A proposal to use the inverse problem for determining parameters in a constitutive model for concrete

Concrete is a material of utmost importance in construction, and therefore various numerical models have been developed to satisfactorily represent its observed behavior. Within this perspective, damage models stand out, especially isotropic damage ones that can satisfactorily represent the nonlinearity characteristics resulting from stiffness deterioration having the added advantages of a few parameters to be identified. However, these models are dependent on experimental observations to determine the constitutive variables which greatly influence the accuracy of the obtained response. Therefore, using more robust identification techniques can improve the process of determining these internal variables. In this sense, this paper intends to contribute to the literature by using swarm intelligence bioinspired optimization techniques combined with a concrete damage model to determine its constitutive variables. Observable sources are obtained from complete stress-strain curve proposals obtained from the literature, and the optimization firefly and bee colony algorithms are used. The efficiency of these optimization algorithms is verified, as well as that of the damage model in uniaxial tension and compression situations. Willmott's correlation indexes reveal values close to 0.97, which indicates an excellent correlation between the generated numerical data and reference responses.