This chapter starts the description of the primary research work conducted: the hybrid CFRP / Sprayed FRP system. The following two chapters will describe the corresponding theoretical analysis and design method for this system. Simply speaking, this strengthening system is comprised of CFRP laminates bonded to the soffit of beams and Sprayed FRP applied to both vertical sides of the beams. Although CFRP beam strengthening and Sprayed FRP beam strengthening have been studied separately by others, their combination into a hybrid system is a new approach. The experimental work to be detailed here focused only on the flexural strengthening of reinforced concrete (RC) beams. Both undamaged beams and damaged beams with simulated steel corrosion were investigated.
Preparation of Reinforced Concrete Beams Beam Formwork
In total, eight beam forms were made with the beam dimensions described in the following beam design section. 9.5 mm thick plywood and 50 mm x 100 mm wood studs that were purchased at a local Home Depot were used to assemble the forms with nails and screws. Figure 4-1 shows schematically the components and configuration of a
form. It should be noted the figure shows only one end panel, one side panel and the bottom board for clarity. The actual form looks like those shown in Figure 4-1.
Figure 4-1 Partial assembly of a beam form.
Beam Casting
All of the beams were cast at the Coastal Lab of the Department of Civil and
Coastal Engineering at the University of Florida. Since the concrete was ready-mixed, the delivery truck poured the concrete directly into the eight wooden forms built beforehand one by one. Before casting, the reinforcing steel cages had been set in place. Two hand- held vibrators were used to consolidate the poured concrete. Figure 4-2 and 4-3 show pictures of the wooden molds and the concrete truck taken on the day of casting.
Figure 4-2. Beam casting site layout.
Figure 4-3. Concrete truck delivering concrete to wooden forms.
Beam Testing Program Concrete Beam Design
To simulate real-life repair situations, large-scale beam testing was chosen over small-scale specimens. The beam size shown in Figure 4-4 was used due to the limited lab facilities (i.e. the beam size was determined by the biggest size that could be tested with available facilities). All of the beams were designed in accordance with the ACI design specifications (ACI, 2002b), with steel-tension yielding/concrete crushing as the target failure mode because only FRP flexural strengthening will be studied herein. The design load capacity of these beams was 222 kN under three-point loading. Figure 4-4 shows the reinforcement details.
Figure 4-4. Concrete beam details.
Strengthening Plan
In total, eight concrete beams were cast. The reinforcing steels of four beams were damaged to simulate corrosion before casting, which will be detailed later. Five of the beams were strengthened with the CFRP laminates and sprayed GFRP using the schemes summarized in Table 4-1, while the other three served as control specimens.
Table 4-1. Beam configuration.
Group A (No Damage)
Group B
(Simulated Corrosion Damage)
Beam Label Strengthening Schemec Beam Label Strengthening Schemec
A-1 None (Control) B-1a None (Control)
A-2 CFRP Laminate Only B-2a CFRP / Sprayed GFRP System
A-3 Sprayed GFRP Only B-3b None (Control)
A-4 CFRP / Sprayed GFRP System B-4b CFRP / Sprayed GFRP System Note: a – Two “corroded” reinforcement bars; b – Three “corroded” reinforcement bars;
c – Beam faces where FRP will be applied were sand-blasted before FRP application 2800mm
305mm 460mm
2 # 4
3 # 6 12 @ 7”
# 3 U, grade 60 stirrups
15 @ 178mm
Program Objectives
The entire arrangement was intended to investigate the effectiveness of the proposed CFRP/sprayed GFRP system for addressing the two problems often
encountered in real-life repair scenarios; namely strengthening deficiencies in capacity and repairing corrosion damage. The approach taken in this research was to subject the strengthened beams to three-point loading to capacity; and compare the results of the hybrid specimen with those of the control, CFRP-only and sprayed GFRP-only specimen.
Group A was designed to examine the performance of the CFRP/sprayed GFRP hybrid system in cases such as; satisfying new structural functional needs or modified design requirements where the load-carrying capacity of structural members needs to be increased. One unretrofitted beam acted as the control, while two other beams, one strengthened with the CFRP laminates alone and one strengthened with the sprayed GFRP alone, were fabricated to compare with the final beam strengthened with the
hybrid CFRP/sprayed GFRP system. The ACI FRP strengthening design code (ACI 2002) indicates that 5% to 40% is the reasonable range of increase in flexural strength provided by FRP when taking into account ductility and serviceability limits. With this statement in mind, the hybrid CFRP/sprayed GFRP system was designed in such a way as to improve the flexural strength of the beams by 20%, around the average of the recommended range. It should be noted that this number does not involve any consideration of premature failure, such as CFRP laminate debonding as would be anticipated without anchorage of the retrofitting material.
Concrete crushing was chosen as the target failure mode when determining the amount of CFRP laminates and sprayed GFRP to be used. Structural problems occurring in concrete beams are often associated with the corrosion of reinforcing steel, not the
concrete. Therefore, concrete crushing of a beam not only indicates the ultimate state of the member but complete usage of the concrete strength as well.
Group B was designed to investigate the ability of the CFRP/sprayed GFRP hybrid system to repair reinforcing steel corrosion damage that is often associated with aged RC beams. Two configurations of typical steel corrosion related damage, the loss of effective steel cross-section and concrete cover spalling, were simulated. The critical area
reduction of the steel cross section of a rebar was assumed to be 25%. This degree of deterioration is believed to be significantly severe, beyond which the replacement of damaged structural members would be a better alternative than FRP strengthening. As the location and extent of concrete spalling is dependent on reinforcing steel corrosion, only the concrete cover directly below the simulated corrosion was removed.
Moreover, two subgroups of Group B were formed, corresponding to two different degrees of reinforcing steel corrosion. Within each subgroup, one beam was used as a control and the other repaired with the CFRP/Sprayed FRP hybrid system.
Strengthening Details
Figures 4-5a through 4-5c illustrate each of the strengthening schemes, where the CFRP reinforcement consisted of two Sika Carbodur® laminates of 50 mm x 1.2 mm and the Sprayed GFRP was applied in a single layer approximately 7.62 mm thick. Although the original design target thickness was 13.7 mm, lack of operator experience resulted in the lower actual thickness for the first beam sprayed. For consistency purposes, the remaining beams were prepared with the same thickness of the Sprayed GFRP.
Figure 4-5. Beam strengthening schemes. a) Control beam.
Figure 4-5. Beam strengthening schemes. b) CFRP laminates only.
Figure 4-5. Beam strengthening schemes. c) Sprayed GFRP only.
12 @ 7”
CFRP Laminates
12 @ 7”
12 @ 7”
Sprayed GFRP
Figure 4-5. Beam strengthening schemes. d) CFRP/sprayed GFRP hybrid system.
Simulated Steel Corrosion
As mentioned previously, the beams in Group B were subjected to simulated corrosion damage, including loss of section of the reinforcing bars and spalling of the concrete cover. Using a milling machine, a length of 305mm at the mid-portion of the reinforcing steel was milled off approximately 6mm deep along the diameter on one side (see Figure 4-6a). This treatment was used to simulate the worst possible corrosion damage that occurs at the mid-span, the most severely stressed region, as well as a loss of cross-section equivalent to 25%. After the reinforcing cages were placed into the beam forms, wooden prisms of 305mm long and 50mm x 50mm cross-section were inserted under the rebars with simulated corrosion damage: two for the case of two damaged rebars and a single bigger piece under the tension steel layer for the case of three
damaged rebars. As a result, concrete at these locations was missing, simulating concrete spalling, after the casting of concrete and removal of the forms (see Figure 4-6b and 6c).
The simulated concrete spalling was patched with grout to restore the original cross- section before FRP application.
12 @ 7”
CFRP Laminates Sprayed GFRP
Figure 4-6. Simulated damages. a) Reinforcement steel.
Figure 4-6. Simulated damages. b) Concrete spalling with two corroded rebars.
Figure 4-6. Simulated damages. c) Concrete spalling with three corroded rebars.
Two “Corroded” Bars
12 @ 7”
305mm Concrete Spalling A
A A-A
A
Three “Corroded” Bars 305mm Concrete Spalling
A-A A
A
Steel Milled off Tension Steel
A
A-A
2745mm
305mm 6mm
FRP Strengthening of Beams
As described in the previous section, there were in total three different
strengthening schemes: namely CFRP only, Sprayed FRP only and the CFRP-Sprayed FRP hybrid system. Thus, the preparations of the strengthened beams consisted of bonding CFRP laminates to the underside of the beams and spraying Sprayed FRP to the sides of the beams. However, before applying any FRP materials, the concrete surfaces had to be properly prepared first to assure a good bond between the concrete and the FRPs. The necessity of this practice has been confirmed in previous investigations introduced in the first chapter. In this study, a sandblaster was used to roughen the concrete surfaces, as shown in Figure 4-7.
Figure 4-7. Sandblasting concrete beam surfaces.
After sandblasting, CFRP laminates were then bonded and Sprayed FRP was applied. Figure 4-8 shows one beam and two CFRP laminates that are to be bonded. First, the bonding face of the laminates were cleaned using acetone to remove any dust or other foreign material on their surfaces, then sufficient adhesive was applied, and finally two persons, holding each end of a laminate underneath the beam, lined it up with one beam edge and pushed it on. Adequate squeezing is needed to force excessive adhesive out from beneath the laminate, making a dense and strong bond.
Figure 4-8. Setup for CFRP laminate bonding.
As for the Sprayed FRP, a three-step procedure was followed, The first step was to spray the concrete surfaces with a ‘tack coat’ of resin alone and let it cure for 15 ~ 20 minutes, thus forming a sticky layer to help the sprayed fibers stay attached before the
resin has a chance to harden. Figure 4-9a shows this process. The second step was the normal spraying procedure, applying resin and fiber simultaneously. Here, as discussed in Chapter 2, the ratio between fiber and resin and the pumping pressure that depends on ambient temperature are extremely important in order to achieve a good spray. Figure 4- 9b illustrates this step. The final step was the rollout, wherein the goal is to obtain a Sprayed FRP plate as uniform as possible by applying pressure and forcing out air bubbles.
It was noted that achieving the target thickness in one non-stop spray application is not feasible. Since the spraying was directed on vertical surfaces, it had to be paused for at least 10 ~ 15 minutes after several passes to allow the sprayed resin to develop sufficient strength to hold the fibers already sprayed in place. Without sufficient curing time, the whole wet sprayed FRP layer would just slide off the concrete. The length of intermission is affected by ambient temperature and must be determined based on the experience of the operator.
To avoid detrimental effects on bonding due to shaking and vibration during moving, the beams were not moved after being sprayed until the next day to allow good bond strength to develop between the concrete and the resin. However, it was during the rainy season and these beams were sprayed outside the lab facility. One beam was rained on when it was left overnight. Although cover was used, some water still leaked to the beam and to the interface of the concrete and resin. The deteriorating effects on the bond were shown during the beam testing, when the weak bond strength was revealed for the beam under loading. As an afterthought, it would always be good practice to keep
sprayed GFRP repaired beams away from water for a certain amount of time, so that a strong bond between the concrete and resin can be formed.
Figure 4-9. Application of sprayed GFRP. a) Applying tack coat to concrete surface.
Figure 4-9. Application of sprayed GFRP. a) Spraying concrete beam.
Load Test Test Setup and Procedure
In this test program, only beam behavior under static loading was investigated. To this end, a three-point static loading setup was chosen, as shown schematically by Figure 4-10a. Figure 4-10b shows the actual setup in the lab. During the testing, load was applied until the instant when the load could not be increased further and an abrupt drop was observed. The rate of loading was controlled manually at approximate 12 kN/min., so that the loading time for each beam was roughly 30 minutes on average.
Figure 4-10. Beam lab testing. a) Schematic of beam test setup.
Figure 4-10. Beam lab testing. b) Photograph of beam test setup.
Concrete Floor
2248mm
Actuator A
A A-A
Instrumentation
The applied load, beam deflection, beam end rotations and material strains in the concrete, FRPs and steel were continuously monitored throughout the loading tests.
Figure 4-11 shows the overall instrumentation schematic. Some instruments used are discernable in Figure 4-10b.
LVDTs were used to measure beam deflection. Five LVDTs were installed at 508 mm on center on the top surface of the beams. Two LVDTs were also installed on the two steel I-beam restraint of the loading frame, each 108 mm away from the nearest LVDT at the corresponding end.
Electrical resistance strain gauges were used to measure the strains in the concrete, FRPs, and steel. Before beam casting, three strain gauges placed at 153 mm on center were bonded to the center segment of the middle tension rebar (see Figure 4-11). Two strain gauges were attached to the sprayed FRP at the mid-span on each side of the beams, one 50mm from the top and the other 50mm from the bottom. One strain gauge was bonded to each of the two CFRP laminates at the midspan. On each side of the beams with no sprayed FRP, two strain gauges were placed on the concrete at the mid-span, one 50 mm from the top and the other 50 mm from the bottom. No strain gage was bonded to the underside of the beams.
In addition, one tilt sensor, 146 mm away from the LVDT on the steel I-beam restraint, was attached to each end of the beams to measure end rotation. A load cell was placed between the top of the actuator and the point where the beams would be pushed up in order to monitor the applied load.
All test data was collected by a computer through a data logger. The strain signals were calibrated first by a strain conditioner and then transmitted to the data logger. A
LabView interface was developed to visually monitor the testing, and automatically compile the receiving data in the background as well.
Figure 4-11. Instrumentation on reinforcement steel.
Figure 4-12. Instrumentation on beams strengthened with CFRP / sprayed FRP hybrid system.
A
A A-A
The Middle Tension Steel
153mm 153mm
A
508mm
A 50mm
50mm 360mm
CFRP Laminates Sprayed FRP
A-A
Legend: Strain Gages; LVDT Tilt sensor
508mm
508mm 508mm
108mm 108mm 146mm
146mm
Results and Discussion
Four sets of results were obtained directly from the tests, namely; load, beam deflection, material strain, and beam end rotation, with time as the common variable.
Further, initial stiffness and energy absorption were also derived from the experimental results. Table 4-2 shows a partial summary of the results of the beam tests, which in turn infers the effect of the CFRP/sprayed GFRP system. The results of the strains and end rotations will be covered later.
67 Table 4-2. Summary of beam test results.
Beam Label (Table
4-1)
Initial Stiffness
(10-6 kN/mm )
Initial Stiffnes Increase
(%)
Ultimate Load (kN)
Load Increase
(%)
Ultimate Mid-span Deflection
(mm)
Deflection Increase
(%)
Engergy Absorbed (kN-mm)
Engergy Increase
(%)
Failure Mode
A-1 22.7 0 386.0 0 29.9 0 7850 0 SY/CC1
A-2 47.7 110.1 347.1 -10.1 10.3 -65.6 1590 -79.7 CD1 A-3 34.7 52.9 351.7 -8.9 11.5 -61.5 4450 -43.3 SR&SD1 A-4 26.4 16.3 398.9 3.3 15.0 -49.8 3600 -54.1 CD1
B-1 32.5 0 343.0 0 21.1 0 4760 0 SY/CC1
B-2 32.2 -0.92 414.5 20.8 11.6 -45.0 1620 -66.0 CD1
B-3 33.9 0 300.4 0 26.4 0 7000 0 SB1
B-4 22.1 -34.8 322.6 7.2 11.7 -55.7 2730 -61.0 SR&CD1
1 – SY/CC: steel yielding followed by concrete crushing; CD: CFRP debonding; SR&SD: one side sprayed GFRP rupture and the other side debonding; SB: steel rupture; SR&CD: sprayed GFRP rupture followed by CFRP debonding.
In comparison with the control beam, the initial stiffness of the strengthened beams in Group A increased but the ultimate deflection, ultimate load and energy absorption all decreased.
Only beam A-4 showed a 3.3% capacity increase while beams A-2 and A-3 both decreased by 10.1% and 8.9%, respectively. The reason for this anomaly might be that the loading was terminated at the point when FRP debonding was observed as well as an apparent drop in load. Actually, at that moment there was still capacity reserve left in the concrete and steel. If loading had continued, the load formerly carried by either the CFRP laminates or sprayed GFRP would be transferred to the concrete and steel and finally fail the beams in the concrete-crushing mode as exhibited by the control beam.
However, in real-life applications it would be risky to assume that the concrete beams were able to carry all of the applied loads even if the FRP had debonded. One reasonable solution to this case would be to prevent FRP debonding with anchorages or the like so that FRP, concrete and steel could all function to their full potential. In beam A-4, the composite action of the CFRP laminates and sprayed GFRP might delay the occurrence of the CFRP debonding, leading to a positive increment in the ultimate load.
Beam A-4 also showed the least improvement in initial stiffness at 16.3%, in contrast with 52.9% and 110.1% for beams A-2 and A-3. One explanation for the
increased initial stiffness is that either singular approach, attaching CFRP laminates to the soffit of the beam or spraying FRP on the sides of the beam, delayed the appearance of concrete tensile cracking by sustaining a load at the strain level equivalent to that of concrete and thus increased the load at which microcracks initiated, resulting in higher
initial stiffness. It is interesting to note, however, that the hybrid strengthening system produced the lowest increase in stiffness of the three strengthening approaches.
In addition, beam A-4 had the least mid-span decrease of 49.8% whereas beams A- 2 and A-4 had 65.6% and 61.5%, respectively. This observation is closely related to the increased initial stiffness.
In summary, these discussions indicate that in Group A the hybrid CFRP/sprayed GFRP system was not only effective in improving beam load-carrying capacity but also in preserving ductility.
Beams B-2 and B-4, strengthened with the CFRP/sprayed GFRP system, both exhibited positive gains in ultimate load but loss of deflection when compared to beams B-1 and B-3, the control beams. It is interesting to note, however, that both strengthened beams showed either negligible (0.92%) or negative (34.8%) effects on the initial stiffness, which is contrary to what was anticipated.
It can be concluded for the case of repairing corrosion damaged beams that the hybrid CFRP/sprayed system will improve beam load-carrying capacity, but at a lower ductility.
Comparing B-2 and B-4 with A-1 reveals that the hybrid system used is able to restore the load-carrying capacity of the beam with two damaged rebars to 107.4% that of the undamaged beam, but is unable to repeat it for the beam with three damaged rebars, bringing its load-carrying capacity to only 83.5%. It implies that there exists a degree of damage between that of two damaged rebars and that of three damaged rebars where this hybrid system would be able to completely restore the original strength. Beyond that