Weak-Post, W-Beam Guardrail Attachment to Culvert …docs.trb.org/prp/14-3930.pdfPaper No. 14-3930 Duplication for ... Culvert Headwalls by Scott K. Rosenbaugh, M.S.C.E., ... Concrete

Paper No. 14-3930

Duplication for publication or sale is strictly prohibited without prior written permission of the Transportation Research Board

Weak-Post, W-Beam Guardrail Attachment to Culvert Headwalls

by

Scott K. Rosenbaugh, M.S.C.E., E.I.T. Midwest Roadside Safety Facility University of Nebraska-Lincoln

130 Whittier Building 2200 Vine Street

Lincoln, NE 68583-0853 Phone: (402) 472-9324 Fax: (402) 472-2022

Email: [email protected] (Corresponding Author)

Karla A. Lechtenberg, M.S.M.E., E.I.T. Midwest Roadside Safety Facility University of Nebraska-Lincoln


Lincoln, Nebraska 68583-0853 Phone: (402) 472-9070 Fax: (402) 472-2022

Email: [email protected]

Ronald K. Faller, Ph.D., P.E. Midwest Roadside Safety Facility University of Nebraska-Lincoln


Lincoln, NE 68583-0853 Phone: (402) 472-6864 Fax: (402) 472-2022


Robert W. Bielenberg, M.S.M.E., E.I.T. Midwest Roadside Safety Facility University of Nebraska-Lincoln


Lincoln, Nebraska 68583-0853 Phone: (402) 472-9064 Fax: (402) 472-2022


submitted to

Transportation Research Board 93rd Annual Meeting January 12-16, 2014 Washington, D.C.

November 1, 2013

Length of Paper: 5,697 (text) + 2,000 (8 figures) = 7,697 words

TRB 2014 Annual Meeting Paper revised from original submittal.

Rosenbaugh, Faller, Lechtenberg, and Bielenberg 2

ABSTRACT A new W-beam guardrail system for use on low-fill culverts was developed and evaluated. The system was adapted from the MGS bridge railing for attachment to the outside face of culvert headwalls. Four attachment concepts were developed and evaluated through dynamic component testing. Both lateral and longitudinal impacts were conducted on the design concepts while mounted to a simulated concrete culvert headwall. The resulting damage from each test was confined to post bending only. Although all four designs prevented damage to the socket assembly and culvert headwall, the top-mounted, single-anchor design and the side-mounted design were recommended for used based on ease of fabrication and installation. The new W-beam guardrail system for attachment to low-fill culverts was designed with multiple advantages over current culvert treatments. The guardrail system is mounted to the outside face of the headwall, thereby minimizing intrusion over the culvert and maximizing the traversable roadway width. The barrier has an unrestricted system length and does not require a transition when attached to standard MGS. Additionally, the attachment configurations were designed utilizing epoxy anchors enabling the system to be installed on new or existing culverts. Finally, implementation guidance was provided for new system installations. Keywords: Roadside Safety, W-beam Guardrail, Guardrail on Culverts, Weak-Post Guardrail, Non-Blocked Guardrail, and Component Testing INTRODUCTION Concrete box culverts are routinely installed under roadways in order to allow water drainage without affecting the motoring public. Unfortunately, these box culverts can also represent a hazard on the roadside when they do not extend outside of the clear zone and often require safety treatments in the form of roadside barriers. The most common safety barriers utilized to shield these areas are W-beam guardrail systems. However, low-fill culverts with less than 40 in. of soil fill prevent the proper installation of standard guardrail posts due to a lack of available embedment depth. Previous crash testing has shown that W-beam installations with shallow post embedment perform inadequately and are prone to vehicle override (1). Therefore, low-fill culverts require specialized guardrail systems to safely treat the hazard. Currently, two different types of guardrail systems are being used to treat cross-drainage, box culverts: 1) guardrail systems anchored to the top slab of the culvert and 2) long-span guardrail systems. Top-Mounted, Culvert Guardrail Systems

Top-mounted guardrail systems typically consist of W6x9 steel posts welded to base plates which are bolted to the the top slab of the culvert, as shown in Figure 1. Anchoring the guardrail posts to the culvert top ensures that the post will provide the lateral stiffness necessary for the barrier to contain and safely redirect errant vehicles. One system utilizes a 27¾-in. top rail height, a 37½-in. post spacing, a deformable ½-in. base plate, and four 1-in. diameter through bolts (2-3). This system was designed and successfully crash tested to the safety performance criteria of NCHRP Report 350 (4). Recent testing at the Midwest Roadside Safety Facility (MwRSF) led to the development of an epoxy anchoring option for this system utilizing 1-in. diameter, ASTM A307 threaded rods and a minimum 8-in. embedment depth (5). Another


Rosenbaug

system in(6). The performaguardrail

FIGURE(2, 3) and

gh, Faller, Lech

ncorporates rail height

ance criteria l systems we

E 1 (a) NCHd (b) MASH

htenberg, and B

a 75-in. poswas increa

from AASHere designed

HRP ReportH-compliant

Bielenberg

t spacing, a ased to 31 iHTO’s Manu

for use with

t 350-complt, culvert-m

thicker, ⅞-iin. in order

ual for Assessh a minimum

(a)

(b)

liant, culvermounted, W-

in. base plater to satisfy ssing Safety Hm fill depth o

rt-mounted-beam guar

e, and a 31-ithe more dHardware (Mof 9 in. on th

, W-beam grdrail system

in. top rail hdemanding sMASH) (7).

he culverts.

guardrail sym (6).

3

height safety Both

ystem



Although top-mounted guardrail designs provide a crashworthy treatment for culvert

openings, they do have disadvantages. Both of the crashworthy systems were crash tested with an 18-in. offset between the back of the post and the inside of the culvert headwall. MwRSF later recommended a 10-in. minimum offset after analysis of the test high-speed video. This offset is necessary to allow the post to rotate back freely without contacting the headwall. If rotation is restricted by placing the post too close to the headwall, the posts can become snag points and may result in vehicle instabilities (2). However, this 10-in. offset coupled with the footprint of the system itself results in the loss of over 4.5 ft of traversable roadway width. Extending the culvert length another 4.5 ft to gain back this loss in roadway width can drastically increase costs. Additionally, when these systems are impacted and the damaged posts must be replaced, the fill soil must be removed around the posts to gain access to the anchor bolts. This soil removal and replacement after the new post is installed adds to repair time and labor costs.

Long-Span Guardrail Systems

Long-span guardrail systems contain unsupported lengths of W-beam rail that span over the top of culverts, as shown in Figure 2. These barrier systems do not require attachment to the culvert, thus allowing the culvert and the barrier system to operate independently. One crashworthy system consists of 100 ft of nested, 12-gauge W-beam guardrail centered over a 25-ft unsupported span length. A 27¾ in. top rail height was utilized for the entire system. Three wooden CRT posts were placed adjacent to and on both sides of the unsupported span length in order to prevent vehicle pocketing and snagging. This system was designed and successfully crash tested to NCHRP Report 350 safety performance criteria (8-10). The Midwest Guardrail System (MGS) Long-Span is an updated version of the original system and was designed to satisfy MASH safety standards. The MGS Long-Span maintained the 25-ft unsupported span length and the use of six CRT posts. However, only a single layer of W-beam is utilized, the rail height was increased to 31 in., and the rail splices were moved to post mid-spans (11-12).

Since long-span guardrail systems do not require additional components for attachment to the culvert, they provide a cost-effective method for shielding culverts. Further, long-span systems do not require an offset from the culvert and can be installed with the back of the post even with the interior face of the culvert headwall. Thus, long-span systems do not intrude into the roadway width as much as top-mounted systems. However, the NCHRP Report 350 long-span system utilizes double blockouts for a 16-in. total depth, while the MGS Long-Span system utilizes 12-in. deep blockouts. These blockout depths in addition to the 8-in. deep post still results in a loss of nearly 4 ft of traversable roadway width. Finally, long-span systems are limited to a maximum unsupported span length of 25 ft. Thus, culverts with a width, or roadway length, over 25 ft cannot be treated with current long-span W-beam systems.


Rosenbaug

FIGUREMASH-c

Weak-Po Althoughsimilaritiin. tall Wculverts)make thioutside eASTM Aposts andbackup pthe traver

The deck damwhile ondeck or damagedsegments

Finaa weakerweak-posnot requi

gh, Faller, Lech

E 2 (a) NCHcompliant lo

ost, MGS B

h it was noties to culvertW-beam gu. The use ofis system uedge of the bA307 throughd sockets ofplates insteadrsable widthuse of wea

mage. Duringnly minor cra

attachment d posts and s. Thus, repaally, the postr post and rest, MGS bridired between

htenberg, and B

HRP Reportong-span gu

Bridge Rail

t designed ft-mounted b

uardrail and f weak, S3x5unique. The bridge deck. h-bolt and aff the edge od of blockouh (13-14). ak S3x5.7 pog the successacking was sockets, reprail segmen

air to the systs were spaceduced post sdge rail very

n the bridge r

Bielenberg

t 350-compluardrail sys

for use on carrier systemattaches to

5.7 posts andposts are iEach socket bottom steeof the bridgeuts allows fo

osts limits thsful MASH Tobserved in pairs to an nts, droppintem should bed at half-pospacing maky similar to trail and the a

(a)

(b)

liant long-sptem (11-12)

culverts, thems. The weao concrete d the methonstalled intot is attachedel angle, as se deck coupr minimal in

he load tranTL-3 crash tthe bridge dimpacted sy

ng in new pbe quick andost spacing, okes the laterathose of the adjacent MG

pan guardr).

e weak-post,ak-post, MGSbridge deckd used to ato HSS4x4x⅜

d to the bridgshown in Figpled with thentrusion into

nsferred to thtesting progrdeck. Withoystem requiposts, and ad easy. or 37½ in. oal stiffness astandard MG

GS installatio

rail system (

, MGS bridgS bridge railks (similar ttach them to⅜ sockets pge deck withgure 3. The e use of 6-in

o the roadwa

he bridge deram, the posout significaire only theattachment

on center. Thand dynamicGS. Therefoons.

(9-10) and (b

ge rail has l incorporateto concrete

o the bridge placed alongh a 1-in. diamplacement on. long, W-b

ay and maxim

eck and prests were bentant damage te removal oof new W-b

he combinati deflection o

ore, a transiti

5

b)

some es 31-e box

deck g the meter of the beam mizes

events t over to the of the beam

ion of of the ion is


Rosenbaug

FIGURE

RESEAR The mainfill, culvmaximiztime andupstreamMASH tfor concrattachme DESIGN In order componebe used fdiameterdid not csquare tuthe post. within thto alterat

gh, Faller, Lech

E 3 Weak-p

RCH OBJE

n objective oerts that adde the travers

d effort, andm and downs

est level 3 (rete bridge dent to the out

N CRITERI

to avoid coent instead ofor the new s and locatiochange, mucube was utiliz

Finally, thehe socket. Alions.

htenberg, and B

post, MGS b

ECTIVES

of this studydresses the dsable roadwad maintain thstream guard(TL-3). Sincdecks, this stutside face of

IA

onfusion betof two, the sa

guardrail-toons, and stach of the sozed to maint same ⅝-in.ll other sock

Bielenberg

bridge rail (

y was to desidisadvantageay width, prohe ability todrails. Finale the weak-udy was foc

f culvert head

tween similaame post ass

o-culvert attaandoff platesocket had to tain installat ASTM A32

ket features,

(13-14).

ign a new Wes of currenovide an unro be utilizedlly, the systepost, MGS bused on adapdwalls.

ar systems asembly fromachment syss would rem

remain unction toleranc25 diameter components

W-beam guarnt culvert trerestricted syd without a em was to bbridge rail p

apting the we

and allow Sm the weak-ptem. Thus, t

main unchangchanged as wces and ensur

bolt was uss, and attach

rdrail systemeatments. Thystem length

stiffness trbe crashworprovides theeak-post, MG

State DOTs tpost, MGS bthe post sectged. Since thwell. The sare a relativeled to hold th

hment hardw

m for use onhe system w, minimize ransition betrthy accordinese characterGS bridge ra

to stock a sbridge rail wtion, length,he post asseame HSS 4xly tight fit arhe post verti

ware were su

6

n low-was to repair tween ng to ristics ail for

single was to , hole embly x4x⅜ round ically ubject



As stated in the objectives of this study, it was desired to maximize the traversable roadway width over the culvert. Thus, similar to the original bridge rail system, the sockets were to be placed along the outside face of the culvert headwall. Attachment hardware could utilize the top, bottom, or inside surfaces of the headwall, but the socket and post had to remain adjacent to the outside face. Additionally, damage to the system was to be confined to the rail and post components. The culvert and socket were to sustain no damage that would affect the structural integrity of the culvert or the strength and rigidity of the socket assembly. Thus, repairs to an impacted system would be limited to posts and rail segments.

It was also desired to design the new system such that it could be attached to existing structures. One of the disadvantages of the weak-post, MGS bridge rail is that it requires a heavily-reinforced concrete deck which would typically be designed for that specific bridge rail. Researchers wanted to eliminate this disadvantage and allow for installations on standard culvert designs and even existing structures. Therefore, all design concepts were developed and evaluated using a critical culvert design.

In order to identify a critical culvert design, a review of the standard culvert plans from states within the Midwest States Pooled Fund Research Program was conducted. Through this process, the smallest dimensions for top slab thickness, headwall height, and headwall width were identified. Additionally, reinforcement patterns providing the lowest strength in both the top slab and the headwall were identified. Together, these characteristics represented the structurally weakest culvert and the design most susceptible to damage during a vehicle-barrier impact. Subsequently, the critical culvert design was found to be an 18-in. tall by 12-in. wide headwall reinforced with four #4 longitudinal rebar and a #4 stirrup at 12-in. spacings. The critical top slab was 9 in. thick and reinforced with top and bottom steel consisting of #4 rebar at 18-in. spacing. A simulated culvert with these dimensions and reinforcement configurations was constructed at the MwRSF testing site for use during the component testing portion of the project. Additional details for the simulated culvert are provided in the research report (15). DESIGN CONCEPTS Through brainstorming and preliminary design calculations, four socket-to-culvert attachment concepts were developed and subjected to dynamic testing and evaluation. These concepts were: 1) a top-mounted, single-anchor concept; 2) a top-mounted, double-anchor concept; 3) a wrap-around concept; and 4) a side-mounted concept, as shown in Figure 4. All four concepts utilized the same post assembly and steel socket as the original MGS bridge rail. However, each concept has a unique way of transferring impact loads to the culvert in hopes of minimizing attachment and culvert damage. These design concepts are described in the following sections. Top-Mounted, Single-Anchor Concept The top-mounted, single-anchor design was developed to be as similar as possible to the original MGS bridge rail attachment. It utilized a top mounting plate, gusset, and single vertical anchor similar to the original system, as shown in Figure 4(a). Thus, impact loads would be transferred into the culvert as a tensile force through the plate, or shear force through the vertical anchor, and a compression force at the bottom of the socket as it bears against the face of the headwall. However, small changes were implemented to minimize the risk of damaging the culvert and/or socket assembly. The top mounting plate was extended 2 in. to reduce potential concrete cracking by moving the threaded anchor farther away from the edge of the headwall.



Additionally, the plate thickness was increased from 7/16 in. to ½ in. to prevent plate tearing, and the anchor rod diameter was increased to 1⅛ in. to reduce concerns for bearing failure. Finally, the length of the socket was extended 2 in. to 16½ in. in order to increase the moment arm distance from the top mounting plate to the bottom attachment plate, thus resulting in reduced tension and compression forces under a constant bending moment.

The original MGS bridge rail system utilized a through-bolt to anchor the top mounting plate to the bridge deck. In an effort to make the new system attachment applicable to existing structures, the bolt was replaced with a threaded rod embedded 10 in. into the top of the culvert headwall using an epoxy with a bond strength of 1,300 psi. During installation, the socket assembly would be lowered into position over the threaded rod.

A ½-in. thick bottom mounting plate was welded to the lower-front face of the socket. Two ½-in. diameter, ASTM A307 thread rods, one on each side of the socket, were utilized to attach the bottom mounting plate to the headwall. The rods were embedded 4½ in. into the headwall using the same 1,300-psi epoxy adhesive as used for the top threaded rod. Two ⅝-in. wide slots were cut into the bottom mounting plate so that the socket assembly could be lowered into place over the threaded rods. Standard washers and nuts were used to attach the socket to the headwall. The socket, mounting plates, and gusset plate were all fabricated with 50-ksi steel. Top-Mounted, Double-Anchor Concept Due to the design similarities with the original weak-post, MGS bridge rail, concerns arose that a single-anchor design could result in concrete cracking similar to that observed during full-scale testing of the MGS bridge rail. Therefore, the top-mounted, double-anchor concept was developed to better distribute the tensile force from the top mounting plate to the headwall and prevent concrete shear cracking.

The only differences between the top-mounted, double-anchor design and the top-mounted, single-anchor design are the top mounting plate dimensions and the use of a second top anchor rod. Two ¾-in. diameter, ASTM A307 threaded rods spaced 6 in. apart were used to anchor the socket assembly, as shown in Figure 4(b). The anchor rods were embedded 4½ in. into the headwall using an epoxy adhesive with a 1,300-psi bond strength. To accommodate the double anchors, the top mounting plate was flared from a 3 in. width adjacent to the socket to a 9 in. width around the anchors.

Similar to the previous design concept, the top-mounted, double-anchor concept was installed by lowering the socket assembly over the epoxy-embedded, threaded rods. Standard washers and nuts were used to attach the socket to the headwall. The socket, mounting plates, and gusset plate were all fabricated with 50 ksi steel. Wrap-Around Concept The wrap-around design concept was developed to further reduce the risk of concrete cracking and failure of the culvert headwall. The wrap-around concept incorporates an elongated top mounting plate that extends over the top of the headwall and continues down the inside face, as shown in Figure 4(c). This concept removed all anchor hardware from the top of the culvert headwall. Although not prevalent during full-scale testing of the original MGS bridge rail system, preventing possible interactions between vehicle tires and the attachment hardware was considered a positive design aspect.



The ½-in. thick top mounting plate maintained a 3-in. width throughout its length and was attached to the inside face of the headwall utilizing a ⅝-in. diameter ASTM A307 threaded rod. The threaded anchor was necessary to keep the top plate in tension and prevent it from unfolding and releasing from the headwall. The bottom plate, bottom anchor rods, and socket tube all remained the same as used in the top-mounted designs. Standard washers and nuts were utilized and the socket, mounting plates, and gusset plate were all fabricated with 50-ksi steel.

For the test installation, the top anchor rod was embedded 4½ in. into the headwall using an epoxy adhesive with a bond strength of 1,300 psi. Consequently, the socket assembly had to be lowered into place before the top anchor was epoxied into the headwall. During removal, the threaded rod had to be cut, and the top mounting plate had to be bent open slightly. However, either a mechanical anchor or an epoxied-anchored threaded insert could have been used to make the installation and removal of the socket assembly easier and the attachment location reusable. Finally, the wrap-around design requires that the inside face of the culvert headwall be exposed for proper installation, similar to the existing guardrail designs that mount to the culvert top slab. This additional soil movement would add to installation and repair costs. Side-Mounted Concept The fourth design concept aimed to keep all attachment hardware on the outside face of the culvert headwall and prevent interactions between vehicle components and attachment hardware. The side-mounted design concept utilized a ½-in. thick top mounting plate, two ¼-in. thick gusset plates, and two ¾-in. diameter ASTM A307 threaded rods to anchor the top of the socket assembly, as shown in Figure 4(d). The top mounting plate was ½ in. thick, matching the bottom mounting plate and keeping the socket vertical. Gusset plates were added between the socket and the top mounting plate to prevent the plate from bending outward when subjected to high lateral loads. The threaded rods were centered 4½ in. from the top of the headwall to avoid interference with the internal steel reinforcing bars that are typically placed near the top of the headwall.

Two options were explored for the anchorage of the top threaded rods: 1) embedding the rods 9 in. using epoxy with a 1,300-psi bond strength and 2) extending the rods through the headwall and using plate washers and nuts on the inside face. Both of these options were evaluated during the dynamic testing portion of this study.

The bottom mounting plate and threaded rods remained largely unchanged from the previous design concepts. However, since the socket assembly was installed laterally instead of dropped in vertically, slotting the bottom mounting plate was unnecessary. Therefore, only ⅝-in. diameter holes were drilled into the bottom plate.


Rosenbaug

FIGUREdouble-a

gh, Faller, Lech

E 4 The fouanchor; (c) w

htenberg, and B

(a)

(b)

ur design cowrap aroun

Bielenberg

ncepts: (a) tnd; and (d) s

top-mounteside-mounte

ed, single-aned.

(c)

(d)

nchor; (b) toop-mounted

10

d,



DYNAMIC BOGIE TESTING Evaluation Criteria and Testing Conditions New highway barriers must typically be subjected to full-scale crash testing and satisfy the MASH safety performance criteria in order to be deemed crashworthy. However, the original weak-post, MGS bridge rail had already satisfied the MASH TL-3 criteria, and this study focused only on adapting the original system for use on culvert headwalls. In fact, the W-beam rail, rail-to-post attachment hardware, mounting height, post assembly, and socket tube all remained unchanged from the original bridge rail. The only new components in these concepts were the attachment hardware utilized to mount the socket flush with the outside face of the culvert headwall. Further, the new attachments and anchorage pieces were designed to withstand impact loads and remain undamaged, while the post and rail components deform and absorb energy. If these new components were shown to withstand extreme loading conditions without damage to the socket assembly or the culvert headwall, the new weak-post guardrail attached to concrete box culvert systems would perform similarly to the original weak-post bridge rail. Thus, full-scale testing was deemed unnecessary, and the evaluation of the new design concepts was limited to dynamic component testing.

As previously discussed, a critical concrete box culvert design was identified and constructed at the MwRSF test site. Seven dynamic component tests were conducted with the test articles mounted to the simulated culvert. The test articles for each component test consisted of an S3x5.7 steel post installed in one of the four socket design configurations. A minimum distance of 3 ft was used between the test component and the nearest support wall in order to fully evaluate possible damage to the culvert headwall and top slab.

The design concepts were subjected to two critical loading conditions. The first involved a lateral impact on the post at a height of 24⅞ in., subjecting it to strong-axis bending. These impact conditions were selected to match the height to the center of the W-beam rail and are similar to the conditions typically used to observe the performance of guardrail posts installed in soil. The second test was a longitudinal impact where the post was subjected to weak-axis bending. The longitudinal impacts were conducted with a load height of 12 in. to simulate a small car bumper impacting posts during a redirection. This second impact was deemed critical because it induces higher shear load into the socket and may cause the socket to rotate downstream.

All testing was performed with a rigid-frame bogie vehicle weighing approximately 1,800 lb impacting at a target speed of 20 mph. The impact head of the bogie was covered with a ¾-in. thick neoprene pad to dampen out high frequency vibrations during the impact. Accelerometers were mounted near the bogie’s center of gravity to determine the dynamic response during the event. Dynamic Bogie Testing Results Test nos. CP-1 through CP-5 were lateral impact tests conducted on each of the socket attachment concepts. Test nos. CP-3 and CP-5 were conducted on variations of the side-mounted concept with CP-3 utilizing through-bolts for the top anchor rods and test no. CP-5 utilizing epoxy-anchored threaded rods. All lateral impact tests resulted in the posts bending about the strong axis at a location adjacent to the top-back edge of the socket. Plastic bending continued until the bogie vehicle eventually overrode the post.



No damage was evident on the culvert in the form of concrete cracking or spalling. Plastic deformations to the socket assemblies or anchor rods were also not found. From an analysis of the high-speed video of the impacts, only slight lateral movements of the socket were documented for the two top-mounted concepts and the wrap-around concept. These small lateral translations at the top of the sockets were attributed to the construction tolerance given to the attachment hardware (i.e., holes in the top mounting plates being slightly oversized compared to the diameter of the threaded rods, and the wrap-around plate being slightly longer than the width of the headwall). None of the sockets shifted enough to affect the installment of a replacement post.

Accelerometer data from each test was processed and analyzed to calculate force and displacement data as a function of time. Force vs. deflection plots for the lateral impacts are shown in Figure 5. All force curves were very similar, which was expected given each test resulted in post bending. Each test article loaded quickly before providing a relatively steady resistance around 6 kips through 13 in. of deflection. Interestingly, the two top-mounted concepts and the wrap-around concept each had a dip in resistance at about 3 in. of deflection. This would coincide with the slight shifting of the top mounting plates described previously and explains why the two side-mounted concept tests did not result in similar force dips.

FIGURE 5 Force vs. deflection curves for the lateral impact tests.

After the completion of the lateral impact testing, it was clear that the weak-post system would not transfer enough load to the culvert headwall to cause significant concrete damage. Recall, the top-mounted, double-anchor concept and the wrap-around concept were developed due to concerns for possible damage to the culvert. With these concerns alleviated, only the top-mounted, single-anchor and the side-mounted concepts were recommended for continued testing.

0

2

4

6

8

10

0 5 10 15 20 25 30

Force (kips)

Deflection (in.)

Force vs. Deflection ‐ Lateral Impacts

CP‐1: Wrap‐Around

CP‐2: Top ‐Mount, Single‐Anchor

CP‐3: Side‐Mount, Through Bolt

CP‐4: Top‐Mount, Double‐Anchor

CP‐5: Side‐Mount, Epoxy Anchor



Further testing of the through-anchor, side-mounted concept was also discontinued as the epoxy anchor variation was viewed as easier to install and provided very similar results during the lateral impact tests.

Test nos. CP-6 and CP-7 were longitudinal impact tests conducted on the epoxy-anchored side-mounted design and the top-mounted, single-anchor design, respectively. The longitudinal impacts resulted in similar results. Upon impact, the posts began bending about the weak axis at a location adjacent to the top-downstream edge of the socket. Plastic bending continued until the bogie vehicle overrode the post. No damage was evident in the form of concrete cracking or spalling, the epoxied anchor rods remained undamaged, and plastic deformations to the socket assembly were not found. Force vs. deflection curves for both longitudinal tests are shown in Figure 6. The force curves are similar in magnitude and duration. However, significantly more ringing vibrations occurred in the bogie vehicle during the longitudinal impact tests than during the lateral impact tests.

FIGURE 6 Force vs. deflection curves for the longitudinal impact tests.

Post-test photographs of the top-mounted, single-anchor and the side-mounted design concepts for both lateral and longitudinal impacts are shown in Figure 7. All damage and plastic deformations sustained during the impacts were confined to the S3x5.7 posts. The damaged posts were easily removed from the sockets and new posts installed in their place. Therefore, both the top-mounted, single-anchor and side-mounted design configurations were recommended for use in the new W-beam guardrail system for attachment to culvert headwalls.

‐4

‐2

0

2

4

6

8

10

0 5 10 15 20 25 30

Force (kips)

Deflection (in.)

Force vs. Deflection ‐ Longitudinal Impacts

CP‐6: Side‐Mount, Epoxy Anchor

CP‐7: Top‐Mount, Single‐Anchor


Rosenbaug

FIGURE

CONCL A new wevaluatedsingle-anattachmeof 37½ idrawings

BothThe S3x5blockoutthe traver

Unliis unrestrAdditioninstalled the culve

gh, Faller, Lech

(a) CP-2: La

(c) CP-5: La

E 7 Post-tes

LUSIONS A

weak-post, Wd through dnchor and sient configurain. on centes can be founh attachment5.7 posts pos results in arsable roadwike long-sparicted in term

nally, the socon new or e

ert itself, was

htenberg, and B

Top

ateral Impac

ateral Impac

st damage p

ND RECOM

W-beam gudynamic comide-mountedation, the over and a 31nd in the testt configuratiositioned offa minimal baway width\cuan guardrail ms of system

ckets are attaexisting culvs damaged d

Bielenberg

p-Mounted,

ct

Side-M

ct

photographs

MMENDAT

uardrail systemponent tesd attachmentverall layout in. top rait report (15)ions mount tf the edge oarrier intrusiulvert lengthsystems, them length anached to the vert structur

during the cri

Single-Anc

Mounted Des

s from both

TIONS

em for attasting. Systemt configuratof the systel mounting . the socket toof the culverion over the

h is maximizee new W-beand can be us

headwall usres. Since noitical impact

chor Design

(b) CP-7:

sign

(d) CP-6:

lateral and

chment to cm drawingstions are shoem remains t

height. A m

o the outside rt coupled wculvert and ed. am guardrai

sed to treat csing epoxy aone of the sot tests, repair

n

Longitudina

Longitudina

d longitudin

culverts was for both town in Figuthe same wimore detaile

face of the with the syst

into the roa

il system attculverts oveanchors, so tocket assemr to a damag

al Impact

al Impact

nal impacts.

s developedthe top-mouure 8. For eith a post sped set of sy

culvert headtem not requadway. There

tached to culer 25 ft in lethe system c

mbly hardwarged system w

14

d and unted, either acing ystem

dwall. uiring efore,

lverts ength. an be re, or

would



consist of simply removing damaged posts and dropping replacement posts into the undamaged sockets.

The side-mounted attachment configuration shown in Figure 8 illustrates the top anchor rods embedded 9 in. into the headwall with an epoxy adhesive. However, the end user may utilize a through-bolt design variation for the top anchor rods. The drilled holes would be extended through the entire headwall, and ¼-in. thick plate washers would be installed along the inside face of the headwall. This through-bolt configuration was evaluated through lateral impact testing during test no. CP-3, and it performed nearly identical to the epoxied anchor variation.

The test installations evaluated during this study utilized an epoxy adhesive with a specified minimum bond strength of 1,300 psi. Therefore, the W-beam, guardrail system attached to culverts can be installed using a wide variety of epoxy adhesives given the specified bond strength remains at least 1,300 psi.

This barrier system was designed as part of a family of non-proprietary, 31-in. high W-beam guardrail systems commonly referred to as the MGS. This new guardrail system attached to culverts was designed with a similar lateral stiffness and overall system performance as the original MGS. Therefore, a transition between the new guardrail attached to culvert system and adjacent standard MGS installations is unnecessary. A 75-in. spacing is recommended between the last S3x5.7 culvert post and the first standard guardrail post of the adjacent MGS installation, as shown in Figure 8.

Although a critical culvert headwall was selected for use in the dynamic impact tests, care should be taken not to install the W-beam guardrail attached to culverts on headwalls of significantly smaller size or reduced internal reinforcement. Installations on weaker structures may result in unwanted damage to the headwall in the form of concrete cracking and/or spalling. Additionally, the system was designed and evaluated for use on low-fill culverts with a relatively flat grading. It is recommended that the system only be used with approach slopes of 10H:1V or flatter.

Finally, installations should be installed with the guardrail terminals and/or end anchorages at a sufficient distance from the culvert to prevent the two systems from interfering with the proper performance of one another. As such, the following implementation guidelines should be considered:

1. A recommended minimum length of 12 ft – 6 in. for standard MGS between the first S3x5.7 weak post and the interior end of an acceptable TL-3 guardrail end terminal.

2. A recommended minimum barrier length of 50 ft before the first S3x5.7 weak post, which includes standard MGS, a crashworthy guardrail end terminal, and an acceptable anchorage system. This guidance applies to the downstream end as well.

3. For flared guardrail applications, a recommended minimum length of 25 ft between the first S3x5.7 weak post and the start of the flared section (i.e. bend between flared and tangent sections).


Ro

FI

osenbaugh, Faller, L

IGURE 8 Wea

Lechtenberg, and B

ak-post guardra

Bielenberg

ail attachment to culvert headdwalls, system llayout and desi

16

ign options.



ACKNOWLEDGEMENTS The authors acknowledge two major sources that made contributions to this project: the Midwest States Regional Pooled Fund Program for sponsoring the research project, and Midwest Roadside Safety Facility personnel for constructing the test articles, conducting the impact tests, and documenting the results. REFERENCES 1. Hirsch, T.J., and Beggs, D., Use of Guardrails on Low Fill Bridge Length Culverts,

Transportation Research Record No. 1198, Transportation Research Board, National Research Council, Washington D.C., 1988.

2. Polivka, K.A., Faller, R.K., Sicking, D.L., Rohde, J.R., Reid, J.D., and Holloway, J.C.,

NCHRP 350 Development and Testing of a Guardrail Connection to Low-Fill Culverts, Research Report No. TRP-03-114-02, Midwest Roadside Safety Facilty, University of Nebraska-Lincoln, Lincoln, Nebraska, November 1, 2002.

3. Polivka, K.A., Faller, R.K., and Rohde, J.R., Guardrail Connection to Low-Fill Culverts,

Transportation Research Record No. 1851, Transportation Research Board, National Research Council, Washington D.C., 2003.

4. Ross, H.E., Sicking, D.L., Zimmer, R.A., and Michie, J.D., Recommended Procedures for the Safety Performance Evaluation of Highway Features, National Cooperative Highway Research Program (NCHRP) Report 350, Transportation Research Board, Washington, D.C., 1993.

5. Price, C.W., Rosenbaugh, S.K., Faller, R.K., Sicking, D.L., Reid, J.D., and Bielenberg, R.W.,

Post Weld and Epoxy Anchorage Variations for W-Beam Guardrail Attached to Low-Fill Culverts, Draft Research Report No. TRP-03-278-13, Midwest Roadside Safety Facilty, University of Nebraska-Lincoln, Lincoln, Nebraska, May 2013.

6. Williams, W.F., and Menges, W.L., MASH Test 3-11 of the W-beam Guardrail on Low-Fill

Box Culvert, Test Report No. 405160-23-2, Texas A&M Transportation Institute, Texas A&M Univeristy, College Station, Texas, November 11, 2011.

7. Manual for Assessing Safety Hardware (MASH), American Association of State and

Highway Transportation Officials (AASHTO), Washington, D.C., 2009.

8. Polivka, K.A., Bielenberg, B.W., Sicking, D.L., Faller, R.K., and Rohde, J.R., Development of a 7.62-m Long Span Guardrail System, Research Report No. TRP-03-72-99, Midwest Roadside Safety Facility, University of Nebraska-Lincoln, April 6, 1999.

9. Polivka, K.A., Bielenberg, B.W., Sicking, D.L., Faller, R.K., Rohde, J.R., and Keller, E.A.,

Development of a 7.62-m Long Span Guardrail System - Phase II, Research Report No. TRP-03-88-99, Midwest Roadside Safety Facility, University of Nebraska-Lincoln, August 13, 1999.



10. Faller, R.K., Sicking, D.L., Polivka, K.A., Rohde, J.R., and Bielenberg, R.W., A Long-Span

Guardrail System for Culvert Applications, Paper No. 00-0598, Transportation Research Record No. 1720, Transportation Research Board, Washington, D.C., 2000.

11. Bielenberg, B.W., Faller, R.K., Rohde, J.R., Reid, J.D., Sicking, D.L., Holloway, J.C.,

Allison, E.M., and Polivka, K.A., Midwest Guardrail System for Long-Span Culvert Applications, Research Report No. TRP-03-187-07, Midwest Roadside Safety Facility, University of Nebraska-Lincoln, November 16, 2007.

12. Bielenberg, R.W., Faller, R.K., Sicking, D.L., Rhode, J.R., and Reid, J.D., Midwest

Guardrail System for Long-Span Culvert Applications, Transportation Research Record No. 2025, Transportation Research Board, National Research Council, Washington D.C., 2007.

13. Thiele, J.C., Sicking, D.L., Faller, R.K., Bielenberg, R.W., Lechtenberg, K.A., Reid, J.D.,

and Rosenbaugh, S.K., Development of a Low-Cost, Energy-Absorbing Bridge Rail, Research Report No. TRP-03-226-10, Midwest Roadside Safety Facility, University of Nebraska-Lincoln, Lincoln, Nebraska, August 11, 2010.

14. Thiele, J.C., Sicking, D.L., Lechtenberg, K.A., Reid, J.D., Faller, R.K., Bielenberg, R.W.,

and Rosenbaugh, S.K., Development of a Low-Cost, Energy-Absorbing Bridge Rail, Transportation Research Record No. 2262, Transportation Research Board, National Research Council, Washington D.C., 2011.

15. Schneider, A.J., Rosenbaugh, S.K., Faller, R.K., Sicking, D.L., Lechtenberg, K.A., and Reid,

J.D., Safety Evaluation Performance of MGS Attached to Culverts, Draft Research Report No. TRP-03-277-13, Midwest Roadside Safety Facility, University of Nebraska-Lincoln, Lincoln, Nebraska, July, 2013.


Documents

Weak-Post, W-Beam Guardrail Attachment to Culvert …docs.trb.org/prp/14-3930.pdfPaper No. 14-3930 Duplication for ... Culvert Headwalls by Scott K. Rosenbaugh, M.S.C.E., ... Concrete