Criteria for Restoration of Longitudinal Barriers, Phase II (2021)

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.

255 CHAPTER 10 – COMBINATION DAMAGE MODE OF RAIL DEFLECTION AND RAIL-POST CONNECTION FOR THE G4(2W) In this chapter guardrail deflection in combination with rail-to-post connection strength are evaluated using finite element analysis to quantify their effects on the crash performance of the G4(2W) guardrail. The results of the study are then used to develop damage assessment and repair criteria for the guardrail system. The research approach was similar to that used by Gabler in Phase I for the evaluation of the G4(1S) guardrail with crash-induced damage. Finite element analysis was used to simulate impact of a 4568-lb pickup into the guardrail at low-impact speeds to obtain various levels of crash-induced damage. The damage resulting from these low-speed impacts also included various other non-critical damage modes such as rail flattening, slight anchor movement, slipping of post-bolt in the rail slot, moderate splice damage, etc. When additional critical damage modes resulted during the low-speed impacts, they were reported as such herein. For this study, the end-terminal was modeled using the baseline anchor response for a standard two-post anchor system as defined in Chapter 9. Finite element analysis was then used to simulate high-speed impact into the damaged guardrail to evaluate the performance of the system with low-severity, crash-induced rail deflections. The basic methodology involved using the finite element model validated in Chapter 7 to (1) simulate impact of the guardrail at low velocities (i.e., 30 mph, 35 mph and 40 mph); (2) save all deformations and residual stresses for the barrier components and soil; and (3) simulate high-speed impact on the damaged system under NCHRP Report 350 Test 3-11 conditions. Evaluate Effects of Rail Deflection and Rail-to-Post Connection Strength The finite element model of the standard G4(2W) guardrail system developed in Chapter 7 was used to evaluate the effects of crash-induced rail deflection on the crash performance of the G4(2W) guardrail system. The baseline (undamaged) condition for the guardrail included the DL0 guardrail post model and anchor strength corresponding to the results of the end-terminal in Test 13011B with force-deflection response scaled by 1.5 to account for rate effects. Low Severity Crash-Induced Damage A total of six low severity, crash-induced damage cases were created by simulating the impact of a 4568-lb ¾-ton pickup into the G4(2W) guardrail at an impact angle of 25 degrees using LS-DYNA. The simulation matrix is shown in Table 61 and includes three impact speeds (i.e., 30 mph, 35 mph and 40 mph) and two impact locations. The FEA models for these analysis cases are shown in Figure 203. Impact Point 1 (IP01) was located at 2 feet upstream of Post 9. This impact point was chosen to create maximum deflection of the rail just upstream of the non-splice connection at Post 10, for investigating the effects of a rail-post connection at a non-splice location (e.g., at this connection point the bolt head only has a single layer of w-beam to pull-through in order to release). Impact Point 2 (IP02) was located at 2 feet upstream of Post 10. This impact point was chosen to create maximum rail deflection just upstream of the splice connection at Post 11, for evaluating the effects of the stronger rail-post connection at the splice where the bolt head must pull through two layers of w-beam to release. To facilitate reporting and data analysis, an

256 alternative reference system was used for labeling the posts within the impact region of the guardrail. This section of posts was labeled A through G, with Post A being the first post upstream of the impact point, as illustrated in Figure 203. This labeling system resulted in the rail-post connections at Posts C and D being the most critical for evaluating proper release of the rail from the posts for all impact scenarios. Recall that the greatest potential for vehicle override occurs when the rail is pulled down by the posts during deflection. Table 61. Simulation matrix for creating low-severity guardrail deflection damage cases. Figure 203. Impact locations, IP01 and IP02, for low-speed impact cases. Table 62 provides a summary of barrier damages for the low-severity impact cases regarding lateral rail deflections, longitudinal rail deflection at the anchors, tensile forces in the w-beam rail, and final position of post-bolts in the slotted hole. Sequential views of the FE analysis results for each case are provided in Appendix K. 30 mph 35 mph 40 mph IP01 x x x IP02 x x x Impact Point Impact Speed Non-Splice at Post 10 Splice at Post 11 25⁰ IP01 IP02 25⁰ Post 10 Post 3 G F E D C B A Post 9 Post 3 G F E D C B A

257 Table 62. Summary of results for the low-severity impact analyses. The maximum lateral dynamic deflections of the guardrail for the 30 mph, 35 mph and 40 mph impact cases were approximately 13 inches, 17 inches and 19 inches, respectively; the resulting maximum permanent deflections were approximately 9 inches, 13.5 inches, and 14 inches, respectively. Regarding the location of maximum deflection in Table 62, a negative number indicates that the maximum deflection occurred upstream of Post C, while a positive number indicates that maximum deflection occurred downstream of the post. In all cases, the maximum deflection occurred just upstream of the critical post-bolt connection at Post C, which was located at a w-beam splice for impact cases IP01 and at a non-splice location for IP02. The rail was flattened during the impact such that the maximum cross-section height of the w-beam ranged 16 to 17 inches for the analysis cases involving Impact Point 1, and 17 to 18 inches for the analysis cases involving Impact Point 2. According to Gabler et al., when the w-beam rail is flattened such that the cross-section height is greater than 17 inches, the damage severity rating is 30 mph 35 mph 40 mph 30 mph 35 mph 40 mph Maximum Dynamic (in) 13.7 17.0 18.8 12.2 16.6 18.9 Maximum Permanent (in) 8.7 13.5 13.9 N.C. 13.3 13.3 Location of Max Deflection (in) (Relative to Post C) -37.0 -35.4 -35.0 -44.3 -30.5 -27.6 Dynamic Deflection at Post A (in) 0.6 0.8 0.7 0.4 0.6 0.6 Dynamic Deflection at Postt B (in) 8.3 9.7 10.7 7.2 9.1 10.0 Dynamic Deflection at Post C (in) 8.1 13.3 16.7 7.2 13.8 17.7 Dynamic Deflection at Post D (in) 0.7 1.7 5.9 0.5 3.3 6.9 Rail Flattening Maximum Rail X-Section Height (in) 16.0 16.9 17.3 16.9 17.8 17.9 Upstream Anchor Permanent (in) 0.72 0.94 1.04 0.57 0.89 1.05 Downstream Anchor Permanent (in) 0.07 0.18 0.33 0.04 0.24 0.36 Maximum Plastic Strain in Splice 0.19 0.24 0.83 0.32 0.58 0.57 Maximum Rail Tension (kips) N.C. N.C. N.C. N.C. N.C. N.C. Final Position of Post-Bolt at Post B USE USE USE USE USE USE Final Position of Post-Bolt at Post C DSE CEN DSE CEN CEN CEN Final Position of Post-Bolt at Post D DSE DSE DSE DSE DSE DSE R ai l D ef le ct io n A n ch o r D e fl e ct io n R ai l Lo ad P o st B o lt P o si ti o n Event USE - Upstream end of post-bolt slot DSE - Downstream end of post-bolt slot CEN - Center of post-bolt slot N.A. - Not Applicable N.C. - Not Computed Impact Point 01 Impact Point 02

258 considered “medium.” Thus, this damage mode was relatively significant for many of these cases. The permanent longitudinal deflection of the rail element at the upstream anchor ranged from 0.6 inch for the 30 mph case to just over 1 inch for the 40 mph case. The movement of the rail at the downstream anchor was negligible for all cases. The maximum plastic strain in the rail element occurred at a splice-bolt location in all cases (typically at the downstream end of the splice at the lowest bolt hole). For simulation cases at IP01, the splice with highest strains was located at Post C; for simulation cases at IP02, the splice damage occurred at Post B. Unfortunately, the maximum tensile forces in the rail element were not recorded for these analysis cases. At the start of the analyses, the post-bolts were positioned at the center of the slotted hole in the w-beam at all rail-to-post connections points. Due to the tension in the rail during impact, the post-bolts located upstream of the impact point tended to pull toward the upstream end of the slot, while the post-bolts downstream of the impact tended to pull toward the downstream end of the slot. At Post C, however, the post-bolt was first pulled to the downstream end of the slot as the vehicle approaches from the upstream direction; then as the vehicle passed by the post during redirection, the tension in the rail tended to pull the post-bolt back toward the center. In some cases, however, the bolt remained at, or very near to, the downstream end of the slot. The post- bolt positioned at the ends of the slotted hole result in a stronger, more critical connection for the study, since it results in a higher probability of the rail to remain attached to, and be pulled down with, the posts as they deflect. Figure 204. Example of low-severity impact damage illustrating final position of post-bolts relative to the slotted hole in w-beam. The low-severity impact cases performed in this study involved low speeds and a high impact angle which resulted in damages to a relatively localized section of the guardrail, as illustrated in Figure 205. The length of damage shown in the figure includes only the section of rail with visually discernable lateral deflection. The full extent of damage actually involved a much longer section of the system and included low levels of permanent anchor deflections. It may be possible to achieve similar magnitudes of lateral deflection, spread over a greater length of guardrail, by using higher impact speeds and smaller impact angles. In such cases, the resulting capacity of the damaged guardrail may differ from those presented herein and should be considered in future evaluations of damage modes related to guardrail deflections. Post D Post C

259 Figure 205. Extent of damage resulting from impact on G4(2W) at 30 mph at Impact Point. High-Speed Impact into Pre-Damaged Guardrail Finite element analysis was then used to simulate high-speed impact into the damaged guardrail to evaluate the performance of the system with various degrees of low-severity, crash- induced rail damage. The nodal deformations and residual stresses of the barrier components resulting from the low-speed impact cases were used as initial conditions for this phase of the study. The impact conditions for the high-speed impact simulations were set to those specified in NCHRP Report 350 for Test 3-11 (i.e., 4409-lb pickup impacting at 62.2 mph and 25 degrees). The impact point for each analysis case was set to those of its corresponding low-severity impact analysis case (e.g., either IP01 or IP02). Figure 206 shows the FE model for the six analysis cases, denoting the impact point, extent of initial damage, and location of splice and non-splice positions relative to the critical post-bolt connection at Post C. Sequential views of the FE analysis results for each of the high-speed impact cases are provided in Appendix L. Table 56 provides a summary of guardrail damages resulting from the analyses of the high-speed impacts, including rail deflections, anchor movement, splice damage, and release of post-bolt connections at Posts C and D. Portions of this information are also presented graphically in Figure 188 and Figure 189. The maximum lateral deflection of the rail during the high-speed impact cases increased as the severity of the pre-existing crash-induced deflection increased, with slightly higher deflections resulting from Impact Point 2. This trend was also prominent regarding maximum deflection of the rail at the post locations. The maximum rail deflection for the six cases ranged from 35.5 inches for the analysis case involving 8.7 inches initial deflection and Impact Point 1 to 41.3 inches for the analysis case involving 14 inches initial deflection and Impact Point 2. The maximum deflection for the baseline analysis of the undamaged G4(2W) was 32 inches (refer to Chapter 8). In all analysis cases, the maximum deflection of the rail occurred just upstream of Post D; thus Impact Point 1 was considered to be representative of the critical impact point (CIP) for the system.

260 Figure 206. Initial conditions used in evaluating effects of low-level guardrail deflection on the performance of the G4(2W). Non-Splice at Post 11 Post 10 Case 4: Impact Point - IP02 Critical post-bolt connection at splice Pre-Damage: • Speed / angle = 35 mph / 25 deg. • Deflection = 13.3 in. • Length of damage = 24.5 ft. G F E D C B A Case 1: Impact Point - IP01 Critical post-bolt connection at non-splice Pre-Damage: • Speed / angle = 30 mph / 25 deg. • Deflection = 8.7 in. • Length of damage = 19 ft. G F E D C B A Non-Splice at Post 10 Post 9 G F E D C B A Non-Splice at Post 10 Post 9 Case 3: Impact Point - IP01 Critical post-bolt connection at non-splice Pre-Damage: • Speed / angle = 35 mph / 25 deg. • Deflection = 13.5 in. • Length of damage = 20 ft. Case 2: Impact Point - IP02 Critical post-bolt connection at splice Pre-Damage: • Speed / angle = 30 mph / 25 deg. • Deflection = 8.7 in. (estimated) • Length of damage = 20 ft. Non-Splice at Post 11 Post 10 G F E D C B A Case 5: Impact Point - IP01 Critical post-bolt connection at non-splice Pre-Damage: • Speed / angle = 40 mph / 25 deg. • Deflection = 13.9in. • Length of damage = 26 ft. G F E D C B A Non-Splice at Post 10 Post 9 Case 6: Impact Point - IP02 Critical post-bolt connection at splice Pre-Damage: • Speed / angle = 40 mph / 25 deg. • Deflection = 13.3 in. • Length of damage = 28 ft. Non-Splice at Post 11 Post 10 G F E D C B A

261 Table 63. Summary of barrier damage resulting from analysis of high-speed impact into the G4(2W) with pre-existing low-severity rail deflection. The values presented in Table 56 for the “maximum deflection at time of post-bolt release” correspond to the lateral deflection of the rail measured at the post-bolt hole at the time when the connection separates. The shaded cells for this category in Table 56 represent those cases where the post-bolt is located at a w-beam splice, and the non-shaded cells correspond to cases where the post-bolt is at a non-splice location. For example, for IP01 the splice is located at Post D, while for IP02 cases the splice is located at Post C. The post-bolt connection at Post D was considered to be the most critical connection for these impact cases. This assumption was based on the fact that the lateral deflection of Post C did not reach critical magnitude until the vehicle was already at or past the post; whereas Post D reached critical deflections while the vehicle was still upstream of the post. For cases involving a w-beam splice at Post D, the release of the post-bolt connection tended to occur at higher post deflections, as the severity of the pre- existing crash-induced rail deflections increased. While for cases with a single w-beam layer at Post D, the connection failed rather consistently at 29-31 inches of post deflection, regardless of the severity of the pre-existing rail deflections. 8.7 inches 13.5 inches 14 inches 8.7 inches 13.5 inches 14 inches Maximum Rail Deflection (in) 35.5 36.3 40.4 38.4 39.1 41.3 Location of Max Deflection (in) (Relative to Post D) -23.8 -8.5 -6.5 -14.6 -14.9 -7.9 Rail Deflection at Post A (in) 3.1 3.9 4.0 3.9 3.7 4.5 Rail Deflection at Post B (in) 15.0 16.7 17.7 17.2 16.5 17.7 Rail Deflection at Post C (in) 31.9 33.4 35.4 33.2 33.7 35.0 Rail Deflection at Post D (in) 34.3 36.2 40.0 37.9 38.7 40.1 Rail Deflection at Post E (in) 21.2 22.7 30.5 25.4 27.3 27.4 Rail Deflection at Post F (in) 2.0 2.5 9.0 3.8 4.4 4.9 Rail Deflection at Post G (in) 0.0 0.0 0.9 0.0 0.0 0.0 Upstream Anchor Deflection (in) 2.6 3.1 3.1 2.6 2.6 3.3 Downstream Anchor Deflection (in) 0.9 0.9 1.0 0.9 1.1 1.2 Maximum Strain in splice 1.08 1.02 1.50 1.09 1.19 1.37 Maximum Rail Tension (kips) 46.5 54.2 53.1 48.3 49.2 68.1 Maximum Deflection at Time of Post-Bolt Release - Post C (in) N.A. 32.1 27.5 27.1 28.5 32.2 Maximum Deflection at Time of Post-Bolt Release - Post D (in) 32.2 33.5 37.3 29.6 28.7 31.3 R ai l D ef le ct io n Event Impact Point 01 Impact Point 02 Initial Crash-Induced Deflection A n ch o r D e fl e ct io n R ai l Lo ad P o st B o lt R e le as e

262 Figure 207. Summary of barrier damage resulting from analysis of high-speed impact into the G4(2W) with pre-existing low-severity rail deflection. Figure 208. Summary of anchor displacement at rail height from analysis of high-speed impact into the G4(2W) with pre-existing low-severity rail deflection. -30.0 -20.0 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 IP01 - 8.7 in IP01 - 13.5 in IP01 - 14 in IP02 - 8.7 in IP02 - 13.5 in IP02 - 14 in 2.61 3.11 3.07 2.58 2.61 3.34 0.92 0.92 1.05 0.94 1.11 1.19 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 8.7 inches 13.5 inches 14 inches 8.7 inches 13.5 inches 14 inches Impact Point 01 Impact Point 02 Initial Crash-Induced Deflection A n ch o r D ef le ct io n ( in ) Upstream Anchor Deflection (in) Downstream Anchor Deflection (in)

263 The movement of the anchors and the maximum tensile forces in the rail tended to increase as the severity of the pre-existing crash-induced rail deflections increased. The anchor deflections also tended to be higher for cases involving Impact Point 1, which was in part attributed to the fact that Impact Point 1 was closer to the downstream end-terminal (e.g., tensile force in rail were distributed to fewer upstream posts for IP01 cases). A summary of the maximum effective plastic strains around the splice-bolt holes in the w-beam is shown in Figure 209 for each analysis case. The results show that the potential for splice rupture tended to increase as the severity of the pre-existing crash-induced rail deflections increased. For all cases, the maximum splice damage occurred at Post 11; which, for those analysis cases involving Impact Point 2, was at Post C located upstream of point of maximum deflection. For reference, the maximum effective plastic strains in the splice for the undamaged baseline case was 0.84 (see Chapter 8). Figure 209. Summary of maximum effective plastic strains occurring at splice-bolt locations. A summary of occupant risk measures computed from the acceleration and angular rate time-histories at the vehicle’s center of gravity is provided in Table 64. This data is also presented graphically in Figures 190 through 192. The results indicate that occupant impact velocity (OIV) is not significantly affected by pre-existing low-level guardrail deflections. The maximum 10-millisecond occupant ridedown accelerations (ORA) in the longitudinal direction tended to increase as pre-existing guardrail deflections increased; however, they remained below critical limits of 20 g’s for all cases. The ORA values in the lateral directions were less affected. 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 8.7 inches 13.5 inches 14 inches 8.7 inches 13.5 inches 14 inches Impact Point 01 Impact Point 02 Ef fe ct iv e P la st ic S tr ai n

264 Table 64. Summary of occupant risk measures from analysis of high-speed impact into the G4(2W) with pre-existing low- severity rail deflection. 8.7 inches 13.5 inches 14 inches 8.7 inches 13.5 inches 14 inches Occupant Impact Velocity x-direction 5.1 5.4 4.1 5.5 4.5 5.1 (m/s) y-direction 5.9 5.8 5.7 5.4 5.4 5.6 at time (0.1512 sec) (0.1504 sec) (0.1515 sec) (0.1534 sec) (0.1523 sec) (0.1572 sec) 7.6 7.3 6.9 7.3 7.1 7.1 (0.1472 sec) (0.1468 sec) (0.1480 sec) (0.1487 sec) (0.1483 sec) (0.1534 sec) Ridedown Acceleration 8.8 10.3 15.4 10.1 12.9 10.9 (g's) (0.1665 - 0.1765 sec) (0.1694 - 0.1794 sec) (0.1986 - 0.2086 sec) (0.2789 - 0.2889 sec) (0.1551 - 0.1651 sec) (0.1704 - 0.1804 sec) 10.7 7.6 9 8.3 10.8 11.2 (0.2301 - 0.2401 sec) (0.1538 - 0.1638 sec) (0.2149 - 0.2249 sec) (0.2149 - 0.2249 sec) (0.2390 - 0.2490 sec) (0.2268 - 0.2368 sec) 11.4 11.3 17.3 11.3 14.5 12.6 (0.2300 - 0.2400 sec) (0.1695 - 0.1795 sec) (0.1985 - 0.2085 sec) (0.1599 - 0.1699 sec) (0.1552 - 0.1652 sec) (0.1705 - 0.1805 sec) 1.03 0.93 0.89 0.93 0.92 0.9 (0.1231 - 0.1731 sec) (0.1280 - 0.1780 sec) (0.1958 - 0.2458 sec) (0.1224 - 0.1724 sec) (0.1212 - 0.1712 sec) (0.1300 - 0.1800 sec) Max 50-ms moving avg. acc. 7.8 7.8 6.4 7.5 7.5 7.0 (g's) (0.1259 - 0.1759 sec) (0.1289 - 0.1789 sec) (0.1231 - 0.1731 sec) (0.1229 - 0.1729 sec) (0.1233 - 0.1733 sec) (0.1395 - 0.1895 sec) 7.2 6.6 6.8 6.2 6.3 6.5 (0.1979- 0.2479 sec) (0.0925- 0.1425 sec) (0.1966- 0.2466 sec) (0.1216 - 0.1716 sec) (0.0925 - 0.1425 sec) (0.2060 - 0.2560 sec) 3 2.9 2.8 2.5 3.3 2.9 (0.2414 - 0.2914 sec) (0.3358 - 0.3858 sec) (0.3243 - 0.3743 sec) (0.3479 - 0.3979 sec) (0.3221 - 0.3721 sec) (0.3374 - 0.3874 sec) Impact Point 02 Undamaged Posts (DL0) / Baseline Anchor Strength (g's) x-direction Occupant Risk Factors Impact Point 01 Initial Crash-Induced Deflection y-direction z-direction ASI PHD THIV (m/s) x-direction y-direction

265 Figure 210. Summary of occupant impact velocity (OIV) from analysis of high-speed impact into the G4(2W) with pre-existing low-severity rail deflection. Figure 211. Summary of occupant ridedown accelerations (ORA) from analysis of high- speed impact into the G4(2W) with pre-existing low-severity rail deflection. 0 1 2 3 4 5 6 7 8.7 inches 13.5 inches 14 inches 8.7 inches 13.5 inches 14 inches Impact Point 01 Impact Point 02 Initial Crash-Induced Deflection O IV ( m /s ) Occupant Impact Velocity x-dir y-dir 0 2 4 6 8 10 12 8.7 inches 13.5 inches 14 inches 8.7 inches 13.5 inches 14 inches Impact Point 01 Impact Point 02 Initial Crash-Induced Deflection O R A ( g) Occupant Ridedown Acceleration x-dir y-dir

266 Figure 212. Summary of 50-ms average acceleration from analysis of high-speed impact into the G4(2W) with pre-existing low-severity rail deflection. Summary and Discussion The purpose of this task was to quantify the effects of crash-induced rail deflection in combination with rail-to-post connection strength on the crash performance of the G4(2W) guardrail system, and to use this information to develop assessment criteria for this very common guardrail damage mode. Finite element analysis was used to evaluate the impact performance of the G4(2W) guardrail system with various levels of pre-existing crash-induced rail deflection. The impact conditions used for the performance evaluations corresponded to those of NCHRP Report 350 Test 3-11 (i.e., 4400 lb vehicle impacting at 62.2 mph and 25 degrees, nominally). The initial crash-induced damage for the guardrail model was created by simulating low- speed impacts into the guardrail with the 4,568-lb pickup truck model at an impact angle of 25 degrees. Three impact speeds were used (i.e., 30 mph, 35 mph and 40 mph) which resulted in three levels of guardrail deflections (i.e., 8.7 inches, 13.5 inches, and 14 inches). Lower levels of damage were investigated by Gabler et al. in Phase I of this project and were therefore not considered in this study. The results from these low-speed impacts, which included guardrail component deformations and residual stresses, were then used as initial conditions for secondary high-speed impact simulations (i.e., Report 350 Test 3-11) into the damaged guardrail system. Two rail-to-post connection strength cases were investigated. In one case a w-beam splice was located at the critical connection point, in which the post-bolt head has to pull through two layers of w-beam in order to release; and in the second case the critical connection point was at a non- splice location, in which the post-bolt has to pull through only a single layer of w-beam in order to release. Finite element analysis was then used to simulate a secondary impact into the damaged guardrail to evaluate the performance of the system with various degrees of low-severity, crash- induced rail damage. The results of the high-speed analyses indicated that as the severity of the pre-existing crash-induced rail deflections increased: 0 1 2 3 4 5 6 7 8 8.7 inches 13.5 inches 14 inches 8.7 inches 13.5 inches 14 inches Impact Point 01 Impact Point 02 Initial Crash-Induced Deflection 5 0 -m s A ve ra ge A cc . ( g) 50-ms Avg. Acceleration x-dir y-dir

267  The maximum lateral deflections of the rail increased,  The tensile load in the rail increased,  The deflection of the upstream anchor increased,  The post-bolt connection at w-beam splice locations allowed greater post deflections before releasing,  The post-bolt connection at non-splice locations released consistently regardless of pre-damage deflections,  Potential for splice rupture increased,  Occupant ridedown accelerations in the longitudinal direction increased (primarily due to increased wheel snag on posts), and  All other occupant risk measures were less affected. The results of the analyses also indicated that the potential for override was relatively low for all cases investigated. This is contradictory to the full-scale crash test results from Gabler’s study, where a post-bolt did not release properly and allowed the rail to be pulled down with the post, resulting in the vehicle overriding the guardrail.[Gabler10] In that test the guardrail system was the G4(1S) (i.e., steel post w-beam guardrail) and the pre-existing crash-induced rail deflection was 14.5 inches. A possible explanation is that the G4(2W) and the G4(1S) respond differently to this particular damage mode. These two systems differ only by the type of guardrail post; i.e., the G4(2W) uses rectangular or round wood posts while the G4(1S) uses W6x9 structural steel section posts - all other aspects of these two systems are identical. Further, bogie impact tests have shown that the lateral force-deflection response for these two posts is essentially equivalent.[Hascall07] However, the results of full-scale crash tests have shown that the steel-post system results in lower impact forces. For example, tests on the modified G4(1S) with wood blockouts resulted in longitudinal ORA values of 7.9 G and 7.6 G [Bullard96; Bligh97]; while tests on the G4(2W) resulted in longitudinal ORA values of 10.2 G, 10.9 G and 11.6 G.[Bullard09; Bligh95; Mak99a] In order to understand the differences in the performance of these two seemingly identical systems, it is first important to understand the differences in the response of the two different types of posts under loading conditions typical of real-world collisions. In particular, guardrail posts generally experience both lateral and torsional loads during vehicle collisions – whereas bogie impact tests induce only lateral loads onto the post. The lateral load from vehicle collision arises directly from the lateral deflection of the w-beam rail against the posts. The torsional loads on the posts result from the tensile load in the w-beam rail acting at the front of the blockouts; the blockouts are offset 8 inches from the front face of the posts, thus creating a torsional moment about the vertical axis of the posts. The wood posts tend to resist this torsion, while the W6x9 steel posts do not. The torsion of the steel posts not only reduces the blockout distance between the rail and the post, but also significantly reduces its resistance to lateral deflection. Figure 213 shows the results of full-scale Test C08C3-027.2 on the G4(1S) guardrail system conducted by the MGA Research Corporation in Gabler’s study, illustrating the torsional deformation of the posts.[Fleck08b] The photo on the right in Figure 213 further illustrates the apparent reduction in lateral stiffness of the “twisted” post, where the post buckled at the groundline rather than displacing the soil. The combined effects of (1) reduced blockout

268 distance, (2) the post buckling at the groundline, and (3) loss of rail tension as the anchor failed resulted in the rail being pulled down with the post rather than releasing from the post as it was intended. In contrast, Figure 214 shows the results of full-scale Test 471470-26 on the G4(2W) guardrail conducted by TTI showing the general deformation mode of the posts during vehicle collision. In this case the posts do not show any apparent torsional deflections and tend to deflect only in the lateral direction. When wood posts fail under the torsional load, it is usually a brittle failure in which the posts split or fracture, as illustrated in Figure 215. Figure 213. Results of Test MGA C08C3-027.2 illustrating torsional deformation of the W6x8 steel guardrail posts during vehicle collision.[Fleck08b] Figure 214. Results of Test 471470-26 illustrating the response of the wooden guardrail posts during vehicle collision.[Mak99]

269 Figure 215. Results of Test 404201-1 illustrating brittle fracture of wood posts due to tensile forces in the w-beam rail.[Bullard00] The research team is not aware of any tests on a wood post guardrail system in which a penetration occurred due to the vehicle overriding the barrier. Penetration of wood post guardrails tend to result from rupture of a w-beam splice; while penetrations of steel post guardrails are sometimes due to splice rupture and sometimes due to override. This is consistent with the results obtained in the current study, in which the most significant effect of pre-existing crash-induced rail deflection on the performance of the G4(2W) guardrail system is an increased potential for rail rupture. Recommendations The analysis results indicated that the performance of the G4(2W) with crash-induced rail damage was compromised when the initial damage included 8.7 inches or greater deflections. The capacity of the splice connection was the aspect of the system that was most affected by this damage mode. For example, the plastic strains around the splice-bolt holes resulting from NCHRP Report 350 Test 3-11 impact was 0.84 for the baseline (undamaged) G4(2W) guardrail. The magnitude of effective plastic strains then increased as the magnitude of pre-damage deflections increased. At pre-damaged guardrail deflections of 8.7, 13.5 inches and 14 inches, the effective plastic strains in the splice-bolt holes reached magnitudes of 1.09, 1.19 and 1.5, respectively. As a result of this study, the research team recommends that guardrail damage with rail deflections of greater than 9 inches constitute a high priority for repair for the G4(2W). Although the effects of rail damage differ between the G4(1S) steel post guardrail and the G4(2W) wood post guardrail, the threshold of damage that constitute need for repair are essentially the same for these two systems; thus the relative repair thresholds defined by Gabler et al. in Phase I are considered valid for the G4(2W) system as well and are adopted here, as shown in Table 65.

270 Table 65. Recommendations for post and rail deflection damage. Future work should include analyses of both the G4(2W) and G4(1S) with pre-existing crash-induced rail deflections in combination with varying anchor strength. Due to the differences in the torsional rigidity of the W6x9 steel posts of the G4(1S) guardrail and from the results of Test MGA C08C3-027-2 on that system which resulted in large anchor deflection and vehicle override, it is expected that the G4(1S) may have a greater sensitivity anchor strength. Also, the low-severity impact cases evaluated herein involved low speeds and a high impact angle which resulted in damages to a relatively localized section of the guardrail; i.e., the damaged area generally spanned only 3 to 4 posts, which increased the potential for pocketing. It is not known how the guardrail will respond to subsequent impacts when the rail deflections are spread over a longer length of the guardrail. It is recommended that future studies on the effects of pre-existing crash-induced rail deflections include higher impact speeds and smaller impact angles to create initial guardrail damage with similar magnitudes of rail deflection spread over a longer length of the guardrail. In the following chapter, FEA is used to simulate Test C08C3-027.2 to reevaluate the cause of the vehicle overriding the barrier in that test.