i

DEPARTMENT OF CIVIL ENGINEERING FACULTY OF ENGINEERING AND ARCHITECTURAL SCIENCE

DEVELOPMENT OF PL-3 BARRIER-DECK DETAILS INCORPORATING V-RODS WITH HEADED ENDS, BENTS AND 180˚

HOOKS

Prepared by Khaled Sennah, Ph.D., P.E., P.Eng. Hamidreza Khederzadeh, M.A.SC.

Professor, Civil Engineering Department, Ph.D candidate, Civil Engineering Department, Ryerson University, Toronto, Ryerson University, Toronto, Ontario, Canada Ontario, Canada E-mail: ksennah@ryerson.ca Email: hkhederz@ryerson.ca

Submitted to:

Pultrall Canada Inc. 700, 9th Street North

Thetford Mines, Quebec, Canada G6G 6Z5

February 2013

______________________________________________________________________________ 350 Victoria Street, Toronto, Ontario, Canada M5B 2K3 Tel: 416-979-5345 Fax: 416-979-5122 www.ryerson.ca/civil

ii

ABSTRACT Corrosion of steel reinforcement due to environmental effects is a major cause of deterioration problems in bridge barriers. Glass fibre reinforced polymers (GFRP), not only addresses this durability problem but also provides exceptionally high tensile strength. A recent design work conducted at Ryerson University on PL-3 bridge barrier has led to an economical glass fibre reinforced polymer (GFRP) bar detailing for sustainable construction. Such design incorporated the special sand-coated GFRP bars and end anchorage heads to ensure optimal bond between concrete deck slab and barrier wall. Also, this design conducted at Ryerson University on PL-3 bridge barrier proposed the use of 15 mm (#5) and 12 mm (#4) diameter GFRP bars as vertical reinforcement in the barrier front and back faces, respectively, with 12 mm (#4) diameter GFRP bars as horizontal reinforcement in PL-3 barrier wall in both front and back face of the barriers. Given the fact that the material cost of the GFRP bars differs based on the type of the bar (i.e. high-modulus versus standard modulus, straight bars versus bent bars and headed-end bars versus bars with 180° hooks), Ryerson University research team developed three different GFRP bar detailing for barrier wall reinforcement and barrier-deck-anchorage. Proposal No. 1 incorporated the use of HM bars with headed ends embedded in the deck slab. While proposal No. 2 incorporated SM bars with bent bars. Proposal No. 3 incorporated SM bars with 180° hooks embedded in the deck slab. Since established design criteria for GFRP-reinforced barriers is as yet unavailable, it was decided to build and test to-collapse few barrier models to qualify the proposed barrier details for ultimate strength. Five full-scale barrier models of 1000 mm length, representing PL-3 barriers were erected and tested to-collapse to determine their ultimate load carrying capacities and failure modes. Four barriers were erected using GFRP bars. Two of these barrier models were built using high modulus (HM) GFRP bars with end anchorage heads, while the other two models were built using GFRP bent and 180˚ hook, respectively, with standard modulus (SM) V-Rod bars. The fifth barrier model was made of conventional steel reinforcements as per Ministry of Transportation of Ontario (MTO) barrier drawings that was used as reference sample. This paper presents the results from ultimate load tests in the form of crack pattern, deflection history, and ultimate load carrying capacity as compared to the design values specified in the Canadian Highway Bridge Design Code for barrier anchorage into the deck slabs. The experimental ultimate load carrying capacities of the barriers were observed to be greater than the factored design loads specified in Canadian Highway Bridge Design Code. Thus, the experimental findings will allow the GFRP manufacturer to decide, with confidence, on the most economical GFRP detailing out of the three proposals.

ACKNOWLEDGEMENTS

The authors acknowledge the financial support and in-kind contribution to this research project by Pultrall Canada Inc. of Quebec, and Trancels-Pultrall Inc. of Barrie, Ontario. The matching financial support by Ontario Centres of Excellence (OCE) is greatly appreciated. The continuous support, commitment and dedication of Mr. Nidal Jaalouk, the senior technical officer of Ryerson University, were an integral part of the experimental work reported in this study.

iii

TABLES OF CONTENTS

ABSTRACT .................................................................................................................................... ii

ACKNOWLEDGEMENTS ............................................................................................................ ii

1. INTRODUCTION ..................................................................................................................... 1

2. GFRP TECHNOLOGY .............................................................................................................. 1

3. BACKGROUND OF THE DEVELOPED GFRP-REINFORCED BARRIER ......................... 3

4. DESCRIPTION OF THE EXPERIMENTAL PROGRAM ON PL-3 BARRIER MODELS .... 7

5. EXPERIMENTAL RESULTS FOR THE TESTED PL-3 BARRIER MODELS ................... 12

6. CONCLUSIONS....................................................................................................................... 28

7. REFERENCES ......................................................................................................................... 29

APPENDIX A: RESULTS FROM CONCRETE CORE SAMPLING ........................................ 30

1

1. INTRODUCTION Conventional Bridges built prior to the 1970’s did not use air-entrained concrete and coated reinforcing steel bars to protect from the effects of freeze-thaw cycles and the application of winter de-icing salt. This leads to corrosion-induced degradation in bridge elements as shown in Fig. 1. Accordingly, exposed bridge elements are all likely candidates for expensive replacement on the majority of these older bridges. In November 2007, The Residential and Civil Construction Alliance of Ontario, Canada, (RCCAO 2007) released a report on the state of Ontario bridges, entitled “Ontario’s Bridges: Bridging the Gap.” The report warns that the integrity of Ontario’s municipal bridge infrastructure and public safety are at risk after years of deferred maintenance, irregular inspections, and lack of government oversight. The RCCAO report stated some recommendations to be made to promote the public’s safety and the sustainability of Ontario’s bridges. One of these recommendations includes promoting bridge engineering designs that improve the life expectancy and reduce maintenance costs of bridges. This can be achieved by using fibre reinforced polymer bars.

Fig.1 Corrosion-induced degradation of steel-reinforced bridge barrier wall

2. GFRP TECHNOLOGY Fibre-reinforced polymers (FRPs), as non-corrodible materials, are considered as excellent alternative to reinforcing steel bars in bridge decks to overcome steel corrosion-related problems. Since it is less expensive than carbon and aramid FRPs, Glass FRP (GFRP) bars are more attractive to bridge deck and barrier applications. Two types of GFRP bars were considered in used in this study, namely: (i) GFRP bars with high modulus (HM) and (ii) GFRP bars with standard modulus (SM). The sand-coated surface profile of these bars, shown in Fig. 2, ensure optimal bond between concrete and the bar. Table 1 summarizes the material properties of the bars considered in

2

this study (Pultrall, 2012). It should be noted that the bent bars and bars with 180° hook are manufactured using standard modulus material. The modulus of elasticity of such bent bars is determined to be 50 GPa and the tensile strength of the bar at the bent of hook locations is considered 473 MPa in this study.

Fig. 2 View of GFRP bar types

Table 1 GFRP Material properties (Pultrall)

Product Type Bar Size Guaranteed Tensile

Strength (MPa) Modulus of

Elasticity (GPa)

Strain at

Failure

Cross-Sectional

Area (mm2)

V-Rod-HM #4 (M12) 1312 65.6 2.5 2.0% 126.7 #5 (M15) 1184 62.5 2.5 1.89% 197.9 #6 (M19) 1105 64.7 2.5 1.70% 285

V-Rod-Standard

#4 (M12) 941 53.6 1.76% 126.7 #5 (M15) 934 55.4 1.69% 197.9 #6 (M19) 807 56.6 1.43% 285

V-Rod-Bent #5 (M15) 473 (bent portion)

50 1% 197.9 1051(Straight Portion)

V-Rod-180˚ hook

#5 (M15) 473 (bent portion)

50 1% 197.9 1051(Straight Portion)

* HM = High modulus ** SM = Standard modulus

3

3. BACKGROUND OF THE DEVELOPED GFRP-REINFORCED BARRIER The design process of bridge barrier walls specified in the Canadian Highway Bridge Design Code (CSA, 2006a; CSA, 2006b) is based on the AASHTO Guide Specification for bridge railings (AASHTO, 1989) and the AASHTO Guide for Selecting, Locating and Designing Traffic Barriers (AASHTO, 1977). CHBDC Clause 12.4.3.5 specifies that the suitability of a traffic barrier anchorage to the deck slab shall be based on its performance during crash testing of the traffic barrier. For an anchorage to be considered acceptable, significant damage shall not occur in the anchorage or deck during crash testing. It also specifies that if crash testing results for the anchorage are not available, the anchorage and deck shall be designed to resist the maximum bending, shear and punching loads that can be transmitted to them by the barrier wall. As such, the initial design of the proposed PL-3 precast bridge barrier (Sennah et al., 2010) was carried out to meet the CHBDC design criteria specified for static loading at the anchorage between the deck slab and the barrier wall. CHBDC specifies transverse, longitudinal and vertical loads of 210, 70 and 90 kN, respectively, that can be applied simultaneously over a certain barrier length. CHBDC specifies that transverse load shall be applied over a barrier length of 2400 mm for PL-3 barriers. Since transverse loading creates the critical load carrying capacity, both the longitudinal and vertical loads were not considered in the design of barrier wall reinforcement and anchorages between the deck slab and the barrier wall. It should be noted that CHDBC specifies a live load factor of 1.7. Thus, the design impact load on PL-3 barrier wall over 2.4 m length is 357 kN. As for the design of the vertical and horizontal reinforcement in the barrier wall, AASHTO-LRFD Bridge Design Specifications (AASHTO, 2010) specifies the yield-line analysis for the ultimate flexural capacity of the barrier wall under vehicle impact. In the analysis, it was assumed that the yield-line failure pattern occurs within the barrier wall only and does not extend into the deck slab. This means that the deck slab must have sufficient resistance to force the yield-line failure pattern to remain within the barrier wall. The LRFD yield-line analysis is also based on the assumption that sufficient longitudinal length of barrier wall exists to result in the desired yield-line failure pattern. It should be noted that the yield-line analysis is typically invoked to evaluate the nominal strength of concrete bridge barriers with steel reinforcement. Due to the linear elastic behavior of GFRP reinforcing bars up to failure, the ASSHTO-LRFD yield-line equations cannot be used in the design of GFRP reinforcement in the barrier wall. However, both equilibrium of forces and compatibility of deformation conditions should be explicitly accounted for by implementing suitable analytical or numerical methods. Since such analytical or numerical modeling is as yet unavailable, evaluation of a full-scale structure or a prototype by loading testing is the only means to qualify a proposed barrier detailing for use in Canada’s bridges. CHBDC prescribes guidelines for the railing loads to be considered for the design of barrier anchorage and the edge of the deck slab in the bridge cantilever barriers. According to the maximum lateral loads and dispersion angle of the load effects along the height of the barrier (see Table 2), factored applied moments at the barrier-deck junction are given in CHBDC driven from Finite-Element (FE) Analysis. These are the resulting moments at the face of the barrier-to-deck joint used in the design of the barrier anchorage. Lateral load also exerts shear at the barrier anchorage which usually does not govern. CHBDC specifies that the design of the barrier-deck

4

anchorage capacity can be achieved by code-approved manual calculations or experimental testing under static loads to failure in lieu of conducting vehicle crash testing. Table 2: Transverse moments in cantilever slabs due to horizontal railing loads in selected PL-3

and PL-2 barriers. (Table C5.4 in CHBDC Commentary, CSA 2006b)

A preliminary design work conducted at Ryerson University for the PL-3 bridge barrier led to the first world-wide vehicle crash test recently conducted by Ryerson University research team (Sennah et al, 2011). This crash test resulted in a MTO Standard Drawing SS110-92 released by the Ontario Ministry of Transportation in May 2011. It should be noted that Ryerson University

5

research team conducted the second world-wide vehicle crash testing on PL-3 barrier using V-Rod-HM bars with headed end in December 2011 (Sennah and Khederzadeh, 2012). As shown in Fig. 3a, 15M and 12M GFRP bars were used as vertical reinforcement in the barrier front and back face, respectively, with 12M GFRP bars as horizontal reinforcement, all at 300 mm spacing. The connection between the deck slab and the barrier wall utilized the GFRP headed end bars for proper anchorage. It should be noted that this new arrangement of the GFRP bars would significantly reduce the cost, given the increased spacing of the vertical bars to 300 mm. Also, the traditional lap splice of the vertical bars at the back face of the barrier wall as well as the bents of these bars in the bridge deck slab, shown in Fig. 3b, were eliminated to reduce the material and labor cost. This decision was reached due the fact that the vertical bars at the back face of the barrier walls are always in compression when the barrier is subjected to vehicle impact.

Fig. 3a Barrier detailing from MTO Standard Fig. 3b Traditional barrier reinforcement Drawing SS110-92 (MTO, May 2011) with bent bars Given the fact that the material cost of the GFRP bars differs based on the type of the bar (i.e. high-modulus versus standard modulus, straight bars versus bent bars and headed-end bars versus bars with 180° hooks), Ryerson University research team developed three different GFRP bar detailing as shown in Fig. 4. Proposal No. 1 in this figure incorporated HM bars with headed ends embedded in the deck slab. While proposal No. 2 incorporated SM bars with bents. Proposal No. 3 incorporated SM bars with 180° hooks embedded in the deck slab. The following section explains the details of GFRP bars for each proposed barrier-deck system. It should be noted that Ryerson University research team has built and tested to collapse full-scale barrier segments of 1 m length to determine their anchorage capacity with the deck slab cantilever. Based on the results from these tests and cost estimates of GFRP material content, other crash test will be conducted on either proposal No. 2 or 3 to qualify it for use in Canada’s bridges.

6

a) Proposal No. 1 (b) Proposal No. 2

c) Proposal No. 3

Fig. 4 Proposed Ryerson University Barrier Details

7

4. DESCRIPTION OF THE EXPERIMENTAL PROGRAM ON PL-3 BARRIER MODELS

Five full-scale PL-3 barrier models were erected and tested to-collapse to determine their ultimate load carrying capacities and failure modes. The length of the barriers was taken 900 mm for barrier models 1 and 2 and 1000 mm for barrier models of 3, 4 and 5 to investigate the barrier-deck anchorage strength. Concrete cover to front and back GFRP bars in the barrier walls and top layer of the bars in the deck slab was taken as 50 mm. The PL-3 barrier walls had a thickness of 225 mm from top surface, which is proportionally increased to 305mm at a depth of 800 mm and tapered to 475 mm at depth of 1025 to the deck slab-barrier joint. The concrete deck slab had an average thickness of 300 mm under the barrier wall which is reduced to 250 mm in the slab portion, and reinforced with M20 Steel bars at the top reinforcement layer with 100 mm spacing and M15 Steel bars at 300 mm spacing at the bottom reinforcement layer in the direction normal and parallel to the wall, respectively. In barrier models 1 to 4, sand-coated GFRP bars were used as vertical and horizontal reinforcement in front and back face of the barrier, while the fifth barrier was constructed using reinforcing steel bars only as a reference specimen. The characteristic strength of the concrete forming the barrier models was obtained from core sampling as 25.4 MPa. More details can be found in Appendix A. Model 1 was constructed using High-Modulus (HM) GFRP bars as shown in Fig. 4(a). 15M HM-GFRP bars with head end were placed at 300 mm c/c spacing in front face of the barrier at deck-barrier joint. In addition, 15M HM-GFRP bars were placed in the front face of the barrier wall extending into the deck slab at 300 mm spacing. 12M HM-GFRP bars were placed as horizontal reinforcement as well as vertical reinforcement in back face of the wall at 300 mm spacing. Figure 5 shows view of GFRP bar arrangement in Model 1.

Fig. 5 Views of GFRP bars in Model 1

8

Model 2 incorporated Standard Modulus (SM) GFRP bars in the barrier wall as depicted in Fig. 4(2). At the junction of barrier to the deck slab, M15 GFRP bent bars were placed at 200 mm spacing in the barrier front face. In addition, M15 GFRP bars were placed at 200 mm spacing in the top tapered portion of the front face of the barrier, with the bars extended into the deck slab as shown in Fig. 4(2). M12 GFRP bars at 200 mm spacing were used as horizontal bars as well as vertical bars in back face of the barrier wall. Figure 6 shows view of GFRP bars in Model 2 before concrete casting.

Fig. 6 Views of GFRP bars in Model 2 Similar to barrier model 2, barrier model 3 was built using SM GFRP bars but with 180˚ hook at the junction of deck slab to the barrier wall. The SM GFRP bars placed on other locations in this model were identical to model 3. Figure 7 shows view of the placement of GFRP of bars in the formwork. Model 4 was constructed similar to model 1 except that the HM GFRP vertical bars were placed at 150 mm center-to-center spacing representing the barrier wall at its ends or at the construction joint locations. However, the spacing for GFRP horizontal bars as well as vertical bars in the back face of the wall was kept as 300 mm. Figure 8 shows a view of formwork with GFRP installation prior to concrete casting. Model 5 was constructed as reference sample using conventional steel reinforcement, as per MTO standard drawings for reinforced steel bridge barriers. Figure 9 shows a photo of this barrier model with steel reinforcement in front and back face of the wall as vertical and horizontal reinforcements. The M15 steel bar with 180° hook was placed to reinforce the top tapered portion of the barrier front face, while M15 bent bars were placed on the bottom tapered portion, all placed at 200 mm center-to-center spacing. M12 steel bars were placed in the model as horizontal bars as well as vertical bars in back face of the wall at 300 mm spacing. Table 3 summarizes description of the tested barrier models.

9

Fig. 7 Views of barrier model 3

Fig. 8 View of barrier model 4

Fig. 9 View of barrier model 5

10

Table 3 Barrier Designations used in this study for each model

Model No. Description of Models

1 Specimen No. 1 with V-Rod-HM with headed-end bars at 300 mm spacing

2 Specimen No. 2 with V-Rod-SM with 90˚ bent bars at 200 mm spacing

3 Specimen No. 3 with V-Rod-SM with 180˚ hook bars at 200 mm spacing

4 Specimen No. 4 with V-Rod-HM with headed-end bars at 150 mm spacing

5 Specimen No. 5 with conventional steel reinforcement at 200 mm bar spacing

All barrier models were cast in the same day using the same concrete batch. The casting was scheduled so that the deck portion was casted first, followed by casting the wall portion of the barrier walls. At each time step during casting, concrete was vibrated uniformly. To prevent the presence of void at deck-wall junction or wall tapered portion joint, the wall and the deck at these locations were hampered for the ease of movement of concrete in the joints. Figure 10 illustrates images of barrier wall during casting.

a) Deck casting b) Deck after casting

Fig. 10 Photo of casting barrier models It should be noted that the cast barrier models have the barrier wall connected to projecting cantilever deck slab to allow for the rotation at the barrier-deck junction. The cantilever deck-slab was fixed to an end concrete block that was anchored to the laboratory concrete floor using

11

tie-down anchoring system. Figure 11 shows a schematic diagram of the test setup and the location of LVDT’s. Each barrier specimen was supported over the structures laboratory floor, then, tied down to the floor using 50-mm diameter threaded rods. Each rod was placed @ 600 mm c-c and tightened by applying a specified torque to control the slab uplift during testing. A 900- kN hydraulic jack was used to apply horizontal load to the barrier wall. A universal flat load cell of 900-kN capacity was used to measure the applied loads on barrier models. SYSTEM 6000 data acquisition unit was used to record readings from all sensors.

a) Schematics diagram of the test set-up b) View of the test set-up for model No. 1

c) Schemactic diamgram showing LVDT locations

Fig. 11 Test setup and sensor locations

12

Each specimen was tested under increasing monotonic load up-to-collapse. During the test, jacking load was applied in increments of 10 kN. At each load increment, the load was maintained for a few minutes to observe crack initiation and propagation as well as change in barrier geometry as depicted from LVDT readings. Failure of the model was attained when the readings from sensors were increasing while the model did not take any further increase in load. Figure 11-c shows LVDT’s locations which were (i) at the top of the back side of the barrier wall oriented in the direction of the applied load; (ii) at the bottom of the deck slab to measure slab movement under transverse load; (iii) at the top of the back side of the deck slab to measure any possible uplift under load; and (iv) at back side of the wall to capture the possible vertical displacement of the barrier wall.

5. EXPERIMENTAL RESULTS FOR THE TESTED PL-3 BARRIER MODELS Figure 12 depicts view of the test setup for the tested barrier walls. Barrier model 1 with HM- GFRP bars of headed end spaced at 300 mm c/c was tested to-complete-collapse. It should be noted that the length of this barrier was 900 mm in longitudinal direction. The load was applied at 990 mm above the deck slab per CHBDC for PL-3 barriers tested under static test. The load was applied manually using hydraulic jack and load increments were captured by load cell attached to the system. The load was applied in increments of 10 kN to observe the crack initiation in the barrier wall and the slab. The first visible flexural crack appeared at the junction of barrier wall-to-deck slab at load of 20 kN as depicted in Figs. 13 and 14. With load increase, cracks were further developed at the corner of deck-to-slab junction down into the deck slab. Similarly, cracks were initiated in the deck slab at 25 to 30 kN due to combined flexure and tension. At a load of 40 kN, flexural crack was observed at the junction of the upper and lower portions of the tapered wall. At this loading stage, cracks started to propagate horizontally through the thickness at the intersection of deck-to-barrier in addition to the cracks observed earlier in the slab. The cracks propagated significantly till the barrier reached the failure load of 95.49 kN. The barrier primary failure took place at the corner of deck-to-slab interface due to flexural mode of failure. As such, the ultimate moment resisted per meter length of the barrier wall was 95.49 kN x0.99 m height /0.9 m width = 105.04 kN.m/m which is greater than the CHBDC factored design moment of 83 kN.m/m shown in Table 2 for interior locations of the barrier wall. At failure stage, the average ultimate lateral deflection of barrier wall was recorded as 41.61 mm, while the deck slab showed an average horizontal movement of 1.63 mm. At failure, an average deck uplift of 2.87 mm was observed. Figure 15 shows load-deformation curves recorded for model 1.

13

Fig. 12 View of the barrier test setup for model No. 1

a) General view of cracks in the deck and barrier wall

14

b) Close-up view of crack pattern at barrier-deck junction

Fig. 13 View of the crack pattern in barrier model 1 with HM-GFRP bars at 300-mm spacing (right side)

a) View of cracks in the barrier and deck from the left side

15

b) Close-up view of crack pattern at the barrier-deck junction

Fig. 14 View of the crack pattern in barrier model 1 with HM-GFRP bars at 300-mm spacing (left side)

Fig. 15 Load-deflection curves of the barrier wall, deck movement and deck uplift of barrier model 1

16

In barrier model 2, Standard Modulus GFRP bent bars were placed vertically in the front face of the barrier wall at 200 mm spacing. The barrier had a length of 1000 mm in the longitudinal direction. The load was applied at 990 mm above the deck slab similar to barrier model 1. Figure 16 shows view of the crack pattern at failure. In this barrier model, the first flexural crack was observed at the fixed end of the deck slab at load of 40 kN. By increasing the load to 50 kN, cracks were further developed in the deck slab and vertical cracks appeared on each side of the deck slab due to combined tension and flexure loading. At 60 kN load, horizontal flexural crack was observed at the deck-to-barrier junction. A similar flexural crack was initiated at junction of the upper and lower portion of the tapered wall at load increment of 65 kN which was further developed into the wall thickness at higher load. At the load range of 70 to 100 kN, cracks propagated further into the slab and barrier thicknesses. At 100 kN load, a flexural crack was observed in the top tapered portion of the barrier wall. With increase in applied load, cracks propagated further into the deck thickness at the barrier-deck junction that led to failure at 116.32 kN primarily due to flexural crack barrier-deck region. As such, the maximum moment reached at failure of deck-to-barrier junction was 116.32 kN x 0.99 m = 115.2 kN.m/m. This ultimate load carrying capacity was greater than the factored design moment of 83 kN.m/m specified in CHBDC at interior location as shown in Table 2. The barrier wall exhibited an average lateral deformation of 46.43 mm with deck horizontal movement of 4.37 mm and deck uplift of 5.3 mm at failure. The load-deformation curves captured by LVDT’s are displayed in Fig. 17.

a) View of cracks in the barrier and deck from right side

17

b) Close-up view of crack pattern at barrier-deck junction from the right side

c) View of cracks in the deck and barrier wall from the left side

18

d) Close-up view of crack pattern at the barrier-deck junction from the left side

Fig. 16 View of the crack pattern in barrier model 2 with SM-GFRP bent bars at 200 mm spacing

Fig. 17 Load-deflection curve of the barrier wall, deck movement and deck uplift of barrier model 2

19

Barrier model 3 was constructed using Standard Modulus GFRP bars by placing GFRP bars with 180˚ hooks at 200 mm spacing as vertical bars at front face of the barrier wall. It is worth to mention the barrier has a length of 1000 mm in the longitudinal direction. Similar to barrier model 1 and 2, the load was applied at 990 mm above the deck. Figure 18 shows view of the crack pattern at failure of model 3. The first visible flexural crack was observed at deck-barrier junction at an applied load of 30 kN. At this load, other flexural crack was also observed at junction of the top and bottom tapered portions of the barrier wall as well as in the deck slab. With increase in applied load to 45 kN, cracks started to propagate into the barrier-deck corner down into the slab thickness. Also, vertical cracks were detected in the deck slab when the load increased from 50 to 70 kN due combined tension and flexural loading on the deck slab. At this loading stage, cracks were further developed in corner junction of barrier-to-deck as well as through thickness in both the barrier wall and the deck slab. At 100 kN, a second flexural crack was observed in the top tapered portion of the barrier wall, extending significantly through thickness at the same load increment. Moreover, several flexural cracks were formed on the top surface of the deck slab at load ranging from 50 to 100 kN. Consequently, the barrier failed at load of 107.13 KN at the corner of barrier-to-deck junction due to combined tension and flexure in the deck slab. The resulting maximum moment at the barrier-deck junction was calculated as 107.13 kN x .99 m = 106.06 kN.m/m which is far greater that the CHBDC factored design moment of 83 kN.m/m for barrier tested at interior location. It should be noted that at failure, the barrier had an average deflection of 36.26 mm, with deck slab movement of 1.57 mm in horizontal direction and deck slab uplift of 4.24 mm. Figure 19 shows load-deformation curves of the barrier and the deck slab under load.

a) View of cracks in the deck and barrier wall from the right side

20

b) Close-up view of the crack pattern at the barrier-deck junction from the right side

c) View of cracks at the deck and barrier wall from the left side

21

d) Close-up view of crack pattern at the barrier-deck junction from the left side

Fig. 18 View of the crack pattern in model 3 with SM-GFRP bars with 180° hooks at 200 mm spacing

Fig. 19 Load-deformation curve of the barrier wall, deck movement and deck uplift of barrier model 3

22

In barrier model 4, the vertical High-Modulus GFRP bars were placed at 150 mm spacing in the front face of the barrier wall. Similar to other barriers, the load was applied 990 mm above the deck. Figure 20 shows view of the crack pattern of model 4 at failure. The first visible flexural crack was observed at the barrier-deck joint at a load of 25 kN. Further cracks due to combined flexural and tension appeared in the deck slab with further increase in applied load. At 60 kN, flexural crack was observed at the junction of the top and bottom tapered portion of the barrier wall. At 70 kN load, a similar flexural crack was observed right above the junction between the top and bottom tapered portion of the wall and propagated significantly through the thickness at the same load increment. At this load increment, cracks were also propagated through thickness at the corner of the barrier-to-deck junction. At 95 kN load, a second flexural crack was observed in the top tapered portion of the wall barrier-deck corner, leading to barrier failure at 153.3 kN. The failure of the barrier occurred at lower portion of the deck slab due to concrete splitting at the compression side of the slab under the barrier wall. The experimental results led to a maximum applied moment at the barrier-deck junction of 153.32 kN x 0.99 m / 0.9 m = 168.6 kN.m/m which is far greater that the CHBDC factored design moment of 102 kN.m/m at barrier end location as shown in Table 2. It should be noted that the average lateral deformation of the barrier wall was 50.13 mm, with an average deck horizontal movement of 5.53 mm. and an average deck uplift of 2.47 mm at failure. Figure 21 shows load-deformation curves of model 4.

a) View of crack in the barrier and deck from right side

23

b) Close-up view of crack pattern at barrier-deck junction from right side

c) View of cracks at the barrier and deck from left side

24

d) Close-up view of crack pattern at barrier-deck junction from left side

Fig. 20 View of the crack pattern in barrier model 4 with HM-GFRP bars at 150 mm spacing

Fig. 21 Load-deformation curves of the barrier wall, deck movement and deck uplift of model 4

25

Barrier model 5 was constructed as reference model to compare the results with barrier models reinforced with GFRP bars. In this model, M15 steel bars were placed vertically in front face of the barrier wall at 200 mm spacing. Similar to GFRP-reinforced models, the load was applied at 990 mm above the deck slab. Figure 22 shows view of the crack pattern observed in barrier model 5 at failure. At load step of 20kN, the first visible flexural crack was observed at the barrier-to-deck junction, which was extended down into the slab at the corner of deck-barrier joint. With load increase to 50 kN, few flexural cracks were observed in the deck slab extending through the deck thickness. In addition, at load step of 50 kN, a flexural crack was developed at junction of the top and bottom tapered portion of the wall, which was extended into the wall thickness with load increase. With increase in load to 85 kN, few cracks were developed horizontally at the corner of the deck-barrier wall. With increase in the load to 105 kN, other flexural crack was observed in top portion of the tapered wall and propagated through the wall thickness with load increase. In addition, more flexural cracks appeared at the top surface of the deck slab. With further increase in load, cracks widened, leading to barrier failure at 128.92 kN at corner junction of deck-to-barrier wall under combined tension and flexure. Results led to a maximum moment at the barrier-deck junction of 128.92 kN x 0.99 m = 127.63 kN.m/m which is far greater that the CHBDC factored design moment of 83 kN.m/m at interior location. The barrier wall displayed an average maximum lateral deformation of 32.81 mm, with deck slab horizontal movement of 5.03 mm and deck uplift of 2.67 mm at failure. Figure 23 depicts the load-deformation curved for of barrier model 5. Table 4 summarizes the experimental findings for the tested models.

a) View of cracks in the barrier and deck slab from the right side

26

b) View of crack pattern at the barrier-deck junction from the left side

c) View of cracks in the deck and barrier from left side

27

d) Close-up view of crack pattern at the barrier-deck junction from the left

Fig. 22 View of the crack pattern in barrier model 5 with steel bars at 200 mm spacing

Fig. 23 Load-deformation curve of the barrier wall, deck movement and deck uplift of barrier model 5

28

Table 4 Experimental findings of the tested models

Model 1 Model 2 Model 3 Model 4 Model 5

Barrier configuration

Proposal No. 1 V-Rod-HM With headed end, 300 mm bar spacing

Proposal No. 2 V-Rod-SM With bents, 200 mm bar spacing

Proposal No. 3 V-Rod-SM With 180° hooks, 200 mm bar spacing

Proposal No. 1 V-Rod-HM With headed end, 150 mm bar spacing

Steel-reinforced barrier, 20 mm bar spacing

Failure load (kN/m) 106.1 116.3 107.1 170.3 128.9

Height of load application (m) 0.99 0.99 0.99 0.99 0.99

Moment at barrier-deck joint -Mexp (kN.m/m)

105.0 115.1 106.06 168.6 127.3

CHBDC Moments -

Mesign (kN.m/m)

Interior Location 83 83 83 - 83

Exterior Location

- - - 102 -

Safety factor = Mexp / Mdesign 1.27 1.39 1.28 1.65 1.53 Barrier wall total lateral

deflection (mm) 41.61 46.43 36.26 50.13 32.81

Deck movement (mm) 1.63 4.37 1.57 5.53 5.03 Net deflection of barrier wall

(mm) 39.98 42.06 35.00 44.60 27.78

Deck uplift (mm) 2.87 5.30 4.24 2.47 2.67

6. CONCLUSIONS Based on the data generated from the experimental study, it can be concluded that the proposed barrier reinforcement details incorporating GFRP bars with headed anchorage, GFRP bents and GFRP bars with 180˚ hook FOR PL-3 barrier configuration can be safely used in bridge barrier walls to resist the applied vehicle impact load specified in the Canadian Highway Bridge Design Code at the barrier wall-deck slab anchorage. Experimental findings showed that the proposed GFRP-reinforced barriers are as good as reinforced steel barrier with respect to strength at the barrier-deck junction. In addition, the maximum barrier deflection exhibited in the proposed GFRP-reinforced barriers is more than that observed for steel-reinforced barrier as attributed to the low flexural modulus of elasticity of the GFRP bars. Such increase in deflection would have a favorable effect in absorbing energy resulting from vehicle impact.

29

7. REFERENCES AASHTO. 1977. AASHTO Guide for Selecting, Locating, and Designing Traffic Barriers.

American Association of State Highway and Transportation Officials, Washington D.C., USA.

AASHTO. 1989. AASHTO Guide Specifications for Bridge Railings. American Association of State Highway and Transportation Officials, Washington D.C., USA.

AASHTO. 2010. AASHTO-LRFD Bridge Design Specifications. Third Edition, American Association of State Highway and Transportation Officials, Washington, DC.

CSA. 2006a. Canadian Highway Bridge Design Code. CAN/CSA-S6-06. Canadian Standard Association, Toronto, Ontario, Canada.

CSA. 2006b. Commentary on CAN/CSA-S6-06, Canadian Highway Bridge Design Code. Canadian Standard Association, Toronto, Ontario, Canada.

MacGregor, J., and Bartlett, M. 2000. Reinforced Concrete: Mechanics and Design. Prentice Hall Canada.

RCCAO. 2007. Ontario’s Bridges Bridging the Gap. Report prepared by MMM Group for Residential and Construction Alliance of Ontario, RCCAO, 56 pages.

Pultrall Inc., Canada. 2012. www.pultrall.com. Sennah, K., Tropynina, E., Goremykin, S., Lam, M., and Lucic, S. 2010. Concrete Bridge

Barriers Reinforced with GFRP Bars with Headed Ends. Proceedings of the 8th International Conference on Short and Medium Span Bridges, Niagara Falls, Ontario, pp. 1-10.

Sennah, K., Juette, B., Weber, A., and Witt, C. 2011. Vehicle Crash Testing of a GFRP-Reinforced PL-3 Concrete Bridge Barrier. Proceedings of the 4th International Conference on Durability & Sustainability of Fibre Reinforced Polymer Composites for Construction and Rehabilitation, Quebec City, Quebec, pp. 417-424.

Sennah, K and Kederzadeh, H. 2012. Vehicle Crash Testing of a PL-3 Concrete Bridge Barrier Reinforced with GFRP Bars with Sand-Coated Surface and Headed Ends. Technical Report submitted to Pultrall Inc. and Ontario Ministry of Transportation, 86 pages.

30

APPENDIX A

RESULTS FROM CONCRETE CORE SAMPLING

Table 3 Concrete core sample detailing of the tested barrier walls

Sample No.

Ultimate Load (kN)

fcoreX

(MPa)

No. of bars

inside

Remarks (type of

bars)

Correction Factors** feq*

Fl/d Fdia Fr Fmc Fd

1 263.86 33.6 - No bars 0.87 1 1 1.08 1 31.57 2 189.28 24.1 - No bars 0.87 1 1 1.08 1 22.64 3 248.74 31.67 - No bars 0.87 1 1 1.08 1 29.76 4 212.51 27.05 - No bars 0.87 1 1 1.08 1 25.41 5 229.73 29.25 - No bars 0.87 1 1 1.08 1 27.48 6 209.71 26.7 - No bars 0.87 1 1 1.08 1 25.09 7 282.12 35.92 - No bars 0.87 1 1 1.08 1 33.75 8 304.39 38.75 - No bars 0.87 1 1 1.08 1 36.41 9 212.45 27.05 2 Steel bars 0.87 1 1.13 1.08 1 28.72 10 180.95 23.4 1 Steel bars 0.87 1 1.08 1.08 1 23.74 11 222.27 28.3 1 Steel bars 0.87 1 1.08 1.08 1 28.72 12 186.55 23.75 1 Steel bars 0.87 1 1.08 1.08 1 24.10 13 230.12 29.3 1 Steel bars 0.87 1 1.08 1.08 1 29.73 14 212.47 27.05 1 FRP bars 0.87 1 1.08 1.08 1 27.45 15 221.6 28.2 1 FRP bars 0.87 1 1.08 1.08 1 28.62 16 208.2 26.51 1 FRP bars 0.87 1 1.08 1.08 1 26.90 17 236.94 30.17 2 FRP bars 0.87 1 1.13 1.08 1 32.03 18 240.39 30.6 2 FRP bars 0.87 1 1.13 1.08 1 32.48 19 229.44 29.2 1 FRP bars 0.87 1 1.08 1.08 1 29.63 20 303.39 38.63 2 FRP bars 0.87 1 1.13 1.08 1 41.02 21 339.14 43.18 2 FRP bars 0.87 1 1.13 1.08 1 45.84 22 233.02 29.67 2 FRP bars 0.87 1 1.13 1.08 1 31.50 23 218.27 27.79 1 FRP bars 0.87 1 1.08 1.08 1 28.20 24 300.58 38.2 1 FRP bars 0.87 1 1.08 1.08 1 38.76 25 231.7 29.5 1 FRP bars 0.87 1 1.08 1.08 1 29.94

* feq = fcore(Fl/d. Fdia. Fr. Fmc. Fd) = equivalent compressive strength of tested core sample as influenced by strength correction factors (MacGregor and Bartlett, 2000) ** Fl/d = Correction factor for length/diameter ratio Fdia = Correction factor for diameter of core Fr = Correction factor for the presence of reinforcement Fmc = Correction factor to account for the effect of moisture content Fd = Correction factor for the effect of the damage of the core surface during drilling X fcore = Compressive strength of tested core sample

31

Table 4 Calculated characteristic compressive strength of the barrier walls

n Kc f̅eq

(MPa) Standard

Deviation (SD) Coefficient of Variation (V)

f'c*** (MPa)

All samples included 25 1.02 30.38 5.575 0.1881 25.43

*** f'c = 0.9 f̅eq{1- 1.28 } = characteristic strength as affected by the number of samples tested

and the change of the strength value of each sample as compared to the average value (CSA, 2006b).

Fig. A1 Photo of core sampling machine

32

Fig. A2 Core sampling of the tested barrier wall

Fig. A3 Image of a core sample taken from the barriers

33

Fig. A4 Image of core samples after cutting and grinding

Fig. A5 Concrete cap grinding machine

34

Fig. A6 Concrete compressive strength testing machine during testing

Fig. A7 Close-up view of tested concrete core showing presence of steel bar in the core sample