CHAPTER 6

Research Conclusions and Guidelines for Longitudinal Barriers on CSORs

A significant amount of information was derived in this project from the vehicle dynamics analyses, crash simulation analyses, and crash testing relative to vehicle-to-barrier interfaces and crashworthiness of longitudinal barriers used on CSORs. This effort extended insights on the effectiveness of barriers when deployed on typical highway curves to those on the shorter radius curves that characterize major highway interchange ramps. In these situations, space is often inadequate to construct ramps with longer radii to facilitate driving and enhance safety. Drivers exiting high-speed highways often do not appropriately lower their speeds from mainline levels as they negotiate off-ramps, and they find themselves at risk of running off the road.

Longitudinal barriers are often placed on ramps, but doing so requires consideration of varying shoulder and roadside conditions and the implications of these on barrier selection and placement. The effects of these conditions were identified in earlier research on barrier placement for highway horizontal curves (3). This report documents the research efforts undertaken and the insights derived for the selection, design, and placement of longitudinal barriers for use on CSORs. A summary of the findings of this research and thoughts on applying them are noted in this chapter. These set the stage for groups such as AASHTO to consider the safety implications for CSORs and promote efforts to reflect these findings into current guidelines for appropriate design, selection, and installation of longitudinal barriers on CSORs.

6.1 Analyses of Vehicle Dynamics Effects of Barrier Interface on CSOR Ramps

Various CSOR scenarios were evaluated using VDA and crash simulation to assess the effects of various CSOR features on barrier effectiveness. These efforts considered the effects of impacts by small cars, large pickup trucks, and SUTs negotiating CSOR ramps. VDA and crash simulations were conducted to reflect various parameters, including the following conditions that characterize CSOR ramps:

• CSOR ramp radii: 150 to 450 ft
• Ramp superelevation: 4% to 8%
• Ramp grade: 0% to −8%
• Shoulder width: 4 to 12 ft
• Shoulder angle: 0% to 8%
• Impact speed: 50 to 120 km/h
• Impact angles: 5 to 25 degrees
• Vehicle weights: 1,100 kg, 2,270 kg, and 10,000 kg

In this effort, VDA was first used to focus on interface variations for vehicles running off the road for four common types of barriers used on CSORs, namely W-beam guardrails (G4(1S))

and W-beam (MGS), Thrie beam (SGR09b), and concrete barriers. Interface envelopes were determined and tabulated to provide the basic minimum and maximum barrier heights for effectiveness on different CSORs. Effectiveness was shown to be a function of the grade and superelevation of the CSORs, as well as shoulder features and barrier type and position. Errant vehicles on CSORs take paths that can lead to sharper impact angles over varying surface slopes, and roadside slope can influence barrier installation. Project efforts included comprehensive analyses of the dynamics of vehicles leaving the ramp roadway and traversing the shoulder before encountering the roadside barrier. An understanding of these factors was needed to address the simple question: Will the barrier provide an effective interface with an errant vehicle?

The degree of slope change between the ramp superelevation and the shoulder was found to change the effective vehicle-to-barrier interface area on CSOR ramps. An immediately useful product of this effort is the interface tables that reflect a broad range of curvature, shoulder widths and angles, grades and superelevation, and barrier placement guidelines. The minimum and maximum measures for vehicle-to-barrier impacts allow an agency to analytically evaluate CSOR location and determine whether the variations in the effects of CSOR design parameters can lead to inadequate interfaces for the type of barrier deployed or considered. The indicated “effective” vertical coverage of the barriers for an errant vehicle on a given CSOR ramp can be evaluated by comparing measurements of the top and the bottom of the barrier with the maximum and minimum effective heights. Where the potential exists for a poor vehicle-to-barrier interface, actions can be initiated to make changes. In the tables, those situations shaded in red flag conditions that need attention to promote safety for typical vehicles. This information provides useful guidelines to practitioners for the selection, design, and installation of the barriers for the features of a CSOR location. The tables cover a broad range of conditions to provide usefulness in design, installation, and maintenance efforts. This information can also support safety investigations. The results can be extended to other barriers that have similar dimensions and placement guidelines.

Crash simulation analyses were used to determine whether the barrier would provide adequate strength to capture or redirect an errant vehicle across various combinations of CSOR geometrics and barrier impact conditions. The simulation software generated detailed metrics for each crash scenario over the duration of the crash event. The numerical results were stored in separate computer files to allow various types of analyses, but the critical end-state metrics were captured to generate a basic visual performance summary for each case. These summaries show the effects of the impact on the vehicles’ position as well as provide the metrics associated with crash physics. The computed metrics were also compared with acceptance requirements to indicate whether safety requirements were met.

Performance summaries were generated to provide a diagram showing the crash and its aftermath as well as the comparison of critical metrics with acceptability. These visualized results are a summary of multiple simulations that reflect the vehicle dynamics and crash physics of the impact with the barrier during the 2- to 3-s crash event. On each summary, the graphic shows views of the vehicle’s trajectory for a given ramp curve and surface, as well as the barrier installation. Each summary also shows the conditions from different views. Although the details are limited for these page-size views, these impacts can be viewed as a video to provide more clarity. These views were generated to show the variations in the impact effects and barrier deformations for the various conditions.

6.2 CSOR Simulations

Using a basic simulation model for a vehicle impacting a barrier on a superelevated curve, the research team executed more than 500 simulations reflecting the impact of a vehicle with a particular type of barrier for various ramp configurations. The conditions reflected the various combinations of barriers, CSOR designs, and barrier placement guidelines for typical MASH

analyses with small and large vehicles. The simulation software undertook the computations of the movement of the vehicle into contact with the barrier and then interactions of the crushing or deformations of the vehicle and barrier for every second for a crash event period of about 2 s. These runs each took about 5 h to complete, but the many micro-level outputs provided the basis to visualize the details of the crash event, follow the trajectory of the vehicle, and monitor the impact effects on the barrier. For most combinations of the factors, a finite element simulation was performed. The conditions that occurred as a result of the impact of the vehicle as influenced by the road and shoulder features, the barrier type, and placement were computed.

The simulation runs generated detailed data reflecting changes in elements, the overall forces acting on the barrier and the vehicle, the effects of the crash, material strengths, the trace of the vehicle during the crash event, and many other factors. Ancillary software allowed the crash event to be visualized. The vehicle’s position heading into impact with the barrier was noted, as was the point of contact with the barrier. The visualization function allowed the roll, pitch, and yaw effects to be observed for each crash simulation. The visualization display also allowed the crash event to be viewed from different sides, as shown. This helps observers to understand the effects on the vehicle as well as the barrier. The visualization tools allowed these views to be expanded to evaluate detailed aspects, such as the bumper snagging on a post or wheels climbing the face of a concrete barrier. The full set of these results is available from the research team.

The performance summary charts included in Appendices D through F report the analysis results from each simulation reflecting a particular barrier, CSOR design, and vehicle type. Across the top of the chart, the vehicle type and barrier setup are noted. In the second line, the features of the curve and barrier installation are noted. The lower parts of these individual crash simulation summaries show the results of a standard MASH evaluation of the efficacy of the specific barrier. A large number of conditions were considered, and based on the cumulative results, various barriers were determined to be effective for most CSOR ramp situations. The following sections offer recommendations for adopting the findings of these efforts, their application to the highway design process, and the needs for future research.

6.3 Proposed CSOR Guidelines

The derived findings of the various aspects of the project provide useful insights and examples for performance envelopes that may be evaluated and considered for an update to the AASHTO Roadside Design Guide (2). The intent of this effort was to analyze concerns and problems related to highway safety for a given aspect of the highway system. The efforts considered various questions for installations of longitudinal barriers on CSORs. The results cover a wide range of CSOR situations. The research team addressed a representative cross section of conditions, undertook a rigorous set of analyses, presented detailed results, and formulated recommendations to improve standards and practices for further discussion and debate. The research team made every effort to ensure that the recommendations are comprehensive, concise, and well supported.

AASHTO and others may carefully review the findings in the context of their considerable direct experience to consider the results of these efforts, tempered by experiences, and the reality of agency resources and priorities. The results of this effort may provide some agencies insights and support for addressing lingering needs by applying the results to problem locations. Over time, if evaluated and considered appropriate, the results may be generalized and considered for inclusion as an update to the highway safety guidelines for states—as well as for the nation—provided by the Roadside Design Guide (2).

Table 19 summarizes the significant implications and guidelines derived for the barriers and CSOR conditions analyzed. The guidelines imply an understanding of the implications of

Table 19. CSOR implications and guidelines derived from NCHRP Project 22-29B efforts and results.

vehicle-to-barrier impacts on CSORs. These implications are included, along with the critical elements of guidelines (in bold) that evolved from this research. It is hoped that this construct offers a useful means to summarize the findings of the multi-faceted analyses and those related findings that support the recommended guidelines for barrier design, selection, and installation.

6.4 Conclusions

In continuing research to enhance efforts to ensure that highways are designed and maintained to provide a high level of safety, the research team recognized that understanding of the influences of CSOR features on safety was limited. The research team found that physics-based criteria had been created to determine appropriate curvature and banking parameters to allow vehicles to safely negotiate curves under varying surface conditions. Criteria for basic curve design are found in the AASHTO Green Book (1). It was noted, however, that limited guidelines were available for addressing concerns about vehicles leaving the roadway under CSOR conditions.

While it is basic understanding that crashes occur more often on curves than tangent sections, the influences of CSOR features on crash propensity were unclear. It was noted that a fundamental issue exists with the level of details associated with crash reporting that limit analysis options. The usual data captured for crashes falls short on details about the features of the road at or upstream of the crash location. In some cases, basic features are provided on crash reports (e.g., pavement condition), but rarely are details on grade, curvature, or basic features captured. The limited capability to analyze CSOR crashes is understandable, as needed data items are not routinely captured. The problem occurs even if an agency has data on road features but cannot link it to specific crash sites.

Understanding has been growing about the dynamics of vehicles as they traverse specific surfaces, but such analyses have not typically been undertaken in crash analysis efforts, despite the availability of software tools for the purpose. It is also understandable that the sophisticated simulation tools that allow the physics of vehicle dynamics and vehicle-to-barrier impacts to be analyzed are not applied because funds and in-house capabilities may be limited. The interest in understanding the safety performance of barriers on CSORs provides an impetus for using advanced tools when ordinary research approaches are limited.

This effort was undertaken in three phases to rigorously generate the insights needed to enhance the understanding of the safety performance of barriers on CSORs and develop guidelines for their effective design, selection, and installation. The following insights resulted from this research:

• Previously, little effort had been made specifically to determine whether longitudinal barriers adjacent to CSORs perform the same as on tangent sections.
• Current guidelines for barrier design, selection, and maintenance are unclear but are assumed to be the same as for tangent sections.
• VDA using commercially available tools provides a means to study the effects of speed, surface features, and vehicle type on the trajectory and orientations of a vehicle departing the traveled way on a CSOR.
• Vehicle trajectories for two types of vehicles on roads at different speeds were studied to relate them to the nature of the interface with barriers at varying positions along the road, as well as on interchange ramps.
• VDA provided useful information on vehicle-to-barrier interfaces for a range of CSOR conditions. These analyses can serve many useful functions for DOTs looking to improve safety on CSORs.

• VDA results were used to determine situations that warranted deeper analyses using simulation.
• Finite element simulations were undertaken to investigate the impact performance (i.e., physics) of selected vehicles actually impacting one of three types of barriers placed on a CSOR. This effort generated many simulations that may be a useful resource for studying problem locations.
• The simulation analyses focused on varied impact conditions to evaluate the performance of New Jersey concrete, G4(1S) W-beam, and MGS barriers using MASH criteria.
• The results indicated some potential for failures, but options for addressing the problems exist.
• Full-scale crash tests were conducted, which were deemed to validate the simulation analyses. These also demonstrated approaches for conducting future tests of barriers that may be used on CSORs.

The findings from all three aspects of the research were summarized and translated into proposed actions that could increase barrier safety on CSORs. Needs for future research were also defined. This final report documents the analyses and results from the project. These provide the necessary understanding of barrier effectiveness that can be applied to generate or update agency recommendations for effective design, selection, and installation of longitudinal barriers on CSORs.

6.5 Needs for Future Research

The findings presented in this report provide a solid basis for agencies to assess and improve their guidelines and practices for deploying longitudinal barriers on CSORs. The results suggest that these findings can enhance safety, but as things change, the guidelines will need to be updated. The following list of topics may warrant future efforts to expand the guidelines, address changing conditions, and consider new barrier treatments and vehicle capabilities:

• Assess other conditions at CSOR sites to expand guidelines and understanding of impact likelihood and consequences:
  • Formulate tables structured to reflect other conditions or related aspects, such as transitions.
  • Assess varied hardware.
  • Consider new or emerging vehicle types.
  • Consider other factors (e.g., vehicle loading, driver input, braking, slippery roads) and a combination of these factors (e.g., yawing and braking).
• Aid agencies in addressing the adequacy of their design guidelines.
• Consider the implications of new and alternative barriers in testing and approval routines for their applicability on CSORs.
• Address crash testing needs:
  • Consider testing conditions that improve representation of ramp conditions.
  • Consider needs for additional imagery, data items, and impact conditions.
  • Define testing requirements specific to barriers on CSORs to include in MASH.
• Incorporate new research findings:
  • VDA and crash simulation tools provide important perspectives and data on CSOR crashes and a means to assess and compare specific conditions or agency standards.
  • Tabular summaries of effective interface areas provide fundamental guidelines for barrier selection and deployment. Shaded areas highlight critical conditions.
  • Establish a basis for developing standards for barriers on ramps that consider ramp features.
  • Demonstrate viability of testing protocols for basic conditions.
  • Identify the importance of and needs for improved barrier installation guidelines for CSORs.

Overall, the findings provide valuable insights for improving the design, selection, and installation of longitudinal barriers on CSORs. However, additional research and updates to guidelines may be needed as conditions change, new barrier treatments are developed, and vehicle capabilities evolve.