Materials, Specimens, and Test Methods
Three commercial CFRP composite systems (A, B, and C) were supplied by manufacturers. CFRP System A consists of unidirectional fabric sheet and a two-part adhesive resin with a mixture ratio 2.9 to 1.0 by mass. CFRP System B consists of unidirectional fabric sheet and adhesive resin system which is composed of primer, putty, and saturant, The mix ratio is 100 to 30 (Part A to Part B) by mass for primer, 100 to 30 (Part A to Part B) by mass for putty and 100 to 34 (Part A to Part B) by mass. CFRP System C consists of precured plate laminate and a two-part adhesive resin, Part A and Part B with a mixture ratio 3 to 1 by volume. According to the technical specifications of manufacturers, the general properties of the above CFRP materials and adhesive resins are listed in Table 1 and Table 2.
The concrete beams used in this study were supplied by Rocky Mountain Prestress, Denver, CO, with dimensions of 100 by 100 by 380 mm (4 by 4 by 15 in.) and 28-day compressive strength 70.9 MPa (10 130 psi). The cement used was Type I and the ratio of cement: sand: coarse aggregate (pea gravel)
TABLE 2—Properties of Adhesive resins.
was 1:2.07:2.30 by weight and the water/cement ratio was 3:2.
Prior to applying the CFRP, a 50 mm (2 in.) deep saw cut was made at mid-span of each beam to maximize environmental exposure at the point of flexural failure. Next, CFRP fabric or laminate strip with dimensions 200 mm by 25 mm (8 in. by 1 in.) was bonded to the beam using adhesive resin of each system. This external reinforcement was centered on the tension side of the flexural specimen as shown in Fig. 1. The development length of the CFRP bonding system is supported by the findings of other investigators. Chajes et al.  tested different development lengths of FRP by performing force-transfer tests. An approximate 90 mm (3.5 in.) effective length could develop the failure load of the CFRP, beyond which no further increase in failure load could be achieved. Dai  conducted a comprehensive literature review on effective development length of FRP externally bonded concrete applications and summarized that an appropriate development length should range from 75 mm (3 in.) to 200 mm (8 in.). The 200 mm long CFRP strip used in this study had a development length of 100 mm (4 in.) on each side. A 90 mm by 90 mm (3.5 in. by 3.5 in.) fabric square was bonded to one end of the concrete beam for Systems A and B for a direct tension test. For System C, the size of the laminate square is 50 mm by 50 mm (2 in. by 2 in.). Both flexural and tension testing are available from a single specimen.
The surface preparation of concrete is another important factor affecting the ultimate strength. The bonding surface must be sound, free from dust, chlorides, and other contaminants. Sandblasting, water jet, mechanical abra-
FIG. 2—Three-point bending test setup.
sion, and needle gun chiseling are common methods to increase strength of FRP composite system [18-20]. The concrete surface preparation can remove weak material, surface laitance, and other contaminants. The roughened concrete surface effectively increases the contact area and mechanical interlocking between adhesive and concrete. In this study, after 28-day standard curing, all the concrete beams were dried and sandblasted to a minimum ICRI profile of 3 , and cleaned using compressed air. CFRP strips and squares were applied to the prepared surface according to the manufacturer’s specification.
After curing 14 days in ambient conditions, CFRP System A, B, and C specimens were placed into steel tanks filled with water at elevated temperatures ranging from 30 ° C to 60 ° C. Commercial heaters were installed in each tank to maintain a constant temperature within ±4 ° C. The top of these tanks were sealed with plastic sheets and covered with 50 mm (2 in.) thick insulating foam board. The sides of each tank were surrounded by 150 mm (6 in.) thick insulation to maintain constant temperature. The maximum and minimum water temperatures in each tank were recorded daily using a commercial thermometer whose sensor was immersed in each tank. Before testing, the specimens were removed from tanks and immediately placed into a room temperature water bath for one day to cool. After cooling for one day, the specimens were allowed to air-dry for one day in the laboratory.
A three-point bending test setup modified from ASTM C78-02  was used for flexural testing to determine flexural strength of CFRP composite specimens. The specimens were loaded using a servo-controlled Instron 1332 testing machine with a 45 kN (10 kip) load cell as shown in Fig. 2 (www. instron. com).
FIG. 3—James 007 Bond Test apparatus, James Instruments Inc., Chicago, IL.
Load was applied at a constant rate of 0.84 mm/min (0.033 in./min) to cause failure in three to five minutes.
After beam flexural testing, the half of the specimen with a CFRP square bonded on the end was tested in direct tension to evaluate the tensile strength of CFRP systems based on ASTM D4541-02 . The CFRP square was cored 2.5 mm (0.1 in.) deep with a 50 mm (2 in.) diameter coring bit. A James Instruments bond testing apparatus (Fig. 3) was used to test the tensile bond strength between the CFRP composite and concrete substrate (www. ndtja – mes. com). A quick setting adhesive resin was prequalified to ensure that adhesive bond strength was greater than tensile bond strength of the concrete substrate and CFRP composite. The observed failure modes were recorded along with the tensile strength results.
In this study, flexural strength from three-point bending tests and direct tension tests were used as a baseline measurement to quantify the degradation of exposed specimens. The exposed specimens were submerged in elevated water baths for up to 18 months. Groups of three specimens were evaluated for exposure times of twelve months or less and the number of specimens was increased to five for 18 months of exposure.