Probabilistic considerations on the effect of specimen size on the critical chloride content in reinforced concrete

► A probabilistic model demonstrates the effect of specimen size on C crit. ► Test methods: a high number of samples is required when using common dimensions. ► Service life modelling: C crit must be related to the rebar length. ► Literature scatter and high laboratory values might be explained by sample size. A probabilistic model was used to predict the critical chloride content ( C crit) for reinforcement corrosion as a function of specimen size. The specimen size is likely to be a reason for the large scatter of C crit values as well as for the high threshold levels often observed in laboratory setups. Regarding test methods, it is discussed that the common combination of small specimens and low numbers of parallel samples results in poor reproducibility. If experimental data is to be used for service life modelling, the geometrical dimensions of the specimens on which C crit was measured have to be taken into account.