LONG-SPAN GUARDRAIL SYSTEM FOR CULVERT APPLICATIONS

Transportation Research Record 1720 ■ Paper No. 00 - 0598 19 Long-Span Guardrail System for Culvert Applications Ronald K. Faller, Dean L. Sicking, Karla A. Polivka, John R. Rohde, and Bob W. Bielenberg A long-span guardrail for use over low-fill culverts was developed and successfully crash tested. The guardrail system was configured with 30.48 m of nested, 12-gauge W-beam rail and centered around a 7.62-m-long unsupported span. The nested W-beam rail was supported by 16 W152 × 13.4 steel posts and 6 standard CRT posts, each with two 150-mm × 200 × 360 mm wood block-outs. Each post was 1830 mm long. Post spacings were 1905 mm on center, except for the 7.62-m spacing between the two CRT posts surrounding the long span. The research study included computer simulation modeling with Barrier VII and full-scale vehicle crash testing, using 3/4-ton (680-kg) pickup trucks in accordance with the Test Level 3 (TL-3) requirements specified in NCHRP Report 350. Three full-scale vehicle crash tests were performed. The first test was unsuccessful because of severe vehicle penetration into the guardrail system. This penetration resulted from a loss of rail tensile capacity during vehicle redirection when the swagged fitting on the cable anchor assembly failed. A second test was performed on the same design, which contained a new cable anchor assembly. During vehicle redirection, the pickup truck rolled over and the test was considered a failure. The long-span system was subsequently redesigned to incorporate double block-outs on the CRT posts and crash tested again. Following the successful third test, the long-span guardrail system was determined to meet TL-3 criteria. Highway designers continually face the task of diverting water or runoff from one side of the road to the other. Cross-drainage structures are commonly used in these situations and are designed to carry water underneath the roadway embankment. These structures generally consist of small concrete or corrugated metal pipes, but often they may be multibarreled, concrete box culverts or structural plate pipes (1). For the smaller pipe structures, the inlets and outlets generally consist of beveled or sloped end sections; however, for the larger box culverts, the end sections are constructed with concrete headwalls and wing walls. Consequently, these cross-drainage structures may be a serious safety concern for motorists who inadvertently leave the roadway and impact or pass over these obstructions. Over the years, several alternatives have been used to prevent serious accidents from impacts with these cross-drainage structures. These alternatives have included extending the culvert, using a traversable design, or shielding the structure. The preferred treatment for any obstacle is removal, eliminating the potential for an impact between a vehicle and hazard; however, removal is often not an option. If so, the obstruction must be placed as far from the roadway as possible, commonly at or beyond the clear zone. This alternaR. K. Faller, K. A. Polivka, and B. W. Bielenberg, Midwest Roadside Safety Facility, University of Nebraska–Lincoln, 1901 Y Street, Building C, Lincoln, NE 68588-0601. D. L. Sicking and J. R. Rohde, Midwest Roadside Safety Facility, University of Nebraska–Lincoln, W328.1 Nebraska Hall, Lincoln, NE 685880529. tive may not always be feasible because of costs in either purchasing additional right-of-way or increased earthwork requirements. For cross-drainage structures placed on traversable slopes, the preferred treatment is to adjust the structure length to intersect with the roadway embankment and adapt the bevel of the culvert’s inlet and outlet to the embankment’s slope. For smaller culverts with a single round pipe 1000 mm or less in diameter or multiple round pipes 750 mm or less in diameter, no additional treatment is required (1). However, for larger culvert openings wider than 1 m, steel bar grates or pipes are required to prevent a vehicle and its wheel assembly from significantly penetrating the opening, potentially resulting in vehicle instability or increased decelerations. Full-scale crash tests have shown that automobiles can safely traverse cross-grated culvert end sections located on slopes as steep as 1:3 and at speeds between 32.2 and 96.5 km/h using steel Schedule 40 pipes spaced 762 mm on center (2, 3). For larger culvert structures, extending the culvert or making it traversable is often not feasible. Therefore, the culvert obstruction must be shielded with an effective, crashworthy barrier system. Barrier systems used in these situations have included strong-post guardrails with steel posts bolted to the top of the concrete culvert, nesting of guardrail and reducing post spacing combined with the rigid steel post option, and using long, unsupported guardrail spans over the culvert (4–7 ). The use of barrier systems is often hindered by the shallowness of soil fill over the culvert. Crash testing has demonstrated that posts with shallow embedment depths can easily be pulled out of the ground, resulting in vehicle snagging or vaulting and causing potentially disastrous results (4). The Texas Transportation Institute (TTI) successfully developed and crash tested a design that alleviates the diminished performance of the guardrail with shallow embedded posts using a passenger-size sedan (4). This design involved welding base plates to the short W152 × 13.4 steel posts and bolting them to the top surface of the concrete culvert. For this design, the back side of the steel posts was placed approximately 483 mm from the traffic-side face of the culvert headwall, allowing space for guardrail and post deflection during impact. This design sometimes required extending the culvert outward from the roadway, which increased the cost of the structure, especially in rehabilitation projects with no other culvert work required. In 1992, an alternative steel post design was developed for the Kansas Department of Transportation (KsDOT) to provide a stiffer barrier and reduced lateral deflection over the culvert (5). The successfully crashtested design consisted of a nested W-beam, half-post spacing, and steel posts bolted to the top of the concrete culvert and installed next to the culvert headwall. For an impact with a passenger-size sedan, lateral dynamic guardrail deflections were reduced from 820 to 473 mm for the TTI design compared with the KsDOT design. These rigid steel post were severely deformed and often pulled loose 20 Paper No. 00 - 0598 when impacted by vehicles, significantly damaging the culvert and incurring expensive repairs. To eliminate the need for a physical attachment to the concrete culvert and to develop a wood post guardrail option, the TTI designed a crashworthy system that included an unsupported guardrail segment or long span across the culvert (6, 7). Designs were developed for 3.81- and 5.72-m long, unsupported lengths of nested W-beam guardrail and successfully crash tested using passenger-size sedans. These designs are simpler and less expensive than the steel post alternatives, which require an attachment to the top of the culvert. These designs also have been recommended for use with wood and steel post guardrail systems because of the compatible strengths of wood and steel posts (6). Although an improved and economical long-span guardrail system was developed to shield culvert obstructions, several state departments of transportation encountered situations requiring unsupported lengths exceeding 5.72 m and up to 7.62 m. In addition, the previously described designs were crash tested according to NCHRP Report 230 evaluation criteria (8). Consequently, these existing designs can no longer be used on federal-aid highways unless they meet current impact safety standards. Also, any new designs with unsupported lengths exceeding 5.72 m must be crash tested. RESEARCH OBJECTIVE The objective of this research project was to develop a new guardrail system that could traverse culverts with an approximately 7.62-mlong span. The new guardrail system was designed to meet the NCHRP Report 350 Test Level 3 (TL-3) safety performance criteria (9). The study was performed by the Midwest Roadside Safety Facility (MwRSF) in cooperation with the Midwest States Regional Pooled Fund Program, the Washington State Department of Transportation, CONTECH Construction Products, Inc., and Lane Enterprises. Additional details on this study are presented in two recently published MwRSF research reports (10, 11). TEST REQUIREMENTS AND EVALUATION CRITERIA Longitudinal barriers, such as long-span guardrail systems traversing culverts, must satisfy NCHRP Report 350 requirements before use on new construction projects or as a replacement for existing designs not meeting current safety standards. According to TL-3, long-span guardrail systems must be subjected to two full-scale vehicle crash tests—(a) a 2000-kg pickup truck impacting at a speed of 100 km/h and at an angle of 25 degrees and (b) an 820-kg small car impacting at a speed of 100 km/h and at an angle of 20 degrees. W-beam barriers struck by small cars, however, have met safety performance standards, being essentially rigid with no significant potential for occupant risk from vehicle pocketing or severe wheel snagging on the post at the downstream end of the long span (12–14). Therefore, the 820-kg small car was not crash tested for this study. Evaluation criteria for full-scale vehicle crash testing are based on three appraisal areas—structural adequacy, occupant risk, and vehicle trajectory after collision. Criteria for structural adequacy evaluate the ability of the barrier to contain, redirect, or allow controlled vehicle penetration in a predictable manner. Occupant risk evaluates the degree of hazard to occupants of the impacting vehi- Transportation Research Record 1720 cle. Vehicle trajectory after collision is a measure of the potential for a vehicle’s postimpact trajectory to result in subsequent multivehicle accidents. This criterion also indicates the potential safety hazard for the occupants of other vehicles or the occupants of the impacting vehicle when subjected to secondary collisions with other fixed objects. These three evaluation criteria are described in greater detail in NCHRP Report 350. The full-scale vehicle crash tests were conducted and are reported in accordance with Report 350 procedures. LONG-SPAN GUARDRAIL DESIGN CONSIDERATIONS The development of the long-span W-beam guardrail system considered three key factors—vehicle capture, rail tensile capacity, and the potential for pocketing and wheel snagging. For a long, unsupported length of guardrail extending across a culvert, vehicle capture is significant. Under NCHRP Report 230 guidelines, previous crash testing of W-beam guardrails demonstrated that unsupported lengths up to 5.72 m were possible. The problem of vehicle capture, however, is intensified under NCHRP Report 350 guidelines because of the increased height of the center of mass for 3⁄4-ton (680-kg) pickup trucks. For a long-span guardrail system, the rail tensile capacity also is a key design factor. Higher tensile loads and longitudinal strain in the rail will occur because dynamic lateral rail deflections are likely to be larger than those observed during an impact into a guardrail system with standard post spacing. Nested W-beam or single thriebeam, combined with increased thickness, was considered as a way to provide the increased tensile capacity required. The final design consideration was the potential for vehicle pocketing and wheel snagging on the posts located on the downstream side of the unsupported length of guardrail. The post type, embedment depth, and spacing were selected to minimize vehicle pocketing and wheel snagging. Computer simulation was used initially to investigate these design considerations to ensure that the most cost-effective guardrail system was implemented. COMPUTER SIMULATION MODELING Computer simulation modeling with Barrier VII was performed to analyze and predict the dynamic performance of various long-span guardrail alternatives before full-scale vehicle crash testing (15). The simulations modeled a 2041-kg sedan and a 1996-kg pickup truck impacting at a speed of 96.6 km/h and at an angle of 25 degrees. Computer simulations were conducted on four long-span W-beam guardrail alternatives, each involving 7.62-m open spans and different total lengths of nested, 12-gauge rail centered about the free span—19.05 m, 22.86 m, 26.67 m, and 30.48 m. Each installation was evaluated in impacts with the 3/4-ton (680-kg) pickup and the full-size sedan. Many simulations were conducted for each vehiclesystem combination to determine the impact locations that maximized wheel snagging and predicted strains in the W-beam rail elements. The maximum longitudinal strain in the W-beam was selected as the best indicator of rail rupture. Although the AASHTO M180 steel used in W-beam guardrails is a relatively ductile material and can sustain significant plastic strain without failure, fullscale crash tests have indicated that guardrails tend to fail at relatively low plastic strains. The cross section of a W-beam rail element is reduced by approximately 15 percent at the rail splice. This cross- Faller et al. Paper No. 00 - 0598 sectional reduction tends to localize strain in the splice region and leads to rail rupture near the point that the full cross section begins to yield. Full cross-sectional yield was therefore chosen as the limiting condition for the design of the long-span guardrail system. This yield would correspond to a limiting strain of approximately 0.0017. The results of the computer simulations are shown in Table 1. For every design evaluated, the sedan produced higher W-beam rail strains than the pickup truck. Further, only the 30.48-m, nested design prevented yielding in the W-beam rail element just upstream of the nested section. Computer simulation was also used to determine the critical impact point (CIP) for the long-span guardrail system. The CIP was based on the impact condition that produced the greatest potential for wheelassembly snagging or vehicle pocketing on the first post at the downstream end of the long-span section (i.e., Post 11 in Figure 1) or the greatest potential of rail rupture. For the 30.48-m, nested W-beam alternative, the computer simulation results indicate that the greatest potential for wheel snagging on the post just downstream of the longspan section would be an impact between Posts 11 and 12. The CIPs were 2134 mm and 2438 mm downstream from Post 12 for the sedan and the pickup truck, respectively. The predicted maximum dynamic lateral rail deflections for the sedan and the pickup truck were 908 mm and 839 mm, respectively, as measured to the center height of the rail. LONG-SPAN GUARDRAIL SYSTEM DESIGN DETAILS: DESIGN A The total length of the test installation was 53.34 m, as shown in Figure 1. Photographs of the test installation are shown in Figure 2. The test installation consisted of 30.48 m of nested, 12-gauge W-beam rail supported by CRT and steel posts, a standard 12-gauge TABLE 1 Barrier VII Computer Simulation Results 21 W-beam guardrail supported by steel posts, and an anchorage system replicating a breakaway cable terminal (BCT) on the upstream and downstream ends but installed tangent to the guardrail system. The system was constructed with 26 guardrail posts. Posts 3 through 8 and 15 through 24 were galvanized ASTM A36 steel W152 × 13.4 sections 1830 mm long. Posts 9 through 14 were CRT 150 mm × 200 mm × 1830 mm timber posts. Posts 1 and 2 and 25 and 26 were 140 mm × 190 mm × 1080 mm timber posts and were placed in steel foundation tubes. The timber posts and foundation tubes were part of an anchor system, similar to a BCT but installed tangent to the system, used to develop the required tensile capacity of the guardrail. Posts 1 through 11 and 12 through 26 were spaced 1905 mm on center. The unsupported span between Posts 11 and 12 was 7.62 m, as shown in Figure 1. For Posts 3 through 24, the soil embedment depth was 1100 mm. The posts were placed in a compacted, coarse, crushed limestone material that met Grading B of AASHTO M147-65 (1990) as found in NCHRP Report 350. For Posts 3 through 8 and 15 through 24, 150 mm × 200 mm × 360 mm, routed wood spacer block-outs were used to block the rail away from the steel post. A standard 2.66-mm-thick W-beam rail, 7.62 m long, spanned Posts 1 through 5. Subsequently, nested W-beam guardrail, 2.66 mm thick and 30.48 m long, spanned Posts 5 through 18. A standard 2.66-mmthick W-beam rail, 7.62 m long, spanned Posts 18 through 22 and another spanned Posts 22 through 26, as shown in Figure 1. The top mounting height of the W-beam rail was 706 mm. CRASH TEST OLS-1 For the first test, a 1999-kg pickup truck impacted the long-span guardrail system (Design A) at a speed of 101.3 km/h and at an FIGURE 1 Long-span guardrail system design details, Design A. Faller et al. Paper No. 00 - 0598 23 CRASH TEST OLS-2 FIGURE 2 Long-span guardrail system, Design A. angle of 25.4 degrees. The initial impact occurred 2.44 m downstream from the center of Post 12 and toward Post 11, as shown in Figure 1. During the crash event at 0.170 s, the BCT cable anchor located on the upstream end of the system unexpectedly failed. This cable failure allowed the W-beam to pull free and release all guardrail tensile capacity, thus causing significant vehicle penetration into the guardrail system. Vehicle trajectory, vehicle damage, barrier damage, and the failed cable anchor hardware are shown in Figure 3. An analysis of the test results determined that the vehicle did not override the long-span guardrail during the impact before the cable anchor failure. An investigation showed that the lower cable anchor stud pulled out of the swagged fitting because of manufacturing inadequacies. Because the system was performing well until the cable failure, the test was rerun in the same configuration. For Test OLS-2, a 1997-kg pickup truck impacted the long-span guardrail system (Design A) at a speed of 102.7 km/h and at an angle of 24.5 degrees. The test results and the sequential photographs are summarized in Figure 4. The initial impact occurred 2.44 m downstream from the center of Post 12 or toward Post 11, as shown in Figures 4 and 5, respectively. At 0.263 s after impact, the vehicle was parallel to the guardrail at a speed of 73.2 km/h. At 0.340 s, the vehicle exited the guardrail at a speed of 66.2 km/h and at an angle of 16.7 degrees. After exiting the guardrail system, the pickup truck rolled over. The vehicle’s instability was attributed to the interaction of the front tire and suspension with the CRT posts immediately beyond the long-span section. Exterior vehicle damage was extensive, as shown in Figure 5. The occupant compartment had minimal deformations with slight deformation of the firewall. The right-front quarter panel was crushed inward, and the right side of the front bumper was also bent back toward the engine compartment. The right-front wheel assembly was deformed slightly. The right-front and right-rear wheels sustained tire holes and rim damage. Longitudinal deformations, from vehicle-rail interlock, were observed along the entire right side of the vehicle. The front, rear, and left-side window glass and the roof’s sheet metal were severely crushed during vehicle rollover. The engine hood popped open after vehicle rollover. As shown in Figure 5, damage to the barrier was moderate, mainly consisting of a deformed W-beam, contact marks on a guardrail section, and deformed and fractured guardrail posts. The W-beam damage consisted of contact marks and moderate deformation and flattening of the lower portion of the impacted section between Posts 9 and 12. The W-beam was released from Posts 2 and 25. Two CRT posts, Posts 10 and 11, completely fractured, while CRT Posts 12 through 14 split longitudinally and remained standing. CRT Post 9 fractured between the breakaway holes but remained upright. Steel Post 8 was twisted slightly. No significant post or guardrail damage occurred upstream of Post 15 nor downstream of Post 7. The maximum lateral permanent set rail and post deflections were approximately 956 mm and 5715 mm downstream from the centerline of Post 12 and 286 mm at Post 12, respectively, as measured in the field. The maximum lateral dynamic rail and post defections were 1342 mm and 6668 mm downstream from the centerline of Post 12 and 802 mm at Post 10, respectively, as determined from the high-speed film analysis. The analysis of Test OLS-2 results showed that the long-span guardrail satisfactorily contained the vehicle but inadequately redirected the vehicle—namely, the vehicle did not remain upright after colliding with the long-span guardrail. Detached elements and debris from the test article did not penetrate or show potential for penetrating the occupant compartment. Deformations of, or intrusion into, the occupant compartment that could have caused serious injury did not occur. The occupant impact velocities (OIV) and occupant ridedown decelerations (ORD) were within the suggested NCHRP Report 350 limits. After the collision, the vehicle intruded into adjacent traffic lanes. Design A was considered unacceptable according to Report 350 safety performance criteria. DESIGN MODIFICATIONS: DESIGN B Although a safety performance evaluation found Design A unacceptable according to Report 350 criteria, the safety performance could FIGURE 3 FIGURE 4 Vehicle damage, barrier damage, and cable anchor failure, Test OLS-1. Summary of test results and sequential photographs, Test OLS-2. Faller et al. FIGURE 5 Paper No. 00 - 0598 25 Impact location, vehicle damage, and barrier damage, Test OLS-2. be significantly improved. However, the vehicular instabilities from the interaction between the front tire and suspension with the CRT posts immediately beyond the long-span section would need to be reduced or eliminated. Subsequently, the long-span guardrail system (Design B) was redesigned to include double block-outs next to the unsupported span to reduce tire-post interaction. For CRT Posts 9 through 14 of Design B, two 150 mm × 200 mm × 360 mm spacer block-outs were used at each post to block the rail away from the posts, as shown in Figures 6 and 7. This configuration contrasts with Design A, previously evaluated in Tests OLS-1 and OLS-2 using single wood spacer block-outs on the six CRT posts next to the long-span section of guardrail. CRASH TEST OLS-3 For Test OLS-3, a 1994-kg pickup truck impacted the long-span guardrail system (Design B) at a speed of 102.9 km/h and at an angle of 24.7 degrees. The test results and the sequential photographs are summarized in Figure 8. The initial impact occurred 2.44 m downstream from the center of Post 12 or toward Post 11, as shown in Figures 8 and 9. At 0.283 s after impact, the vehicle was parallel to the guardrail at a speed of 77.6 km/h. At 0.469 sec, the vehicle exited the guardrail at a speed of 70.2 km/h and at an angle of 9.4 degrees. The pickup exited the guardrail stably and came to rest 57.37 m downstream from the impact and 18.62 m laterally away from the traffic-side face of the rail. Exterior vehicle damage was minimal, as shown in Figure 9. Interior occupant compartment deformations were negligible. The rightfront quarter panel was crushed inward, and the right side of the front bumper was bent back toward the engine compartment. The right-front wheel assembly was deformed slightly, including contact marks on the rim. Small contact marks were found on the lower right-rear fender, right-rear bumper, lower-right side of the truck box, and the right-side door. The right side of the truck box shifted downward and was twisted. As shown in Figure 9, damage to the barrier was moderate, mainly consisting of deformed W-beam, contact marks on a guardrail section, and deformed and fractured guardrail posts. The W-beam damage consisted of contact marks and moderate deformation and flattening of the lower portion of the impacted section between Posts 9 and 12. The W-beam rail was creased on the lower portion 533 mm downstream of Post 11. The W-beam rail was pulled off of Posts 3 and 4. Two CRT posts, Posts 10 and 11, completely fractured, while CRT Posts 9 and 12 through 14 rotated backward. Steel Posts 7 through 8 and 15 through 24 were twisted slightly and pushed backward. No significant post damage or deformations occurred to Posts 3 through 6. No significant guardrail damage occurred upstream of Post 14 or downstream of Post 9. The maximum lateral permanent set rail and post deflections were approximately 1016 mm and 953 mm upstream from the centerline of Post 11 and 362 mm at Post 9, respectively, as measured in the field. The maximum lateral dynamic rail and post deflections were 1450 mm and 1905 mm upstream from the centerline of Post 11 and 894 mm at Post 11, respectively, as determined from the high-speed film analysis. FIGURE 6 Long-span guardrail system design details, Design B. Faller et al. Paper No. 00 - 0598 27 conducted on Design B of the long-span guardrail system was acceptable according to NCHRP Report 350 criteria. SUMMARY AND CONCLUSIONS A long-span guardrail design for use over low-fill culverts was developed and successfully underwent full-scale vehicle crash testing. The long-span guardrail system was configured with a 30.48-m-long, nested W-beam rail and incorporated an unsupported guardrail length equal to 7.62 m. Three full-scale vehicle crash tests were performed according to NCHRP Report 350 TL-3 safety performance criteria. The first crash test, Test OLS-1, failed because of severe vehicle penetration into the guardrail system. This penetration resulted from a loss of rail tensile capacity during vehicle redirection because the swagged fitting on the cable anchor assembly failed. A retest, Test OLS-2, was performed on the same design that had a new cable anchor assembly. During vehicle redirection, the pickup truck rolled over; the test was deemed unacceptable according to TL-3 criteria. The analysis of Test OLS-2 results indicated that the interaction of the impacting vehicle’s front tire and suspension with the CRT posts immediately beyond the long-span section significantly contributed to the vehicle’s instability. Following an analysis of the second test, the long-span guardrail system was redesigned to include double block-outs on the CRT posts next to the free span to reduce tire-post interaction. A retest, Test OLS-3, was performed with a 3⁄4-ton (680-kg) pickup truck on the guardrail system (Design B) and was deemed acceptable according to TL-3 criteria. RECOMMENDATIONS FIGURE 7 Long-span guardrail system, Design B. The analysis of Test OLS-3 results showed that the long-span guardrail adequately contained and redirected the vehicle with controlled lateral displacements of the guardrail. Detached elements and debris from the test article did not penetrate or show potential for penetrating the occupant compartment. Deformations of, or intrusion into, the occupant compartment that could have caused serious injury did not occur. The vehicle remained upright during and after the collision. Vehicle roll, pitch, and yaw angular displacements were noted, but deemed acceptable because they did not adversely influence occupant risk safety criteria nor cause rollover. The OIVs and ORDs were within the suggested limits provided in Report 350. After the collision, the vehicle intruded slightly into adjacent traffic lanes but at an acceptable level. In addition, the vehicle’s exit angle was less than 60 percent of the impact angle. Therefore, Test OLS-3 The Test OLS-3 results indicate that the long-span guardrail design is suitable for use on federal-aid highways. The research described herein could be further developed using the test data to modify future designs of different lengths. However, any design modifications made to the long-span guardrail system may require verification via full-scale vehicle crash testing. The long-span guardrail system (Design B), as shown in Figures 6 and 7, was constructed with a rail splice at the midspan of the 7.62-m unsupported length of nested W-beam. Because crash testing found this design acceptable if a reduced cross section exists in the steel splice, other variations in splice location would also be acceptable, such as using a 7.62-m-long, nested rail in the unsupported region. For Design B, the length-of-need guardrail posts, Posts 3 through 8 and 15 through 24, were configured using steel sections. However, acceptable performance would also be achieved by using any other longitudinal W-beam guardrail systems that comply with Report 350 criteria. The crash tests were performed on a test installation that did not include a concrete box culvert, headwall, and wing wall. In field applications, a concrete headwall would typically extend above the low-fill soil, run parallel to the roadway, and prevent the soil from eroding over the culvert end. In this situation, if the headwall is placed too close to the guardrail, the potential exists for the vehicle’s wheel or fractured CRT posts to contact the headwall. If the wheel significantly contacts the headwall or post debris striking the headwall, 28 Paper No. 00 - 0598 FIGURE 8 Transportation Research Record 1720 Summary of test results and sequential photographs, Test OLS-3. vehicular instabilities or rollover may result. The analysis of OLS-3 crash test results revealed a maximum lateral dynamic rail deflection of 1.45 m. Also during this event, the vehicle’s right-front wheel protruded under the deformed guardrail. To minimize or eliminate the potential for wheel contact on the culvert headwall or post debris wedged between the headwall, the back face of the guardrail should be positioned a minimum of 1.5 m away from the front face of the headwall. As mentioned previously, the final long-span guardrail system was constructed with 30.48 m of nested W-beam rail, as shown in Figure 6. On the crash-tested installation, two 7.62-m-long W-beam rails, or 15.24 m total, were placed upstream of the nested region, while one 7.62-m-long W-beam rail was placed downstream of the nested region. This configuration provided an asymmetrical layout about the centerline of the system that is more common in field installations. Typically, longer guardrail runout lengths would be required on the upstream end of the obstruction. However, the system could be installed symmetrically with a standard guardrail terminal placed beyond each end of nested W-beam rail. For a standard guardrail terminal length of 11.34 m, the total installation length would be approximately 53.34 m, which was also the final length of the asymmetrical crash-tested design. Finally, the guardrail system was configured with the entire length installed tangentially. However, this guardrail system can be installed in the field with either one or two ends flared away from the traveled path. For locations using a guardrail flare, the minimum recommended length of tangent section adjacent to the unsupported length is 7.62 m. Flare rates should follow the recommended guidelines provided in AASHTO’s Roadside Design Guide (1). ACKNOWLEDGMENTS The authors wish to acknowledge several sources contributing to this project—the Midwest States Regional Pooled Fund Program funded by the Iowa, Kansas, Minnesota, Missouri, Nebraska, Ohio, South Dakota, and Wisconsin departments of transportation for sponsoring this project; the Washington State Department of Transportation, CONTECH Construction Products, Inc., and Lane Enterprises for cosponsoring this project; MwRSF personnel for constructing the barriers and conducting the crash tests; the Center for Infrastructure Research, Engineering Research Center, and the University of Nebraska–Lincoln for matching support; and Daniel Mushett, Highway Timber Products Co., a division of Cox Industries, for donating the wood posts and block-outs. Faller et al. FIGURE 9 Paper No. 00 - 0598 29 Impact location, vehicle damage, and barrier damage, Test OLS-3. REFERENCES 1. Roadside Design Guide. AASHTO, Washington, D.C., Jan. 1996. 2. Ross, H. E., T. J. Hirsch, B. Jackson, and D. L. Sicking. Safety Treatment of Roadside Cross-Drainage Structures. Research Report 280-1, Texas Transportation Institute, Texas A&M University, College Station, March l981. 3. Ross, H. E., D. L. Sicking, T. J. Hirsch, H. D. Cooner, J. F. Nixon, S. Fox, and C. P. Damon. Safety Treatment of Roadside Drainage Structures. In Transportation Research Record 868, TRB, National Research Council, Washington, D.C., 1982. 4. Hirsch, T. J., and D. Beggs. Use of Guardrails on Low Fill Bridge Length Culverts. In Transportation Research Record 1198, TRB, National Research Council, Washington, D.C.,1988, pp. 62–75. 5. Pfeifer, B. G., and J. K. Luedke. Safety Performance Evaluation of a Nested W-Beam with Half-Post Spacing Over a Low-Fill Culvert, Final Report to the Kansas Department of Transportation. Transportation Report TRP-03-36-92, Midwest Roadside Safety Facility, University of Nebraska–Lincoln, March 1993. 6. Memorandum on W-Beam Guardrail over Low-Fill Culverts. File Designation HNG-14, FHWA, U.S. Department of Transportation, Sept. 1991. 7. Mak, K. K., R. P. Bligh, D. J. Gripne, and C. F. McDevitt. Long-Span Nested W-Beam Guardrails over Low-Fill Culverts. In Transportation Research Record 1367, TRB, National Research Council, Washington, D.C., 1992. 8. Michie, J. D. NCHRP Report 230: Recommended Procedures for the Safety Performance Evaluation of Highway Appurtenances. TRB, National Research Council, Washington, D.C., March 1981. 9. Ross, H. E., D. L. Sicking, R. A. Zimmer, and J. D. Michie. NCHRP Report 350: Recommended Procedures for the Safety Performance Evaluation of Highway Features. TRB, National Research Council, Washington, D.C., 1993. 10. Polivka, K. A., B. W. Bielenberg, D. L. Sicking, R. K. Faller, and J. R. Rohde. Development of a 7.62-m Long Span Guardrail System, Final 11. 12. 13. 14. 15. 16. 17. Report to the Midwest States Regional Pooled Fund Program. Transportation Report No. TRP-03-72-99, Midwest Roadside Safety Facility, University of Nebraska–Lincoln, April 1999. Polivka, K. A., B. W. Bielenberg, D. L. Sicking, R. K. Faller, J. R. Rohde, and E. A. Keller. Development of a 7.62-m Long Span Guardrail System—Phase II. Final Report to the Midwest States Regional Pooled Fund Program, Transportation Report TRP-03-88-99, Midwest Roadside Safety Facility, University of Nebraska–Lincoln, Aug. 1999. Buth, C. E., W. L. Campise, L. I. Griffin, M. L. Love, and D. L. Sicking. Performance Limits of Longitudinal Barrier Systems—Volume I— Summary Report. Report FHWA/RD-86/153. FHWA, U.S. Department of Transportation, May 1986. Ivey, D. L., R. Robertson, and C. E. Buth. Test and Evaluation of W-Beam and Thrie-Beam Guardrails. Report FHWA/RD-82/071. FHWA, U.S. Department of Transportation, March 1986. Ross, H. E., H. S. Perera, D. L. Sicking, and R. P. Bligh. NCHRP Report 318: Roadside Safety Design for Small Vehicles. TRB, National Research Council, Washington, D.C., May 1989. Powell, G. H. Barrier VII: A Computer Program for Evaluation of Automobile Barrier Systems. Report FHWA RD-73-51. FHWA, U.S. Department of Transportation, April 1973. Vehicle Damage Scale for Traffic Investigators, 2nd ed. Technical Bulletin 1, Traffic Accident Data Project, National Safety Council, Chicago, Ill., 1971. Collision Deformation Classification–Recommended Practice J224 March 1980, Handbook Volume 4. SAE, Warrendale, Pa., 1985. The contents of this report reflect the views of the authors, who are responsible for the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the state highway departments participating in the Midwest States Regional Pooled Fund Research Program, Washington State Department of Transportation, Federal Highway Administration, CONTECH Construction Products, Inc., or Lane Enterprises. This paper does not constitute a standard, specification, or regulation. Publication of this paper sponsored by Committee on Roadside Safety Features.