Identification of cohesive zone model and elastic parameters of fiber-reinforced cementitious composites using digital image correlation and a hybrid inverse technique

Abstract

Traditional methods for the inverse identification of elastic properties and local cohesive zone model (CZM) of solids utilize only global experimental data. In contrast, this paper addresses the inverse identification of elastic properties and CZM of a range of materials, using full-field displacement through an optimization technique in a finite element (FE) framework. The new experimental–numerical hybrid approach has been applied to fiber-reinforced cementitious composites (FRCC). PVA microfibers are used at four volume fractions: 0.5%, 1%, 2% and 3%. Digital image correlation (DIC) technique is used to measure surface displacement fields of the test specimens. Four-point bend tests are carried out for the measurement of the modulus of elasticity, E, and the Poisson’s ratio, ν, while single edge-notched beams (SENB) are used for measurement of mode-I CZM parameters. A finite element update inverse formulation, which is based on minimization of the difference between measured and computed displacement field, is used for both identification problems. For the identification of E and ν, linearized form of the Hooke’s tensor in plane stress condition has been derived for two-dimensional linear elasticity in FE frame, and Newton–Raphson solver is employed for the inverse problem. For the identification of the CZM, generic spline curves have been used for the parameterization of any CZM thus avoiding the need of an assumption of the CZM shape, while derivative-free Nelder–Mead optimization with CZM shape regularization is employed as the solution method, which reduces the complexity of numerical implementation and improves robustness. The computed E and ν are consistent with published results. The computed CZMs of the FRCCs with different fiber volume fractions reveal a strain-hardening characteristic. The computed CZM is used in direct problem simulation, the results of which are consistent with the experimental global response.