Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
78 Chapter 4 Strategies to Improve Vehicle/Roadside Hardware Compatibility Throughout the research period, a number of techniques and available data sources were used to evaluate compatibility between vehicles and roadside hardware systems. Findings presented in previous chapters offer some insight into existing compatibility issues. In this chapter, a number of future strategies to further improve compatibility were synthesized based on research discoveries and findings. Overall, these strategies are as follows. Each of the following four sections overview suggested approaches and ideas to improve research in these areas. These include: ⢠Increasing awareness of roadside safety related organizations including vehicle manufacturers, DOT regulatory groups and roadside safety engineers regarding compatibility issues. This effort requires significant commitment from both automotive and roadside safety engineers to regularly communicate and work together to optimize both components of the roadside/vehicle system simultaneously. ⢠Improve current methods used to collect real world accident data so that future studies may benefit from improved information. Suggested data collection forms targeting roadside features are included here so that NHTSA's NASS Crash Investigators can consider their implementation. ⢠Proposed methods to improve test methodologies so that vehicle to roadside hardware compatibility can be better assessed using test results. This testing should include a wider yet non-excessive sample of vehicle platforms to better characterize vehicle interaction for the entire fleet. Methods to select these vehicles have been proposed using average vehicle characteristics compiled within Chapter 3. ⢠Initiating the use of advanced modeling techniques to isolate occurrences of incompatibility across an expanded group of vehicle platforms. As a supplement to testing or as a more cost effective method for verifying vehicle to roadside hardware, a protocol for the use of advanced simulation techniques should be implemented.
79 4.1 Industry Interaction and Workshop Findings A workshop involving representatives from the automotive industry, Federal and State DOTs, roadside hardware manufacturers and other related groups was held to discuss roadside hardware compatibility issues. This open forum allowed relevant groups to learn about ongoing research and offer valuable suggestions to improve safety in this area. Attendees for the workshop and their affiliations are as follows: 1. Maurice Bronstad, Research Team Contractor- Dynatec Engineering 2. Monique Evans, Ohio DOT 3. Gene Buth, Texas Transportation Institute 4. Chuck Niessner, NCHRP 5. Daniel Godrick, Research Team- GWU 6. Steve Kan, Research Team- GWU 7. Leonard Meczkowski, FHWA 8. Michele McMurtry, NTSB 9. Paul Bedewi, Ford 10. Michael Griffith, FHWA 11. Stephen Maher, 12. Michael Cammisa, Alliance International Automotive Manufacturers 13. Ralph Hitchcock, Honda 14. John Laturner, E Tech 15. Richard Powers, FHWA 16. Harry Taylor, FHWA 17. Kennerly Digges, Research Team- GWU 18. George Bahouth, Research Team- GWU 19. Dhafer Marzougi, Research Team- GWU 20. Azim Eskandarian, Research Team- GWU At the workshop, the nature of future vehicle characteristics in relation to roadside hardware and the expected safety implications were discussed. The workshop agenda also included the discussion of potential changes in technology affecting roadside hardware and design, e.g., new materials, new trends in manufacturability and assembly, and new federal and/or state initiatives affecting transportation policy.
80 This workshop initiated dialog and interaction between groups who do not typically communicate on such topics. A primary feature of this workshop was the exchange of information between the roadside community and vehicle safety researchers/manufacturers. No forum currently exists in which vehicle manufacturers learn about observed performance of roadside systems in relation to their vehicle design features. Additionally, current Federal Motor Vehicle Safety Standards (FMVSS) do not mandate that new vehicles meet any minimum standards in terms of roadside hardware crash performance. One overwhelming response and intended outcome of the workshop was the initiation of communication and further interactions of this type between the roadside safety and vehicle safety communities. Future collaboration between roadside researchers and vehicle manufacturers to address key workshop finding was proposed and efforts to begin this process is underway. The workshop began with an overview of research presented by the GW Team. Following these presentations, a questionnaire was used to guide discussion on a variety of relevant topics. The exchange of ideas related to the proposed questions was found to be very valuable for the research team and other in attendance. Below, a summary of the discussion surrounding the questionnaire is given. Note: The text below has been reproduced based on workshop discussions. This information contains opinions of the workshop attendees; however the accuracy and validity of the ideas have not been verified by the research team. 4.1.1 Accident Data Related Results 1. Midsize and large SUVs far exceed the car market regarding fatal rollovers during impact with guardrails. What is the reason for this? o SUVs differ from cars in center of gravity height, front overhang, width, wheelbase and height and therefore react differently with guardrails. In particular, the static stability factor (track width/(2*cg) is usually lower with these vehicles. i. It is more difficult to keep an SUV from rolling after it hits a guardrail. ii. Car based or Unibody constructed SUVs seem to have lower CGs and structures. Pickup based SUVs seem to have relatively higher CGs and structures.
81 o Driver behavior is a confounding issue. SUV drivers may behave differently than automobile drivers. Also, it may make sense to examine only driver fatalities to control for the fact that SUVs may transport more occupants. This characteristic has not been proven. o It is impossible to say definitively that a barrier caused a rollover. It would be nice to know whether the SUVs rolled over before guardrail engagement or if the guardrail was a contributing factor to the rollover. Currently there is not enough granularity in the data (in some cases there is often not enough cases or data presented). i. In order to address this further, NASS data needs to be examined to see evidence where the outcome would have been different had a different guardrail been installed. These cases would have the most harmful collision with a guardrail. o Impacts with guardrails are often difficult to test and computer simulations can help model the circumstances surrounding rollover. When doing a computer simulation of an SUV into a guardrail, a real world scenario must be modeled. o In terms of frontal compatibility, the structure of the pickup trucks and SUVs needs to be lower to engage the current guardrails properly. Currently there is no bumper standard with SUVs (unlike cars). It is not necessary to lower the CG, just the vehicle structure so the vehicle can engage the guardrail. Ideally the guardrail and vehicle will distribute the forces since a more uniform distribution prevents penetration. o In hardware tests that have been performed, the thrie-beam performed far worse for the pickup than the standard W-beam - although a modified (14 inch block in vertical plane added) thrie-beam performed well. Severe failures were found with G41S - even with wood block-outs. These tests may not have been indicative of real world conditions though. During these tests, it was found that higher barriers were not necessarily safer. Also, if changes are made for the pickup truck, this may lead to serious delta-v problems for small cars. Currently, no tests have been performed with an SUV in a 25-degree 100km impact. o During rollovers, risk of ejection is high. This risk can be reduced with the introduction of side curtain airbags.
82 o Time is an important issue- in order to replace all of the guardrails in operation, it will take many years. 2. Large and compact SUVs were shown to have high fatality rates during impacts with concrete barriers. o Is there any evidence that higher safety shape (NJ or F) barrier might reduce this? i. Passenger cars would not benefit from the use of Jersey or F shape barriers. Taller barriers were not necessarily better since vehicles ride up and roll over these barriers ii. The concrete barrier held no advantage over guardrails when testing with SUVs iii. Higher barriers did have less intrusion into oncoming traffic. o Is there any evidence that the higher constant slope barriers reduce these fatality rates compared to the lower safety shapes? i. Vertical walls were best at preventing rollovers ii. Not much real world accident data exists about the type of barrier impact- either in police data or NASS. Some state DOTâs have information, but police and accident investigators rarely examine the guardrails for failure. iii. To get more information it would be best to work with state highway inspectors and receive data and pictures. Another option would be to include more information in the NASS database. 3. Buckled guardrails have penetrated into passenger compartments of SUVs and pickup trucks. What causes this? o Guardrails have seams that break, and these exposed ends may lead to increased risk of intrusion. o When the end terminals of flexible longitudinal barriers penetrate vehicles, this may not be a compatibility issue. If it is a compatibility issue, it is difficult to determine whether the guardrail penetrated due to a vehicle issue or a guardrail issue. 4. Should guardrails be tested with mid-sized SUVs due to the high fatality rate of occupants in these vehicles involved in rollovers?
83 o Many of the fatalities were caused when people were ejected from the vehicle during a rollover. Also, since SUVs roll more without engaging a guardrail, the fatalities of occupants in mid-sized SUVs might not be a guardrail issue. o The costs of testing roadside hardware with SUVs are potentially high. SUVs are expensive to buy as used vehicles. The SUV category falls within the range of the small car to large pickup. o Installation problems are larger issues. The installation is often performed improperly due to terrain constraints and the training of the construction crew. o Tests of popular guardrails should be run with SUVs - such as a 25-degree impact to see how they perform in these impacts. 5. Due to the significance of the compact SUV fatality rate, should testing with a compact SUVs replace the 820 kg car for redirection tests? o Previous tests show these vehicles pass the redirection tests and during testing have demonstrated that they are not necessarily less stable than a passenger car (vehicle characteristics do not support this claim however). Compact SUVs do have potential snagging issues though. 6. Do the current 2000 kg pickup truck and 820 kg car adequately represent the range of high sales vehicles? o These vehicles are satisfactory given the budget considerations. The categories of small car and large truck encompass most vehicles. However, testing end terminals with different classes may be worthwhile. In addition, computer analysis can be used to reduce the cost and possibly test more vehicles. 4.1.2 Assessment of Vehicle Compatibility 1. What are your thoughts on the available data sources to search for compatibility issues?
84 o More granularity is needed in the accident data in order to know what to test. Anecdotal evidence can be used to determine where the issues lie. The NASS cases can be investigated to show where vehicles are having engagement issues. o Full-scale tests are good for finding compatibility issues, however these are costly therefore more than one platform is rarely tested. o Computer simulation is also good at unearthing issues, and it is cheaper than full-scale tests. 2. How can the accident data be improved to assess the performance of roadside hardware? o Better granularity is needed in the accident data (i.e. type of barrier hit/ end treatment was used?) Pictures are also valuable tools in investigating accidents and should be included whenever possible. This is something that should be brought up to NASS. o It would be interesting to take an inventory of currently used roadside hardware and evaluating the real world performance. 3. What improvements can be made to full scale crash tests? o Dummies could be used, although this increases the number of factors that are being tested and increases that cost of the test. Since cost considerations limit the amount of tests than can be run and analyzed, the inclusion dummies might limit the amount of tests that could be run. o It is theorized that event data recorders would also give more information into what is happening in real world scenarios. With this information more realistic tests can be created. It was indicated that roadside crash events often have a very long duration. For that reason most EDRs do not capture enough information to make their data useful. o Computer simulations can test different configurations. Since these are possibly cheaper than full-scale crash tests, more scenarios can be run.
85 4. Can the design of barrier systems be modified to allow for improvements in airbag systems? o If airbags and dummies are used, is it a vehicle or a guardrail test? o If airbags are considered in guardrail development, the guardrails may be made stiffer rather than softer. This might have unintended negative consequences. However, if the hardware is made too soft, it might complicate airbag firing logic leading to less sub- optimal firing during a collision. 5. What are vehicle characteristics that may have an impact on guardrail performance? o The vehicle height (especially the height of the front structure) influences how the vehicle engages the guardrail. Also the frontal overhang of the vehicle affects the potential for snagging of the front wheel and guardrail engagement. The center of gravity and yaw moment of inertia also have an impact since these characteristics determine how the vehicle yaws and rolls. Side curtain airbags may help the occupants inside of the vehicle. In addition, ABS and other vehicle handling countermeasures like Dynamic Stability Control (DSC) may help vehicles from engaging the guardrails altogether. 4.1.3 Policy Issues Regarding Compatibility 1. What is the best course to take with regards to compatibility? o More communication within the industry of the vehicle manufacturers and hardware creators. If the hardware test is a failure, it is hard to tell if this failure is fault of the vehicle or the hardware. Often the test failures can be attributed to several different causes. Therefore it is important to change the mindset from vehicle versus hardware to vehicle and hardware working together to reduce the severity of these accidents. o The ideal scenario would use a broad range of vehicle models to test compatibility with hardware. o Passenger cars usually have no problems with guardrails since these vehicles have standard frontal structure requirements. Since SUVs do not have standard frontal structure requirements, the guardrail design for these vehicles is made more difficult.
86 2. What is the best way to resolve hardware concerns? o It would be good for FHWA to rate hardware like NHTSA rates new cars. In order to do this, a grading system needs to be developed where devices are judged on a scale opposed to current criteria where each device passes or fails at a given test level. These grades could be arrived at using data including occupant (dummy) injury values in addition to current vehicle acceleration, dynamics and deformation criteria. o There are 2 types of barriers - those that absorb energy and those that break away. Because of this, the designs need to evaluate based on performance. In particular, "In service performance" needs to be monitored to examine how well the designs are performing in real world accidents. Also, since no state wants to pay too much for roadside hardware, cost is of utmost concern. o Formal and informal communication in the industry and government is necessary to ensure that both parties are moving in the same direction regarding vehicle to guardrail impacts. 4.1.4 Computer Simulation 1. How much faith do you have in computer simulation and its ability to identify roadside safety problems? o Models are validated and perform with 80-90% similar responses, but the timing of guardrail rupture is still a shortcoming. o Currently, the validation of wood and soil models is underway. o FHWA has confidence in modeling as a prediction tool. It is a good way to see what should be tested and as a supplemental source of data from the tests. 2. What role should simulation have in the NCHRP report 350 update? o It is a tool that definitely should be used to help choose vehicles for future tests.
87 o It is a less expensive approach to examine the existing vehicle platforms used. o At the very least, it should be used as a first step in the research before full scale tests are performed. 3. Are 6-year-old vehicle models OK for tests? o Yes. The designs of vehicle structures has not changed dramatically within this period of time, however, vehicle fleet populations (relative numbers of small passenger cars vs. large cars vs. SUV's vs. Pickups) can drastically change within this period. 4. What should be the basis of vehicle selection? o Models should be made based on sales of the platform. The more common vehicles should be modeled and tested. o The mid-sized vehicle classes need to be updated and simulated based on characteristics of real world accidents (impact angles, yaw angles, speeds, etc.) 5. What are the challenges for side impact simulation? o Side impacts leave less room for energy absorption. Therefore it is necessary to look in the accident data to determine typical impact speeds that cause injuries. Conclusions and Future Steps This workshop showed that dialog between the automakers and the roadside hardware creators was valuable to share ideas regarding vehicle to guardrail impacts. More workshops like this were seen as an effective way to disseminate information to all parties interested in reducing the severity of vehicle to roadside hardware collisions. A formal committee should be put into place, possibly within SAE, that directly addresses these issues on an on-going basis.
88 In order to determine if there is a true compatibility problem, steps should be taken to not only get more data surrounding accidents with roadside hardware, but also to get more out of the data that is currently available. With increased support, more modeling can be done to address the problems of vehicle compati- bility with guardrail systems. Test methods, vehicles and criteria can be evaluated to understand if the tests truly assess the compatibility of guardrails with the most appropriate segments of the US vehicle fleet. 4.2 Crash Data Collection Accident databases that provide some insight into vehicle compatibility were evaluated in this study. The Fatality Analysis Reporting System (FARS), National Automotive Sampling System / Crashworthiness Data System (NASS/CDS) and the Highway (HSIS) databases were reviewed. It is believed that no database that exists today provides a large and complete enough data set to confidently identify compatibility issues. Clearly, as identified in other sections of this report; another try in this important area is warranted. Of each database reviewed, the NASS/CDS system contains the most complete and relevant information to assess vehicle to roadside hardware compatibility occurrences however a series of shortcomings remain. Current NASS/CDS data collection procedures focus on crash causation, vehicle handling, vehicle crashworthiness, restraint system performance, occupant characteristics and injury outcomes. Currently, over 500 crash variables are collect for the sampled crashes described above, however little attention if given to roadside attributes and safety systems. In order to adequately characterize roadside crash dynamics, suitability of installation configurations and compatibility of vehicles with roadside systems it appears that a large collection of additional variables must be considered to improve the NASS System for use by the roadside safety community for compatibility investigation. In 1983, a research project, known as the Longitudinal Barrier Special Study(LBSS), was initiated to collect additional data for NASS collected barrier crashes[29]. Additional variables were added to the NASS systems and NASS investigators were trained to collect these variables related to roadside systems. The study was successful in determining relevant information for a limited population of roadside events, however the large number of data points collected and corresponding high cost of such collections lead to the eventual termination of the study. During statistical evaluation of NASS data during data collection years following the LBSS, it was found that critical crash attributes necessary to conduct accurate analyzes were missing from the CDS
89 coded variables. These variables include pre-impact dynamics of vehicles that interact with roadside hardware systems, barrier characteristics and resulting barrier interaction/performance. Further, the concept of improper installation of barrier systems must be addressed during accident investigations. It is believed that this additional information should be gathered by NASS investigators if possible. The use of electronic photos with sufficient detail to post process case information collected is one method to significantly increase information collected while limiting time spent in the field by accident investigators. Items including barrier designs, dimensions and deformations may be extracted from photos by a roadside expert at the completion of NASS investigation. This approach would require improved photos with geometric indicators (measurement guides/rulers) and adequate labeling. A number of practical considerations remain regarding roadside investigations as well. These relate to current practices used by NASS investigators and safety concerns related to on-road investigations. Many roadside crash events occur on state roads and busy highways. In order to perform the necessary evaluations of the crash scene, investigators would be exposed to dangerous environments in some cases. This was evident during the investigation of individual cases shown in Chapter 2. Many photos were "drive by" shots of accident scenes because it was not possible to walk to the impact location. Another issue concerning barrier interactions involves the timeliness that the crash scene is reviewed by investigators. Often state DOTs repair barrier sections and impacted devices before the arrival of accident investigators. The process for selection and inclusion is outline in the NASS/CDS coding manual. It involves preliminary review of the police accident report to determine if it is eligible for study inclusion by the PSU. This process may occur in as little as one day but up to two weeks from the time of the crash. On average investigation occurs 1-2 weeks from the crash event and in some cases, limited deformed barrier data can be collected. The following forms have been created based on information found lacking in current NASS/CDS Cases for roadside safety investigations and using the Longitudinal Barrier Special Study (LBSS) data collection forms which were previously developed. A number of variables and sections have been eliminated which were found to be of limited importance for compatibility evaluation. These forms are proposed as a starting point for improved collection strategies for NASS/CDS investigations.
90 NCHRP 22-15 Proposed Roadside Form NASS/CDS Data Collection Format I. Header Variables 1. Primary Sampling Unit Number _________ 2. Case Number- Stratification _________ 3. Record Number _________ 4. Investigator I.D. _________ 5. Accident Year _________ II. Location Data 6. State _________ 7. County _________ 8. Route Number _________ 9. Mile-point _________ 10. Transaction Code _________ 11. Investigator I.D. _________ 12. Accident Year _________ III. Impact Sequence Data *See current NASS/CDS Accident Form Shown in Appendix C of this report 13. Indicate Objects Impacted (from column 2 list) and Associated Event Sequence Number as shown in Form A.2 (Variable 29) Object Number Sequence Number (1)______________ ________________ (2)______________ ________________ (3)______________ ________________ (4)______________ ________________ (5)______________ ________________ (6)______________ ________________ (7)______________ ________________ (8)______________ ________________ (9)______________ ________________ (10)_____________ ________________ 14. Total Number of Longitudinal Barrier Impacts ____ (0)-(6) Code Actual Number of Barrier Impacts ____ (7) 7 or more ____ (9) Unknown Vehicle Number or Object Contacted Codes (01-30) - Vehicle Number Non-collision (31) Overturn - rollover (excludes end-over-end) (32) Rollover - end-over-end (33) Fire or explosion (34) Jackknife (35) Other intra unit damage (specify): ___________ (36) Non-collision injury (38) Other non-collision (specify):________________ (39) Non-collision - details unknown Collision with Fixed Object (41) Tree (< 10 cm in diameter) (42) Tree (> 10 cm in diameter) (43) Shrubbery or bush (44) Embankment (45) Breakaway pole or post (any diameter) Non breakaway Pole or Post (50) Pole or post (< 10 cm in diameter) (51) Pole or post (>10 cm but < 30 cm in diameter) (52) Pole or post (>30 cm in diameter) (53) Pole or post (diameter unknown) (54) Concrete traffic barrier (55) Impact attenuator (57) Fence (58) Wall (59) Building (60) Ditch or culvert (61) Ground (62) Fire hydrant (63) Curb (64) Bridge (68) Other fixed object (specify):________________ (69) Unknown fixed object Collision with Non-fixed Object (70) Passenger car, light truck, van, or other vehicle not in transport (71) Medium/heavy truck or bus not in-transport (72) Pedestrian (73) Cyclist or cycle (74) Other non motorist or conveyance (75) Vehicle occupant (76) Animal (77) Train (78) Trailer, disconnected in-transport (79) Object fell from vehicle in-transport (88) Other non-fixed object (specify):_____________ (89) Unknown non-fixed object (90) Traffic barrier (includes guardrail) (91) Barrier End Terminal (99) Other event (specify):_____________________ Figure 4.1: Proposed General Form, Roadside Crashes
91 NCHRP 22-15 Proposed Roadside Form NASS/CDS Data Collection Format I. General Roadside Form 1. Impacted Device ____ (0) None ____ (1) Deformable Guardrail ____ (2) Other Deformable Barrier ____ (3) Concrete Barrier ____ (4) Bridge Rail ____ (5) Longitudinal Barrier End Terminal ____ (6) Barrier Transition ____ (7) Crash Cushion ____ (8) Other (specify___________________) ____ (9) Unknown 2. Location of Feature (in direction of vehicle travel) ____ (0) Impact conditions not applicable (see manual) ____ (1) Off left side of roadway ____ (2) Off right side of roadway ____ (3) Other (specify___________________) ____ (9) Unknown 3. Impact Angle (è 1- angle formed by longitudinal axis of vehicle and primary axis of feature) ____ (00)-(90) Code Actual Angle in Degrees ____ (99) Unknown 4. Separation Angle (è 2- angle formed by longitudinal axis of vehicle and primary axis of feature at last contact) ____ (00)-(90) Code Actual Angle in Degrees ____ (99) Unknown 5. Vehicle Yawing Angle at Impact(è 3- angle formed by direction of vehicle travel and longitudinal axis of vehicle) ____ (000)-(180) Code Yawing Angle in Degrees ____ (999) Unknown 6. Vehicle Rotation at Impact (ù 1-about vehicle vertical axis) ____ (1) No ____ (2) Yes ____ (9) Unknown 7. Run length of impacted treatment section ____ (00) Not Applicable ____ (01-29)Estimated Distance in Meters ____ (30) Greater than 30 meters ____ (99) Unknown 8. End Treatment Type ____ (0) None ____ (1) BCT ____ (2) Free End ____ (3) Turned Down End ____ (4) Cable with Concrete Anchor ____ (9) Unknown 9. Impact Speed (derive based on vehicle/barrier deformation) ____ (01-98)Code speed in km/h ____ (99) Unknown 10. Treatment Performance ____ (1) Vehicle Redirected by Treatment ____ (2) Vehicle snagged/pocketed by treatment ____ (3) Vehicle overrode treatment ____ (4) Vehicle vaulted treatment ____ (5) Vehicle Penetrated Treatment ____ (6) Vehicle contained by treatment ____ (7) Other (specify___________________) ____ (9) Unknown 11. Post Impact Vehicle Trajectory ____ (1) Vehicle remained on roadside ____ (2) Vehicle returned to roadway ____ (3) Vehicle crossed roadway/ran off opposite side ____ (4) Vehicle crossed median other travel way ____ (5) Vehicle remained on top of, went over or through treatment ____ (6) Other (specify___________________) ____ (9) Unknown 12. Curb Type/Presence ____ (0) No curb present ____ (1) Barrier curb ____ (2) Mountable Curb ____ (3) Other (specify___________________) ____ (9) Unknown 13. Curb Height ____ (0) No curb present ____ (00)-(49) Code actual curb height to nearest cm. ____ (0) 50 cm. or higher 14. Perpendicular Distance from Curb to Struck Feature ____ (0) No curb present ____ (000)-(996) Actual distance to nearest cm ____ (997) 25 meters or greater or greater ____ (999) Unknown 15. Height of Treatment Relative to Roadway Edge ____ (-97) -97 cm or higher ____ (-96)-(96) Code actual height of treatment relative to roadway edge to the nearest cm. ____ (97) +97 cm or higher ____ (+99) Unknown Figure 4.2: Proposed General Form, Continued
92 NCHRP 22-15 Proposed Roadside Form NASS/CDS Data Collection Format 16. Treatment Height ____ (-97) -97 cm or higher ____ (-96)-(96) Code actual height of treatment to the nearest cm. ____ (97) +97 cm or higher ____ (+99) Unknown 17. Normal Treatment Height if different from height at impact point ____ (00) Constant height ____ (00)-(99) Actual height in cm. ____ (99) Unknown 18. Treatment Damage Refer to the diagram below for recording of field measurements on barrier damage. ____ Length of Contact Damage in meters (Ld) ____ Length of Induced Damage in meters (Li) 19. Maximum Depth of Treatment Deformation ____ (0) No deformation (ie. Minor scrapes, paint transfer) ____ (000)-(999) Code actual deformation in cm. Figure 4.3: Proposed Longitudinal Barrier Data Form
93 NCHRP 22-15 Proposed Roadside Form NASS/CDS Data Collection Format Longitudinal Barrier Form Complete this section for each impact involving a longitudinal barrier. (if multiple impacts with a barrier type take place, relative location by vehicle number should be indicated by sequence number for item 1 below.) Code this form for the following Longitudinal Barrier Types a. Guardrails b. Median Barriers c. Bridge Rails 2. Sequence number of Impact with Longitudinal Barrier ____ (01)-(98) Code impact sequence number ____ (99) Unknown 3. Beam Type ____ (0) N/A- No Beam ____ (0) Cable ____ (0) âWâ Beam ____ (0) Box Beam ____ (0) Aluminum Extrusion ____ (0) Thrie Beam ____ (0) Other (specify________________.) ____ (0) Unknown 4. Beam Material 5. Beam Dimensions 6. Post Shape 7. Post Material ____ (0) Wood ____ (0) Steel ____ (0) Aluminum ____ (0) Concrete ____ (0) Fiberglass/Composite ____ (0) Plastic ____ (0) Other (Specify) 8. Blockout Type 9. Blockout Material ____ (0) Wood ____ (0) Steel ____ (0) Aluminum ____ (0) Concrete ____ (0) Fiberglass/Composite ____ (0) Plastic ____ (0) Other (Specify) 10. Post Spacing (center to center) ____ (0) N/A- No Posts ____ (0) Record actual distance from center to center in meters. ____ (0) 30 meters or greater ____ (0) Unknown 11. Post Dimensions 12. If the post spacing at the point of initial impact is different from that of the normal section of the barrier, record the normal spacing below to the nearest cm. Code the post spacing at the point of initial impact for variable B45 13. Concrete Barrier Type ____ (0) N/A- Not a Concrete Barrier ____ (1) Concrete Safety shape (indicate profile dim. cm.) ____ (2) Vertical Wall ____ (3) Constant Slope Barrier ____ (8) Other (provide sketch with dimensions) ____ (8) Unknown 14. Concrete Barrier Dimensions ____ (000) No Concrete Barrier ____ Vertical Rise ____ Lower Slope (999) Indicates unknown quantity 15. Permanent Barrier ____ (0) N/A- Not a Concrete Barrier ____ (1) Moveable Barrier (in workzone) ____ (1) Permanent Barrier 16. Portable/Moveable Barrier Connections ____ (0) N/A- Not a Moveable Barrier ____ (0) No Connections ____ (0) Pin and loop with fastening nut ____ (0) Pin and loop with no nut ____ (0) Pin and loop with fastening nut /w spacer ____ (0) Tongue and groove ____ (0) Fastening Plate ____ (0) Top C-Channel Figure 4.4: Proposed Crash Cushion Data Form
94 NCHRP 22-15 Proposed Roadside Form NASS/CDS Data Collection Format End Treatment/Crash Cushion Complete this section for each impact involving an end treatment or crash cushion. (if multiple impacts with a barrier type take place, relative location by vehicle number should be indicated by sequence number for item 1 below.) 1. Sequence number of Impact with End Treatment/Crash Cushion ____ (01)-(98) Code impact sequence number ____ (99) Unknown Barrier end-treatment/crash cushion Dimensions 2. Upstream End Treatment Type ____ (0) N/A- Impact not with Barrier End ____ (0) Blunt ____ (0) Non-Breakaway Cable Terminal ____ (0) Turndown ____ (0) Breakaway Cable Terminal (BCT) ____ (0) Anchoring to backslope ____ (0) Attached to parapet wall/bridgerail/abutment ____ (0) Best ____ (0) ET-2000/ ET-Plus ____ (0) FLEAT ____ (0) SKT ____ (0) SENTRE ____ (0) TREND ____ (0) MELT ____ (0) ELT ____ (0) REGENT ____ (0) SRT 350 ____ (0) WYBET 350 ____ (0) MELT ____ (0) VT Lowspeed ____ (0) Quad Trend 350 1. Crash Cushion Type ____ (0) Sand Barrels ____ (0) Number of Barrels ____ (0) Great ____ (0) Number of Bays ____ (0) Material (1,2,3) ____ (0) Quad Guard ____ (0) Number of Bays ____ (0) Material (1,2,3) ____ (0) React 350 ____ (0) CAT 350 ____ (0) Brake Master ____ (0) CTAS ____ (0) TRACC ____ (0) ABSORB 350 ____ (0) DRAGNET ____ (0) Other (specify_____________) ____ (0) Unknown 2. Location of End Treatment (in direction of vehicle travel) ____ (0) N/A- Impact not with Barrier End ____ (0) Upstream ____ (0) Downstream ____ (0) Other (specify_____________) ____ (0) Unknown 3. Distance From End Treatment to Initial Point of Impact ____ (0) N/A- Impact not with Barrier End ____ (0) Impact with barrier or within .5 meters of barrier ____ (01)-(96) Code Actual Distance to nearest meter ____ (99) Unknown 4. Length of Flare 5. Flare Offset 6. Performance ____ (0) N/A- Impact not with Barrier End ____ (0) Vehicle came to rest in contact w/ treatment) ____ (0) Vehicle redirected by barrier ____ (0) Energy absorbing stage in mid-stroke indicate stroke amount __________cm Figure 4.5: Proposed End Treatment Data Form
95 In order for NASS/CDS investigators to accurately distinguish specific crash cushions, barriers and end treatments, additional training in this area is necessary. Alternatively, a requirement for clear labeling on each device may facilitate data coding required by proposed Figure 4.5. 4.3 Crash Test Methodology Crash test results allow detailed evaluations of vehicle interaction with impacted roadside devices. The crashworthiness performance of roadside structures is currently is determined through crash testing according to the procedures of NCHRP Report 350 [38]. The number of required tests varies with the device ranging from 2 for longitudinal barriers, supports structures, and TMAs to 7 for terminals/crash cushions. A number of different Test Levels (TLs) are specified in Report 350. The number ranges from 6 for longitudinal barriers to 2 (TL 2 and 3) for terminals and crash cushions, support structures, and truck mounted attenuators. Currently, the basic test level (TL-3) and TL-2 require testing with a 820 kg car and a 2000 kg pickup. Higher test levels also use larger vehicles including an 8000 kg 2 axle truck, a 36,000 kg tractor/van trailer and a 36,000 kg tractor/tanker trailer. Report 350 recommends various geometric property ranges for the test vehicles for each test that provide some uniformity in vehicles used. The crash tests are conducted using a very limited number of vehicles that are not more than 6 years old. For economy of testing, it is not surprising that vehicles near the age cut off are normally used. This practice may be a significant contributor to the lag in roadside device improvement when compared with the rate that new vehicle platforms emerge. Current test methods do not include representations of occupants. Largely the longitudinal and lateral acceleration limits are designed to limit the severity of loading experienced by occupants. As new occupant restraint systems emerge including advanced frontal and side impact airbags, plus pretensioned and force limited belt systems, the crash environment will change significantly. In the case of longitudinal barrier design, increasing stiffness to avoid vehicle pocketing and penetration would have a divergent effect. First, barrier penetrations and deformations such that vehicle override becomes possible will be avoided. However, lateral accelerations during vehicle redirection following impact with a stiff barrier could increase injury risk for occupants. Newly emerging side impact airbags and improved energy absorbing vehicle side structures may mitigate the effects of this increased risk. Without the use of human surrogates during testing and analysis, the true nature of occupant loading and injury risk cannot be quantified. Similarly, airbag systems may not be well designed to trigger during oblique impacts at low angles with longitudinal barriers. If late deployments occur after an occupant has move out of position, the
96 resulting interaction with a deploying airbag could have harmful effects. Human surrogates and/or close attention to airbag deployment timing is necessary to understand this phenomenon. At this time, an effort to update NCHRP Report 350 has begun. Many of these issues discussed above should be addressed by future updates in some way. Based on current indications, specific areas to be addressed include the following: ⢠Test vehicles used ⢠Number of tests ⢠Transition/Temporary Barrier Test Conditions ⢠Higher Test Speed Reflecting 70-75 mph Speed Limits ⢠TMA Crash Test ⢠Occupant Risk ⢠Occupant Compartment Intrusion ⢠Soil Specification ⢠Side Impact Requirement
97 4.3.1 Test Vehicle Selection During this project, definite behavioral trends were observed when comparing vehicle response during impacts with roadside devices. It was determined that pickup trucks do not represent the behavior of SUVs adequately during all impact conditions. Further, the compact car category does not represent a significant population of vehicles on the roads today and therefore its use should be reconsidered. Rather, one vehicle per identifiable class should be selected for crash testing. Although this approach would greatly increase the number of tests required, the benefit in terms of lives saved and injuries reduced would greatly outweigh the financial implication of more tests. Table below provides information regarding vehicle characteristics for a series of vehicle classes. Those classes include compact, mid-size and large cars and SUVs. Compact/full-size pickups as well as mid-size and large vans. The vehicle which most closely resembles the weighted average vehicle for its class should be considered during selection of future test vehicles. This vehicle may not be the most popular, yet behavior during crashes with roadside structures would represent the mean characteristics exhibited by its class. Type Class Length(in.) Width (in.) Height (in.) Wheel Base (in.) Curb Weight (lb) Front Bump Ht. (in.) Front Overhng. (in.) Car Compact 168 65 53 96 2380 11 35 Midsize 187 70 53 104 3160 11 39 Large 206 74 55 114 3832 12 41 Pickup Compact 187 67 64 113 3039 15 31 Large 213 77 71 132 4269 18 34 SUV Compact 158 66 67 95 2849 13 28 Midsize 178 70 69 105 4022 17 31 Large 196 78 73 116 4908 16 34 Van Midsize 187 72 67 112 3548 10 36 Large 200 78 78 121 4427 33 Table 4.1: Average population weighted vehicle characteristics per class A second approach to selecting a set vehicle platforms for future testing would be through a review of vehicle whose characteristics lay at the extremes of each vehicle class. Those extremes may fall at the high or low end depending on the probable worst case per test. Appendix B of this report contains all specifications for all vehicles surveyed. A quick search of these parameters would provide an indication of the vehicle platform whose specifications place it at the extreme of each group. This philosophy resembles the current approach taken during the selection of future test vehicles.
98 4.3.2 Occupant Representation During Crash Testing Federal Motor Vehicle Safety Standards (FMVSS) require that both active and passive restraint systems provide a minimum level of protection for belted and unbelted occupants. These standards have lead to the introduction of frontal driver and passenger airbags in all new vehicles and a rapid growth in the population of side impact airbag equipped vehicles. The presence of these newly emerging restraint systems in roadside crash involved vehicles may lead to vastly different occupant injury potential however their effects have not been studied. Early development of test criteria shown in NCHRP Report 350 considered only unbelted occupants who were not protected by airbag systems. The fail space model, currently used during roadside testing, limits loading through vehicle lateral and longitudinal accelerations. In addition, this model allows an occupant compartment intrusion. Although these criteria may still correspond with reasonable injury thresholds, the effect of countermeasures between the accelerating vehicle and the occupant must be further evaluated using human surrogates (dummies) during crash testing. Due to the nature of off angled impacts with deformable longitudinal barriers, one concern regarding airbag system function is their ability to sense an impact event before any occupant excursion or motion takes place. In other words, if a weak longitudinal crash pulse results following a crash with a barrier system, an occupant may move towards the steering wheel, A-Pillar or side glass before an airbag is triggered. If later during the crash the vehicle is suddenly decelerated, the airbag may then deploy. This would result in an out of position airbag deployment where occupant injury risk may be higher than during typical deployments. The possibility of these conditions must be evaluated. However current test procedures do not provide sufficient information to determine occupant kinematics during crashes. Full scale testing using instrumented human surrogates is required. 4.4 Application of Computer Simulation The maturities of Finite Element (FE) simulation using codes like LS-DYNA now make it possible to use highly complex computer models to investigate compatibility issues. In 1995, FHWA created a consortium of university research centers to develop accurate models of roadside structures. Since that time, developed FE models have been used to evaluate and improve roadside hardware design safety. Similarly, both FHWA and NHTSA have supported the development of highly detailed vehicle models for a variety of uses. These models and those that may be provided by vehicle and roadside safety manufacturers provide a wide ranging opportunity for investigations of vehicle/roadside hardware compatibility. One strategy to
99 recognize potential occurrence of incompatibility would be to exercise all available roadside models with all available vehicle models to assess overall performance. To date, the only vehicles used for roadside hardware simulation studies have been limited to those specified by NCHRP report 350 requirements. An aggressive effort by FHWA and NHTSA is recommended to maximize the number of computer models for vehicles and roadside safety features. A judicious selection process for future models developed will allow continuous monitoring of different classes of vehicles interaction with different roadside safety features. Another approach to improving compatibility would be through joint studies by FHWA and vehicle manufacturers to evaluate the performance of newly emerging vehicle platforms with the most commonly installed roadside devices. NHTSA or FHWA could request that each manufacturer provides FE models for a selection of their passenger vehicles. At the same time, FHWA could require that FE models of each public or proprietary roadside device be delivered upon approval for use in the NHS system. The process for creation and validation of these models has evolved sufficiently that any manufacturer producing roadside hardware features fit for use on US roadways should not find model creation prohibitive. Further, the safety importance of these devices should not be overshadowed by resource constraints of private and public companies developing these devices. The Centers of Excellence could be utilized when necessary to create and/or analyze features and their performance with emerging vehicles. In the event that such an ideal partnership proves to be unattainable, more modest efforts using The Centers of Excellence and other sources could be used to reduce the reliance on expensive crash tests and accident data collection/analysis. These efforts would involve exercising currently available vehicle and roadside hardware models to simulate a variety of impact scenarios and identify potential compatibility issues. Furthermore, some of the previously identified vehicle characteristics, which could potentially lead to incompatibilities, can be changed in the computer models to understand their effects on the crash outcome. A list of currently available vehicle and roadside hardware models, which could be used for such a study are listed below: List of currently or soon to be available computer Models: Vehicle Models: ⢠Geo Metro (1997 year model) ⢠Toyota RAV4 (2000) ⢠Plymouth Neon (1996)
100 ⢠Chevrolet S-10 Pickup Truck (1998) ⢠Ford Taurus (1991) ⢠Ford Taurus (2001) ⢠Honda Accord ⢠Crown Victoria ⢠Chevrolet Lumina ⢠Chevrolet C-1500/C-2500 Pickup Truck (1994) ⢠Dodge Caravan (1997) ⢠Ford Explorer ⢠Ford Econoline (1998) ⢠Ford F800 18,000 lb. Truck (1996) ⢠Freightliner Tractor/Trailer (1991) Roadside Hardware Models: ⢠Slipbase Sign Supports ⢠U-Channel Sign Supports ⢠Dual Support Sign ⢠Portable Concrete Barrier (PCB) Several designs varying in length, shape, connection types, ... etc. ⢠G41S W-Beam Guardrail multiple versions with different posts type, post height, and blockouts ⢠Bullnose ⢠Thrie Beam Guardrail
101 ⢠Three strand Cable Rail Barrier ⢠G42W W-Beam Guardrail ⢠Concrete Median Barriers (CMB) Four Shapes: F-Shape, NJ Shape, Vertical Wall, Single Slope ⢠W-Beam to CMB Transition -Four models: Thrie-beam/W-beam, Wood-post /Steel-post ⢠PCB to CMB Plate Transition ⢠Secure Mailbox
