Preliminary Experimental Results of the Bond between GFRP Bars and Concrete
Mohammod Minhajur Rahman, Xudong Zhao, Tommaso D’Antino, Zahra Ameli, Francesco Focacci, and Christian Carloni
Appears on pages(s):
glass fiber-reinforced polymer (GFRP), bond behavior, pull-out test
Fiber-reinforced polymer (FRP) bars are an alternative solution to traditional steel bars for internal reinforcement of reinforced concrete (RC) structures. The potential reduction of damage in RC structures due to the absence of corrosion and the low weight-to-strength ratio of the FRP bars when compared to steel bars make FRP bars a cost-effective solution when durability is a concern. While a recent ASTM standard (ASTM D7913) has been issued to test the bond of FRP bars, limited work is available in the literature that deals with the determination of the interfacial properties between the FRP bars and concrete and the bond mechanism.
In this paper, experimental results are presented that aim at identifying a suitable setup to study the bond behavior and determining the effect of different bonded lengths on the stress transfer mechanism. Pull-out tests present some advantages to studying the bond mechanism without the complication of flexural stresses. Once the mechanism of the bond is studied at the small scale, and the interfacial cohesive material law is obtained, it is possible to simulate the behavior of full-scale members. The majority of the pull-out tests are performed with a short bonded length, which does not allow to fully establish the stress transfer between the bar and the surrounding concrete. In this paper, bars are embedded in concrete cylinders and pull-out tests are performed in displacement control with four bonded lengths. The first bonded length is equal to 5 times the bar diameter in order to consider the case of ASTM D7913. The other three bonded lengths are equal to 10, 20, and 40 times the diameter of the bar, respectively. Loaded-end displacement is obtained from the measurements of three linear variable displacement transformers (LVDTs). For some specimens, the free-end displacement was measured by two additional LVDTs. The load responses in terms of applied load versus machine stroke, loaded-end slip, and free-end slip are plotted and compared for the different bonded lengths. The results show that the average shear stress calculated according to ASTM D7913 is not constant for the different bonded lengths. In addition, the slip at the free end is activated at a different percentage of the peak load as the bonded length increases, which indicates that the bond phenomenon requires a certain bonded length to be fully established. The experimental peak stress versus bonded length and the stress level in the bar as a function of the embedded (bonded) length, according to ACI 440.1R-15, are compared. The results indicate that for the bar type studied in this paper, the provisions of ACI 440.1R do not match the results of the pull-out tests.