An innovative method for mesoscale modelling of moisture diffusion in concrete

Moisture diffusion influences the durability and long-term performance of concrete and whilst it predominantly occurs via the cement matrix and Interfacial Transition Zone, most existing models consider concrete to be homogeneous. This paper introduces a novel micro-meso model that employs random packing and Voronoi tessellation. Rayleigh-Ritz pore distribution and Brunauer-Skalny-Bodor models are combined to determine the radius and fraction of various pores. The results indicate that relative humidity diffuses faster with increasing temperature, decreasing ambient relative humidity and tortuosity. Ambient relative humidity has a greater influence on diffusion compared to temperature and tortuosity. Numerical and experimental comparisons demonstrate that the proposed methodology effectively captures relative humidity distribution across various scenarios. Furthermore, explicit pore network modelling incorporates key parameters for a more accurate analysis. Integrating the proposed methodology into a fully coupled hygro-mechanical framework can potentially yield more accurate predictions of mechanical behaviour; enhancing the reliability of long-term performance assessments and enabling more durable concrete design. •A comprehensive meso-micro model for moisture diffusion in concrete is developed.•The mesoscopic model of concrete is generated using a novel geometric model.•An accurate pore size distribution model is employed to calibrate the diffusion coefficient of liquid and vapour water.•The meso-micro model accurately describes the distribution and evolution of relative humidity.