Statistical analysis of time dependent variation of diffusion coefficient for various binary and ternary based concrete mixtures

The numerical modelling of chloride penetration into concrete requires sound description of input parameters. One of the crucial inputs is the diffusion coefficient of concrete. The concrete is heterogeneous material and its parameters depend on the level of maturity. Therefore, the objective of this work is to describe the time dependent variability of the diffusion coefficient of chlorides into concrete. Computation of the diffusion coefficient is based on the non-destructive testing of bulk electrical resistivity following ASTM C1760 standard and Nernst-Einstein relationship. The presented results show the variation of diffusion coefficient of control OPC mixture and 32 various binary and ternary concrete mixtures where the effect of prolonged aging is even more significant. Two possibilities of the time dependent diffusion coefficient probabilistic model based on coefficient variation (mean value and regression via linear approximation) are proposed. The quality of both approaches is evaluated using Root of Mean Squared Error approach. Overall, this study provides guidance to select proper statistical tool for the assessing the time dependent variation of diffusion coefficient.