 Ford Mustang | 244 | 1.63 |
| 2002 Ford Crown Victoria | 245 | 1.63 |
| 2002 Dodge Ram 1500 PU | 249 | 1.66 |
| 2003 Chevy Cavalier | 264 | 1.76 |
| 2003 Toyota Tacoma | 265 | 1.77 |
| 1997 Dodge Grand Caravan | 265 | 1.77 |
| 1998 Chevy S-10 PU | 275 | 1.83 |
| 2003 Ford Focus | 277 | 1.85 |
| 2003 Ford F-150 | 289 | 1.93 |
| 2001 Chevy Tahoe | 290 | 1.93 |
| 2002 Honda CRV | 305 | 2.03 |
| 2002 Toyota Camry | 315 | 2.10 |
| 2003 Chevy Impala | 316 | 2.11 |
| 2003 Nissan Xterra | 346 | 2.31 |
| 2003 Subaru Forester | 480 | 3.20 |
| Average | 275 | 1.83 |
The force displacement curves for seven tested vehicles were averaged to produce the deflection curve for the 2.5 alternative and the force displacement curves for fifteen vehicles were averaged to produce the curve for the 3.0 alternatives. These average curves are shown in Figure IV-D. A third order polynomial regression was then run for each curve in Figure D to produce a model of expected displacement impact from higher force levels. The resulting models for the two alternatives were as follows:
For 2.5 times vehicle weight y = -0.000005 x3 + 0.0021 x2 + 0.0482 x (r2 = 0.9984)
For 3 times vehicle weight y = -0.000005 x3 + 0.0021 x2 + 0.0802 x (r2 = 0.9990)
Where:
y = intrusion prevented (in mm)
x = baseline intrusion (in mm)
These models were applied to baseline intrusion levels for each case that meets the definition of a relevant injury in our target population to estimate the hypothetical intrusion prevented by the increased force requirement. This was then added to the measured post-crash headroom to produce an estimate of the resulting hypothetical post crash headroom under each alternative. If the measured post-crash headroom was negative and the hypothetical post crash headroom was positive or zero, then that case would benefit from the new requirement. The weighted value of all such cases was then used as the target population and the effectiveness estimates discussed previously were applied to these totals to produce an estimate of benefits within the full vehicle fleet.
Due to the nature of the third order equation, there is an inflection point at which the curve begins to slope downward. This inflection point is realized after 254 mm, and therefore cannot be visualized on our curve fit. With very large baseline intrusions, represented by "x" in the equations, the intrusion prevented ("y") might become a negative value. Since the upgrade would never increase overall intrusion, we calculated the largest "y" value, and with the intrusion corresponding "x" value assumed that any baseline intrusion greater than "x" would have an intrusion prevented value of the maximum "y." This assumption did not alter the benefits number because it only affected the most severe cases, which had little or no influence on the benefits calculations.
Table IV-3 contains results for each alternative using two different analytical approaches. The first approach (Approach A) analyzed specific cases in NHTSA’s NASS-CDS database to determine the impact that would occur in each case from added roof strength. However, due to a combination of a small sample of relevant cases and unknown data elements, this analysis suffered from gaps in data and spikes in case weights. An alternative analysis was thus performed based on a construct, which assumes a probability of occupant height in each vehicle equal to the national distribution of occupant heights (Approach B). This was done to mitigate the impact that specific driver characteristics had in determining the inclusion of the case. The theory behind this is that any size occupant might have been involved in a crash of each case’s specific intrusion magnitude. By reflecting the probability that the occupant was of a height that would benefit from specific reductions in roof crush, the spikes in case weights would be minimized. The agency recognizes that this method assumes a random relationship between the height of drivers and the headroom that exists in vehicles that they purchase. To the extent that drivers height and vehicle headroom are actually related, this second approach loses credibility. However, the agency did examine the relationship between vehicle headroom and occupant size and found no perceptible trends [19]. Table IV-5 displays the results of this analysis. Overall, taller drivers seem no more likely to be in vehicles with less headroom than do shorter drivers. It is likely that passengers adjust the seat backs and positions in their vehicles to prevent uncomfortable proximity to internal vehicle structures such as the roof. The agency seeks comment on these two approaches.
| Height in cm | <159 | 160-167 | 168-174 | 175-182 | 183-187 | 188+ | Total |
| Height in feet and inches | <5’3” | 5’3” to <5’6” |
5’6” to <5’9” |
5’9” to <6’0” |
6’0” to <6’2” |
6’2”+ | |
| <3” Headroom | 13% | 8% | 34% | 16% | 57% | 9% | 26% |
| (8) | (22) | (22) | (15) | (11) | (4) | (82) | |
| 3 to3.5” | 16% | 22% | 13% | 18% | 15% | 18% | 17% |
| (17) | (38) | (50) | (57) | (26) | (11) | (199) | |
| 4” + Headroom | 71% | 70% | 53% | 66% | 28% | 73% | 57% |
| (56) | (90) | (142) | (116) | (62) | (35) | (501) | |
| Total | 100% | 100% | 100% | 100% | 100% | 100% | 100% |
| (81) | (150) | (214) | (188) | (99) | (50) | (782) | |
| Missing Consumers Union Vehicle Headroom | 25 | ||||||
| Note: These occupants from 1997 through 2001 NASS-CDS crash database were belted non-ejected front outboard occupants over 12 years old. They were in single-vehicle rollovers of more then one-quarter turn to the side or end-over-end. The vehicles were model years 1987 or later, not convertibles, not certified altered vehicles, not towing a trailing unit, and did not have a fixed object collision with the top of the vehicle. Also, cases with unknown values of occupant height and intrusion have been excluded. | |||||||
To adjust for occupant height probability in Approach B, the agency calculated the lower and upper height bounds of occupants that would benefit from the proposal. These bounds are a function of actual intrusion in the crash, the level of improvement that would be prevented in the crash, and the probability that occupants would be of a height that falls within these bounds. Actual intrusion was derived from NHTSA’s data files and the improvement in intrusion was estimated using the force displacement curves previously discussed.
The lower bound is a person for whom the amount of initial intrusion just reaches the top of their head. Therefore a marginal amount of prevented intrusion would give the occupant positive post-crash headroom. This is a lower bound because if the occupant was shorter, the initial intrusion would be above their head (positive post-crash headroom) and would not count for benefit purposes, i.e., they would be assumed to be uninjured by the roof intrusion in the crash. The equation that satisfies this condition would be:
Post-crash Headroom = 0
or
Pre-crash Headroom – Intrusion = 0
Where:
Pre-crash Headroom = (CU Headroom - (Occupant Height - 69)/0.48)
CU Headroom = CU headroom measurement for specific make/model vehicle in crash
69 = height in inches of CU tester
.48 = ratio of seated to standing height.
Intrusion = original intrusion recorded in the CDS data base converted from centimeters to inches
This equation is pre-crash headroom (CU headroom adjusted for the difference between the seated height of the actual occupant and the 69 inch tall CU tester) minus actual intrusion equals zero. Solving for occupant height, the lower bound becomes:
Occupant height = 69 + 0.48 * (CU Headroom - Intrusion)
The upper bound is a person for whom the intrusion that would occur under the upgrade just reaches the top of their head. In this case, the prevented intrusion is just enough to turn the case from negative to positive post-crash headroom. This is an upper bound because a taller person would still have negative post-crash headroom and would not count for benefits purposes. (Even though some intrusion was prevented, it would not be enough to create positive post-crash headroom). The equation that satisfies this condition would be:
Post-crash Headroom = -Prevented Intrusion
or
Pre-crash headroom - Intrusion + Prevented Intrusion = 0.
or
(CU Headroom - (Occupant Height - 69)/0.48) – Intrusion + Prevented Intrusion =0
In this case, post-crash headroom (pre-crash minus actual intrusion) would equal the amount of intrusion prevented under the proposal (because by definition initial headroom for an occupant of this height is zero). Note that in the original equation prevented intrusion is expressed as a negative number because it represents a reduction in intrusion levels. Solving this equation for occupant height, the upper bound becomes:
Occupant height = 69 + 0.48 * (CU Headroom – Intrusion + Prevented Intrusion)
Once the lower and upper heights were derived, we calculated the probability that a person randomly drawn from the US (adult) population would have a height between the lower and upper bound under the assumption of a normal distribution with mean height of 66.4 (average of men and women) and standard deviation of 3 from the CDC. This probability was then applied to the case weight for the specific CDS case to estimate a benefit impact.
Full fleet benefits would only be realized if changes were made to all vehicles. However, many vehicles already pass the proposed requirements. In Section III it is noted that 32.4 percent of tested vehicles failed the proposed test at 2.5 x vehicle weight. It is thus assumed that only 32.4 percent of the vehicle fleet will require changes and this failure rate is then applied to the full fleet benefits to estimate the safety impact of the 2.5 requirement on the failed portion of the fleet. For the 3.0 alternative, the failure rate was 84.3 percent. The results are summarized in Table IV-6.
| 2.5 Times Vehicle Weight | ||||||
|---|---|---|---|---|---|---|
| Relevant Population | Failure Rate | Effectiveness | Estimated Benefits | |||
| Approach A | Approach B | Approach A | Approach B | |||
| Fatalities | 36 | 123 | 0.324 | 0.9147 | 11 | 37 |
| MAIS 5 | 0 | 6 | 0.324 | 0.3037 | 0 | 1 |
| MAIS 4 | 5 | 54 | 0.324 | 0.3037 | 0 | 5 |
| MAIS 3 | 10 | 122 | 0.324 | 0.3037 | 1 | 12 |
| MAIS 2 | 2337 | 678 | 0.324 | 0.6802 | 515 | 150 |
| MAIS 1 | 952 | 1290 | 0.324 | 0.6969 | 215 | 291 |
| 3.0 Times Vehicle Weight | ||||||
| Relevant Population | Failure Rate | Effectiveness | Estimated Benefits | |||
| Approach A | Approach B | Approach A | Approach B | |||
| Fatalities | 53 | 146 | 0.843 | 0.9147 | 41 | 112 |
| MAIS 5 | 0 | 7 | 0.843 | 0.3037 | 0 | 2 |
| MAIS 4 | 5 | 60 | 0.843 | 0.3037 | 1 | 15 |
| MAIS 3 | 11 | 138 | 0.843 | 0.3037 | 3 | 35 |
| MAIS 2 | 2447 | 835 | 0.843 | 0.6802 | 1403 | 479 |
| MAIS 1 | 978 | 1510 | 0.843 | 0.6969 | 575 | 887 |
Adjustment for Electronic Stability Control
The data used to determine the target population were from the years 1997 through 2002. Vehicles on the road during that time frame generally were not equipped with electronic stability control (ESC). Evaluations of ESC in existing vehicles have found them to be highly effective in preventing single vehicle crashes. NHTSA estimates that roughly 18% of current MY 2005 vehicle fleets (17% of passenger cars and 19% of LTVs), now come equipped with ESC. An adjustment will be made to the 1997-2002 target population to reflect the impact of ESCs in the newer vehicle fleets that will be subject to the higher roof strength requirements of this proposal.
The adjustment is a function of the increase in ESC penetration in the new vehicle fleet compared to the on-road fleet that is reflected in 1997-2002 NASS and FARS databases, the effectiveness of ESCs in reducing single vehicle crashes, and the portion of rollover crashes that occur in single vehicle crashes. The formula is as follows:
ESCf = (Pn-Pb)*e*s
Where
ESCf = ESC adjustment factorPn = Current (MY 2005) ESC penetration in new vehicle fleet
Pb = ESC penetration in the MY fleets that were on road during 1997-2002.
e = effectiveness of ESC against single vehicle crashes
s = portion of rollover crashes that occur in single vehicle crashes<