Modeling of Chloride Spatial Variability in a Reinforced Concrete Wharf from Onsite Measurements

Chloride ingress by diffusion is the major deterioration process of reinforced concrete (RC) structures exposed to the marine environment. These structures have significant lengths or surfaces exposed to the outside environment. Due to the material variability (different concrete batches and vibration) and exposure variability, the material experiences a spatial variability of the deterioration process. This paper presents the geostatistical analysis of in situ chloride profiles, leading to the assessment of the spatial variability (SV) of both the chloride ingress itself and the parameters of the widely used Fick’s diffusion law (the average surface chloride content, $C_{sa}$, and the average chloride diffusion coefficient, $D_a$). 37 chloride profiles measured on both sides of the same spandrel beam of a RC wharf were studied, as well as the associated estimates of $C_{sa}$ and $D_a$. From an initial selection of random field models, the geostatistical analysis consists in the evaluation of model parameters using a procedure that tests both data and model assumptions on the fly (ergodicity, stationarity, and random field modeling). Combined with the calculation of information criteria for each model, this procedure allows to provide relevant geostatistical models for chloride ingress, $C_{sa}$ and $D_a$, which render SV as well as measurement error. It is noteworthy that the estimation error can be neglected when focusing on the SV for the range of chloride content studied in this paper. The SV of the chloride content seems to depend on the depth, with a large variability within the convection zone, and much less variability and more stability in the diffusion zone with a practical range of about 70 cm. This order of magnitude is consistent with the range of SV calculated for $C_{sa}$ (50–73 cm).