Abstract Concrete aging and deterioration are the result of a complex combination of multiple interacting time-dependent phenomena. Two major ones are shrinkage and creep, which affect and in turn are affected by both environmental conditions and any additional damage mechanisms like Alkali-Silica Reaction (ASR), corrosion, freeze/thaw and others. While many modeling approaches can represent the overall uncoupled contribution of these phenomena, they can not capture their coupled effects especially when large internal self-equilibrated stresses are produced like in the case of ASR free expansions. Macroscopic continuum models can not capture the induced meso-scale creep deformation and stress relaxation due to these conditions. Recently, the Lattice Discrete Particle Model (LDPM) was extended to account for coupled creep, shrinkage and ASR deformations and has shown noticeable success in capturing these meso-scale phenomena. The model uses a multi-physics formulation to evolve temperature, humidity, cement hydration, and alkali ion diffusion in both space and time at the meso-scale between large aggregate pieces. Creep and shrinkage deformations are formulated based on a discrete version of the Micro-prestress Solidification theory. It was shown that using this detailed multi-scale multi-physics framework is essential for the accurate prediction of both plain and reinforced concrete long-term behavior. The major challenge associated with using this comprehensive formulation is the appropriate calibration of its multiple parameters. This requires multiple experiments to be performed on the same material, which is usually not the case in many experimental campaigns. Till today, no established recommendation existed concerning the optimal calibration sequence dependent on the available data. This paper, presents a detailed procedure for the calibration of the hygro-thermo-chemical and creep/shrinkage parameters in an uncoupled manner. The procedure is thoroughly investigated using 9 different experimental datasets that vary in complexity and level of detail. Results show the effectiveness of the procedure and the capabilities of the modeling framework.
Read full abstract