A multi-physics-based approach to predict mechanical behavior of concrete element in a multi-scale framework

Abstract

Concrete is a heterogeneous material whose constituents (e.g., hydrated cement paste, aggregate etc.) ranges from a characteristic length-scale of a few nanometres to metre. Hence, prediction of its realistic mechanical properties needs a deeper study at different length-scale. In this paper, multiple physical and chemical processes that occurs within concrete constituents at different length-scale are considered and a combined multi-scale and multi-physics-based analysis method to characterise the behaviour of concrete has been proposed. Firstly, concrete is described at three different length-scale, micro, meso and macro level based on the information of the constituents at different scale. Thereafter, based on the constituents (e.g., C3S, C2S etc.) hydration at micro-level, cement paste RVE (representative volume element) is formed and analyzed to obtain the homogenized properties of the cement paste. Such homogenized properties are then used as an input for the matrix properties in the next higher scale i.e., at meso-level where aggregates are randomly placed inside the cement matrix. Like micro-scale, homogenisation is performed at meso-level to obtain macro-scale properties of concrete. It is the meso-level of description where coupling of the hygro-thermal-mechanical phenomena is considered. By using such a progressive homogenisation approach at different scale and including coupled hygro-thermal-mechanical phenomena makes the strength prediction model of concrete both multi-physics and multi-scale dependent. A wide range of simulations are afterwards performed, and the proposed model is then validated with the corresponding available experimental results that highlights the robustness of the model. Such an approach of studying concrete would give the designer the flexibility through which an optimum and targeted design of a cementitious material like concrete can be achieved numerically that would otherwise require several time-consuming and costly experiments.