10th International Conference on Short and Medium Span Bridges

Quebec City, Quebec, Canada,

July 31 – August 3, 2018

REHABILITATION OF ONE-WAY SLABS USING PRE-STRESSED CONCRETE PRISMS

Saeed Gerami1 and Dagmar Svecova2,3

1 M.Sc. Student, University of Manitoba, Winnipeg, MB, Canada 2 Professor and Head, Civil Engineering, University of ManitobaWinnipeg, MB, Canada 3 corresponding_author_ Dagmar.Svecova@umanitoba.ca

Abstract: During its service life, the structural elements in bridges are exposed to wear and tear from vehicular loads, overloading and exposure to harsh environmental conditions that slowly deteriorate the structure. Typically, deficiencies in concrete structures are revealed as corrosion of steel reinforcement, cracking and spalling of concrete and excessive deflection. Rehabilitation of bridges is a complex process that includes many considerations. The techniques for rehabilitation of reinforced concrete structural members and bridge decks are very similar. The most practical techniques are concrete jacketing, steel plate bonding, Fiber Reinforced Polymers (FRP) bonding, external post-tensioning and the use of near surface mounted (NSM) reinforcement. In this study, prestressed concrete prisms (PCPs) were utilized as an NSM reinforcement. The PCPs were cast with high strength concrete and had cross-sectional area of 50 mm x 50 mm and/or 35 mm x 35 mm. They were concentrically prestressed by 9 mm diameter seven-wire strands. Six one-way slabs with dimensions of 3000 mm in length, 600 mm width and depth of 200 mm were cast, then cracked and subsequently rehabilitated with PCPs. Deflected profiles and crack widths of the slabs before and after rehabilitation were compared in this research. Utilization of PCPs for rehabilitation diminished the magnitude of deflection and crack width at service load. Slabs strengthened by prisms with higher prestressing level and larger cross section demonstrated slightly smaller deflection and narrower crack widths in comparison with those reinforced with prisms of smaller cross section and prestressing level.

1 INTRODUCTION

Cracking is an inevitable phenomenon in concrete and happens as a result of low tensile strength and extensibility. Generally, reinforced concrete structures experience few cracks under service load except for cracking due to shrinkage and temperature changes (Piyasena 2002). When this occurs, steel reinforcement inside concrete members is at risk of corrosion, and the concrete can become delaminated at the location of the reinforcement. In addition, flexural stiffness of reinforced concrete members reduces which may lead to increased deflection. Limited cracking under service load is normal in reinforced concrete. It is not always applicable or economical to substitute a current dysfunctional structure with a new one. In their research, Klaiber et al. (1987) classified various strengthening procedures, such as

62-1

external pre-stressing, injection methods, shotcrete, hand-applied repairs with concrete mortar, and the application of different concrete castings. The installation of steel plates to concrete using epoxy has been a common repair method for the last 30 years, despite the advancements in technology (Beber et al. 2001). In this technique, a steel plate with a specified thickness is bonded with adhesives and anchored to the existing concrete member to improve its stiffness and strength. Steel plates enhance the moment of inertia of any given cross section however, because they are externally bonded reinforcement, they require protection against environmental factors. The use of steel plates increases the self-weight of the structures. Therefore, lightweight materials, such as fiber-reinforced polymers (FRP) are often used. FRPs are bonded to the external surface of the deficient concrete member in the direction of tensile force or as restraining reinforcement perpendicular to the axis of the member. In other research, concrete jacketing was used. Concrete jacketing includes increasing of existing reinforced concrete cross section by providing more steel reinforcement and concrete. Some bridges can be strengthened using external post-tensioning to offset pre-stress loss or to increase their live load capacity. In this strengthening technique, pre-stressing tendons are positioned on the external surface of the concrete member, anchored at the ends and post-tensioning force is applied using a hydraulic jack. This method is quite effective however, it requires adequate reserved strength in current concrete member to transfer prestressing force. In addition, pre-stressed tendons and anchorages need proper protection system against corrosion.

Pre-stressed concrete prisms (PCPs) are high strength concrete rectangular members concentrically pre-tensioned by steel or FRP tendons. To prevent any camber or cracking under pre-stressing load, the tendons are to be located at the center of the prisms to minimize the effects of eccentricity. On the other hand, high-strength concrete is used to provide sufficient tensile capacity against high forces at the time of pre-stressing release. In the current study, PCPs are embedded in existing concrete to protect against thermal, environmental and mechanical damage. Other advantages of NSM method include improved durability and more optimized stress-sharing mechanism. Grooves are sawed in concrete slabs at a depth that would inhibit any damage to the main internal flexural reinforcements. The application of PCPs in the NSM method is going to decrease final deflection at structural failure, reduce crack width and spacing under service load, and demonstrate more ductile behavior.

2 EXPERIMENTAL PROGRAM

2.1 Test SpecimensSix one-way slabs with length of 3000 mm, 600 mm width and 200 mm height were cast and tested to failure at the University of Manitoba’s structures laboratory. All six slabs were cast using ready-mixed, normal-strength concrete with an average 28 days compressive strength of 40 MPa. Standard cylinder compression tests were performed according to ASTM C39/C39M-14a (ASTM 2014) on cylinders at 3, 7, 14 and 28 days after casting to track the progression of strength gain of the concrete. The average compressive strength varied between 41 and 45 MPa on the day of flexural testing, as shown in Table 1. The first two characters in the slab label identify the dimension of the PCPs installed in the slab i.e. “SP” for small prisms (35 x 35 mm) and “LP” for large prisms (50 x 50 mm). The third character “S” stands for slab. Finally, the last two digits indicate level of effective pre-stress force in the prisms in kN. Grade 400 deformed steel bars with a modulus of elasticity of 200 GPa were used as longitudinal and transverse reinforcements for all six slabs.

62-2

Table 1: Reinforcement details of the slabs before and after rehabilitation

Specimen Concrete Strength ( f ' c ) (MPa)

Longitudinal Reinforcemen

t

AS(mm2)

ρ1

(% )ρ2

(% )SP-S-20 42

4 No 15 800 1.01

1.28SP-S-25 44SP-S-30 45LP-S-20 41LP-S-25 42LP-S-30 43

ρ1 (% ) : Reinforcement ratio before rehabilitationρ2 (% ) : Reinforcement ratio after rehabilitation

A total of 48 PCPs were cast. The PCPs had a cross section of 35 x 35 mm or 50 x 50 mm, and were 2000 mm long. Steel angles were assembled concentrically around the pre-stressed steel strand to create side of the formwork until the time of release to prevent any probable camber in prisms. To minimize the probability of buckling in prisms, the load was released gradually by loosening the nuts and bolts attached between the steel chucks and bulkheads, not by cutting the steel strands, to prevent the application of a sudden high force on the PCPs. High strength, 100 MPa concrete was used to cast these elements, and single 9 mm diameter seven wire grade 1860 MPa steel strands were used for pre-stressing. Three levels of effective pre-stressing were achieved in these prisms: 20 kN, 25 kN and 30 kN. Three of the one-way slabs were strengthened with 35 x 35 mm PCPs, while three others were strengthened with 50 x 50 mm PCPs. In bridge deck, a saw cut method would be used to create grooves for installation of PCPs. Saw cutting is a hard and time consuming method especially when it requires to be performed overhead. Therefore, to expedite and facilitate research project, styrofoam was used before casting inside formwork to keep the location where PCPs were to be inserted hollow. Styrofoam was removed after casting and PCPs were installed inside grooves, using adhesive. The grooves were sand-blasted prior to the application of the prisms to provide a rougher surface area and a better bond mechanism between the concrete and adhesive. The prisms were installed in overhead position. All slabs were tested to failure seven days after the installation of the PCPs to allow sufficient time for the adhesive to be cured. Figure 1 shows the dimension, reinforcement arrangement and internal instrumentation of the tested decks. Table 1 provides the reinforcement details of the samples before and after rehabilitation.

Figure 1: Deck details reinforcement

62-3

2.2 Instrumentation and Testing ProcedureDuring the testing process, the applied load was measured using a load cell attached to an actuator. A total of four Linear Variable Displacement Transducers (LVDTs) were used to measure the deflection of decks at different sections. PI gauges were used to monitor the concrete strain and measure crack-width throughout the test. For each deck, a total of six strain gauges with gauge length of 6 mm were mounted on the longitudinal steel reinforcements. Strain gauges were used to measure strain due to applied load and will be used to establish the moment-curvature relationship for each deck. One electrical strain gauge (ESG) was also installed on the concrete surface at mid span, in the compression zone. In addition, there was one ESG with size of 2 mm -120 Ω located at the mid span of the seven-wire steel tendon in each PCP.

A line load was applied at the mid span of the simply supported deck. The load from the actuator was transmitted to a steel spreader beam to facilitate uniform loading distribution across the width of the slab. The load was applied monotonically, at a rate of 0.5 mm/min. A Data Acquisition System (DAQ) was used to read the data from all strain gauges, PI-gauges, LVDTs and the MTS machine. Figure 2 illustrates experimental set up and detailed location of all sensors attached to the specimens.

3 TEST RESULTS

3.1 General Behaviour and Mode of FailureTable 2 provides experimental ultimate loads for all slabs along with their modes of failures and strain in concrete and longitudinal reinforcements. All six slabs were designed to fail in flexure but experienced shear failure. This is believed to be caused by the increase in the flexural capacity of the slabs due to installation of PCPs. Therefore, concrete compressive strain at failure for all six slabs was below the ultimate strain of 0.0035 defined by CSA A23.3-14 (2014). Slabs were analyzed based on strain compatibility using CSA A23.3-14 (2014) and the actual mechanical properties of the materials. It was assumed that plane sections remain plane in bending, so theoretically all the strain at different heights must lie along a straight line. Moment-curvature diagrams were constructed using the strain profile at each section. A trend line was fitted to the strain values at different heights. The slope of this trend line shows the value of the curvature at that section at load level being considered. Finally based on the theoretical moment curvature diagrams, theoretical failure loads were calculated.

Figure 3 shows similar cracking behavior in terms of initiation and propagation until failure in all tested slabs. Cracking started with the development of a few very fine vertical flexural cracks at the mid span, followed by bond crack along the steel reinforcement close to the supports. Early hairline cracks in the rehabilitated slab initiated at approximately 31, 37, 36, 33, 39 and 41 percent of the failure load for slabs SP-S-20, SP-S-25, SP-S-30, LP-S-20, LP-S-25 and LP-S-30, respectively. Subsequently, several inclined cracks widened in the area close to the section where PCPs were introduced in the cross section. As the cracks stabilized, one of the inclined cracks widened and extended towards the top of the slab, ultimately causing failure of the slab. Except in slabs LP-S-30 and LP-S-25, the steel reinforcement in four remaining slabs yielded prior to the failure. Nearly 85% of the tensile capacity of steel strands inside the PCPs were utilized before failure took place. All slabs failed in sections where PCP embedment started, close to the supports.

62-4

Figure 2: Experimental setup and detailed location of all sensors attached to the specimens (all dimensions in mm)

Table 2: Experimental failure load, strains at failure and modes of failure

SlabFailure Load

(kN) PExp. / PCal.Strain at Failure (10−6)

Mode of FailureExp. Cal. Concrete Steel

SP-S-20 128.0 152.5 0.84 2010 11092

Shear Failure

SP-S-25 133.1 156.0 0.85 2907 8491SP-S-30 137.3 156.2 0.87 2935 4103LP-S-20 120.5 149.1 0.80 2299 3596LP-S-25 126.3 150.5 0.84 2382 2279LP-S-30 132.3 151.2 0.87 2620 2144

3.2 Strain in Reinforcement and ConcreteFigure 4 illustrates the strain in longitudinal steel reinforcement under the line load and concrete strain at the extreme compression fiber for all slabs at mid span. It is indicated that strain in all slabs changed linearly with load up to the cracking loads of the PCPs. The PCP cracking load varied from 50 kN to 65 kN for different levels of pre-stressing. When the prisms cracked, strain in flexural reinforcement experienced a rapid increase with load up to levels close to yielding load. When mild steel yielded at load levels from 100 kN to 139 kN, the strain in mild steel bars changed considerably with a very small change in load and a plateau was formed for most of the slabs. In slabs SP-S-20 and SP-S-25 strains reached strain hardening phase before failure.

Figure 4(b) clearly shows that concrete strain in slabs prior to the cracking of PCPs (50 to 65 kN) is linear and increases slightly with load. Following the appearance of the tension cracks in the slabs after PCP cracking at load levels from 60 to 139 kN, compressive strain of concrete changed with load more considerably followed by yielding of steel. The smaller values of strain for slabs strengthened with larger prisms (50 x 50 mm PCPs) compared with ones reinforced with smaller prisms (35 x 35 mm PCPs) indicate higher stiffness of the slabs with larger prisms. At failure load, the ultimate compressive strain in concrete varied from 2000 to 3000 micro-strain in all slabs.

62-5

Figure 3: Cracking patterns of tested slabs at failure

62-6

PCP

PCP

PCP

PCP

PCP

PCP

(a) (b)Figure 4: Strain at mid span of the slab: (a) Tension steel reinforcement, (b) Maximum concrete

compression strain

Strain in mild steel before and after rehabilitation is plotted in Figure 5 for slabs reinforced with large and small PCPs respectively. It is evident for all slabs that strain values reduced due to presence of PCPs. Lower strain will lead to smaller deflections and crack widths. The installation of PCPs enabled the slabs to regain their stiffness to the levels before cracking of concrete. As can be seen in Figure 5, the stiffness only changed at cracking of PCPs, which varied from 50 kN to 65 kN depending on the pre-stressing load levels of the prisms.

(a) (b)

Figure 5: Strain in tension steel reinforcement: (a) In slabs LP and (b) In slabs SP

3.3 Load-Deflection ResponseFigure 6 provides load-deflection behavior for all slabs at mid span under the line load and location of maximum bending moment. The curve contains the two loading stages of the slabs: before the installation of the prisms and after the installation of the prisms. It can be seen that all slabs followed a general linear trend of load-deflection prior to cracking of concrete. After cracking of concrete and installation of PCPs, compared to the pre-cracking phase, the initial stiffness increased due to the increase in the area of the slabs and presence of prisms. Load-deflection behavior remained linear after cracking of concrete at 20 kN without significant change of slope until cracking of the prisms at load levels of 50 kN to 65 kN depending on the PCP pre-stressing levels.

62-7

PCPs Cracking

Steel Yielding

After PCPs Before PCPs After PCPs Before PCPs

Figure 6: Load-deflection at mid span under load point

All slabs showed a bilinear behavior with a change of stiffness at prism cracking. The non-linear behavior of load-deflection after cracking of PCPs is primarily associated with the presence of steel reinforcement and their yielding at load levels from 100 to 139 kN depending on the level of PCP pre-stressing. Deflection increased at a faster rate afterwards as a result of lower values of stiffness after prism cracking. The yielding of steel reinforcements plays a main role for the non-linearity of the load-deflection relationship after cracking of PCPs. The more ductile behavior of slabs SP-S-25 and SP-S-20 can be seen in the Figure 6 as steel reinforcement fully yielded in the two slabs.

3.4 Moment-Curvature RelationshipFigure 7 represents moment-curvature diagrams at mid span for all tested slabs until failure load. The moment-curvature for all slabs after installation of PCPs is trilinear with a change of slope at prism cracking and at yielding of steel reinforcement. After cracking of concrete and installation of PCPs, the stiffness in all slabs increased slightly due to the presence of prisms. Therefore, the moment of inertia remained close to that of an un-cracked slab. Nevertheless, after PCPs cracked, stiffness was considerably reduced and the effect of tension stiffening for high strength concrete diminished with increment in load. Furthermore, prior to failure, slabs LP-S-20 and SP-S-30 indicated moderate plateau while SP-S-25 and SP-S-20 experienced large plateau as a result of yielding and strain hardening phase in steel reinforcement before failure. On the other hand, slabs LP-S-30 and LP-S-25 due to the higher pre-stress load of PCPs showed stiffer behavior and lower curvature values which led them to smaller deflection and narrower crack widths overall.

Figure 7: Moment-curvature at mid span62-8

Steel Yielding

PCP Cracking

Service Load = 45 kN

3.5 Cracking BehaviourCrack spacing was measured using a tape measure with accuracy of ±1 mm in all tested slabs over the cracked zone. Table 3 provides the average crack spacing and numbers of formed cracks along the length of all slabs. At ultimate load, slabs LP-S-30 and SP-S-30 presented the least numbers of cracks after failure due to the higher pre-stressing force provided by the presence of the PCPs in the cross section.

Table 3: Average crack spacing and number of cracks in all slabs

Slab Number of Cracks Average crack spacing (mm)

SP-S-20 15 128SP-S-25 14 132SP-S-30 11 141LP-S-20 14 130LP-S-25 14 135LP-S-30 12 148

Generally, crack spacing is directly related to transfer length of the reinforcement while inversely proportional to the bond strength. In other words, longer crack spacing translates into longer transfer length and weaker bond mechanism between reinforcement surface and concrete. Based on the experimental values in Table 3, all slabs showed roughly similar crack spacing and a number of cracks which demonstrates the same transfer length and bond mechanism for all of them. However, slabs reinforced with PCPs with higher pre-stressing levels showed fewer number of cracks and longer average crack spacing. Higher pre-stressing force reduces number of cracks in slabs LP-S-30 and SP-S-30 compared to other tested specimens. Figure 8 illustrates development of maximum crack width at mid span with an increase in load for all six tested slabs. LP-S-30 had the narrowest crack prior to the cracking of PCPs at about 50% of ultimate load and afterwards. SP-S-20 showed the widest crack width among all other tested slabs equal to 2.3 mm.

Figure 8: Maximum crack width at mid-span

Crack width is in direct relation with strain in main longitudinal reinforcement. As it was discussed previously, strain in flexural reinforcement and curvature in slabs LP repaired with 50 mm prisms were smaller compared to the slabs SP with 35 mm prisms. Hence, narrower crack widths formed in slabs LP-

62-9

S-30 and LP-S-25 respectively before cracking of the PCPs. Table 4 compares the crack width of all six slabs at mid span before and after PCP installation. It can be seen that at cracking load level of 20 kN, cracks were wider before installation of PCPs, however after rehabilitation, at the same load level of 20 kN the cracks were narrower. In fact, after application of PCPs, the total slab stiffness increased and pre-stress force in the prisms provided sufficient restraint for the surrounding concrete in the slabs to arrest further crack initiation and propagation until higher load levels at which the prisms themselves cracked.

Table 4: Comparison of crack width at mid-span before and after installation of PCPs

Slab W cr 1 (mm)*

W cr 2 (mm)**

PW cr 1

(kN)PW cr 1

R

(kN)***

SP-S-20 0.21 0.12 20.14 50.01SP-S-25 0.27 0.19 20 44.34SP-S-30 0.14 0.07 21.9 57.9LP-S-20 N/A N/A N/A N/ALP-S-25 0.05 0.03 21.66 35.35LP-S-30 0.05 0.04 22.15 33.6

* W cr 1 (mm): Crack width before installation of prisms** W cr 2 (mm): Crack width after installation of prisms*** PW cr 1

R (kN): Load during failure phase when crack width was equal to W cr 1

Table 4 demonstrates that after installation of PCPs in the slabs, stiffness for all six slabs increased and cracks propagated in a slower rate compared to the cracking phase. For instance, crack width for slab SP-S-20 at cracking phase at load level of 20 kN was 0.21 mm while after installation of PCPs the same width of cracks occurred at load level of 50 kN. The same cracking behavior and pattern was observed for other slabs.

4 CONCLUSIONS

The following conclusions were drawn from the work presented in this paper:1. Higher pre-stressing levels in PCPs increase the prism cracking load and improve service behavior of slabs reinforced with these PCPs at service load.2. No slip was observed between seven wire steel strands in the prism and between the PCPs and slabs.3. The moment-curvature relationship for all slabs rehabilitated with PCPs showed a trilinear behavior. 4. After installation of PCPs, stiffness in all slabs increased slightly due to the presence of prisms. Following PCP cracking, the stiffness reduced considerably.5. The load-deflection behavior of all slabs was bilinear with a change of stiffness at prism cracking. Slabs strengthened with larger prisms with higher pre-stressing level experienced lower values of deflection.6. After prism installation, the growth of existing cracks was arrested. Concrete cracks were hairline before initiation of cracking in prisms and did not grow to pass the prisms level. The crack width decreased with the increase in the pre-stressing level in PCP. Throughout all test procedures, slabs reinforced with higher pre-stressing level and larger prisms had narrowest cracks. However, after cracking of prisms, cracks widened at a faster rate and propagated rapidly towards compression zone.

Acknowledgements

The authors would like to acknowledge assistance provided by W.R. McQuade Structures Laboratory technical staff at University of Manitoba. Funding for this work was provided by NSERC Discovery Grantand Manitoba Infrastructure.

62-10

References

ASTM 2014. Standard test method for compressive strength of cylindrical concrete specimens. ASTM C39/C39M-14a, West Conshohocken, PA.

Beber, A. J; Filho, A. C; and Campagnolo, J.L. 2001. CFRP in the strengthening of reinforced concrete beams. Proceedings of the International Conference on FRP Composites in Civil Engineering, Hong Kong, China,

391-398. CSA Standard A23.3-14. 2014. Design of Concrete Structures. Canadian Standards Association, Rexdale, Toronto,

Ontario, CanadaKlaiber, F.W; Dunker, K.F; Wipf, T.J and Sanders, W.W. 1987. Methods of strengthening existing highway bridges.

Technical Report, Transportation Research Board, NCHRP Research Report No. 293.Piyasena, R. 2002. Crack Spacing, Crack Width and Tension Stiffening Effect in reinforced Concrete Beams and One-

Way Slabs. Ph.D. Thesis, Griffith University, Gold Coast Campus, Australia.

62-11