Inverse Analysis of Concrete Meso-constitutive Model Parameters Considering Aggregate Size Effect

Abstract

The meso numerical simulation has become an important method to study the characteristics of materials; however, the key to its further application is determining the parameters of meso-constitutive model. Considering that the meso-scale parameters of materials are hard to measure, this paper took into account the aggregate size effect and proposed a meso-parameter identification method by combining random aggregate numerical simulation and genetic algorithm. First, a random aggregate model of concrete was established, and its meso-model parameters were analyzed. The Morris method was used to analyze the sensitivity of meso-component parameters to the macro-responses, and results showed that the elastic modulus of mortar matrix, interface and large aggregates had a great effect on the peak strain and that the elastic modulus, Poisson’s ratio and tensile strength of interface and mortar matrix, as well as the Poisson’s ratio of large aggregates and the elastic modulus of small aggregates all had an effect on the peak stress, among which the interface tensile strength produced the greatest effect. Second, a parametric inversion and optimization function was established. The uniaxial compression numerical simulation test and genetic algorithm were combined to invert the meso-parameters, and results showed that compared with the single-aggregate parameter inversion curve, the multi-aggregate inversion stress-strain curve was much closer to the measured curve. That was because the aggregates of small size had lower elastic modulus, easing the stress concentration at the interface between aggregates and cement stone, and delaying the formation and growth of cracks.

Keywords

• concrete
• random aggregate model
• parameter estimation
• numerical simulation