Statistical analysis of the carbonation rate of concrete

Abstract The carbonation rate of concrete has implications for the lifecycle carbon emissions of concrete. This paper describes the reported effect of several variables on the rate of concrete carbonation and collates a data set of measurements published in the literature. Many studies producing predictive models for the carbonation rate constant, K, use only small data sets. 1999 measurements of carbonation depth as a function of time and other variables were collected for analysis. Models in the form ln (K) = a + bI1 + cI2 + … have been produced by which the rate of carbonation can be predicted. Hierarchical Models were used to combine different authors' data and introduces a new explanatory variable called ‘origin’, which indicates whether the concrete was taken from a working structure or cast specifically for experiments. Two models of the carbonation rate using concrete properties have been produced, allowing prediction of K over a range of conditions and compositions.