Bond and Development Length of GFRP Bars Embedded in Shotcrete

Abstract

Glass fiber–reinforced polymer (GFRP)–reinforced concrete structures can be efficiently constructed with the low cost and rapid concrete placement of sprayed concrete, termed shotcrete, whether for new construction or retrofit applications. This study is the first to investigate the bond strength and development length of ribbed and sand-coated GFRP rebar embedded in shotcrete using 20 notched beams. Two bar diameters (db) of 16 and 22 mm were tested with embedment lengths (le) of 20, 47, and 74 db. Four control beams were fabricated conventionally by pouring and vibrating the shotcrete mix, and 16 were shotcreted using the wet-mix process. Bond failure in 18 beams was by concrete cover splitting. Two shotcreted beams experienced pullout bond failures at significantly reduced bond strength as a result of large voids around the GFRP bars. Bond strength was, on average, 8% lower in shotcrete than poured concrete due to the presence of small voids in the shotcrete. For shotcrete beams without voids, the bond of shotcrete was equivalent to poured concrete. Ribbed rebar had, on average, a 21% lower bond strength than sand-coated bars. Overall, the bond of GFRP bars to shotcrete displayed similar trends as bond in poured concrete. Normalizing the embedment length with bar diameter (le/db) was found to inaccurately control for the effects of bar diameter and embedment length on bond strength. The bond equation in ACI CODE-440.11-22, overestimated the bond strength and did not accurately reflect the effect of concrete strength. This observation is most likely not limited specifically to shotcrete. A modified formulation of the ACI bond strength equation is presented that more accurately matches the observed trends of GFRP bond behavior, particularly the nonlinear variation of bar stress with le.