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.

303 CHAPTER 12 – EFFECTS OF SOIL EROSION AT GUARDRAIL POSTS FOR THE G4(2W) The effects of various levels of soil loss around a guardrail post were evaluated for the G4(2W) guardrail in this study using a combination of pendulum testing and Finite Element Analysis (FEA). The post-soil assembly is a fundamental aspect of a guardrail and its response during a crash event is important to the overall performance of the system. Soil erosion effectively reduces the stiffness of the post-soil system and may have similar effects on guardrail performance as that of weakened posts described in Chapter 8. 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 does not significantly degrade system performance.[Polivka00b] In such a case, the lateral stiffness of the guardrail is reduced uniformly across the entire system and results in increased rail deflection with reduced potential for pocketing (e.g., analogous to a weak-post guardrail system). A more critical situation may be when the soil is eroded away from one or two isolated posts, as shown in Figures 244 and 245. In this case, the increased deflection of the guardrail when struck at the lower stiffness section may result in pocketing as the vehicle approaches the stiffer downstream posts. Thus, a field- assessment procedure for assessing degradation of guardrail performance as a function of soil erosion around the guardrail posts was warranted to ensure proper performance of the guardrail system. Figure 244. Example of soil erosion at a guardrail post.

304 Figure 245. Example of severe soil erosion at multiple posts. The objectives of this study were to quantify the effects of soil erosion around guardrail posts on the crash performance of the G4(2W) guardrail system, and to develop recommendations for conducting field assessments for this damage mode for determining repair priority. Research Approach The basic research approach involved: (1) pendulum impact tests to measure the force- deflection response of the post-soil system for various levels of “manufactured” soil erosion; (2) developing finite element models for the various soil erosion cases and calibrating/validating the models using the test data; and (3) using FEA to quantify the effects of the various degrees of soil erosion on the crash performance of the G4(2W) guardrail system under Report 350 Test 3- 11 impact conditions. Physical Testing The primary purpose of the physical test program was to measure the reduction in soil resistance as a function of soil erosion around the post for use in calibrating/validating the soil erosion FE models. The pendulum tests involved a W6x16 structural steel post embedded 36 inches in the soil, which was consistent with the embedment depth of the posts for full-scale crash test 471470-26 on the G4(2W) guardrail system.[Mak99] The soil for the tests conformed to Grading B of AASHTO M147-95 and was compacted in 6-inch lifts using a pneumatic tamper. The density, moisture content and degree of compaction of the soil was measured in front of, and behind the post after each compaction process using a Troxler-Model 3440 Surface Moisture-Density Gauge. There were a total of twelve soil readings – six on the front-side of the post and six on the back-side of the post – which were averaged to determine the effective soil

305 properties, as shown in Table 67. The target soil conditions included a dry density of 138 pcf, a moisture content of 3.4 percent, and a soil compaction of 92 percent. The erosion condition was manufactured by removing a layer of soil behind and on the side of the post, as illustrated schematically in Figure 246. Four erosion depths were investigated: 3 inches, 6 inches, 9 inches, and 12 inches. The soil behind the post was flat and level and conformed to the desired erosion depth. The soil in front of the post tapered gradually from grade level down to the specified erosion depth, starting at approximately 10 inches in front of the post and ending at the back-face of the post. There were a total of 7 tests conducted. The complete test matrix is shown in Table 68. Figure 246. Test set-up for pendulum tests. 36 in X = 22 in 3 inches 6 inches 9 inches 12 inches Soil Conditions - dry density = 138 pcf - moisture = 3.4% - compaction = 92% W6x16

306 Table 67. Dry density of soil for each test case measured at 6-inch lifts. Table 68. Pendulum test matrix for erosion study showing post embedment, soil properties and impact conditions. Lift 1 Lift 2 Lift 3 Lift 4 Lift 5 Lift 6 Average Lift 1 Lift 2 Lift 3 Lift 4 Lift 5 Lift 6 Average (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) (pcf) 14003B 0" (Baseline) 143.7 140.1 142.9 143.9 143.8 145.2 143.3 143.3 140.6 146.4 143.7 140.8 141.7 142.8 143.0 14003A 3" Erosion 143.8 139.5 147.1 139.3 148.3 146.2 144.0 147.9 141.1 142.9 142.4 148.0 140.2 143.8 143.9 14003C 6" Erosion 144.0 147.5 145.0 141.5 140.9 147.3 144.4 145.1 140.8 145.8 144.0 142.5 144.9 143.9 144.1 14003D 9" Erosion 144.0 139.5 141.7 144.9 141.7 147.2 143.2 147.2 146.9 139.2 144.5 140.4 145.7 144.0 143.6 14003E 12" Erosion 140.0 140.6 143.4 146.5 140.8 145.2 142.8 144.6 140.7 142.6 145.7 141.0 145.2 143.3 143.0 14003F 0" (Baseline) 149.4 142.6 142.1 145.4 147.9 147.8 145.9 145.1 144.0 141.2 144.5 148.8 147.8 145.2 145.6 14003G Supplemental 146.3 142.7 145.2 144.6 145.3 148.9 145.5 147.0 144.8 143.1 142.8 145.2 145.4 144.7 145.1 Effective Density Dry Density Soil Properties Front-Side of PostTest No. Damage Mode Rear-Side of Post Weight Speed (in) (in) (pcf) (%) (%) (lbs) (mph) (in) 14003B 5/15/2014 Baseline W6x16 36 36 143.0 3.6 92.6 2372 20 22.0 14003A 5/13/2014 3" Erosion W6x16 34 33 143.9 4.1 93.1 2372 20 22.0 14003C 5/19/2014 6" Erosion W6x16 34 30 144.1 4.2 93.3 2372 20 22.0 14003D 5/20/2014 9" Erosion W6x16 32.5 27 143.6 3.8 92.9 2372 20 22.0 14003E 5/22/2014 12" Erosion W6x16 35 24 143.0 3.6 92.6 2372 20 22.0 14003F 5/23/2014 Baseline W6x16 36 36 145.6 3.0 94.2 2372 20 22.0 14003G 5/27/2014 Supplemental W6x16 40 40 145.1 3.0 94.0 2372 20 22.0 Post Data Embedment Depth Test No. Test Date Damage Mode Impact ConditionsEffective Soil Properties Type Front of Post Back of Post Dry Density Moisture Compaction Pendulum Impact Point

307 Equipment and Instrumentation Pendulum Device The striker for the tests was a 2,372-lb concrete pendulum with a semi-rigid nose, which was the same striker used to evaluate the effects of post deterioration in Chapter 8 (see Chapter 8 for more details). Figure 247. 2,372-lb pendulum device with semi-rigid nose. Accelerometers The pendulum was instrumented with three accelerometers mounted onto the backside of the pendulum mass. Accelerometers 1 and 3 recorded data in the x-direction (forward direction) and Accelerometer 2 recorded data in the z-direction (vertical direction). Figure 248 provides a schematic showing the locations of the accelerometers. Figure 248. Schematic of the accelerometer instrumentation for the pendulum tests.

308 Photography Cameras The tests were also recorded using four high-speed cameras with an operating speed of 500 frames per second and a digital video camera (∼30 fps). Figure 249 provides the specifications and the general placement of the high-speed cameras for the tests. The accelerometers and the high-speed video were triggered using pressure tape switches when the pendulum contacted the post. The test setup and results were also documented with photographs taken before and after each test. Figure 249. High-speed camera specifications and placement. Impact Conditions The 2,372-lb pendulum struck the post at 22 inches above grade at an impact speed of 20 mph, which resulted in 380.3 kip-in of kinetic energy for the striker. The posts were oriented in the strong direction such that the impact load was applied perpendicular to the flange of the post. The impact point height corresponded to the post-bolt location for the G4(2W) guardrail.

309 Results The x-channel accelerometer data was processed to obtain the accelerations of the pendulum during impact with the posts. The data was filtered using an SAE Class 60 filter with cutoff frequency of 100 Hz. The impact force-time history response from each case was approximated by multiplying the acceleration-time history curves by the total mass of the pendulum. The displacement-time history of the pendulum was obtained by double integrating the acceleration-time history curve. These results were used to generate the force-deflection response of the post-soil system during impact. The force-deflection curve was then integrated to obtain the energy vs. deflection curves. Table 69 shows a summary of the test results. Plots of force vs. deflection and energy vs. deflection are shown in Figures 251 through 254. Test summary sheets for each erosion case are provided in Appendix M. The data from the primary accelerometer (i.e., Channel 1) resulted in somewhat inconsistent results, as evidenced in the force-deflection plot in Figure 251 and the energy vs deflection plot in Figure 253. The data from the secondary accelerometer were much more consistent, as shown in Figure 252 and Figure 254. The two baseline tests from the secondary accelerometer resulted in essentially identical response; further, the results from the erosion tests showed that the resistance reduced monotonically as the erosion depth increased. Thus, the results from the secondary accelerometer (Channel 3) were considered to be more accurate and were used for the calibration/validation study for the finite element model in the next section. The standard test procedure for measuring soil stiffness, as specified in MASH, entails a W6x16 post embedded 40 inches in the soil and impacted at 20 mph at 24.9 inches above grade. Recall that this test was previously performed in Test 13010F and was presented in Chapter 8. A supplemental test involving a W6x16 post embedded 40 inches with no manufactured erosion (i.e., Test 14003G) was included in the current test series. The impact point for the test was 22 inches (consistent with the mounting height of the G4(2W)), which was approximately 3 inches lower than that for the standard MASH test, and thereby provided a more direct comparison to the tests performed in this study. The peak force in all cases was due to the initial inertial/impulse force between the pendulum head and the steel post. The magnitude of this peak was essentially the same for all cases regardless of post-soil stiffness, as shown in Figure 255. The one exception was Test 14003A corresponding to the 3-inch erosion case, which was the first test performed in this test series. During the “return swing” of the pendulum after striking the post in Test 14003A, the impact head snagged on the top of the post and was detached from the pendulum. The repair to the pendulum head included a steel “slide plate” mounted underneath the pendulum covering the gap between the head and the body of the pendulum, as shown in Figure 250. The plate was welded to the impact head and to the pendulum body. This modification resulted in a more rigid connection between the head and the pendulum mass which, consequently, increased the initial impulse/inertial spike for all the subsequent tests.

310 Figure 250. Repair to the pendulum head included adding a steel “slide Plate” underneath the head. Figure 256 shows a plot of the total energy absorbed by the post-soil system as a function of erosion depth. The data in the plot corresponds to the total energy absorbed at the time when the pendulum overrode the post. As indicated in the plot, the data from the primary accelerometer generally resulted in slightly lower magnitudes than those from the secondary accelerometer. Since it has not been confirmed which of the accelerometers is correct, the data from both accelerometers were averaged and plotted in Figure 257. The data for the 40-inch embedment case was also included in the plot (e.g., the erosion depth was included as -4 inches). The results indicate that the energy capacity of the post-soil system reduces linearly with respect to erosion depth. Using an embedment depth of 36 inches as the baseline, the reduction in energy capacity as a function of erosion depth can be approximated using the following relationship: 𝐸𝑛𝑒𝑟𝑔𝑦 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 = 178.33 − 8.04 ∗ (𝑒𝑟𝑜𝑠𝑖𝑜𝑛 𝑑𝑒𝑝𝑡ℎ)

311 Table 69. Summary of results for Test Series 14003. Peak Peak Weight Speed Force Energy Force Energy (in) (in) (pcf) (%) (%) (lbs) (mph) (in) (kips) (kip-in) (kips) (kip-in) (kip-in) 14003B 5/15/2014 Baseline W6x16 36 36 143.0 3.6 92.6 2372 20 22.0 18.7 149.5 19.2 186.4 167.9 14003A 5/13/2014 3" Erosion W6x16 34 33 143.9 4.1 93.1 2372 20 22.0 13.2 158.0 13.6 181.2 169.6 14003C 5/19/2014 6" Erosion W6x16 34 30 144.1 4.2 93.3 2372 20 22.0 17.2 121.8 17.3 142.1 132.0 14003D 5/20/2014 9" Erosion W6x16 32.5 27 143.6 3.8 92.9 2372 20 22.0 18.7 85.9 19.4 116.0 101.0 14003E 5/22/2014 12" Erosion W6x16 35 24 143.0 3.6 92.6 2372 20 22.0 18.1 85.2 18.3 73.7 79.4 14003F 5/23/2014 Baseline W6x16 36 36 145.6 3.0 94.2 2372 20 22.0 21.3 183.1 21.6 183.7 183.4 14003G 5/27/2014 Supplemental W6x16 40 40 145.1 3.0 94.0 2372 20 22.0 17.7 197.4 18.4 214.8 206.1 ResultsPost Data Embedment Depth Primary Accel. Secondary Accel. Average Energy Test No. Test Date Damage Mode Impact ConditionsEffective Soil Properties Type Front of Post Back of Post Dry Density Moisture Compaction Pendulum Impact Point

312 Figure 251. Force vs. deflection curves for Test Series 14003 from primary accelerometer. Figure 252. Force vs. deflection curves for Test Series 14003 from secondary accelerometer. 0 2 4 6 8 10 12 14 16 18 20 0 5 10 15 20 25 30 35 40 Fo rc e ( ki p ) Displacement (in) MASH Minimum 14003B - Baseline 36" embed. 14003F - Baseline 36" embed. 14003A - 3" erosion 14003C - 6" erosion 14003D - 9" erosion 14003E - 12" erosion 14003G - 40" Embedment Channel 1 Accelerometer 0 2 4 6 8 10 12 14 16 18 20 0 5 10 15 20 25 30 35 40 Fo rc e ( ki p ) Displacement (in) MASH Minimum 14003B - Baseline 36" embed. 14003F - Baseline 36" embed. 14003A - 3" erosion 14003C - 6" erosion 14003D - 9" erosion 14003E - 12" erosion 14003G - 40" Embedment Channel 3 Accelerometer

313 Figure 253. Energy vs. deflection curves for Test Series 14003 from primary accelerometer. Figure 254. Energy vs. deflection curves for Test Series 14003 from secondary accelerometer. 0 50 100 150 200 250 300 0 5 10 15 20 25 30 35 40 En e rg y (k ip -i n ) Displacement (in) 14003B - Baseline 36" embed. 14003F - Baseline 36" embed. 14003A - 3" erosion 14003C - 6" erosion 14003D - 9" erosion 14003E - 12" erosion 14003G - 40" embedment Channel 1 Accelerometer 0 50 100 150 200 250 300 0 5 10 15 20 25 30 35 40 En e rg y (k ip -i n ) Displacement (in) 14003B - Baseline 36" embed. 14003F - Baseline 36" embed. 14003A - 3" erosion 14003C - 6" erosion 14003D - 9" erosion 14003E - 12" erosion 14003G - 40" embedment Channel 3 Accelerometer

314 Figure 255. Peak “impulse” force in Test Series 14003. Figure 256. Total energy vs. erosion for Test Series 14003 – energy values correspond to point when pendulum overrides the post. 0 5 10 15 20 25 -6 -3 0 3 6 9 12 P e ak F o rc e ( ki p ) Erosion Depth (in) Primary Accel. Secondary Accel. 0 50 100 150 200 250 -6 -3 0 3 6 9 12 To ta l E n e rg y (k ip -i n ) Erosion Depth (in) Primary Accel. Secondary Accel.

315 Figure 257. Total energy vs. erosion for Test Series 14003 – average of primary and secondary accelerometers – including linear curve fit to the data. Finite Element Analysis Finite element models with various levels of erosion were developed and included in the G4(2W) guardrail model. FEA was then used to simulate Report 350 Test 3-11 impact conditions to quantify the effects of soil erosion on the performance of the guardrail system. The development and validation of the finite element model of the G4(2W) guardrail was presented in Chapter 7. The various components of the G4(2W) guardrail were modeled according to their baseline conditions, including the wood post model. The soil models representing the various levels of erosion were calibrated/validated through comparison with pendulum test data from test Series 14003 presented earlier. The erosion levels were modeled by simply translating the soil- spring model to the appropriate erosion depth. The following sections discuss the methodology and results for calibrating/validating soil erosion models and evaluation of the effects of soil erosion on performance of the G4(2W) guardrail. Calibration/Validation of FE Soil-Erosion Model The impact conditions for Test Series 14003 were simulated using the finite element model shown in Figure 258. From the results of the pendulum tests, the initial spike at the beginning of the impact, as seen in Figures 251 and 252, was due in part to the impulse load between the pendulum and the post, but also to an increase in soil stiffness from “friction locking” of soil particles under high soil compaction. The soil-spring model developed in Chapter 7 was modified slightly to account for this “friction locking”. In particular, the force- deflection characterization for the soil springs was increased for the first 0.5 inches of soil deflection, and then then reverted back to their original stiffness. The following sections present a comparison of the FE model results compared to the physical tests for the baseline condition (i.e., 36-inch embedment depth) and various erosion depths. y = -8.037x + 178.33 R² = 0.9665 0 50 100 150 200 250 -6 -3 0 3 6 9 12 To ta l E n e rg y (k ip -i n ) Erosion Depth (in) Avg. Energy Linear Fit

316 Figure 258. Finite element model used for validating/calibrating soil-spring stiffness corresponding to various levels of soil erosion. Baseline Case – Damage Level 0 (DL0) The results of the model for the baseline condition are compared to full-scale test results in Figures 259 through 261. The force vs. deflection curves and the energy vs. deflection curves are shown in Figure 259 and Figure 260, respectively. As indicated from these plots, the FE model does not accurately capture the magnitude of the initial force response; otherwise, the overall response of the model matches reasonably well to the physical test. The model shows slightly higher response than the tests at post deflections beyond 20 inches. From the high-speed videos of the physical tests, it is evident that the soil is loosened significantly as the post rotates through the soil at increasingly higher deflections; at that point the post is essentially moving through a loose soil mass. The characterization of the soil in FE model does not include this softening behavior; however, the friction between the post and soil in the model effectively results in a similar response as the post slides up and out of the soil plates. The total energy absorbed at 12 inches displacement in the analysis was 95.4 kip-in compared to 109 kip-in in Test 14003F; the total energy absorbed at 20 inches displacement in the analysis was 149 kip-in compared to 152 kip-in in Test 14003F; and the total energy absorbed at 40 inches displacement in the analysis was 247 kip-in compared to 211 kip-in in Test 14003F. Figure 261 shows sequential views comparing analysis results to dynamic impact Test 14003F.

317 Figure 259. Force vs. deflection for DL0 soil model compared to Tests 14003B and 14003F. Figure 260. Energy vs. deflection for DL0 soil model compared to Tests 14003B and 14003F. 0 2 4 6 8 10 12 14 16 18 20 0 10 20 30 40 Fo rc e ( ki p ) Displacement (in) Test 14003B Test 14003F FEA14003B - Baseline 0 50 100 150 200 250 300 0 10 20 30 40 En e rg y (k ip -i n ) Displacement (in) Test 14003B Test 14003F FEA14003B - Baseline

318 Figure 261. Sequential views of Test 14003F and FE analysis on baseline post-soil case, DL0. 3” Erosion Case The impact conditions for Test 14003A were simulated using the finite element model shown in Figure 258 with the following modifications: In the physical test, the soil depth was at grade level starting at approximately 10 inches in front of the post and tapered gradually to 2 inch below grade at the front-face of the post. The soil continued to taper to three inches below grade at the back-face of the post. The soil then remained at 3 inches below grade level for several feet behind the post. As mentioned previously, the soil-spring model was shifted downward in the vertical direction to account for the erosion. For this case, the soil model was set three inches below grade on the backside of the post, while the soil level on the front side of the post was kept at its original position at grade level. This methodology was based on the fact that the post pushes against the soil much farther below grade on the front-side of the post, where the lateral stiffness of the soil is affected by the overbearing pressure of a greater volume of soil (e.g., considering an angle of internal friction >40 deg.). The volume of soil removed from this region when creating the manufactured erosion was considered negligible; particularly in this case where only 2 inches of soil was removed at the front edge of the post. On the backside the post, on the other hand, the post interacts with the soil nearer to the surface, where a much greater volume of soil was removed. Figure 262 and Figure 263 show the force vs. deflection and energy vs. deflection results, respectively, for the FE model compared to pendulum Test 14003A. The overall response of the model matches reasonably well to the physical test, including the impulsive force spike at the beginning of the impact event. Recall that this test was performed using the original pendulum head (before the head was repaired), which resulted in a lower impulse/inertial spike compared to subsequent tests with the repaired head. 0.08 s / 22.5 in0.04 s / 12.4 in 0.12 s / 30.1 in 0.16 s / 37.9 in 0.08 s / 23.5 in0.04 s / 12.7 in 0.12 s / 33.1 in 0.16 s / 41.8 in

319 Figure 262. Force vs. deflection for soil model with 3” erosion compared to Test 14003A. Figure 263. Energy vs. deflection for soil model with 3” erosion compared to Test 14003A. 0 2 4 6 8 10 12 14 16 18 20 0 10 20 30 40 Fo rc e ( ki p ) Displacement (in) MASH Minimum Test 14003A FEA14003A - 3-inch 0 50 100 150 200 250 300 0 10 20 30 40 En e rg y (k ip -i n ) Displacement (in) Test 14003A FEA14003A - 3-inch

320 The total energy absorbed at 12 inches displacement in the analysis was 99 kip-in, which was the same as that measured in Test 14003A; the total energy absorbed at 20 inches displacement in the analysis was 146 kip-in compared to 147 kip-in in Test 14003A; and the total energy absorbed at 40 inches displacement in the analysis was 199 kip-in compared to 203 kip-in in Test 14003A. Figure 264 shows sequential views comparing analysis results to dynamic impact test 14003A. Figure 264. Sequential views of Test 14003A and FE analysis for 3-inch erosion case. 6” Erosion Case – Damage Level 1 (DL1) The impact conditions for Test 14003C were simulated using the finite element model shown in Figure 258 with the following modifications. In the physical test, the soil depth was at grade level starting at approximately 10 inches in front of the post and tapered gradually to 2 inches below grade at the front-face of the post. The soil continued to taper to 6 inches below grade at the back-face of the post. The soil then remained at 6 inches below grade level for several feet behind the post. For this case, the soil model was set six inches below grade on the backside of the post, while the soil level on the front side of the post was kept at its original position at grade level. Figure 265 and Figure 266 show the force vs. deflection and energy vs. deflection results, respectively, for the FE model compared to pendulum Test 14003C. The overall response of the model matches reasonably well to the physical test, except for the impulse/inertial spike at the beginning of the impact. Note that this test was performed using the repaired pendulum head which resulted in higher impulse/inertial force magnitude compared to the original head. The total energy absorbed at 12 inches displacement in the analysis was 75.2 kip-in compared to 87.6 kip-in in Test 14003C; the total energy absorbed at 20 inches displacement in the analysis was 109 kip-in compared to 125 kip-in in Test 14003C; and the total energy absorbed at 40 inches displacement in the analysis was 124 kip-in compared to 140 kip-in in Test 0.08 s / 23.3in0.04 s / 12.7 in 0.12 s / 32.9 in 0.16 s / 42.1 in 0.08 s / 23.8 in0.04 s / 12.9 in 0.12 s / 33.6 in 0.16 s / 42.7 in

321 14003C. Figure 267 shows sequential views comparing analysis results to dynamic impact Test14003C. Figure 265. Force vs. deflection for soil model with 6” erosion compared to Test 14003C. Figure 266. Energy vs. deflection for soil model with 6” erosion compared to Test 14003C. 0 2 4 6 8 10 12 14 16 18 20 0 10 20 30 40 Fo rc e ( ki p ) Displacement (in) MASH Minimum Test 14003C FEA14003C - 6 inch 0 50 100 150 200 250 300 0 10 20 30 40 En e rg y (k ip -i n ) Displacement (in) Test 14003C FEA14003C - 6 inch

322 Figure 267. Sequential views of Test 14003C and FE analysis for 6-inch erosion case. 9” Erosion Case – Damage Level 2 (DL2) The impact conditions for Test 14003D were simulated using the finite element model shown in Figure 258 with the following modifications. In the physical test, the soil depth was at grade level starting at approximately 10 inches in front of the post and tapered gradually to 3.5 inches below grade at the front-face of the post. The soil continued to taper gradually to 9 inches below grade at the back-face of the post. The soil then remained at 9 inches below grade level for several feet behind the post. For this case, the soil model was set 9 inches below grade on the backside of the post, while the soil level on the front side of the post was set at 2 inches below grade. Figure 268 and Figure 269 show the force vs. deflection and energy vs. deflection results, respectively, for the FE model compared to pendulum Test 14003D. The overall response of the model matches reasonably well to the physical test, except for the impulse/inertial spike at the beginning of the impact. Again, this test was performed using the repaired pendulum head which tended to result in higher impulse/inertial force magnitude compared to the original head. The total energy absorbed at 12 inches displacement in the analysis was 55.9 kip-in compared to 69.8 kip-in in Test 14003D; the total energy absorbed at 20 inches displacement in the analysis was 81.5 kip-in compared to 97.7 kip-in in Test 14003D; and the total energy absorbed at 40 inches displacement in the analysis was 93.1 kip-in compared to 122 kip-in in Test 14003D. Figure 270 shows sequential views comparing analysis results to dynamic impact Test 14003D. 0.08 s / 24.3 in0.04 s / 12.9 in 0.12 s / 35.3 in 0.16 s / 46.3 in 0.08 s / 24.5 in0.04 s / 13.0 in 0.12 s / 35.3 in 0.16 s / 46.1 in

323 Figure 268. Force vs. deflection for soil model with 9” erosion compared to Test 14003D. Figure 269. Energy vs. deflection for soil model with 9” erosion compared to Test 14003D. 0 2 4 6 8 10 12 14 16 18 20 0 10 20 30 40 Fo rc e ( ki p ) Displacement (in) MASH Minimum Test 14003D FEA14003D - 9 inch 0 50 100 150 200 250 300 0 10 20 30 40 En e rg y (k ip -i n ) Displacement (in) Test 14003D FEA14003D - 9 inch

324 Figure 270. Sequential views of Test 14003D and FE analysis for 9-inch erosion case. 12” Erosion Case – Damage Level 3 (DL3) The impact conditions for Test 14003E were simulated using the finite element model shown in Figure 258 with the following modifications: In the physical test, the soil depth was at grade level starting at approximately 10 inches in front of the post and tapered gradually to 4 inches below grade at the front-face of the post. The soil continued to taper gradually to 12 inches below grade at the back-face of the post. The soil then remained at 12 inches below grade level for several feet behind the post. For this case, the soil model was set at 12 inches below grade at the backside of the post, while the soil level on the front side of the post was set at 2 inches below grade. Figure 271 and Figure 272 show the force vs. deflection and energy vs. deflection results, respectively, for the FE model compared to pendulum Test 14003E. This test was performed using the repaired pendulum head, and the overall response of the model matched reasonably well to the physical test, except for the impulse/inertial spike at the beginning of the impact. The total energy absorbed at 12 inches displacement in the analysis was 40.4 kip-in compared to 48.8 kip-in in Test 14003E; the total energy absorbed at 20 inches displacement in the analysis was 58.6 kip-in compared to 67.6 kip-in in Test 14003E; and the total energy absorbed at 40 inches displacement in the analysis was 66.1 kip-in compared to 67.1 kip-in in Test 14003E. Figure 273 shows sequential views comparing analysis results to dynamic impact Test 14003E. 0.08 s / 25.5 in0.04 s / 13.2in 0.12 s / 37.1 in 0.16 s / 48.8 in 0.08 s / 25.0 in0.04 s / 13.1 in 0.12 s / 36.7 in 0.16 s / 48.4 in

325 Figure 271. Force vs. deflection for soil model with 12” erosion compared to Test 14003E. Figure 272. Energy vs. deflection for soil model with 12” erosion compared to Test 14003E. 0 2 4 6 8 10 12 14 16 18 20 0 10 20 30 40 Fo rc e ( ki p ) Displacement (in) MASH Minimum Test 14003E FEA14003E - 12 inch 0 50 100 150 200 250 300 0 10 20 30 40 En e rg y (k ip -i n ) Displacement (in) Test 14003E FEA14003E - 12 inch

326 Figure 273. Sequential views of Test 14003E and FE analysis for 12-inch erosion case. Effects of Soil Erosion on Guardrail Performance Finite element analysis was used to evaluate the effects of soil erosion around guardrail posts on the crash performance of the G4(2W) guardrail system. The baseline (undamaged) condition for the guardrail included the DL0 guardrail post model (refer to Chapter 8); 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 loading rate (refer to Chapter 9). The various levels of erosion were included in the model by simply shifting the soil-spring model in the vertical direction to the appropriate position according to erosion depth, as discussed in the preceding section. Three levels of erosion were considered, including erosion depths of 6, 9 and 12 inches. The study included the effects of erosion at a single post and also erosion at two consecutive posts. The analysis matrix is shown in Table 70. Table 70. Simulation matrix for evaluating soil erosion around creating low-severity guardrail deflection damage cases. Refer to Chapter 7 regarding details of the development and installation length of the G4(2W) guardrail model. The impact conditions were set to those of full-scale crash test 471470- 26 and involved the 4,568-lb C2500D pickup model impacting the guardrail at 62.6 mph (100.8 km/hr) at an angle of 24.3 degrees.[Mak99] The critical impact point (i.e., regarding the maximum potential for pocketing and rail rupture) for the 6-inch erosion case was determined 0.08 s / 25.6 in0.04 s / 13.3 in 0.12 s / 37.9 in 0.16 s / 50.1 in 0.08 s / 25.7 in0.04 s / 13.3in 0.12 s / 37.6 in 0.16 s / 49.5 in 6 inches 9 inches 12 inches Single Post x x x 2 Consecutive Posts x x x Varying Erosion Depth Number of Posts

327 using FE analysis to be 45 inches (1.16 m) upstream of the w-beam rail splice connection at Post 14. Due to time and budget constraints, the CIP for the 6-inch erosion case was used for all subsequent cases in this study. The analyses were conducted for 0.6 seconds of the impact event. The analysis model used for the evaluations is shown in Figure 274. Posts 4 through 21 were modeled using MAT143 in LS-DYNA; while the posts located upstream of the impact zone were modeled using MAT13. The material properties corresponded to undamaged wood posts (i.e., no deterioration) for all posts in the model. Refer to Chapter 8 for more details on wood post model development and validation. Figure 274. Analysis setup for evaluating effects of anchor strength on the performance of the G4(2W). Soil Erosion at a Single Post The following is an evaluation of the G4(2W) guardrail with various levels of soil erosion at a single post location. Sequential views of the FE analysis results for each case are provided in Appendix N. Table 71 provides a summary of barrier damage from the analyses related to rail deflections, anchor movement and splice damage. This information is also presented graphically in Figures 275 and 276. The maximum lateral rail deflection was 32 inches for the baseline case and increased to 35.5 inches at 12 inches of erosion (i.e., 9.2 percent higher than the baseline case). Regarding the “location of maximum deflection” in Table 71, a negative number indicates that the maximum deflection occurred upstream of the w-beam splice located at Post 16, while a positive number indicates that maximum deflection occurred downstream of the splice. For the 6-inch erosion case, the maximum lateral deflection occurred at 29 inches upstream of the splice connection at Post 16. For the 9-inch and 12-inch cases the maximum deflection occurred exactly at Post 16. Figure 276 shows the longitudinal displacement of the upstream and downstream ends of the w-beam at the anchor locations. In all erosion cases, the loading on the upstream anchor was of similar magnitude as the baseline case. The longitudinal rail deflections at the downstream anchor increased slightly with each level of erosion. In general, higher lateral deflections in the impact region were associated with higher anchor deflections. Undamaged Posts (MAT143) Undamaged Posts (MAT13) Non-Linear Spring Anchor Stiffness Non-Linear Spring Anchor Stiffness Soil Erosion

328 Table 71. Summary of barrier damage evaluation from analyses of undamaged G4(2W) guardrail with various levels of soil erosion at a single post. Figure 275. Summary of barrier damage evaluation from analyses of undamaged G4(2W) guardrail with various levels of soil erosion at a single post. Baseline 6 inches 9 inches 12 inches Maximum Rail Deflection (in) 32.0 32.5 34.4 35.5 Location of Max Defl. (in) (Relative to Post 16) -14.8 -28.9 0.0 0.0 Rail Deflection at Post 13 (in) 1.8 1.3 1.7 2.5 Rail Deflection at Post 14 (in) 13.1 11.3 13.9 15.1 Rail Deflection at Post 15 (in) 27.8 27.9 28.1 28.9 Rail Deflection at Post 16 (in) 31.1 31.3 34.4 35.5 Rail Deflection at Post 17 (in) 17.5 22.7 25.5 27.1 Rail Deflection at Post 18 (in) 1.2 1.7 4.6 5.8 Rail Deflection at Post 19 (in) 0.0 0.0 0.0 0.0 Upstream Anchor Deflection (in) 1.4 1.5 1.5 1.7 Downstream Anchor Deflection (in) 0.9 1.1 1.1 1.1 Maximum Strain in splice 0.84 1.05 1.04 1.12 Varying Erosion Depth at 1 Post Event -40.0 -30.0 -20.0 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 Baseline 6" Erosion at 1 Post 9" Erosion at 1 Post 12" Erosion at 1 Post

329 Figure 276. Summary of anchor displacement at rail height from analyses of undamaged G4(2W) guardrail with various levels of soil erosion at a single post. 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 72. This data is also presented graphically in Figures 277 through 279. The results indicated that increased levels of erosion at a single post did not significantly affect occupant risk; and, in general, the values were slightly lower for the erosion cases. A summary of the maximum effective plastic strains around the splice-bolt holes in the w-beam for each analysis case is shown in Figure 280. The results indicated that the potential for splice rupture increased as erosion levels increased. For example, the plastic strains at the splice-bolt holes reached magnitudes of 1.05, 1.04, and 1.12 for cases of 6, 9 and 12 inches of erosion, respectively. The values computed from the 9-inch and 12-inch analysis case likely under-predict the actual strains, since the critical impact point was not achieved. 1.38 1.54 1.55 1.66 0.91 1.12 1.08 1.06 0.00 0.50 1.00 1.50 2.00 2.50 3.00 Baseline 6 inches 9 inches 12 inches Varying Erosion Depth at 1 Post A n ch o r D ef le ct io n ( in ) G4(2W) with Various Erosion Depths Upstream Anchor Deflection (in) Downstream Anchor Deflection (in)

330 Table 72. Summary of occupant risk measures from evaluation of undamaged G4(2W) guardrail with various levels of soil erosion at a single post. Figure 277. Summary of occupant impact velocities for undamaged G4(2W) guardrail with various levels of soil erosion at a single post. Baseline 6 inches 9 inches 12 inches Occupant Impact Velocity x-direction 5.2 4.8 4.1 3.6 (m/s) y-direction 5.3 5.2 5.3 5.3 at time (0.1519 sec) (0.1473 sec) (0.1483 sec) (0.1427 sec) 7.0 6.9 6.7 6.5 (0.1471 sec) (0.1424 sec) (0.1432 sec) (0.1428 sec) Ridedown Acceleration 10.3 10.6 10.4 10.1 (g's) (0.1519 - 0.1619 sec) (0.2179 - 0.2279 sec) (0.2734 - 0.2834 sec) (0.2900 - 0.3000 sec) 10.7 8.6 8.2 9.2 (0.2198 - 0.2298 sec) (0.1488 - 0.1588 sec) (0.2164 - 0.22264 sec) (0.2187 - 0.2287 sec) 13.7 13.5 12.3 12.7 (0.2003 - 0.2103 sec) (0.2174 - 0.2274 sec) (0.2739 - 0.2839 sec) (0.2947 - 0.3047 sec) 0.93 0.84 0.78 0.79 (0.1219 - 0.1719 sec) (0.1812 - 0.2312 sec) (0.1951 - 0.12451sec) (0.2657 - 0.3157 sec) Max 50-ms moving avg. acc. 7.6 5.6 5.5 6.0 (g's) (0.1216 - 0.1716 sec) (0.1798 - 0.2298 sec) (0.1589 - 0.2089 sec) (0.1540 - 0.2040 sec) 6.5 6.6 6.4 6.1 (0.1976 - 0.2476 sec) (0.1099 - 0.1599 sec) (0.1955 - 0.2455 sec) (0.2675 - 0.3175 sec) 2.4 2.2 3.2 3.0 (0.3344 - 0.3844 sec) (0.2320 - 0.2820 sec) (0.3164 - 0.3664 sec) (0.3186 - 0.3686 sec) THIV Occupant Risk Factors Erosion at Single Post Varying Erosion Depth x-direction y-direction z-direction (m/s) x-direction y-direction PHD (g's) ASI 0 1 2 3 4 5 6 7 Baseline 6 inches 9 inches 12 inches O IV ( m /s ) Erosion Depth at 1 Post Occupant Impact Velocity x-dir y-dir

331 Figure 278. Summary of maximum occupant ridedown accelerations for undamaged G4(2W) guardrail with various levels of soil erosion at a single post. Figure 279. Summary of maximum 50-ms running average accelerations for undamaged G4(2W) guardrail with various levels of soil erosion at a single post. 0 2 4 6 8 10 12 Baseline 6 inches 9 inches 12 inches O R A ( g) Erosion Depth at 1 Post Occupant Ridedown Acceleration x-dir y-dir 0 1 2 3 4 5 6 7 8 Baseline 6 inches 9 inches 12 inches 5 0 -m s A ve ra ge A cc . ( g) Erosion Depth at 1 Post 50-ms Avg. Acceleration x-dir y-dir

332 Figure 280. Summary of maximum effective plastic strains occurring at the splice-bolt locations at Post 16. Soil Erosion at Two Consecutive Posts The following is an evaluation of the G4(2W) guardrail with various levels of soil erosion at two consecutive post locations. Sequential views of the FE analysis results for each case are provided in Appendix O. Table 73 provides a summary of barrier damage from the analyses related to rail deflections, anchor movement and splice damage. This information is also presented graphically in Figures 281 and 282. The maximum lateral rail deflection was 32 inches for the baseline case and increased to 40.1 inches at 12 inches of erosion (i.e., 25.3 percent higher than baseline case). As shown in Table 73, the maximum lateral deflection generally occurred within 34 inches upstream of the splice connection at Post 16. Thus, the impact conditions were considered to correspond to critical impact conditions for the various cases. Figure 282 shows the longitudinal displacement of the upstream and downstream ends of the w-beam at the anchor locations. In all erosion cases, the loading on the upstream anchor was of similar magnitude as the baseline case; while the longitudinal rail deflections at the downstream anchor increased slightly with each level of erosion. In general, higher lateral deflections in the impact region were associated with higher anchor deflections. 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 Baseline 6 inches 9 inches 12 inches Varying Erosion Depth at 1 Post Splice Strains

333 Table 73. Summary of barrier damage evaluation from analyses of undamaged G4(2W) guardrail with various levels of soil erosion at two consecutive posts. Figure 281. Summary of barrier damage evaluation from analyses of undamaged G4(2W) guardrail with various levels of soil erosion at two consecutive posts. Baseline 6 inches 9 inches 12 inches Maximum Rail Deflection (in) 32.0 35.5 37.4 40.1 Location of Max Defl. (in) (Relative to Post 16) -14.8 -28.1 -34.4 -21.9 Rail Deflection at Post 13 (in) 1.8 4.4 7.1 10.1 Rail Deflection at Post 14 (in) 13.1 21.7 21.7 25.0 Rail Deflection at Post 15 (in) 27.8 33.7 35.2 37.1 Rail Deflection at Post 16 (in) 31.1 34.9 36.4 39.8 Rail Deflection at Post 17 (in) 17.5 25.4 20.2 26.1 Rail Deflection at Post 18 (in) 1.2 4.1 1.7 4.3 Rail Deflection at Post 19 (in) 0.0 0.0 0.0 0.0 Upstream Anchor Deflection (in) 1.4 1.7 1.7 1.8 Downstream Anchor Deflection (in) 0.9 1.1 1.1 1.1 Maximum Strain in splice 0.84 1.07 1.16 1.22 Event Varying Erosion Depth at 2 Posts -40.0 -30.0 -20.0 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 Baseline 6" Erosion at 2 posts 9" Erosion at 2 Posts 12" Erosion at 2 Posts

334 Figure 282. Summary of anchor displacement at rail height from analyses of undamaged G4(2W) guardrail with various levels of soil erosion at two consecutive posts. 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 74. This data is also presented graphically in Figures 283 through 285. The results indicate that increased levels of erosion at two consecutive posts do not significantly affect occupant risk; and, in general, the values were again slightly lower for the erosion cases. A summary of the maximum effective plastic strains around the splice-bolt holes in the w-beam for each analysis case is shown in Figure 286. The results indicated that the potential for splice rupture increased as erosion levels increased. For example, the plastic strains at the splice-bolt holes reached magnitudes of 1.07, 1.16, and 1.22 for cases of 6, 9 and 12 inches of erosion, respectively. As was discussed in Chapter 11, strains of this magnitude for steel are generally associated with a high potential for material failure. In these cases, however, the strains are restricted to a very localized area at the end of the splice-bolt holes and are compressive (i.e., caused from the bearing load between the splice-bolt and the edge of the w- beam hole). Thus, the potential for tear initiation is much lower than the same magnitude of strain in a tensile region of the w-beam (e.g., on the lateral edge of the upper or lower edge of the splice-bolt holes). 1.38 1.69 1.74 1.82 0.91 1.10 1.12 1.08 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Baseline 6 inches 9 inches 12 inches Varying Erosion Depth at 2 Posts A n ch o r D ef le ct io n ( in ) G4(2W) with Various Erosion Depths Upstream Anchor Deflection (in) Downstream Anchor Deflection (in)

335 Table 74. Summary of occupant risk measures from evaluation of undamaged G4(2W) guardrail with various levels of soil erosion at two consecutive posts. Figure 283. Summary of occupant impact velocities for undamaged G4(2W) guardrail with various levels of soil erosion at two consecutive posts. Baseline 6 inches 9 inches 12 inches Occupant Impact Velocity x-direction 5.2 4.9 4.8 4.4 (m/s) y-direction 5.3 5.4 5.6 5.4 at time (0.1519 sec) (0.1541 sec) (0.1580 sec) (0.1625 sec) 7.0 7.2 7.1 6.8 (0.1471 sec) (0.1493 sec) (0.1532 sec) (0.1579 sec) Ridedown Acceleration 10.3 6.9 9.7 7.0 (g's) (0.1519 - 0.1619 sec) (0.1789 - 0.1889 sec) (0.2713 - 0.2813 sec) (0.3036 - 0.3136 sec) 10.7 8.7 9.2 8.5 (0.2198 - 0.2298 sec) (0.2223 - 0.2323 sec) (0.2333 - 0.2433 sec) (0.2179 - 0.2279 sec) 13.7 10.9 12.4 10.2 (0.2003 - 0.2103 sec) (0.2223 - 0.2323 sec) (0.3209 - 0.3309 sec) (0.3004 - 0.3104 sec) 0.93 0.75 0.8 0.81 (0.1219 - 0.1719 sec) (0.0990 - 0.1490 sec) (0.1244 - 0.1744 sec) (0.1065 - 0.1565 sec) Max 50-ms moving avg. acc. 7.6 4.8 5.1 4.8 (g's) (0.1216 - 0.1716 sec) (0.0980 - 0.1480 sec) (0.1741 - 0.2241 sec) (0.1069 - 0.1569 sec) 6.5 6.0 6.4 6.4 (0.1976 - 0.2476 sec) (0.1166 - 0.1666 sec) (0.1244 - 0.1744 sec) (0.1230 - 0.1730 sec) 2.4 2.0 4.2 3.8 (0.3344 - 0.3844 sec) (0.3460 - 0.3960 sec) (0.3467 - 0.3967 sec) (0.3425 - 0.3925 sec) Occupant Risk Factors Erosion at 2 Posts Varying Erosion Depth (g's) x-direction y-direction z-direction ASI PHD THIV (m/s) x-direction y-direction 0 1 2 3 4 5 6 7 Baseline 6 inches 9 inches 12 inches O IV ( m /s ) Erosion Depth at 2 Posts Occupant Impact Velocity x-dir y-dir

336 Figure 284. Summary of maximum occupant ridedown accelerations for undamaged G4(2W) guardrail with various levels of soil erosion at two consecutive posts. Figure 285. Summary of maximum 50-ms running average accelerations for undamaged G4(2W) guardrail with various levels of soil erosion at two consecutive posts. 0 2 4 6 8 10 12 Baseline 6 inches 9 inches 12 inches O R A ( g) Erosion Depth at 2 Posts Occupant Ridedown Acceleration x-dir y-dir 0 1 2 3 4 5 6 7 8 Baseline 6 inches 9 inches 12 inches 5 0 -m s A ve ra ge A cc . ( g) Erosion Depth at 2 Posts 50-ms Avg. Acceleration x-dir y-dir

337 Figure 286. Summary of maximum effective plastic strains occurring at the splice-bolt locations at Post 16. Summary and Discussion The purpose of this study was to quantify the effects of various levels of soil erosion around guardrail posts on system performance. The study involved: (1) performing pendulum impact tests to measure the force-deflection response of the post-soil system for various levels of “manufactured” soil erosion; (2) developing finite element models for the various soil erosion cases and calibrating/validating the models using the test data; and (3) using FEA to quantify the effects of the various degrees of soil erosion on the crash performance of the G4(2W) guardrail system under Report 350 Test 3-11 impact conditions. The pendulum tests involved a W6x16 structural steel post embedded 36 inches in the soil. The W6x16 post was used because of its rigidity, so that the energy absorption measured in the test would be attributable only to soil deformation. The erosion condition was manufactured by removing a layer of soil behind and on the sides of the posts. Four erosion conditions were evaluated in the physical test program: 3 inches, 6 inches, 9 inches and 12 inches of erosion. The impact conditions involved a 2,372-lb pendulum impacting the post at 22 inches above grade at an impact speed of 20 mph. The posts were oriented in the strong direction. The results showed that the energy capacity of the post-soil system reduced as a linear function of erosion depth, in which: 𝐸𝑛𝑒𝑟𝑔𝑦 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 (𝑘𝑖𝑝 − 𝑖𝑛) = 178.33 − 8.04 ∗ (𝑒𝑟𝑜𝑠𝑖𝑜𝑛 𝑑𝑒𝑝𝑡ℎ) Where the constant,178.33 kip-in corresponds to the energy capacity of the post at an embedment depth of 36 inches with no erosion (i.e., consistent with the post embedment depth used in full-scale crash test 471470-26 which has been used as the baseline system throughout this study.) 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 Baseline 6 inches 9 inches 12 inches Varying Erosion Depth at 2 Posts Splice Strains

338 Finite element models for the various soil erosion cases were then developed and validated based on comparison to the physical tests. In each soil-erosion case, the characterization of the soil model was exactly the same; that is, the same model was used in each case. The various levels of erosion were modeled by simply translating the soil-spring model to the appropriate erosion depth. The finite element model of the standard G4(2W) guardrail system was then used to evaluate the effects of soil erosion around guardrail posts on the crash performance of the guardrail system. The impact conditions were set to those of full-scale crash test 471470-26 and involved the 4,568-lb C2500D pickup model impacting the guardrail at 62.6 mph (100.8 km/hr) at an angle of 24.3 degrees.[Mak99] The critical impact for the 6-inch erosion case was determined using FE analysis to be 45 inches (1.16 m) upstream of the w-beam rail splice connection at post 14. Due to time and budget constraints, the CIP for the 6-inch erosion case was used for all subsequent cases. A total of six analysis cases were evaluated. These included two erosion scenarios (i.e., erosion at a single post and erosion at two consecutive posts) with three erosion depths (i.e., 6, 9 and 12 inches). The results indicated that rail deflection increased as erosion depth increased. For the baseline case (i.e., no erosion) the maximum rail deflection was 32 inches. For the case of erosion at a single post, the lateral deflection increased 7.5 percent at 9 inches erosion and 11 percent at 12 inches erosion. For the case of erosion at two posts, with erosion depths of 6, 9, and 12 inches the lateral deflection increased 11 percent, 17 percent, and 25 percent, respectively. The analyses further indicated that the erosion does not significantly affect occupant risk measures, and in general the values were slightly lower for the erosion cases. The most critical effect of erosion was related to the potential for rupture of a rail-splice connection. For the case of erosion at a single post, the three levels of erosion (i.e., 6, 9 and 12 inches) resulted in an increase of 25 percent, 24 percent, and 33 percent increase in plastic strain levels, respectively, relative to the baseline case. Given that the G4(2W) is near its capacity under these impact conditions, the effective increase in plastic strain in the splice, indicates that the potential for rail rupture is relatively high for all cases, but particularly for the 12-inch erosion case. For the case of erosion at two consecutive posts, the three levels of erosion resulted in an increase of 27 percent, 38 percent, and 45 percent increase in plastic strain levels, respectively, relative to the baseline case. Again, the effective increase in plastic strain indicated that the potential for rail rupture was relatively high for all three erosion levels, but particularly so for the 9-inch and 12-inch erosion cases. Recommendations As a result of this study, the research team recommends that the repair threshold for soil erosion at a single post include depths 9 inches or greater measured at the backside of the posts. For erosion depths ranging from 9 to 12 inches, the recommended priority for repair is medium. For erosion depths of 12 inches or greater, the relative priority for repair is high, based on potential for excessive pocketing and increased potential for rail rupture. For soil erosion at two consecutive posts, the research team recommends that the repair threshold include depths 4 inches or greater measured at the backside of the posts. For erosion depths ranging from 4 to 6 inches, the recommended priority for repair is medium. The lower bound for the medium priority was based on engineering judgment, since the analysis matrix did

339 not include erosion depths less than 6 inches. The upper bound was based on the high magnitude of strain in the splice-bolt holes for the 6-inch erosion analysis case, which was considered borderline regarding high potential for rail rupture. For erosion depths of 6 inches or greater, the relative priority for repair is considered high, based on potential for excessive pocketing and increased potential for rail rupture. A summary of the recommendations regarding soil erosion for the G4(2W) guardrail are presented in Table 75. Table 75. Recommendations for soil erosion around guardrail posts for the G4(2W). Damage Mode Repair Threshold Relative Priority High Medium High MediumErosion depth > 4 inches and < 6 inches Erosion at Two or More Consecutive Posts Erosion depth ≥ 6 inches (erosion depth is measured at the back-face of the posts) Erosion at a Single Post Erosion depth ≥ 12 inches (erosion depth is measured at the back-face of the posts) Erosion depth > 9 inches and < 12 inches