Random field modelling of spatial variability in concrete – a review
Wouter Botte, Eline Vereecken, Robby Caspeele
Abstract
Concrete is a heterogeneous material, consisting of different constituents such as aggregates of different sizes and cement paste. Due to the heterogeneity of the concrete, its properties such as strength, stiffness, permeability, etc. show variations in space. In scientific literature, several investigations have shown the influence of spatial variation of material properties on the reliability of reinforced concrete structures, and it was concluded that spatial variability and spatial correlation should be accounted for to obtain accurate estimations of, e.g. the load-bearing capacity of degrading concrete structures. For this purpose, random fields are often applied in literature. To apply these random fields in a structural reliability analysis, a correlation model and corresponding correlation length need to be defined. Nevertheless, experimental results are not always available to derive the correlation model, and hence some modelling assumptions are required. However, there is no consensus yet in literature on which correlation model to apply for the different concrete properties. This paper provides an overview of different correlation models applied in literature in relation to the analysis of reinforced concrete structures.