On a Poromechanical Approach to Long-Term Autogenous Shrinkage

For cementitious materials, the time-dependent strain of a non-loaded specimen exchanging no water with the outside is called autogenous shrinkage. Whether this autogenous shrinkage is asymptotic or not at large times has not been systematically studied. The first objective of the present work is to study how autogenous shrinkage evolves over time at long term by analyzing experimental data that are available in literature. In these experimental data, autogenous shrinkage showed logarithmic evolution over time at large times, from which we characterized the long-term kinetics by a long-term slope of the deformations plotted on a log scale. In the second part, we are interested in the origin of autogenous shrinkage. Considering concrete as a multiscale material, by three steps of downscaling, we linked the long-term log-slope of autogenous shrinkage to the creep modulus of calcium silicate hydrates (C-S-H) gel. Then we computed at this scale an equivalent mechanical stress that should act in order to explain the measured long-term rate of autogenous shrinkage of the concrete. This mechanical stress is of the same magnitude as the effective pore water pressure due to self-desiccation and computed from the theory of poromechanics. Hence, we may consider autogenous shrinkage as creep induced by capillary forces due to self-desiccation.