Previous articles in Public Roads on the Interactive Highway Safety Design Model (IHSDM) have described a roadside safety structures module. (See figure 1.) This module has been dropped from IHSDM. The design of roadside safety structures is beyond the scope of the IHSDM as it is now envisioned. However, work is continuing as a separate research activity.

This article discusses some of the issues that must be resolved in the development of a design and evaluation methodology for roadside safety structures.

Background

The National Cooperative Highway Research Program (NCHRP) has recently published a number of papers that were presented at the meeting of the Transportation Research Board (TRB) Roadside Safety Features Committee in August 1994. (1)

One of the presenters discussed the evolution of roadside safety features focusing on the major milestones that have occurred in roadside safety in the last 35 years.(2) This presentation should be required reading for all professionals involved in roadside safety issues. It explains how a significant safety problem was identified and the efforts of highway safety professionals to correct the problem.

Although thousands were ultimately involved in the implementation of the roadside safety features, the bulk of the research and development was done by a relatively small group. Use of roadside safety features, which include roadside safety structures (hardware), is one of the primary means of mitigating the injury severity of run-off-road accidents.

Each decade since the 1960s has produced at least one written procedure for evaluating the safety performance of roadside safety hardware. (3-7) These procedures are based upon several assumptions made in the early 1960s.

Roadside safety hardware was supposed to: (1) smoothly redirect the vehicle, (2) breakaway upon vehicle impact, or (3) bring the vehicle to a controlled stop. Evaluation of the performance of the hardware would be based on the results of crash tests.

Since it would be prohibitively expensive to test all vehicles under all impact conditions, "practical" worst case scenarios were developed. Two classes of automobiles were chosen to cover the range of all light motor vehicles. It was felt that by testing vehicles at the extreme ends of the vehicle fleet, all vehicles would be covered. Impact speed and angles chosen for crash tests were also "practical" worst case scenarios.

Most of the safety advances since the 1960s have been evolutionary. They built on and refined the original safety feature concepts.

A number of events have occurred since the initial assumptions were made in the 1960s that affect these early decisions. A short list would include the following:

• Efforts of the National Highway Traffic Safety Administration (NHTSA) to improve vehicle crashworthiness (Federal Motor Vehicle Safety Standards).
• Corporate Annual Fuel Economy (CAFE) standards, which have led to a reduction in vehicle size and weight.(8)
• Emergence of light-duty pickups and vans as a significant part of the fleet.
• Rapid increase in the computational power of desktop computers.
• Partnership for the Next Generation of Vehicle (PNGV) Program, a cooperative research program of the federal government and the U.S. Council for Automotive Research (USCAR) to increase competitiveness and fuel efficiency. (9)

In view of these events and our safety experience with roadside safety hardware over the last 35 years, there is a need to reexamine the philosophy upon which the evaluation of the safety performance of roadside safety hardware is based.

Current Procedure for Developing Roadside Safety Structures

This discussion is limited to the current procedure, discussed in NCHRP Report 350, for approving roadside safety hardware.

The document was prepared by contract under the supervision of an NCHRP committee. The committee consisted of representatives of three state departments of transportation, one county representative, one city representative, two Federal Highway Administration (FHWA) employees, one representative of the hardware manufacturers, one international representative, one member from academia, and two staff members from TRB. There were no representatives from either the automobile industry or NHTSA.

NCHRP Report 350 is an update of NCHRP Report 230. It is a consensus document based largely upon experience and engineering judgment. The report establishes three criteria for evaluating the safety performance of roadside safety hardware: structural adequacy, occupant risk, and post-impact vehicle response.

It is clear from the evaluation criteria that the vehicle/roadside safety hardware is a design system. The vehicle moves through three phases: preimpact, impact, and post-impact. Currently, the evaluation criteria ignores the pre -impact conditions but assumes the vehicle is in a tracking mode. The impact phase deals with the interaction between the vehicle and the hardware and the associated occupant risk. The post-impact looks at occupant risk (based on snagging that's when the vehicle gets caught on/by the hardware) and the vehicle trajectory after it leaves the hardware. The evaluation criteria deal with the impact and post-impact conditions. The final evaluation is somewhat subjective and based largely on the response of the vehicle.

Evaluations of the safety performance are based on crash tests. NCHRP Report 350 describes the vehicles to be used in testing, the test conditions, and the instrumentation that will be used in testing the hardware. The testing criteria are hardware-specific longitudinal barriers, terminals and crash cushions, and support structures and require multiple tests under different impact conditions. (See NCHRP Report 350 for details.)

There are a number of problem areas associated with using full-scale crash tests to evaluate the performance of roadside safety hardware. These include:

• Crash tests are not completely reproducible. The results may vary because of changes in impact speeds, angles, etc. Even under identical test conditions, different vehicles within the same platform may produce different results.
• A systematic method of assigning severity indices to roadside safety hardware does not exist. The outcome of the current procedure depends on the characteristics of the vehicle: weight, center of gravity, aerodynamic profile, etc. According to NCHRP 350, "The safety performance of a highway feature cannot be measured by a series of crash tests only." Unfortunately, real-world performance data on all safety hardware are sadly lacking. On new hardware, "forensic" evidence should be collected on each accident involving the new hardware until its performance in the field has been demonstrated.

Summary of Criteria for Evaluating the Safety Performance of Roadside Safety Hardware

Structural Adequacy

• Test article contains and redirects the vehicle.
• Test article activates in a predictable manner.
• Test article redirects, controls penetration, or brings vehicle to a controlled stop.

Occupant Risk

• Debris from test article should not pose a threat to driver or bystanders.
• Debris from test article should not block the driver's vision.
• Vehicle shall remain upright.
• Preferred and maximum occupant impact velocities in meters per second (m/s) based upon an unrestrained, front-seat occupant calculated from vehicle accelerations.
• Preferred and maximum occupant ridedown accelerations (g's) based upon an unrestrained, front-seat occupant calculated from vehicle accelerations.
• Hybrid III dummy (optional) test for frontal or head-on impacts.

Post-Impact Vehicular Response

• Vehicle does not intrude into adjacent traffic lanes.
• Occupant impact velocity (not to exceed 12 m/s in the longitudinal direction) and occupant ridedown acceleration (not to exceed 20 g's in the longitudinal direction).
• Exit angle from test article (not to exceed 60 percent of impact angle).
• Vehicle trajectory behind the test article is acceptable.

• Accident studies suggest that many vehicle are yawing, rolling, and pitching at the time of impact. Current testing is done in the tracking mode.

• Test vehicles are chosen to cover or bracket most of the vehicle population. The variety of vehicles on the road makes this difficult, if not impossible. The spread in vehicle types is even greater today than in the past.

• The test procedure does not encourage the use of new vehicles for crash tests, and changes in the vehicle fleet can quickly make the roadside safety hardware obsolete. The Breakaway Cable Terminal (BCT) is an example of how the safety hardware can become obsolete as a result of changes in the vehicle fleet. About 500,000 BCTs have been installed, and now we know they do not work well with wedge-shaped vehicles or with light vehicles. BCTs have not passed the NCHRP 350 criteria when tested with the 2000P vehicle (pickup).

• Testing and development of hardware is done in isolation. FHWA and some hardware manufacturers do cooperate, but involvement by the automobile industry and NHTSA has been lacking. Consequently, the responses to changes in the vehicle fleet have been largely reactive.

• Testing is done under "practical" worst case conditions. There is some evidence that the critical impact conditions may be vehicle-specific.

 

Historically, use of finite element analysis (FEA) to design and evaluate roadside safety hardware has attracted considerable interest. The Highway Vehicle Object Simulation Model (HVOSM) was developed in the 1960s, and the BARRIER VII program was developed in the 1970s.(10,11) However, FEA to date has focused on replication of crash tests in an effort to better understand the crash phenomenon.

The use of FEA to analyze specific hardware and identify design changes that will improve the performance has been limited. The use of FEA as a tool for evaluating the safety performance and accepting the hardware for use has not been done. Past FEA models can be divided into two categories: (12)

• Impact Models WRECKER, Barrier VII, GUARD, CRUNCH, NARD.

• Handling Models HVOSM (RD2 and VD2 versions).

These specialized models had several limitations. The limited computational power available required many simplifying assumptions. Due to their specialized nature, there were few users of these models. These programs were not user-friendly. They were model building tools, and the quality of the simulation depended heavily on the skill of the analyst.

In summary, the current procedure for the evaluation of roadside safety hardware is based upon crash tests conducted in accordance with NCHRP 350 and the comparison of the crash test results with the evaluation criteria contained in NCHRP 350. NCHRP 350 is based upon "practical" worst case scenarios. Two vehicles are used to try to bracket the light-duty fleet as a whole, and the impact conditions chosen are for extreme conditions.

This procedure is doomed from the start. The development and installation of roadside safety structures can never keep pace with the changing vehicle fleet.

Future Procedures

Although it is difficult to define which future procedure will be used to design roadside safety hardware, it is possible to identify trends that will continue. Computer power will continue to increase, and predicting vehicle characteristics and the vehicle fleet mix of the future is uncertain.

The future procedure should build on our existing knowledge; to the extent possible, eliminate past mistakes; and ensure that future vehicles and hardware are compatible. The future procedure for evaluating the safety performance of roadside safety hardware will resemble the current program in many respects. It will be based upon assumptions; it will require some sort of performance standards; and it will involve full-scale crash tests and finite element analysis. There are many factors that must be discussed and resolved.

The assumptions of the 1960s need to be reexamined. Are the current assumptions still valid? Are there better assumptions?

Recent work indicates that guardrail ends are 40 percent more hazardous than the line-of-run guardrail.(13) Specific attention needs to be focused on terminals. Currently, terminals are described in NCHRP 350 as either "Terminals and Redirective Crash Cushions" or "Nonredirective Crash Cushions". Which type of terminal is safer? Should there only be one type?

The line-of-run guardrail is designed to redirect the vehicle. Vehicles are either redirected parallel to the barrier or back into the traffic stream. What hazards does this pose to the vehicle occupants? What hazards does this pose to other users of the highway? Is there anything we can learn from accident data that provides insight into these problems? Should all errant vehicles that impact hardware be brought to a controlled stop?

These are key issues that deal with the performance of the hardware. Equally important is the design system the vehicle and the hardware. As noted earlier, NCHRP 350 specifies crash tests that use an 820C vehicle (Honda Civic and Ford Festiva) or a 2000P vehicle. These vehicles were chosen because they appear to bracket the existing vehicle fleet. Are these good choices? The risk of occupant injury during impact depends to a large extent upon the crashworthiness of the impacting vehicle. Should the most popular vehicle be used for evaluation and a relative ranking developed for all other vehicles? Should future vehicles have specific features designed to work with roadside safety structures?

Observation of recent crash test films have raised serious questions about the test vehicles themselves. In recent tests using pickup trucks (2000P), it appears that subsequent rollovers are caused by a damaged wheel system. What is being tested the hardware or the test vehicles? Should crash tests be used to evaluate roadside safety hardware? Should a standardized test (similar to NHTSA's deformable barrier test) be developed? Should we develop surrogate vehicles and use them to test the system?

It now appears that one of the most promising techniques for evaluating roadside safety hardware is FEA. Today, FHWA and NHTSA use Lawrence Livermore National Laboratory's DYNA3D and Livermore Software Technologies' LSDYNA3D to study crash impacts. The motor vehicle industry also uses (among other methods), these same tools to design and evaluate motor vehicles. Preliminary findings would indicate that FEA has the potential to both improve the design of roadside safety hardware and evaluate safety performance.

Given the difficulties associated with crash tests, is FEA a better technique? Is it affordable? Does it provide consistent and accurate data? How should NHTSA's program on crashworthiness be factored into the development of roadside safety hardware?

 

 

One of the major problems associated with FEA is the development of finite element models of motor vehicles. A limited number of finite element vehicle models have recently been developed to replicate the 820C vehicle. A model of a 2000P vehicle is under development at the National Crash Analysis Center at The George Washington University, Virginia Campus. These are very complicated models. Should FHWA and NHTSA use the same vehicle models? Can the automobile manufacturers supply finite element models for testing? Should testing be done with future prototype models, perhaps from the PNGV program?

Finally, in the development of a new procedure for the evaluation of the safety performance of roadside safety hardware, input must be sought from all of those involved in the motor vehicle/roadside safety hardware design problem. The vehicle manufacturer must develop safer vehicles that can compete in a global economy. NHTSA is involved in research to improve the crashworthiness of the motor vehicle, FHWA and the states develop standards for highway design and operation. Manufacturers of roadside safety hardware are challenged to develop hardware that provides safe operation for a multitude of vehicle platforms.

Any future program should recognize the contributions that each group makes and build upon the strengths of each group.

References:

1. "Roadside Safety Issues," Transportation Research Circular 435, Transportation Research Board/National Research Council, Washington, D.C., January 1995.

2. H.E. Ross Jr. "Evolution of Roadside Safety Features," Transportation Research Circular 435, Transportation Research Board/National Research Council, Washington, D.C., January 1995.

3. "Proposed Full-Scale Testing Procedures for Guardrails," Circular 482, Highway Research Board, Washington, D.C., September 1962.

4. M.E. Bronstad and J.D. Michie. "Recommended Procedures for Vehicle Crash Testing Highway Appurtenances," NCHRP Report 153, Transportation Research Board, Washington, D.C., 1974.

5. "Recommended Procedures for Vehicle Crash Testing of Highway Appurtenances," Transportation Research Circular 191, Transportation Research Board/National Research Council, Washington, D.C., 1978.

6. J.D. Michie. "Recommended Procedures for the Safety Performance Evaluation of Highway Appurtenances," NCHRP Report 230, Transportation Research Board, Washington, D.C., 1981.

7. H.E. Ross, D.L. Sicking, R.A. Zimmer, and J.D. Michie. "Recommended Procedures for the Safety Performance Evaluation of Highway Features," NCHRP Report 350, Transportation Research Board, Washington, D.C., 1993.

8. Report to Congress on CAFE Standards.

9. Review of the Research Program of the Partnership for the Next Generation of Vehicles, Transportation Research Board/National Research Council, Washington, D.C., 1994

10. R.R. McHenry and N.J. Delays. "Automobile Dynamics A Computer Simulation of Three-Dimensional Motions for Use in Studies of Braking Systems and the Driving Task," Calspan Report No. VJ-2251-V-7, August 1970.

11. G.H. Powell. "General Computer Program for Analysis of Automobile Barriers," Record No. 343, Highway Research Board, Washington, D.C., 1971.

12. An Upgrade Plan for FHWA Roadway Safety Simulation Models , Publication No. FHWA-RD-93-189, Federal Highway Administration, Washington, D.C., December 1994.

13. J.G. Viner. "The Roadside Safety Problem," Transportation Research Circular 435, Transportation Research Board/National Research Council, Washington, D.C., January 1995.

Jerry Reagan is the chief of the Design Concepts Research Division. Prior to this assignment, he served as chief of the Safety Traffic Implementation Division. He has had a variety of experiences with FHWA, beginning in 1967 as a materials engineer. He has a bachelor's degree and a master's degree in civil engineering from the University of Tennessee. He is a licensed professional engineer in Tennessee and Virginia.