807 Sixth International Conference on Durability of Concrete Structures Paper Number DSM06 18 - 20 July 2018 University of Leeds, Leeds, West Yorkshire, LS2 9JT, United Kingdom An image-based local homogenization method to model mass transport at the steel-concrete interface Z. Zhang and U. Angst ETH Zurich, Institute for Building Materials, Zurich, Switzerland A. Michel and M. Jensen Technical University of Denmark, Department of Civil Engineering, Lyngby, Denmark ABSTRACT Mass transport (moisture and ions) at the steel-concrete interface is closely related to corrosion of rebar in reinforced concrete structures. Thus, in the model simulating mass transport, the structure of the steel-concrete interface must be well represented. In this study, an image-based local homogenization method is proposed based on images taken in the scanning electron microscope (SEM) under the backscattered electron (BSE) detector. According to the gray level of the image, porosity can be calculated and then proposed equations are used to associate the transport properties (sorption isotherms, diffusion coefficients and permeability) with porosity. Experimental data of chloride concentration taken from the literature are used to validate the proposed method and a good agreement with simulated results them is found. Keywords: Mass transport; Back-scattered electrons (BSE); Porosity; Local homogenization; Steel-concrete interface; Corrosion 1.0 INTRODUCTION The degradation of reinforced concrete structures subjected to the natural environment is one of the major concerns about the durability issues. When pH in reinforced concrete decreases caused by carbonation, or when chloride penetrates into concrete with liquid water, corrosion is apt to occur at the steel-concrete interface (SCI) (Angst et al., 2017). To evaluate the degradation of reinforced concrete structures, mass transport behavior at the SCI must be well understood. In this study, we are primarily interested in using numerical tools to simulate mass transport at the SCI. SEM images show that the thickness of the SCI, including mill scale, water bleeding zone and corrosion product-filled paste, is generally smaller than 1 mm (Wong et al., 2010; Zhao et al., 2017). At this scale, the presence of aggregates can significantly affect the mass transport and the growth of corrosion products, and therefore the SCI cannot be viewed as a homogeneous domain. Instead, its heterogeneity has to be considered in numerical simulations. To construct the structure for numerical modelling at the mesoscale, concrete can be separated into several individual components, such coarse aggregates, interfacial transition zone (ITZ), and mortar matrix (cement paste with fine aggregates). Coarse aggregates are commonly considered as impermeable and excluded in the simulations. The mortar matrix is treated as a homogeneous medium with a certain size of fine aggregates being assumed to uniformly distribute in cement paste. In the study of Du et al. (2014), all aggregates smaller than 2 mm are included in the mortar matrix. The ITZ was viewed as a uniform thin layer around each aggregate and the thickness was assigned as 500 µm which is much wider than the actual thickness of ITZ (e.g., 20 - 45 µm according to Horne et al. (2007)). The purpose of using a relatively large thickness is to ensure that ITZ can be properly meshed for simulations. Similarly, Abyaneh et al. (2013) also took the bulk mortar as a homogeneous material, by they did not use a single thickness for the ITZ. Alternatively, they set an exponential function for the porosity distribution in the ITZ showing that ITZ porosity becomes the same as the bulk porosity at about 40 µm from the aggregate. The choice of method to heterogenize the material is always dependent on the size of the domain and the resolution of the simulations. The domain size in Du et al. (2014) was several centimetres, so a 500 µm thick ITZ can be reasonably meshed. Abyaneh et al. (2013) simulated water transport in a cube of 7.5 mm and a 40 µm ITZ could be captured. They found that to consider the effect of ITZ, the voxel size must be smaller than 16.7 µm. However, these methods are much idealized because the ITZ is not uniform around the aggregate. In contrast, it is always dependent on the size of cement particles, the orientation of the aggregate and bleeding effects. In addition, the brought to you by CORE View metadata, citation and similar papers at core.ac.uk provided by Purdue E-Pubs