Experimental and numerical study of strain distributions and shear lag effects in FRP tendon pull-out experiments

In the present study, sand-coated carbon fiber reinforced polymer (CFRP) tendons were instrumented with integrated optical fibers. The direct integration during manufacturing did not affect the bond between tendon and concrete and resulted in an excellent measurement quality. The tendons were cast into concrete cubes with different embedment lengths of 5, 10 and 15 times the diameter. The bond between tendon and concrete was tested using direct pull-out tests. Not only the force and the end-slip were measured but also the strain distribution in the tendon. This allowed to derive a position-dependent local bond–slip curve. Significant deviations from the commonly measured average bond-stress vs. end-slip relationship were observed for long and short embedment lengths. A complementary finite element study confirmed that average bond-stress vs. end-slip curves cannot properly predict the strain distribution in pull-out experiments, especially for longer embedment lengths. This allows to conclude that strain distributions are highly affected by local effects, e.g. partial pull-out failure of the concrete. In addition, shear lag was identified as additional parameter affecting the measurements of optical fibers when integrated in materials with low shear stiffness. The shear lag is highly dependent on the radial and axial position and further affected by the tendon material, the diameter of the tendon and the magnitude of the shear stress.