A post-installed insert for pull-out tests on concrete up to 70 MPa

• A new split sleeve for post-installed inserts for pull-out tests on concrete.
• Wide experimental calibration shows the post-installed insert to be reliable and efficient.
• The calibration takes into account different stress states, a unique feature of this research.
• The new sleeve is efficient also in cracked concrete and in high strength concrete.
• A theoretical approach is provided fitting the experimental calibration curves.

Even though limited to the outer concrete layer of structures, pull-out tests have been proved to be a reliable technique for the in situ estimation of concrete strength. The most studied post-installed wedge anchor, the CAPO test, is not widely used due to some complexity in its use. A simplified procedure, resembling the CAPO test, the Zykon technology, has been used in Italy for more than 20 years with substantially unsatisfactory results, mainly for tensile stress states and for high stress gradients. In both the cases, the calibrations curve, relating the pull-out load to concrete strength, is related to unstressed concrete, which is troublesome when applied to in-service structures, i.e. with not vanishing stress states. A new post-installed wedge anchor is discussed in this paper that, based on a careful choice of all the technical details, has been proved to be reliable for all stress states since: (i) the drilling and expansion procedure avoids much of the uncertainties that affect the commonly used procedure; (ii) the calibration of the procedure has been performed on 6 classes of concrete and considering 5 different stress distributions (medium and low compression, vanishing stress states, compressive and tensile stresses with strong gradients). The correlation curves are nonlinear, as for pre-installed inserts, and depend on the stress state, with statistical scattering never exceeding 7–8% of the average values. A theoretical analysis fits rather well the test data and provides explanation to several aspects of pull-out tests.