Click for next page ( 17

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 16
16 Report 350 Test 3-11 impact (100 km/hr and 25 degrees). The As shown in Figure 7, this task actually conducted two constraining factor was the maximum speed of the FOIL pen- crash tests. In the first test, a length of guard rail was purpose- dulum, which is 32.2 km/hr (20 mph). The speed is limited by fully damaged in a low-severity crash, i.e., a low-speed angled the maximum height to which the pendulum can be raised. impact. This test produced a realistic profile of minor damage The pendulum impacts, however, are more severe than a to the barrier before the second test. Finite element modeling full-scale crash test for two primary reasons: (1) the pendulum predicted that minor deflection could be produced through test section is a more rigid system and (2) the impact energy is an impact speed of 47 km/hr (30 mph). In the second test, distributed over a smaller area. The end fixtures attaching each a Chevy C-2500 pickup truck was impacted into the dam- end of the w-beam test section to the rigid posts allow only aged section of the barrier at NCHRP Report 350 conditions minimal longitudinal translation of the rail section in contrast (100 km/hr and an impact angle of 25 degrees). The result to the full-scale test where the posts surrounding the impact of the crash performance was a laboratory assessment of the area deflect, reducing the tension in the rail and splices. Sec- performance of a barrier with minor post rail deflection ond, in the pendulum test, all the impact energy is absorbed damage. by a single two post (1,905 mm) barrier section. In full-scale Instrumentation for these tests included a tri-axial ac- tests, however, the lateral energy is primarily distributed over celerometer at the vehicle center of gravity, yaw, roll, and two to four of these 1,905 mm (6 foot) barrier sections. To pitch sensors, as well as high-speed photography of the tests. account for this distributed loading, the pendulum impact Detailed pre-test and post-test photographs were taken of speed was reduced to 28.2 km/hr. This impact speed conser- both the guardrail system and the pickup trucks (Figures 8 vatively assumes that the lateral impact energy in a full-scale and 9). The first lower severity test was documented by a total test is absorbed by two 1,905 mm barrier sections, with each of six high-speed cameras recording at 500 frames per second section absorbing half the vehicle lateral kinetic energy. and one real-time camera. High-speed cameras were placed The pendulum tests were primarily intended to test the alongside the guardrail to obtain both a front and rear over- structural adequacy of barrier and damaged barrier sections all view and a single camera was suspended over the impact site based on representative lateral forces induced by a perpendi- to collect an overhead overall view. There were also three high- cular impact to the barrier section. Other relevant barrier per- speed cameras mounted behind the right side of the guardrail formance factors, such as wheel snagging, vehicle rollover, and at varying distance to record the guardrail behavior. The occupant risk, cannot be evaluated using this test methodology. final real-time camera was located behind the left side of the guardrail and panned to capture the full impact. The second full-scale test was documented with an almost identical setup, 3.2.2 Test Plan and Barrier Damage Modes except that the front overall high-speed camera was removed. A total of 3 pendulum tests were conducted of undamaged The remaining high-speed cameras and the real-time camera two-post barrier section to serve as a baseline against which were placed in the same location and recorded at the same to compare the impact performance test sections with minor rates as for the first crash. damage. Eleven tests of damaged barriers were conducted to This test also served as a validation case for the finite ele- test five different barrier damage modes. In each test, a flaw was ment model of vehicle-to-guardrail impact. An LS-DYNA artificially introduced into the test article prior to the pendu- simulation of both crash tests was conducted of both tests lum impact. Table 12 presents a field example of each damage prior to the tests themselves. The accuracy of the LS-DYNA mode and the analogous pendulum test setup. Pendulum tests model was assessed by comparing the results of the simula- were conducted either at 32.2 km/hr (20 mph) or 28.2 km/hr tion with the results of the crash tests. Parameters which were (17.5 mph). compared included vehicle acceleration at the vehicle center of gravity (x, y, and z-axes), vehicle yaw rate, vehicle depar- ture angle from the barrier, and vehicle stability. 3.3 Full-Scale Crash Test Plan This section describes the configuration of a crash test 3.4 Finite Element series to evaluate the crash performance of a damaged longitu- Modeling Approach dinal barrier. The crash performance of the deflected post/rail, the most common damage mode, was evaluated in a full-scale The ideal method to test the safety of strong-post w-beam crash test. The plan was to conduct a full-scale crash test of a guardrail with minor damage would be to perform crash tests. large pickup truck (2000P) into a damaged strong-post w-beam However, the cost of evaluating large numbers of different barrier at NCHRP Report 350 test level 3 conditions (Test 3-11). damage modes would be prohibitive. As an alternative ap- Both tests were conducted by MGA Research in Burlington, WI. proach, finite element modeling was used to evaluate the

OCR for page 16
17 Table 12. Barrier damage modes evaluated in pendulum tests. Damage Mode Field Example Pendulum Test Setup Vertical Tear Horizontal Tear Splice Damage Twisted Blockout Missing Blockout Hole in Rail

OCR for page 16
18 2000 kg Pickup 2000 kg Pickup 30 mph 60 mph 25 25 Crash Test 1 (30 mph) "dents" the rail Crash Test 2 (62 mph) - second impact to the "dent" Figure 7. Full-scale crash testing plan of minor post/rail deflection. crashworthiness of damaged guardrail. Four damage modes were evaluated using finite element modeling: (1) post and Figure 9. Guardrail for MGA crash tests. rail deflection, (2) missing or damaged posts, (3) post and rail separation, and (4) rail flattening. This chapter describes the development and validation of the model used to evalu- models was built using roughly 172,000 elements. Running ate each of these damage modes. the simulations on the system described previously, each sim- The LS-DYNA code was used to develop a finite element ulation took approximately one day of real time to calculate model of the damaged longitudinal barrier systems. LS-DYNA 1,000 ms of simulated time. is used extensively by the roadside safety community to study the impact performance of roadside safety features, and by 3.4.1 The Vehicle-Guardrail Model the automotive industry to study the crashworthiness of passenger vehicles. LS-DYNA is well suited to model the A full-scale finite element model was created from two parts: large deformations and high strain rates which are char- (1) a model of a 175.8-foot (53.6 meters) length of strong-post acteristic of vehicle crashes into roadside features. It is a w-beam guardrail and (2) a model of a Chevrolet 2500 pickup general-purpose, explicit finite element program used to ana- truck. Each model is described in more detail below. All of lyze the nonlinear dynamic response of three-dimensional the initial conditions for the full scale model were adjusted structures (LSTC, 2003). to match the values specified by NCHRP Report 350, i.e., the All of the LS-DYNA finite element models were run on a vehicle was given an initial velocity of 62.1 mph (100 km/hr) SGI Altix parallel system with 120 processors and 512 GB of and angle of impact was set to 25 degrees. An example of a com- memory. Each simulation was run using four processors, pleted full-scale model with 6 inches of rail and post deflec- with multiple simulations being run in parallel to decrease the tion is shown in Figure 10. time needed to complete the study. Each of the finite element 3.4.2 Strong-Post W-Beam Guardrail Model This research program focused on the modified G4 (1S) strong-post w-beam guardrail that uses steel posts with plastic Figure 8. Vehicle orientation prior to first MGA Figure 10. Simulated guardrail with rail and crash test. post deflection.

OCR for page 16
19 Figure 12. The NCAC finite element model of a 1994 Chevrolet 2500 pickup truck. Figure 11. The NCAC strong-post w-beam guardrail model. used Version 0.7, published by the NCAC finite element library on November 3, 2008 (NCAC, 2009b). Like the guardrail blockouts. A guardrail model with steel posts was selected model, this vehicle model was designed to be used with the because the steel posts represent the worst case scenario for LS-DYNA finite element solver. The vehicle is shown below both snagging of the vehicle tires during impact and the devel- in Figure 12. opment of localized stress concentrations on the edges of the The detailed Chevrolet 2500 pickup model was selected for post flanges. While the results using a steel post system will be a number of reasons. First, the model was already subjected to conservative, it was felt better to err on the side of caution a thorough validation effort to ensure the fidelity of the suspen- than to allow a borderline hazardous condition to be consid- sion of structural stiffness (NCAC, 2009a). The detailed model ered an acceptable amount of damage. also incorporates many interior parts that would not be pres- The basic modified steel strong-post w-beam guardrail ent in a reduced model, such as the seating, steering column, model was a publicly available model from the National Crash bearings, fuel tank, and battery. The higher mesh density for Analysis Center (NCAC) finite element library (NCAC, 2009a). the detailed pickup model also improved the accuracy and The basic guardrail model used for this study is shown in Fig- contact stability during simulation. ure 11. The model was designed to be used with the LS-DYNA A limitation of the finite element model of the Chevrolet finite element simulation software (LSTC, 2003). The guardrail 2500 was that the dimensions of the vehicle were fixed. Most system was 53.6 meters (175.8 feet) in length from end to end real pickup trucks have adjustable suspensions, which allow with 29 posts. Routed plastic blockouts were used instead of the front and rear bumper height of the vehicle to vary by as wood blockouts. The soil supporting the guardrail system was much as 100 mm (3.9 inches). However, even changes of a few modeled as individual buckets around each post, rather than centimeters in the relative height of the vehicle and guardrail as a continuum body. Each steel post was embedded in a cylin- have had been shown to have dramatic effects on the crash test drical volume of soil 2.1 meters (6.9 feet) deep and 1.6 meters results (Marzougui et al., 2007). (5.25 feet) in diameter. The success or failure of a crash test can depend greatly on the Since the vehicle and guardrail models selected for use in relative height of the vehicle and guardrail (Marzougui et al., this research were validated against test data, there was little 2007). It was critical that the finite element vehicle model need to make changes to the models. The only alteration to match the recorded dimensions of the real test vehicles as the guardrail model was an increase in the stiffness of the closely as possible. The three crash tests that were used for this springs holding the splice bolts together. The increase in stiff- study had drastically different bumper heights, as shown in ness from 66.5 to 2,400 kN (15 to 540 kip) was needed to keep Table 13. A modified version of the original finite element the splice bolts from unrealistically separating during impact. vehicle was developed to match these alternate dimensions. The increase in stiffness reflected the bolt strength used in a The majority of the simulations in this study were performed model developed for a study on guardrails encased in paved with the vehicle matching the dimensions for the Texas Trans- strips (Bligh et al., 2004). portation Institute (TTI) crash test. 3.4.3 Pickup Truck Model 3.4.4 Matrix of Finite Element Simulations As a test vehicle, the finite element simulations used the Table 14 shows the finite element simulation matrix. In each detailed model of a 1994 Chevrolet 2500 pickup truck avail- case, a flaw was artificially introduced into the basic guardrail able from the NCAC library. Specifically, the simulations model prior to impact.