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.

71 CHAPTER 5 – PRIORITIZATION OF DAMAGE MODES Introduction The objectives of this task were to: (1) identify which additional guardrail systems should be included in the field guide, (2) evaluate the importance of addressing each damage mode associated with each of those systems, (3) identify the research methods that could be used in quantifying the effects of various levels of damage for each case, and (4) prioritize the list of damage modes and system elements for inclusion in this study. The basic research approach used in Report 656 for assessing performance degradation of damage to the Modified G4(1S) guardrail involved a combination of pendulum tests, computational analyses and full-scale crash testing. Finite element analysis was used to investigate damage modes that were cost prohibitive to investigate using full-scale tests and that could not be adequately assessed with pendulum testing. In particular, finite element analysis was used to assess performance degradation due to (1) post and rail deflection, (2) missing or damaged posts, (3) post and rail separation, and (4) rail flattening. Physical testing was used to assess the effects of damage modes for cases where finite element analysis was not well suited, such as fracture of system elements (e.g., rail tears). Physical tests were also used to validate the accuracy of the finite element models and gain confidence in their results. Crash performance of the damaged systems was assessed based on NCHRP Report 350 Test 3-11 test conditions (i.e., 2000-kg pickup impacting at 100 km/hr and 25 degrees) and evaluation criteria. This study resumed where Report 656 left off, expanding on the information already in the field guide by including additional guardrail systems and/or additional damage assessment criteria for the modified G4(1S). For example, some of the damage modes evaluated in Report 656 warranted additional investigation so that damage thresholds could be better defined, or to further confirm and strengthen the already-established guidelines. Alternative methods of analysis may be used to confirm previous analysis results, or additional analyses may be conducted to encompass a broader range of damage levels in the assessments. The evaluations may also be expanded to investigate damage characterized by multiple damage modes. When a guardrail system is damaged, for example in a low-speed impact, the damage is usually characterized by several minor damage modes such as a flattened and deflected rail, deflected and twisted posts, loose soil around posts, damaged bolt connection, etc. As such, the guidelines should be expanded to include guardrails with multiple combined damage modes to better aid highway maintenance personnel faced with making repair decisions. Methods for Evaluating Damage Modes Full-Scale Testing While full-scale crash tests are the most conclusive method for determining crashworthy performance of roadside safety hardware, they tend to be expensive and yield relatively little information about the contribution of each specific component to overall performance. In other words, full-scale crash tests are definitive in terms of assessing performance but information poor in terms of knowing the stresses, strains and failure conditions of individual components. Since this project is concerned with detailed damage and failure mechanisms, full-scale crash testing is of limited use. For example, Report 656 demonstrated that pre-existing guardrail

72 deflection can affect the performance of the system in subsequent collisions. If one were only using full-scale tests to determine the threshold of lateral deflection where performance is compromised, a series of three, four or five tests would be required until the performance limit was identified. This would cost well over $100,000 to quantify this single damage mode. Using pendulum testing or finite element simulations, on the other hand, can provide the same basic information at dramatically lower cost. The research team believes it is more important to provide as much quantitative guidance for the broadest possible range of systems and damage modes than it is to definitively establish only one or two damage modes with full-scale crash testing. For this reason, full-scale crash testing is not the preferred method for studying damage effects, although one or two tests may be warranted for confirming results from pendulum testing or simulation. Full-scale testing of the undamaged systems will not be required since that data should be available from the original acceptance testing. The most recent crash test guidelines for assessing crash performance of roadside safety hardware are detailed in the AASHTO Manual for Assessing Safety Hardware (MASH).[MASH09] Most strong-post guardrail systems, however, have been around for many years and were tested under NCHRP Report 350, the then-current crash testing procedures. As such, the evaluations of the damaged systems should be consistent with crash testing procedures used in evaluating the original undamaged systems, so that performance degradation can be directly assessed. Full-scale testing was not within the scope of this study. Pendulum Testing As mentioned previously, the advantage of pendulum testing is that it allows for more precise control of boundary and loading conditions so that response behavior can be more accurately quantified. The disadvantage is that it is difficult (when at all possible) to extrapolate these results to overall crash performance of the system. In the context of this project, pendulum testing was used to (1) quantify strength degradation due to various damage modes (e.g., reduction in stiffness or energy absorption capacity of a post-and-rail system due to a damage mode or combination of damage modes), (2) evaluate failure modes that were not well suited for finite element analysis (e.g., tears and ruptures) and (3) for sub-system validation of the finite element models (e.g., pendulum test conditions were simulated to verify that specific aspects of the model were providing accurate response for those particular loading and boundary conditions). Computer Simulation Whereas full-scale crash testing is definitive but information poor, finite element simulation, in comparison, is less definitive because it is a mathematical simulation but extremely rich in detailed structural mechanical information. Stresses and strains of each part of the structure can be examined at every point in time and space of the crash allowing the analyst to determine exactly how close the system is to failure. Finite element analysis has become a reliable and widely accepted tool for investigating crashworthiness of roadside safety structures in simulated collision events. LSDYNA is a nonlinear, dynamic, explicit finite element code that is very efficient for the analysis of vehicular impact and is used extensively by the automotive industry to analyze vehicle crashworthiness.[LSDYNA03]

73 Guardrail Systems General descriptions of several common non-proprietary strong-post guardrail systems and their crash test results were provided in the “Background” section of this report. A summary of each of those system’s components is shown below in Table 17. As indicated in the table, many of these systems have similar designs and share several of the same components and are therefore susceptible to similar damage modes as the modified G4(1S) examined in Report 656. In many of these cases, the performance degradation due to a particular damage mode should likewise be similar; thus, analysis results for a particular damage mode may be used to infer the response for systems that share those same system elements. For example, the G4(1W), G4(2W) and the G4(2W) with round posts all use slightly different shaped wooden posts but otherwise share the same system components and over-all system design (e.g., rail height, post spacing, etc.). Although the posts are different (e.g., square, rectangular, and circular, respectively), they are similar in that they are all wood and are of similar dimensions (i.e., 8-10 inches in width/diameter and 6 ft in length); thus it could be inferred that each of these systems incur similar performance degradation due to each type and combination of damage mode(s). Similarly, all three systems use the same w-beam guardrail, blockout type and connection hardware, so it is likely that the effects of horizontal or vertical tears and holes will be very similar in all three systems as well. The results from the survey of practitioners indicated a strong desire for including damage assessment criteria for wood strong-post guardrail systems in the field guide (refer to Appendix B). Combining the responses from both survey groups resulted in the wood-post systems having the highest rating and the weak-post w-beam guardrail having the second highest rating; however, given the small number of respondents and the similarity in the weighted average scores for the thrie-beam guardrail, the modified thrie-beam guardrail and the weak-post w-beam guardrail, a case could be made for including any or all systems in the study. Critical Damage Modes As mentioned in the introduction, this study will expand on the information already in the field guide by including additional guardrail systems and/or additional damage assessment criteria for the modified G4(1S). The following sections discuss the common damage modes for several of the guardrail systems listed in Table 17, the importance of addressing them in this study, and the research methods that could be used in quantifying the effects of various levels of damage for each case.

74 Table 17. Summary of common strong and weak post guardrail systems. Material Type Height (inches) Gauge Material Type Length (inches) Embed- ment (inches) Spacing (inches) Material Type Depth (inches) Diameter (inches) Length (inches) Modified G4(1S) Steel W-beam 27 12 Steel W6x9 72 43.3 42 Wood 6 x 8 x 14 8 0.625 10 G4(2W) Steel W-beam 27 12 Wood 6 x 8 72 43.3 42 Wood 6 x 8 x 14 8 0.625 18 G4(2W) - Round Wood Steel W-beam 27 12 Wood 8-in Dia. 72 43.3 42 Wood 6 x 8 x 14 8 0.625 18 G4(1W) Steel W-beam 27 12 Wood 8 x 8 72 43.3 42 Wood 8 x 8 x 14 8 0.625 18 MGS Steel W-beam 31 12 Steel W6x9 69 37.2 42 Wood 8 x 12 x 14 12 0.625 18 MGS - Wood Post Steel W-beam 31 12 Wood 8-in Dia. 69 37.2 42 Wood 8 x 12 x 14 12 0.625 22 G9 Thrie-Beam Steel Thrie-beam 31.6 12 Steel W6x9 78 45.4 42 Steel W6x9 5.9 0.625 1.38 G9 Thrie-Beam - Wood post Steel Thrie-beam 32.5 12 Wood 6x8 78 45.4 42 Wood 8 x 8 x 14 8 0.625 10 Modified Thrie-Beam Steel Thrie-beam 34 12 Steel W6x8 81 46 42 Steel M14x17.2 14 0.625 1.38 Modified G2 (Weak-Post W-Beam) Steel W-beam 32.3 12 Steel S3x5.7 63 30.7 84 - - - 0.775 1.38 Rail-to-BlockRail Post Blockout Guardrail Type

75 G4(1W) and G4(2W) Wood-Post W-Beam Guardrails The G4(1W) and G4(2W) guardrail systems contain many of the same components and dimensions as the modified G4(1S) guardrail. Thus, many of the common damage modes for these systems are very similar to those of the modified G4(1S), such as w-beam damage, splice damage, twisted or missing blockouts, missing posts, post-rail separation, rail flattening and end terminal damage. As such, the damage assessment criteria developed in Report 656 for those specific damage modes on the G4(1S) should be applicable to the G4(1W) and G4(2W) systems. Damage modes related to the guardrail posts, on the other hand, may result in a different response for wood post systems when compared to the response for steel post systems. For example, under relatively severe impact conditions, as shown in Figure 59, wooden posts tend to either fracture at the ground line, split along the vertical direction of the post, or remain essentially undamaged (i.e., the posts simply rotate in the soil); whereas the steel W6x9 posts of the G4(1S) system, which have relatively low torsional rigidity, tend to twist as they are pushed back during impacts and subsequently bend about the weak axis of the post, as illustrated in Figure 60. Although the impact performance may be somewhat similar for these two systems when struck either in their undamaged state or when damage levels are relatively low, their performance under somewhat higher levels of damage may change due to the differing damage modes associated with the two very different guardrail post types. Figure 59. Typical damage modes for wood post guardrails.[Bullard00;Mak95] Figure 60. Typical damage mode for steel wide-flange posts.[Fleck08b]

76 Posts There are five basic damage modes for wood guardrail posts: (1) deflection, (2) rotted or weakened, (3) soil eroded away from post, (4) split post and (5) missing or fractured post. Based on the survey of practitioners, the most common damage mode for wooden posts in strong-post guardrails is post deflection, followed by rotted (or weakened) posts (see Appendix B). The effects of these two damage modes were evaluated and quantified for the G4(2W) wood-post guardrail in this study (see Chapter 8). Post Deflection Recall that this damage mode was evaluated in full-scale Crash Test C08C3-027.2 in Report 656 which showed that the damaged guardrail was unable to contain the vehicle and the vehicle overrode the guardrail.[Fleck08b; Gabler10] It was stated in Report 656 that, “The first MGA impact, a low-speed collision intended to cause a minor amount of deflection, was successfully reproduced. A simulation speed of 32 mph (52 km/hr) was required to reproduce the 14.5 inches (368 mm) of deflection observed in the 30 mph (48.3 km/hr) crash test. For the second MGA crash test, initial attempts at reproducing the results were unsuccessful. After an investigation, a critical factor in the outcome of the crash test was found to be the failure of a single post, located roughly 12.8 feet (3.9 meters) downstream of the impact point, to separate from the rail during both the first and second impacts. The addition of a constraint on the same post in the finite element model resulted in a drastic change in the predicted outcome of the impact, changing a successful crash into a failure with the vehicle vaulting over the guardrail.”[Gabler10] It was thus implied in Report 656 that the reason for the override was the result of a post- rail connection that did not release properly, as shown in Figure 61. As the guardrail post located downstream of the impact point deflected and rotated back, the undetached rail was pulled down with the post, allowing the vehicle to override the system. Figure 61. Post-test photo for Test MGA C08C3-027.2 showing the un-failed rail-post connection which resulted in the post pulling the rail down during impact.[Fleck08b]

77 Since the rail-post connection is “designed” to release as post deflection increases, it was not clear that the test was successful in confirming the effects of rail deflection on guardrail performance; but rather in confirming the effects of improper release of the rail-post connection. So, the real question was then, “What caused the rail-post connection to function improperly in the test?” The most likely answer was a loss of tension in the rail resulting from the excessive movement of the anchor system, as shown in Figure 62. Figure 62. Post-test photo for Test MGA C08C3-027.2 showing the excessive anchor movement that occurred during the test.[Fleck08b] The most cost effective method for quantifying the effects of post/rail deflection on system performance is to conduct finite element analyses of crash events. The basic methodology would be to (1) simulate impact of the guardrail at a low velocity; (2) save all nodal deformations and residual stresses in the barrier components; (3) re-initialize the guardrail under gravity; and (4) simulate impact on the damaged system at MASH or Report 350 TL-3 conditions. FEA should provide reasonably accurate information regarding vehicle stability, potential for barrier override, the relative increase in rail forces due to increased pocketing, the relative increase in stress concentration around connection points (e.g., splice bolt holes) due to increased rail forces, etc. The effects of post-rail deflection for the modified G4(1S) guardrail was investigated using finite element analysis in Report 656. It was shown that, when other damage modes were eliminated, there was very little difference in guardrail performance when guardrail deflections were less than 11 inches. Although it is expected that low levels of post/rail deflection will have similar effect for the G4(1W) and G4(2W) as it did for the G4(1S), it was necessary to investigate this damage mode for the G4(2W) to verify this assumption and to better quantify rail-post deflection for wood post systems as will be discussed in a later section. Rotted or Weakened Post One challenge regarding evaluation of rotted or weakened posts was how to quantify the degree of rot or insect damage in terms of post strength. Figure 63 shows a photo of a wooden guardrail post that fractured just below the groundline due to rot. Although deterioration at the

78 location of the break seems evident in the photo, the fact that the damage was below grade and not visible makes this type of damage difficult to identify. Figure 63. Post fracture below groundline due to rot. [Photo provided by Mark Bloshock] As stated by Dave Olson of Washington State DOT, “perhaps the answer is that any detectable rot or insect damage is sufficient reason to replace the posts, knowing that these conditions would be expected to progress and worsen over time. (Experience has taught that) visual inspection is not sufficient as rot typically starts below ground and often starts from the inside of the post and works outward ” [Olson12] The Washington State DOT is currently promoting research into methods for identification of in situ test methods for determining degradation of wood posts.[Olson12] There have been studies conducted to evaluate the effects of different strengths of posts on guardrail performance, such as wood grade (refer to the literature review), but those studies involved cases in which all the posts in the guardrail system were of the same type and grade of wood (i.e., same strength). A more critical situation may be when a damaged post (including impact damage such as split posts) is upstream and adjacent to an undamaged post, which would increase the chance for pocketing during impact. The suggested methodology for evaluating this damage mode was to use pendulum testing to quantify the loss of post strength as a function of rot or weakness of the post. Crash analyses could then be performed using FEA to quantify the effects of the various degrees of post damage on system performance. The pendulum tests would also be used to validate the FEA models of the damaged posts. Soil Eroded Away From Posts A similar approach could be used to investigate the effects of soil eroded away from the posts (e.g., environmental or crash induced soil loss). This type of damage would effectively reduce the stiffness of the post-soil system and may have similar effects on guardrail

79 performance as that of weakened posts described above. For cases in which soil confinement is reduced by the same degree at every post, such as when posts are installed at the edge of a foreslope, the increased deflection is not expected to significantly degrade system performance. Recall from the literature review that the w-beam guardrail adjacent to 2:1 foreslope was successfully tested to NCHRP Report 350 TL-3. A more critical situation may be when the soil is eroded away from one or two isolated posts. In this case, the increased deflection of the guardrail when impacted at the lower stiffness section may result in pocketing as the vehicle approaches the stiffer downstream posts. The suggested methodology for evaluating this damage mode would be to use pendulum testing to quantify the loss of post strength as a function of the depth of soil erosion behind the post. Then conduct finite element analyses to simulate the effects of the various degrees of soil erosion on system performance – using the pendulum tests to validate the FEA models of the post-soil models. Missing Posts The damage modes involving missing posts are considered to be of lower priority for this study. In particular, the effects of missing posts in the G4(2W) system are expected to be similar to that for the modified G4(1S) studied in Report 656. Thus the assessment criteria for the modified G4(1S) should be applicable to the G4(2W) and G4(1W) for that particular damage mode. W-Beam Rail There are six basic damage modes for the rail element in a guardrail system: (1) deflection, (2) flattened, (3) vertical crush, (4) vertical tear, (5) horizontal tear and (6) hole in the rail. Each of these damage modes were evaluated for the modified G4(1S) in Report 656, except for the vertical crush damage mode. Based on the survey of practitioners, the most common damage mode for the rail element for the wood-post w-beam guardrail is rail deflection, followed by rail flattening and vertical crush (see Appendix B). The similarity in the overall design of the G4(1W) and G4(2W) wood-post guardrails to the design of the modified G4(1S) suggests that the effects of damage to the w-beam rail element would be similar for these systems. Thus the assessment criteria for the modified G4(1S) developed in Report 656 should be applicable to the G4(2W) and G4(1W) for all w-beam rail damage modes. Rail-Only Deflection As discussed in the literature review, the potential for rail rupture increases when the w- beam experiences excessive deflection leading up to a relatively stiff guardrail post (i.e., pocketing). Report 656 provides guidance for guardrail repair priority for this damage mode based on analyses at 3 and 6 inches of rail deflection with standard soil conditions. Higher rail deflections were not investigated since it was assumed that “since larger rail deflections generally do not occur without also deflecting the posts.”[Gabler10] However, there are some potential crash scenarios that may lead to such conditions, including low-speed impact on the mid-span between posts under high impact angle or rigid ground conditions (e.g., frozen soil, posts driven through asphalt, etc.). In each of these cases there could be significant rail deflections with little or no post deflections, other than slight twisting of the posts.

80 The damage mode of rail-only deflection was considered to be of low priority for this study. It did not make sense to evaluate the effects of pocketing on guardrail with rigid soil conditions unless the response of the undamaged guardrail under these soil conditions was known. The authors agree with the Report 656 conclusion that larger rail deflections are not likely to occur without some post deflection for the standard soil condition. However, it was expected that the evaluation criteria for this damage mode would be applicable to both the G4(2W) and G4(1S) system due to the similarity between these systems. Vertical Rail Crush Although it was not expected that vertical rail crush would significantly affect the structural capacity of the rail, the reduction in the cross-sectional height of the rail could affect the ability of the rail to “capture” and contain the vehicle (i.e., increase the potential for underride or override). Since this damage mode was not investigated in Report 656, it was therefore considered for the current study. The most efficient and effective analysis method for evaluating this damage mode would be to use finite element analysis to quantify the increase in the potential for override/underride (i.e., for pickup and small car, respectively) for various degrees of vertical crush. Wooden Blockouts There are four basic damage modes for guardrail blockouts: (1) twisted, (2) split, (3) rotted, and (4) missing. These damage modes are common to all strong-post guardrail systems that use wooden blockouts. Based on the survey of practitioners, the most common damage mode for blockouts on the wood-post guardrails was “twisted blockouts,” followed fairly closely by “split blockouts.” (see Appendix B) Twisted Blockouts Twisted blockouts were evaluated in Report 656 for the modified G4(1S) guardrail. It was determined that a twisted blockout was of low priority for repair based on the results of pendulum tests that showed that “the performance of (the) damaged barrier section was virtually identical to that of the undamaged strong-post barrier section.”[Gabler10] Recall that the overall design of the G4(1W) and G4(2W) are similar to the modified G4(1S) and that all three systems share the same 6x8x14-inch wooden blockout. The effects of a twisted blockout on the wood post guardrail systems are expected to be similar to the effects of this same damage mode on the steel post guardrail. Thus the assessment criteria for the modified G4(1S) developed in Report 656 should be applicable to the G4(2W) and G4(1W) for twisted blockouts. Split Blockouts As discussed in the literature section, the primary function of blockouts is to create separation between the post and rail, which (1) reduces the possibility of a vehicle’s tire impacting against the post and (2) helps to maintain critical rail height during guardrail deflection. In order to perform this function, the blockout must effectively transfer the loads from the rail element to the post. A split blockout should not significantly affect the ability of the blockout to transfer load to the post, but the reduction in width of the split blockout may affect the contact area between the rail and blockout, which may result in a different deformation mode for the splice. It does not seem likely that a split blockout would adversely affect the potential for splice rupture, based on the illustration in Figure 64; however, many of the full-scale crash tests that resulted in splice rupture also involved split blockouts, as shown in Figure 65.

81 The most appropriate analysis method for this damage mode would be to use pendulum tests to quantify the potential for splice rupture as a function of blockout width (measured at the interface between the block and the w-beam). Figure 64. Illustration of possible splice deformation mode resulting from a split blockout. Figure 65. Blockout split during Test NEC-1 (left) and in Test 405160-1-1 (right).[Polivka00a; Buth06] Missing Blockouts The effects of missing blockouts for the modified G4(1S) guardrail was addressed in Report 656. It was shown that the missing blockout on the steel wide-flange post could lead to rail tear as the w-beam pushes back against the post during impact. Recall from the literature review that Test No. 473750-1 of the weak-post guardrail resulted in rail rupture after a small “nick” formed at the bottom edge of the w-beam rail as the rail was pulled over the top of a post downstream of the impact point (refer to Figure 50). The 9-inch tear in the rail that occurred during the pendulum test in Report 656 would likely have resulted in complete rupture of the rail if such a tear occurred during a vehicle collision event. That is, the high tensile force in the rail Split Blockout

82 upstream of the vehicle would likely propagate the tear as the vehicle passed down-stream of the tear location. The dangers of the rail directly contacting steel wide-flange posts are well known, which is why backup plates are generally used in guardrail designs that do not use blockouts (e.g., weak-post w-beam guardrail) and for those that use wide-flange steel sections as blockouts (e.g., standard G4(1S) and modified thrie-beam). However, as mentioned in Report 656, there is less of a propensity for tearing when the w-beam interacts directly with solid wooden posts. FEA was used in Report 656 to evaluate vehicle stability due to missing blockout(s), and it was determined that, “the missing blockout case does result in elevated vehicle instability, but not to the extent of the missing post case.” As such, the damage modes of “missing blockouts” and “rotted blockouts” on wood-post guardrail systems are also considered to be of low priority in this study. There seems to be little evidence for an increased chance of rail rupture and it is expected that the vehicle kinematic response would be similar to that of the modified G4(1S) evaluated in Report 656. Thus the assessment criteria for the modified G4(1S) developed in Report 656 should be applicable to the G4(2W) and G4(1W) for missing (or rotted) blockouts. Connections There are three basic damage modes for guardrail connections in wood-post guardrails: (1) post-rail separation, (2) splice damage, and (3) end-anchor damage. Based on the survey of practitioners, each of these damage modes have similar rate of occurrence (see Appendix B). Post-Rail Separation The effects of a separated rail-to-post connection on guardrail performance was evaluated in Report 656 for the modified G4(1S) guardrail using finite element analysis. The similarity in the overall design of the G4(1W) and G4(2W) wood-post guardrails to that of the modified G4(1S) suggests that the effects of post-rail separation would be similar for these systems. Thus, the assessment criteria for the modified G4(1S) in report 656 should be applicable to the G4(2W) and G4(1W) regarding rail-to-post separation. Report 656 indicates that the separation of the rail from the post actually improved vehicle stability during impact and redirection. When only a single connection was broken, the deflection of the system was essentially identical to the undamaged guardrail performance; whereas when two posts were detached from the rail with a separation distance of three inches, guardrail deflection was increased 5.6 percent. The improvement in vehicle stability should be expected since early release of the rail from the post as the vehicle approaches the post is a fundamental part of the design of the post-rail connection, which is necessary for reducing the potential for severe snag on the post as well as for reducing the risk of the post pulling the rail down as the post rotates back during impact (refer to the literature review for more discussion). In fact, a more critical case would be a non-failed connection on a deflected system, such as occurred in Test C08C3-027.2 and shown in Figure 61. As mentioned earlier, the effects of post-rail separation was sufficiently addressed in Report 656. If further study on post-rail connection is considered, the focus should be on how much post-rail deflection is permissible before a non-released connection becomes critical. FEA may prove to be an effective analysis method for evaluating this damage mode. The challenge will be how to set up the analysis in such a way that the non-released connection is accurately

83 modeled (e.g., correct pre-damage to bolt and to the w-beam slot). There may be several factors that cause the connection not to release, such as:  Low deflection of the guardrail (i.e., insufficient deflection to fail the connection),  Bolt head installed too close to the edge of the w-beam slot (see further discussion in the literature review), or  Loss of tension in the rail which causes the rail to slip, resulting in the bolt head “jamming” into the corner of the slot, as shown earlier in Figure 61. Recall from the literature review that the force required to pull the bolt head through the slot at a splice connection when the bolt head is located at the edge of the slot is almost four times the amount of force required to pull the bolt head through the slot when the bolt head is located at the center of the slot at a non-splice connection.[Plaxico03] In fact, the MGS design includes moving the splices to the mid-span and increasing the slot-length in the rail specifically to alleviate this problem.[Sicking02] Further, as shown in the literature review, the force magnitude required to detach the rail from the post can vary from 4,050 lb to 14,500 lb depending on the position of the bolt head in the slotted rail, and if the connection is at a splice (refer to Figure 44). Therefore, analyses may indicate that the criteria for assessing rail-post connections at a splice may differ from that for connections at non-splice locations. Rail Splice Although w-beam splice damage was addressed in Report 656, additional splice damage modes should be considered for further investigation in this study to more thoroughly quantify their effects on rail capacity. For example, Report 656 evaluated the damage mode shown in Figure 66, which involved a simulated tear-out of a splice bolt at the lower up-stream section of the splice. Note, however, that the field installation photo in Figure 66 (left photo) shows the damage being located at the lower, downstream splice bolt, which would result in different loading (or stress concentrations around the bolt holes) than the one studied in Report 656 (right photo). For example, impact on a guardrail either upstream or downstream of the splice connection will cause the two rail elements to tend to separate at the downstream splice bolts; whereas the two sections of rail at the upstream splice bolts tend to compress together. Thus, rail rupture should be more likely when the downstream splice bolt holes are damaged, compared to similar damage in the upstream splice bolt location. This is particularly true for reverse impact cases, such as when a vehicle crosses the roadway and strikes the guardrail on the opposing roadside (e.g., vehicle approaches the splice connection in Figure 66(a) from the right). In such a case, the rail element on the right side of the splice in Figure 66(a) would bend back and “tend” to tear the splice bolt through the w-beam. Further, note that the rupture of the w-beam splice in full-scale crash Tests NEC-1, 405160-1-1 and C08C3-027 involved tearing along the bolt-line at the downstream splice bolts of the back-side rail, as shown in Figure 67.[Polivka00a; Buth06; MGA08b] Additional full-scale tests that resulted in w-beam rupture at the downstream splice bolts include TTI Tests 471470-23 and 405421-2.[Buth99a; Mak96c] Although the potential for splice rupture at the upstream splice bolts exists, the research team has only identified one full- scale test which resulted in rupture at that location – Test RF476460-1-5.[Bullard10]

84 Figure 66. Splice damage mode investigated in Report 656.[Gabler10] Figure 67. Splice rupture initiated at downstream splice bolts in (a) Test NEC-1, (b) Test 405160-1-1 and (c) Test C08C3-027.1. [Polivka00a; Buth06; MGA08b] To evaluate splice damage, the damage modes for the splice connection should be imposed onto the test sample by a pre-impact test, since this type of damage is difficult (if not impossible) to achieve otherwise. Since small tears in the splice bolt holes are likely to be hidden by the bolt head and/or nut, it may be reasonable to quantify the degree of damage by such parameters as degree of flattening, angle of bend in the splice, measure of separation between the two rail elements at the downstream end of the splice, etc. The evaluation for quantifying the reduction in rail capacity as a function of splice damage could be achieved through pendulum testing or a combination of pendulum testing and finite element analysis. Pendulum tests will obviously indicate when failure has occurred, but the relative increase in stress levels at the critical locations in the splice would be better assessed through FEA, as illustrated in the effective-plastic-strain contour plot of the weak-post w-beam system in Figure 68. (a) Field Installation damage (b) Simulation damage (a) (b) (c)

85 Methods for Damage Assessment for the G4(1W) and G4(2W) Table 18 provides a summary of possible methods for evaluating the various damage modes for the strong-post w-beam guardrail systems (i.e., G4(1W) and G4(2W)). An “x” indicates that the analysis method is applicable; an “x” in a shaded cell indicates the preferred method for analysis; and a “blank” cell indicates that the analysis method is not applicable. The table also includes a rating scale which indicates the damage modes considered most common for this system, based on the weighted average summary of the survey results. Figure 68. Effective plastic strain contour plot at a splice connection. [Ray01b] Thrie-Beam Steel Post Guardrail The primary components of the standard G9 thrie-beam guardrail system include 12- gauge thrie-beam rails, W6x9 steel posts, W6x9 steel blockouts, and standard connection hardware. Like the G4(1S) w-beam, the steel-post thrie-beam guardrail was modified from its original design by replacing the 6-inch deep W6x9 blockout with an 8-inch deep wooden blockout. This design change increased the blockout distance and reduced the tendency for torsional failure of the posts as the posts deflected back during impacts, resulting in improved crashworthiness of the system.[Bullard10] Although the current FHWA approved system includes 8-inch wooden blockouts, installations of the original G9 design with W6x9 blockouts likely still exist throughout the country. According to the survey of practitioners, the interest in including the strong-post thrie- beam guardrail in the update to the field guide was rated “medium” based on the weighted average of responses (i.e., weighted average value = 3.08 on a scale of 1 to 5). Approximately 1/5 of the respondents rated it as “medium”, 2/5 of the respondents rated it as “high” or “very high” and the remaining 2/5 of respondents rated it as “low” or “very low,” which indicates that there were very mixed opinions among the groups (see Appendix B). Based on the fact that the thrie-beam guardrail is expected to have different sensitivities to damage than strong-post w-beam guardrails, and the fact that current funding levels may not permit development of repair evaluation guidelines that are sufficiently comprehensive for Down-stream splice bolts

86 immediate field use, this system was considered to be of medium priority for inclusion in this study. Table 18. Summary of possible methods for assessing damage modes for the G4(1W) and G4(2W) guardrail systems. Posts Based on the survey of practitioners, the most common damage mode for guardrail posts in strong-post thrie-beam guardrails is post deflection, followed by soil erosion around the post and twisted posts (see Appendix B). Post-Rail Deflection The suggested methods for evaluating post-rail deflection for the G4(1W) and G4(2W) guardrails also apply to the thrie-beam guardrail systems. However, due to the additional height of the thrie-beam guardrail, compared to the common strong-post w-beam guardrail systems (i.e., 32.5 inches compared to 27 ¾ inches, respectively), deflection damage will not likely result in the same level of performance degradation for the thrie-beam guardrail. Also, as shown in Figure 69, the rail-to-post connection for the standard thrie-beam guardrail consists of two bolts at each post, which increases the chances for the connection to remain intact during low severity Extrapolate from Report 656 Pendulum Test FEA Deflection 4.26 x x Rotted or Weakened 3.65 x x Soil Eroded from Post 3.13 x x Split Posts 3.09 x x Missing Posts 2.73 x Twisted Posts 2.48 Deflection 4.25 x x Flattening 3.54 x Vertical Crush 2.91 x Vertical Tear 2.65 x Horizontal Tear 2.52 x Hole in Rail 2.39 x Twisted Blockouts 3.82 x Split Blockouts 3.26 x x Rotted Blockouts 2.61 x x Missing Blockouts 2.48 x Post-Rail Separation 2.82 x x x Splice Damage 2.81 x x x End Anchor Damage 2.77 x x * Priority is rated on a scale of 1 (very low) to 5 (very high) Legend x Preferred method for analysis. x Analysis method is applicable. N.A. Not applicable Connections Blockouts Component Damage Mode Weighted Avg. Priority from Survey (Scale 1-5)* Posts Rail Element Possible Methods for Evaluation

87 impacts. In such cases, the rail element would rotate back at an angle with the undetached post and effectively act as a ramp for the vehicle to climb and vault over the guardrail. As a side note, a possible design improvement for this system may be to exclude or weaken the upper bolt connection. Recall that one of the design changes for the modified thrie-beam guardrail was the exclusion of the lower bolt connection. Figure 69. Photo of the standard thrie-beam guardrail. This damage mode was considered to be of high priority for the G9 thrie-beam guardrail system due to the fact that post-rail deflection is a common damage mode for thrie-beam guardrail in low severity impacts; and also due to the increased potential for vaulting the deflected system - especially if the connections do not fail correctly or consistently. Finite element analysis would be the most practical method for evaluating performance degradation of deflected guardrail systems. Although full-scale testing is not in the scope of this study, the conclusions made from the computational analyses should be confirmed through full-scale testing if future work is performed. Soil Eroded Away From Posts This damage mode was considered to be of medium priority for evaluation in this study. The approach used to investigate the effects of soil eroded away from the posts for the G4(1W) and G4(2W) wood post guardrail systems also applies here for the G9 thrie-beam guardrail. Pendulum tests could be used to quantify the loss of post strength as a function of the depth of soil erosion behind the post. FEA could then be used to simulate the effects of the various degrees of soil erosion on system performance – using the pendulum tests to validate the FEA models of the post-soil models. Twisted Posts Steel wide-flange posts, unlike rectangular wooden posts, tend to twist and bend during impacts, as opposed to fracturing. As such, “twisted posts” are an inherent part of the damage mode of a deflected system and were thus considered to be of low priority for additional study beyond that of post-rail deflection. However, if the twisted post damage mode is included, then the most effective analysis method would be FEA.

88 Thrie-Beam Rail Based on the survey of practitioners, the most common damage mode for the rail element in a strong-post thrie-beam guardrail is deflection, followed by rail flattening (see Appendix B). Rail deflection was included as part of the post-deflection damage mode since it is not likely that one can exist without the other. Thus, rail deflection is considered to be of low priority for additional study beyond that of the combination post-rail deflection damage mode. Although rail tearing was not considered to be a common damage mode for thrie-beam rail elements, evaluation of vertical tears in thrie-beam rail may be warranted since that particular damage mode has been shown to be critical for w-beam guardrail systems. Alternatively, the damage assessment and repair criteria developed in Report 656 for the w-beam rail may serve as conservative assessment criteria for the thrie-beam rail. Evaluation of horizontal tears and holes in the thrie-beam rail are considered to be of low priority based on the pendulum test results from Report 656 which showed that those damage modes were of lower significance than a vertical tear for the w-beam rail. Rail crush is also considered to be of low priority in this study since the cross-section height of the thrie-beam is not likely to be reduced enough to significantly affect the ability of the rail to capture and contain the vehicle. In other words, even with substantial vertical crush the thrie-beam would still be taller than the undeformed w-beam, and thus should have sufficient vertical height to capture an impacting vehicle unless other damage modes are also present. In Report 656 it was shown that vehicle instability and vaulting were likely when the w- beam rail was 75% or more flattened. The flattened rail increased the probability for the vehicle’s tire to mount and climb the rail. The bottom edge of the thrie-beam rail is lower to the ground than the w-beam, thus rail flattening may pose a greater risk for the thrie-beam systems. Blockouts Based on the survey of practitioners, the most common damage mode for blockouts in the thrie-beam steel post guardrail is twisted blockout, followed by split blockout and missing blockout (see Appendix B). Based on the responses of the survey respondents, it is assumed that these damage modes were rated based on the thrie-beam guardrail with wood blockouts. Since there may be many G9 guardrails with steel blockouts still in existence, the evaluation criteria should include both the steel and wood blockout versions of the system. Twisted Blockouts Since the blockout is fastened to the post with two bolts in the G9 guardrail, it seems unlikely that the blockout could rotate unless one of the bolted connections failed. It is difficult to say for certain that a twisted blockout in the G9 guardrail would be of higher or lower risk for system failure than it was for the G4(1S) guardrail studied in Report 656, but the risk was considered to be relatively low for the wood blockouts compared to other critical damage modes. For the standard G9 with wide-flange steel blockouts, on the other hand, there may be an increased risk for tearing the rail against the sharp edges at the top and bottom of the wide-flange blockout when the blockout is rotated on the post. Finite element analysis could be used to investigate the potential for increased stress concentrations and overall crash performance; then if the results indicate that twisted blockouts may lead to poor performance (e.g., rail rupture), pendulum tests could be used to confirm the results.

89 The research team considers the evaluation of the twisted blockout to be low priority in this study, since this damage mode is unlikely to occur unless (1) the rail is detached from the blockout and (2) the blockout is detached from the post. For the wood blockout system, it is probable that the rail could separate from the block, but much less probable that the blockout would also be detached from the post at the same location. This seems to be a highly unlikely event unless bolts were either left out during installation or there was significant damage to the system. Split Blockouts The split blockout damage mode only pertains to wooden blockouts. As discussed for the G4(2W) system, a split blockout should not significantly affect the ability of the blockout to transfer load to the post, but the reduction in width of the split blockout may affect the contact area between the rail and blockout, which may result in a different deformation mode for the splice. If this damage mode is determined to be critical to the performance of the G4(2W) guardrail, then it would likewise be considered to be of higher priority for the G9 as well. Otherwise, this damage mode was considered to be of low priority in this study. Missing Blockouts As was shown in the Report 656, a missing blockout on the steel wide-flange post could lead to rail tear as the w-beam pushes back against the post during impact. The effects of missing and rotted blockouts for the thrie-beam guardrail should be similar to, or less severe than, that for the G4(1S). Thus the assessment criteria for the modified G4(1S) developed in Report 656 may be extrapolated to the thrie-beam guardrail systems. The damage modes of “missing blockouts” and “rotted blockouts” were considered to be of low priority in this study. Connections The post-rail connection was considered to be high priority for evaluation of the G9 system since, as mentioned previously, proper release of the connection may prove to be a critical aspect of system performance. The release of the connection is not only a function of deflection but is also a function of bolt-position in the rail-slot and rail tension (e.g., anchoring). Further, these aspects are not likely independent of each other. For example, if the anchor system yields and releases the tension in the rail, the forces on the connection would reduce, which could result in large deflections, as well as undetached rail-post connections. Pendulum testing may be appropriate for evaluating the release of the post-rail connection as a function of (1) rail deflection, (2) bolt-position relative to slots in the thrie-beam rail, and (3) longitudinal movement of the rail element during lateral deflection. Care should be taken to ensure that the bolts are properly position in the thrie-beam slot so that the effects of bolt position can be adequately quantified. Finite element analysis would be the most effective method of quantifying the effects of various levels of damage on guardrail performance, given that the finite element models can be adequately validated with results of the pendulum tests. Methods for Damage Assessment for the G9 Table 19 provides a summary of possible methods for evaluating the various damage modes for G9 thrie-beam guardrail.

90 Table 19. Summary of possible methods for assessing damage modes for thrie-beam guardrail systems. Modified Thrie-Beam Guardrail The primary components of the modified thrie-beam guardrail system include the same 12-gauge thrie-beam rails, W6x9 steel posts, and standard connection hardware as the standard G9 thrie-beam guardrail. In addition to a taller rail height, the modified thrie-beam guardrail also has a unique deep structural steel blockout, which is designed to help keep the face of the rail vertical during impacts, but is unfortunately susceptible to bending through the web (see Figure 42). So, although many damage modes will be similar for these two systems, the differences in the damage modes for the blockout may result in significant differences in guardrail performance. Extrapolate from Report 656 Pendulum Test FEA Deflection 3.80 x Soil Eroded from Post 2.93 x x Twisted Posts 2.87 x Missing Posts 2.43 x Rotted Posts 1.93 x Split posts 1.50 x x Deflection 3.83 x Flattening 3.08 x x Horizontal Tear 2.50 x x Vertical Crush 2.45 x Vertical Tear 2.42 x x Hole in Rail 2.17 x x Twisted Blockouts 3.17 x x x Missing Blockouts 2.17 x x Bent Blockouts 2.00 x x Twisted Blockouts 3.17 x x x Split Blockouts 2.42 x x Rotted Blockouts 1.83 x x Missing Blockouts 2.17 x x End Anchor Damage 2.90 x x Splice Damage 2.73 x Post-Rail Separation 2.27 x x * Priority is rated on a scale of 1 (very low) to 5 (very high) Legend x Preferred method for analysis. x Analysis method is applicable. N.A. Not applicable Wood Blockouts Connections Steel Blockouts Component Damage Mode Weighted Avg. Priority from Survey (scale of 1-5) Possible Methods for Evaluation Posts Rail Element

91 The modified thrie beam is not widely used; therefore, this system was considered to be of low priority in this study. The G9 guardrail was considered to be of higher priority than the modified thrie-beam due to the fact that (1) the G9 is more widely used than the modified thrie- beam and (2) many of the damage assessment criteria developed for the G9 would also apply to the modified thrie-beam guardrail. The damage mode evaluations for the modified thrie-beam guardrail would be similar to those discussed for the standard G9 guardrail in the previous section. The primary differences in damage assessment would arise from the differences in the blockout design and the reduced number of connections for the rail-to-blockout. Posts Based on the survey of practitioners, the damage mode for posts that is most common with the modified thrie-beam is post deflection, followed by posts with soil erosion around them, and posts that are missing (see Appendix B). Refer to the earlier section of standard thrie beam guardrail posts for discussion of damage types and proposed evaluation methods. It is expected, however, that additional damage modes will be introduced for impacts on the modified thrie- beam system at relatively low rail deflections due to the low collapse load of the blockouts. Thus, it may not be possible to fully evaluate the post-rail deflection without considering a more involved multi-damage mode evaluation study. Rail Based on the survey of practitioners, the damage mode for rail elements that is most common for the modified thrie-beam guardrail is rail deflection. Tearing and crushing of the rail both vertically and horizontally were all rated roughly the same by the respondents, and rated much lower than rail deflection. (see Appendix B). Blockouts Based on the survey of practitioners, the damage mode for blockouts most common in the modified thrie-beam guardrail is a twisted blockout, followed by missing blockout and bent steel blockout (see Appendix B). However, as was the case with the standard G9 thrie-beam guardrail, it does not seem likely that the blockout could twist unless one of the bolts connecting the blockout to the post was broken or missing. It is also unlikely that a low-speed impact would result in breaking one of these bolts, which are 5/8-in diameter high-strength steel bolts. In order for a blockout to be missing, both bolts would need to be broken. It is likely, however, that the blockout could be bent as the result of a low-speed impact. A bent blockout would result in a reduced blockout distance between the rail and post, which would increase the chance for vehicle impact against the posts. The bending of the blockout also induces a torsional moment onto the posts, which may reduce its effective stiffness. Based on these observations, the bent blockout was considered to be of high priority for the modified thrie-beam guardrail system. Connections Based on the survey of practitioners, the most common damage mode regarding connection elements for the modified thrie-beam guardrail is splice damage, followed by damage to the guardrail cable anchor (see Appendix B). These damage modes are similar to those for the standard thrie-beam guardrail discussed in a previous section. The effects of rail-to-post connection will be different for the modified G9 system compared to the standard G9, since the modified thrie-beam guardrail only uses a single bolt to connect the rail to the blockout at each

92 post. In this case, the evaluation methods would be similar to those described for evaluating the rail-to-post connection for the G4(2W). Methods for Damage Assessment for the Modified Thrie-Beam Table 20 provides a summary of possible methods for evaluating the various damage modes for the modified G9 thrie-beam guardrail. Table 20. Summary of possible methods for assessing damage modes for the modified thrie-beam guardrail system. MGS – Steel Post W-Beam Guardrail Although the MGS guardrail consists of the same w-beam rail, W6x9 steel posts, and connection hardware used in the modified G4(1S), there are several fundamental design differences between these two systems. The primary differences being that the MGS is four inches higher, the splices are moved to the mid-span, the posts have shallower embedment depth, and the blockout distance is 50% greater. In minor impacts, it is expected that the damage modes for the MGS would be similar to that of the modified G4(1S); however, the response of these two systems, given the same level of damage, may differ considerably due to the fundamental differences of their designs. For example, the increased height of the MGS makes it less sensitive to small decreases in rail height, but may make it more sensitive to small increases Extrapolate from Report 656 Pendulum Test FEA Deflection 2.89 x Soil Eroded from Post 2.56 x x Missing Posts 2.25 x Twisted Posts 2.11 x Deflection 3.50 x Flattening 2.63 x Horizontal Tear 2.50 x x Vertical Tear 2.50 x x Vertical Crush 2.38 x Hole in Rail 2.00 x x Twisted Blockouts 3.00 x x Missing Blockouts 2.29 x Bent Blockouts 2.14 x Splice Damage 2.71 x End Anchor Damage 2.33 x Post-Rail Separation 1.86 x x * Priority is rated on a scale of 1 (very low) to 5 (very high) Legend x Preferred method for analysis. x Analysis method is applicable. N.A. Not applicable Posts Rail Element Steel Blockouts Connections Component Damage Mode Weighted Avg. Priority from Survey (scale of 1-5) Possible Methods for Evaluation

93 in rail height (e.g., increase potential for under-ride); the splice connection located at the mid- span between posts results in less severe loading on the splice (e.g., no bending), which means less sensitivity to splice damage. Likewise, a missing post in combination with a deflected rail may result in a higher probability for a vehicle with low profile (e.g., Geo Metro) to push underneath the higher rail element and impact with the adjacent downstream post. The MGS guardrail was not included in the survey of practitioners, and none of the survey respondents included the MGS system when asked “which additional guardrail systems should be included in this study.” The MGS was therefore considered to be of lower priority for this study; however, given the success of the MGS in crash testing, it is likely that the MGS (or systems with similar design) will see increased use in the near future and may be considered with higher priority in future projects. Methods for Damage Assessment for the MGS with Steel Post Table 21 provides a summary of possible methods for evaluating the various damage modes for the MGS with steel posts. Table 21. Summary of possible methods for assessing damage modes for the MGS with steel posts. Extrapolate from Report 656 Pendulum Test FEA Deflection N.A. x Soil Eroded from Post N.A. x x Missing Posts N.A. x Twisted Posts N.A. x Deflection N.A. x Flattening N.A. x Vertical Crush N.A. x Vertical Tear N.A. x Horizontal Tear N.A. x Hole in Rail N.A. x Twisted Blockouts N.A. x x Split Blockouts N.A. x x Rotted Blockouts N.A. x x Missing Blockouts N.A. x x Post-Rail Separation N.A. x Splice Damage N.A. x x End Anchor Damage N.A. x * Priority is rated on a scale of 1 (very low) to 5 (very high) Legend x Preferred method for analysis. x Analysis method is applicable. N.A. Not applicable Possible Methods for Evaluation Posts Rail Element Blockouts Connections Component Damage Mode Weighted Avg. Priority from Survey (scale of 1-5)

94 MGS – Wood Post W-Beam Guardrail The wood-post version of the MGS is expected to experience the same basic damage modes under low speed impact as the standard steel-post MGS system. As such, the assessment criteria for damage modes that are common to both systems should be applicable to both. However, damage modes specific to the guardrail posts, as discussed previously, may result in a different response between the wood post MGS and the steel post MGS. Thus, the additional damage modes that would need to be included for assessing the wood-post version would essentially include quantification of performance degradation due to (1) post/rail deflection and (2) rotted or weakened posts. Methods for Damage Assessment for the MGS with Wood Posts Table 22 provides a summary of possible methods for evaluating the various damage modes for the MGS guardrail with wood posts. Table 22. Summary of possible methods for assessing damage modes for MGS with wood posts. Extrapolate from Report 656 Pendulum Test FEA Deflection N.A. x Rotted or Weakened N.A. x x Soil Eroded from Post N.A. x x Split Posts N.A. x Missing Posts N.A. x Twisted Posts N.A. x Deflection N.A. x Flattening N.A. x Vertical Crush N.A. x Vertical Tear N.A. x Horizontal Tear N.A. x Hole in Rail N.A. x Twisted Blockouts N.A. x x Split Blockouts N.A. x x Rotted Blockouts N.A. x x Missing Blockouts N.A. x x Post-Rail Separation N.A. x Splice Damage N.A. x x End Anchor Damage N.A. x * Priority is rated on a scale of 1 (very low) to 5 (very high) Legend x Preferred method for analysis. x Analysis method is applicable. N.A. Not applicable Possible Methods for Evaluation Posts Rail Element Blockouts Connections Component Damage Mode Weighted Avg. Priority from Survey (scale of 1-5)

95 G2 – Weak-Post W-Beam Guardrail The G2 weak-post guardrail was described in the Background section and is repeated here for convenience. The weak-post w-beam guardrail is composed of w-beam rails supported on weak S3x5.7 steel posts with rectangular soil plates. The system performs much like the cable guardrail in that the posts hold up the rail at the proper height until the guardrail is struck by an errant vehicle. The posts are spaced at 12.5 feet and the rail is connected to the posts using 5/16- inch diameter bolts with 1-3/4 inch square washers under the head. The bolts are designed to fail in an impact allowing the rail to separate easily from the post. The rail separation from the post is an important feature of the design since this action allows the rail to remain in contact with the vehicle instead of being pulled to the ground by the post. Once the rail is separated from the post, the posts bend at the ground-line allowing the w-beam to slip over the top of the posts. As the rail continues to deflect, tension develops and the rail effectively behaves as a cable (or ribbon) anchored at the ends. Ray et al. demonstrated that relatively small changes in several important design details can significantly affect performance of the weak-post w-beam guardrail.[Ray01a] Although the weak post w-beam guardrail shares many of the same components as the strong-post systems, the basic function of weak post systems is quite different and is expected to have different sensitivities to various levels of damage. As discussed in the literature review, the performance of the weak-post guardrail is likely to be much more sensitive to: (1) post-rail connection damage due to weaker posts and no blockouts; (2) rail tears due to higher tensile forces in rail during impact; (3) guardrail height due to the potential for over-ride and underride given the large rail deflections and length of unsupported section of rail at the point of maximum deflection; (4) and missing backup plates which protect the w-beam rail from the sharp edges of the small-flange posts. Recall from the literature review that rail rupture occurred in Test 473750-1 as a result of a small nick in the rail as the w-beam was pulled over the top of one of the guardrail posts.[Buth00a] Finite element analysis would be the most cost effective analysis method for studying/quantifying many of the damage effects for the G2. For example, a low severity crash could be simulated (e.g., low-speed low angle) for the purpose of imposing realistic damage modes onto the system, including the associated residual stresses in the system components. Then, additional specific damage modes could be directly imposed onto the model such as:  Additional detached rail-post connections (which under gravity would result in lower rail height over a relatively long span of the rail),  Undetached post-rail connections (which would result in lower rail height due to post pulling the rail down),  Various levels of rail-tear damage and  End-terminal damage which would allow greater rail deflections. Pendulum testing may be applicable for evaluating rail tears on the w-beam rail element, if appropriate boundary conditions can be achieved for the sub-system tests. Also, weak post guardrails have much larger deflections than strong post guardrails, and it may not be possible to achieve a constant loading height on the rail as the pendulum swings into its arc. These issues will have to be considered when developing the testing plan for the G2 system. Based on the fact that the G2 weak-post guardrail is expected to have significantly different sensitivities to damage than strong-post guardrails, and the fact that current funding

96 levels may not permit development of repair evaluation guidelines sufficiently comprehensive for immediate field use, the G2 guardrail was considered to be of medium priority for this study. Although the modified G2 guardrail (i.e., 32.25 inch tall, splice at mid-span, modified rail-post connection, and backup plates at posts) is the only weak-post w-beam guardrail that currently meets crashworthiness criteria for both Report 350 and MASH, installations of the original G2 are still common in the field. It is suggested that any damage to an original G2 system should warrant high priority for repair, due to its susceptibility to unacceptable performance even in its undamaged state. Methods for Damage Assessment for the Modified G2 Weak-Post Guardrail Table 23 provides a summary of possible methods for evaluating the various damage modes for the modified G2 weak-post guardrail. Table 23. Summary of possible methods for assessing damage modes for the Modified G2 guardrail. End Terminals The primary purpose of end-terminals is to anchor the ends of a guardrail, but end- terminals must also be crashworthy themselves. In fact, end-terminals are much more complex systems than guardrails. Guardrails are designed for one basic type of loading, (i.e., lateral impact on the traffic-facing side of the system). Accordingly, only two tests are required in MASH to assess the crashworthiness of guardrails: Test 3-10 (small car impact at 62 mph and 25 degrees) and Test 3-11 (pickup truck impact at 62 mph and 25 degrees). Extrapolate from Report 656 Pendulum Test FEA Deflection 4.00 x Twisted Posts 3.00 x Soil Eroded from Post 2.94 x Missing Posts 2.82 x Deflection 4.13 x Flattening 3.19 x Vertical Crush 2.69 x Vertical Tear 2.60 x x Hole in Rail 2.20 x x Horizontal Tear 2.13 x x Post-Rail Separation 3.07 x Splice Damage 3.00 x x Missing Back-Up Plate N.A. x x End Anchor Damage 2.14 x * Priority is rated on a scale of 1 (very low) to 5 (very high) Legend x Preferred method for analysis. x Analysis method is applicable. N.A. Not applicable Possible Methods for Evaluation Posts Rail Element Connections Component Damage Mode Weighted Avg. Priority from Survey (scale of 1-5)

97 End-terminals, on the other hand, can be impacted on the side or on the end. The loading of an end-terminal’s components, and hence their intended function, will vary depending on several factors including: (1) which end of the guardrail the end-terminal is located (e.g., down- stream or upstream), (2) where the end-terminal is struck (side hit or end-on), (3) the orientation of the vehicle (frontal impact or side impact), (4) impact direction relative to up-stream or down- stream placement (forward or reverse hit), and (5) the end-release function of the end-terminal (i.e., gating or non-gating). These additional functions of an end-terminal are reflected in the number of tests and impact conditions that are required in the crash testing guidelines.[Ross93; AASHTO09] Because of these basic differences in function, the damage criteria assessments derived from guardrail performance studies may not correspond directly to end-terminals; thus damage modes for end-terminals will need to be evaluated using criteria appropriate for the various functions of those systems. The buried-in-backslope end-terminal is the only non-proprietary end-terminal that meets FHWA eligibility for NCHRP Report 350 TL-3; this system, however, has limited application since it can only be installed in locations where there is a backslope. Although several older, largely obsolete end-terminal installations are still in service, it is expected that most of these will eventually be replaced by accepted systems; thus, any damage assessment criteria developed in this study for end-terminals would likely be focused on proprietary systems. Unfortunately, there are several of these systems and they all function somewhat differently (e.g., some are gating, some are non-gating, each has a unique energy absorbing mechanism, etc.). Because of the complexity and the proprietary nature of these systems, it would be difficult to conduct comprehensive assessments without the consent and participation from manufacturers. Concerning their function as part of a guardrail system, however, essentially all proprietary end-terminals share many of the same basic features and components that function to ensure that the guardrail maintains proper tension during impacts (e.g., cable anchor mechanism, soil foundation tubes, ground-line strut, bearing plates, etc.). End-terminals also function in much the same way as guardrails when hit on the side. Thus, evaluation criteria for damage modes which affect an end-terminal’s function as a guardrail anchor system or which affect its performance as a guardrail (i.e., side hits) were considered to be of high importance. Report 656 did not evaluate damage modes for guardrail end-terminals, but did provide rationale for repair criteria based on an Ohio Department of Transportation Energy Absorbing End Terminal Maintenance Checklist. The assessment criteria for end-terminals included damaged end-post, missing or slack anchor cable, improper stub height, missing or failed lag bolts on impact head, and bearing plate. A summary of generic end terminal repair guidance from Report 656 is shown in Table 24. Evaluation criteria for additional damage modes should be included, such as soil eroded away from anchor posts and missing or damaged ground-line strut. These damage modes could affect the end-terminal’s ability to properly anchor the guardrail during impacts, which could in turn lead to excessive rail deflection, improper release of rail-post connection, and pocketing. Further, whereas local damage to a guardrail results in a relatively isolated damage-affected section of the guardrail (i.e., limited exposure to future impacts in the damage region), a damaged anchor will affect the performance of the entire guardrail system (i.e., increased risk of exposure). Due to the importance of the end-terminal to guardrail performance, the research team considers that the additional damage evaluations for end- terminals to be of high priority.

98 Table 24. Summary of generic end terminal repair guidance. [Gabler10] Methods for Damage Assessment for Generic End-Terminal Table 25 provides a summary of possible methods for evaluating the various damage modes for a generic end-terminal. Table 25. Summary of possible methods for assessing damage modes for generic end terminals. Included in Report 656 Pendulum Test FEA Deflection x Rotted or Weakened x x x Split Posts x x x Stub Height x x x Bearing Plate x x x Soil Eroded from Post x x Groundline Strut x x Anchor Cable x x x Cable Anchor Bracket x x Legend x Preferred method for analysis. x Analysis method is applicable. N.A. Not applicable Posts Foundation Connections Component Damage Mode Possible Methods for Evaluation

99 Guardrail Transitions Transition systems are used to join two barriers of different stiffness, such as a strong- post guardrail to a rigid bridge rail. The function of these systems is to provide a gradual transition from the more flexible system to the stiffer system to prevent “pocketing” and “snagging” on the approach end of the stiffer system. These systems generally share the same (or similar) components that are used in strong-post guardrail (e.g., w-beam or thrie-beam rail, steel or wood posts, rail-to-post fasteners, splice connections, etc.). The increase in stiffness is accomplished simply by using smaller spacing of the posts and stronger posts as the system approaches the more rigid structure. So, although they share the same components, the resulting response of the components may be quite different. For example, as the transition approaches the stiffer barrier, the w-beam rail functions more as a shear-beam transferring load to the posts, which results in less tensile force being developed in the rail element compared to the same impact on a guardrail. Thus transition systems will have different sensitivities to w-beam damage modes compared to typical strong-post guardrails. Guardrail transitions were not included in this study; however, the results from the survey of practitioners indicated that the “steel-post w-beam to rigid barrier transition” and the “thrie- beam to rigid barrier transition” was of greater interest than other transition types and should therefore be considered in future studies involving damage assessment of roadside safety barriers (see Appendix B). Combination Damage Modes The damage analyses in Report 656 consisted primarily of isolated damage modes. In most real-world impacts any given damage mode is usually associated with several other damage modes. For instance, rail/post deflection is usually accompanied at a minimum by a deformed w- beam rail and post-rail separation. The net effect for a given group of damage modes will not always be the sum of the individual effects. As mentioned earlier, rail-post separation is generally considered to be a damage mode, but when rail deflection is in excess of some critical amount, an undetached rail-post connection would likely be a more hazardous situation than a detached connection. As stated in Report 656, “A critical contribution to the vaulting of the vehicle in the MGA crash test was believed to be the failure of some of the posts to detach from the guardrail.” All post and beam guardrails can experience any of the damage modes listed in Table 26. Any of these modes can be observed in combination with any of the others. The net effect on barrier performance may be somewhat different for each type of guardrail system but in principal each system has the same potential range of combinations. Several specific combinations of damage modes that may be of interest in this study are discussed below. Combination of Post and Rail Deflection and Rail-to-Post Connection Recall that FEA was used in Report 656 to evaluate guardrail response as a function of rail-post deflections; rail deflections of 3, 6, 9, 11 and 14 inches were investigated in that study. Two series of analyses were conducted in order to bracket the crash performance. In the first series, the rail was allowed to detach normally from the post during the impact event. In the second series, the rail was constrained from releasing at a critical post location. The first series indicated that even with 14 inches of pre-deflection of the rail the guardrail performance was acceptable when the rail detached normally from the post. On the other hand, when the rail-post

100 connection was prevented from releasing, vehicle instability resulted due to severe wheel snags on the undetached post, increased potential for pocketing, and increased potential for override as the rail was pulled down by the post. The second series of analyses provided interesting information about the dangers of a rigidly constrained rail-post connection; however, the connection should not be expected to remain fixed in all cases. If so, then essentially all rail damage would need to be repaired, because even an undamaged system will not perform successfully with a fixed post-rail connection. Table 26. Guardrail damage modes. Combination of Rail Deflection and Splice Damage Because the potential for rail rupture increases when excessive rail deflections occur just upstream of a post (i.e., pocketing) at splice connections, the combination of pocketing and splice damage should warrant higher consideration for repair. Although this combination mode was not addressed in NCHRP Report 656, evaluation was considered to be of low priority for this study. Combination of Post and Rail Deflection and Anchor Damage The anchor is a critical component of a guardrail system, and a damaged section of guardrail may be further compromised if the anchor system is compromised in any way. Quantification of this combination damage mode was considered to be of low to medium priority Damage Region Damage Mode 1. Post/Rail Deflection 2. Missing Posts 3. Split Posts 4. Rotted/Weakened Posts 5. Twisted Posts 6. Soil Erroded Away From Posts 7. Rail Deflection 8. Vertical Tear in W-Beam 9. Horizontal Tear in W-Beam 10. Hole in Rail 11. Rail Flattening 12. Rail Crushing (vertical) 13. Twisted Blockout 14. Split Blockout 14. Rotted Blockout 15. Missing Blockout 16. Bent Blockout (steel) 17. Splice Damage 18. Post-Rail Separation 19. End Anchor Damage 20. Missing Backup Plates Post Damage Rail Damage Blockout Damage Connection Damage

101 for the study. The individual assessment criteria for post/rail deflection and for anchor damage should be sufficient for establishing repair priority for this combination mode. Combination of Post/Rail Deflection and Missing Post(s) A damaged guardrail may sometimes exhibit broken or missing posts. Quantification of this combination damage mode was considered to be of low priority for this study. The individual assessment criteria for post/rail deflection and for damage or missing posts should be sufficient for establishing repair priority for this combination mode. Particularly, since Report 656 recommended that even one missing post in an otherwise undamaged guardrail should be given a high priority for repair. Other Considerations Traffic Exposure When assessing the need to repair a damaged section of guardrail, it may be important to consider its exposure to future impacts. For example, if the damaged section is located on the outside of a curve and there is a history of frequent impacts at that location, then the damaged section may warrant a higher priority for repair. On the other hand, if the damaged section were located along a tangent section of a low-volume, low-speed roadway which had little or no history of impacts at that specific location, then the priority for repair may be deferred to other guardrails with equal or less damage but higher exposure. From the Survey of States in Report 656 it was reported that “… the survey respondents could provide almost no documented cases of vehicles impacting previously damaged barriers.”[Gabler10] Evaluating Impact Upstream or Downstream of Damaged Section: Another important factor to consider when evaluating guardrail damage is its effect on performance when the rail is struck at locations upstream or downstream of the damaged section. For example, repair of a ruptured rail element should have increased priority, since the damage will not only affect subsequent impacts at the rupture location, but will also effect impacts over a significant length of the guardrail both up- and down-stream of the rupture (i.e., a ruptured rail is essentially an unanchored system). The same is true for damage to the end-terminals which are necessary for maintaining proper tension in the rail and limiting guardrail deflections, regardless of where an impact occurs along the length of the guardrail. The two above scenarios are relatively obvious signs for guardrail repair; however, it is not known how lower levels of damage to a local section of guardrail will affect performance of the system when struck at points up- and down-stream of the damaged area. For example, impact down-stream of a damaged section, may result in higher rail deflection due to reduced tension in the rail element as the previously damaged section “straightens out” during the impact. As discussed earlier, excessive deflections of an unsupported section of rail results in an increased potential for pocketing. Thus, when assessing damage to guardrails, one must include the total damage affected region of the guardrail. Hazard Being Shielded The level of crash severity for the hazard being shielded should also be considered when deciding repair priority. For example, if a guardrail is shielding a hazard is likely to result in

102 severe injuries if struck, then that damage may warrant higher priority for repair compared to other damaged guardrails that are shielding hazards of lower risk severities. Recommendations The ultimate goal of this study is to develop damage assessment and repair criteria that will lead to a comprehensive and highly valued field guide for assessing guardrail damage and establishing priority for repair. So it was necessary to carefully select which of the damage modes were to be included in the study to make the most effective use of the available funding. The G4(2W) guardrail was rated the highest for inclusion in this study. The wood-post w- beam guardrail is the second most commonly installed guardrail system used in the U.S. and was also rated the highest by the survey respondents for inclusion in this study. As indicated in Table 17, the G4(2W) is composed of the same components and same basic design as the G4(1S), with the exception of the guardrail posts and is thus expected to be susceptible to similar damage modes. In many of these cases the performance degradation due to a particular damage mode will also be similar between these two systems; therefore evaluation criteria for several damage modes for the G4(1S) should be directly applicable to the G4(2W) guardrail. Further, some of the additional damage modes of the G4(1S) that warrant more investigation could be studied as part of the assessment for the G4(2W) and then “extrapolated” to the G4(1S), or vice versa. The assessment criteria for the G4(2W) should apply to all wood post w-beam guardrails that are of the same basic design and that use guardrail posts of similar dimensions to those of the G4(2W). For example, the G4(1W), G4(2W), and the ODOT Type 5 guardrail all use slightly different shaped wooden posts, but otherwise share the same system components. Although the posts are different (e.g., square, rectangular, and round, respectively), they are similar in that they are all wood and are of similar dimensions (i.e., 8-10 inches in width/diameter and 6 ft in length); thus it could be inferred that each of these systems incur similar performance degradation due to each type and combination of damage mode(s). The G2 weak-post guardrail was considered to be the next highest rated system for evaluation in this study. This system was also rated second highest for inclusion in the study by the survey respondents. The thrie-beam guardrail was considered to be the third highest rated system for inclusion in the study. The thrie-beam is not as popular as the strong-post w-beam, but most states have limited installations of these systems. Although the G2 weak-post guardrail received the second highest rating from the survey respondents for inclusion in the study, the rating for the thrie-beam guardrail was very similar (see Appendix B), thus a case could be made for selecting either system. Consideration should also be given to the fact that the thrie-beam guardrail will require fewer damage mode evaluations for developing a comprehensive set of assessment criteria compared to the weak-post system. The guardrail systems and damage modes were prioritized for evaluation in this study, as shown in Table 27, based on considerations of (1) Has the damage mode been quantitatively assessed before? (2) Applicability of damage assessment criteria to multiple systems. (3) How widely used is the guardrail system? (4) Can a sufficient number damage modes for a given system be evaluated (e.g., considering budget constraints) to ensure comprehensive damage assessment and repair guidance?

103 Table 27. Prioritization of damage modes. Guardrail Type Priority Order Damage Mode Evaluation Method Comments A1 Validate G4(2W) Model FEA Simulate tests 471470-26 and/or RF476640-1-5 to validate the model for use in subsequent analyses A2 Weak Anchor FEA Evaluate the effects of anchor stiffness and strength on guardrail performance. Results for this system should also correspond to the G4(1S) . A3 Rotted/Weakened Posts Pendulum / FEA Pendulum test to quantify loss of post strength and validate FEA model. FEA to evaluate system performance A4 Combination Mode: Post/Rail Deflection and rail-post separation Pendulum / FEA / Full-Scale Test? Combination Modes: Pendulum tests to determine amount of rail deflection at which rail detachment should occur. FEA to evaluate barrier performance for a range of barrier deflections w.r.t. varying connection strength. Full-scale test to validate model. A5 Splice Damage Pendulum Evaluate damage to downstream splice bolt holes (damage to front and back layers). Results should also be applicable to the G4(1S) and vice versa. A6 Soil Eroded Away from Posts Pendulum / FEA Pendulum test to quantify loss of post strength and FEA to evaluate system performance. A7 Split Blockouts Pendulum Pendulum tests to quantify the potential for splice rupture. A8 Post/Rail Deflection and Flattened Rail FEA Combination Modes: Rail flattening was shown to increase probabability for system vaulting in R656. A deflected system in combination with rail flattening may further increase potential for vaulting. B1 Reduced Embedment of Anchor Foundation Tubes Static Testing / Pendulum / FEA B2 Missing Grounline Strut Static Testing / Pendulum B3 Slack Cable Static Testing / Pendulum / FEA B4 Rotted / Weakened Anchor Posts Pendulum C1 Validate G4(1S) Model FEA Simulate tests C08C3-027 using appropriate anchor strength to verify research team's hypothesis that the weak anchor was the cause of failed test. C2 Weak Anchor FEA Use FEA to evaluate effects of anchor strength on guardrail performance. D1 Validate G2 Model FEA Simulate full-scale crash test to validate the model for use in subsequent analyses D2 Post/Rail Deflection FEA + Full-Scale Test(?) FEA to simulate low-speed impacts followed by high-speed impact (analogous to Report 656 methodology). Full-scale test to validate model? D3 Anchor Damage FEA The weak-post system will have different sensitivity to anchor strength than strong- post systems. D4 Horizontal Tear Pendulum / FEA This damage mode may be more critical for weak-post systems than strong-post systems due to higher tensile loads in rail D5 Hole in rail Pendulum / FEA This damage mode may be more critical for weak-post systems than strong-post systems due to higher tensile loads in rail D6 Soil Eroded Away from Posts Pendulum / FEA The weak-post system will have a diferrent sensitivity to this damage mode compared to strong post systems. D7 Splice Damage FEA Splice damage will be less critical for the modified G2 (e.g., splice at mid-span) then the original G2 D8 Rail-Post Separation FEA FEA could be used to evaluate this mode; however, a detached rail-post connection would also indicate a missing back-up plate which is not acceptable. D9 Missing Back-up Plate Pendulum / FEA FEA or Pendulum testing could be used to evaluate this damage mode; however, test results have already proven that a missing back-up plate is not acceptable. G4(2W) Measure force-deflection response (i.e., stiffness and strength) of anchor for each damage mode and level of damage. Combine with A2, C2, D2 and E2 to establish repair criteria. G4(1S) Generic Anchor G2

104 Table 27. Prioritization of damage modes. (continued) Guardrail Type Priority Order Damage Mode Evaluation Method Comments E1 Validate G9 Model FEA Simulate full-scale crash test to validate the model for use in subsequent analyses E2 Anchor Damage FEA Evaluate the effects of anchor stiffness and strength on guardrail performance. These results cold be combined with the damage vs stiffness evaluations above. E3 Combination Mode: Post/Rail Deflection and rail-post separation Pendulum / FEA / Full-Scale Test? Combination Modes: Pendulum tests to determine amount of rail deflection at which rail detachment should occur. FEA to evaluate barrier performance for a range of barrier deflections w.r.t. varying connection strength. Full-scale test to validate model. E4 Combination Mode: Post/Rail Deflection and rail flattening FEA Parametric study to determine critical degree of rail flattening for varying degrees of rail deflection. E5 Soil Eroded Away from Posts Pendulum / FEA Pendulum test to quantify loss of post strength and FEA to evaluate system performance. E6 Twisted Posts FEA FEA to evaluate system performance due to various degrees of twisted posts. E7 Rotted/Weakened Posts Pendulum / FEA Pendulum test to quantify loss of post strength and validate FEA model (note: May be able to extrapolate from G4(2W) study). FEA to evaluate system performance. F1 Validate MGS Model FEA Simulate full-scale crash test to validate the model for use in subsequent analyses F2 Combination Mode: Post/Rail Deflection and rail-post separation Pendulum / FEA / Full-Scale Test? Combination Modes: Pendulum tests to determine amount of rail deflection at which rail detachment should occur. FEA to evaluate barrier performance for a range of barrier deflections w.r.t. varying connection strength. Full-scale test to validate model. F3 Anchor Damage FEA Evaluate the effects of anchor stiffness and strength on guardrail performance. Combine with results of anchor study to develop repair criteria. F4 Soil Eroded Away from Posts FEA Results from the pendulum tests in Tasks 4A-6 and 4E-5 will be used to calibrate the soil model for use in the finite element analyses. F5 Combination Mode: Post/Rail Deflection and rail flattening FEA Parametric study to determine critical degree of rail flattening for varying degrees of rail deflection. F6 Twisted Posts FEA FEA to evaluate system performance due to various degrees of twisted posts. F10 Rotted/Weakened Posts FEA Pendulum test to quantify loss of post strength and validate FEA model (note: May be able to extrapolate from G4(2W) study). FEA to evaluate system performance. MGS G9