Research Team: S. Dehghanpoor Abyaneh, H.S. Wong, N.R. Buenfeld

Funding: European Union Seventh Framework Programme (FP7 / 2007-2013)

Diffusion and capillary absorption are important mechanisms by which aggressive species such as chloride and sulphate ions penetrate concrete. This study aims to develop numerical approaches to model these mechanisms from a three-dimensional mesostructure of concrete.

The models are then applied to increase our understanding of the role of various phases on transport properties. To enable this, a three-dimensional representative volume element of concrete mesostructure (Figure 1a) was generated by randomly placing aggregate particles in a heterogeneous porous cement paste matrix consisting of ‘interfacial transition zone’ (ITZ) and bulk cement paste.

The volume fraction, particle shape and size distribution of aggregates are defined variables, while the thickness of ITZ and porosity of the cement paste vary depending on the w/c ratio and degree of hydration [1].

A finite difference method coupled with mass conservation was used to simulate diffusion through the mesostructure. A similar approach was used to model capillary absorption [2]. Here, the mesostructure was discretised into a cubic lattice where lattice elements act as conductive "pipes" with transport properties assigned based on the phase they represent. A non-linear finite element method was used to solve the governing diffusion equation for capillary absorption according to unsaturated flow theory.

The distribution of absorbed water content in the mesostructure was calculated as a function of space and time (Figure 2), based on the hydraulic diffusivity, moisture content and sorptivity of the porous matrix. The models were validated against available experimental data and compared with analytical relationships for ideal cases (Figure 1b and Figure 3). The models were then applied to study the effect of w/ ratio, degree of hydration, aggregate size, volume fraction, shape and orientation, ITZ width and percolation on transport.

It was found that the most significant parameters influencing diffusivity were w/c ratio, degree of hydration and aggregate content, while the ITZ width and aggregate size have less influence. The percolation of ITZ when aggregate volume fraction exceeds 30% did not result in an increase in diffusivity.

The simulations also showed that aggregate shape and orientation have significant influence on diffusivity and capillary absorption. The absorbed water front advances in an uneven manner that is influenced by the amount, spatial distribution and shape of the aggregate particles.

In all cases, diffusivity and sorptivity decreased when spherical aggregate particles are replaced with ellipsoidal particles due the increase in tortuosity of the cement paste. Furthermore, the effect is more pronounced at higher aspect ratio and aggregate volume fraction. However, the size of aggregate particle appears to have an insignificant influence because the increased tortuosity due to flow around larger aggregate particles is balanced by the reduction in the number of aggregate particles.

From the simulations, it is evident that the tortuosity and dilution effects are more important compared to that of the ITZ. The modelling approaches developed in this study are particularly useful for isolating and evaluating the influence of various parameters on transport properties that are otherwise difficult to achieve through laboratory-based testing alone.

Acknowledgements: The research leading to these results has received funding the European Union Seventh Framework Programme (FP7 / 2007-2013) under grant agreement 264448.

References

  1. S. Dehghanpoor Abyaneh, H.S. Wong, N.R. Buenfeld (2013), Modelling the diffusivity of mortar and concrete using a three-dimensional mesostructure with several aggregate shapes, Computational Materials Science, 78, 63-73.
  2. S. Dehghanpoor Abyaneh, H.S. Wong, N.R. Buenfeld (2013), Computational investigation of capillary absorption in concrete using a three-dimensional mesostructure mesoscale approach, Computational Materials Science (Under review)