Characterising interfaces for reinforced concrete: Experiments and multiplane cohesive zone modelling for titanium alloy rebars

Abstract

Abstract This paper addresses the experimental and numerical characterisation of the structural interface between titanium alloy Ti6Al4V plain bars and Normal (NWC) and Light-Weight Concrete (LWC) mixtures. Results of pull-out tests on ϕ 8 and ϕ 16\u202fmm rebars from NWC and LWC batches and SEM analyses show that, although the employed rebars are plain, the debonding process is strongly affected by defects-induced surface roughness still present at the microscopic level, which activates mechanical interlocking responsible for the dilatant behaviour of the interface. Experimental tests are supported by Finite Element (FE) analyses employing Cohesive Zone Models (CZMs) for simulating the interfacial delamination. To this end, the micromechanics-based CZM proposed by Serpieri et al., accounting for damage, friction and interlocking, is employed, upon extending this formulation by addressing the degradation of the depth of asperities as a novel mechanical feature, Sensitivity analyses permit to assess the proposed modelling strategy and to devise a procedure for the numerical identification of model parameters. Experimental pull-out curves are fitted by employing a single set of material parameters for each concrete batch, achieving reasonable numerical-experimental agreement for tests with both ϕ 8 and ϕ 16\u202fmm, thus showing good predictivity of the proposed modelling strategy. • Pull-out tests on plain titanium alloy bars from different concrete specimens are carried out. • Results of the tests suggest an interlocking mechanism at microscale level and a dilatant behaviour of the interface. • Bond strength values of titanium-concrete interface are comparable to those of other reinforcement-concrete interfaces. • Finite element models simulate the pull-out tests by employing a CZM accounting for damage, friction, mechanical interlocking and dilatancy. • Sensitivity analysis, parameters identification and model validation are performed, achieving very satisfactory agreement between numerical and experimental curves.