There is an increasing demand from asset owners for service life prediction of existing reinforced concrete structures. This requires assessment of the current condition and modelling to allow for a prediction. This paper critically discusses a number of subjects relevant for service life prediction with respect to durability related to chloride-induced reinforcement corrosion. The subjects include the physical meaning and variability of the end-of-service-life criterion, the validity of the deterioration models, the availability, variability and reality level of input values for some model parameters, as well as some practical issues concerning site investigations. The findings are exemplified by calculation examples using both a deterministic as well as a full probabilistic approach. It is anticipated that in the future a full probabilistic approach will be adopted which makes service life predictions more prone to manipulation of input values, as literature provides a wide spectrum of values to choose from. Although a probabilistic approach seems very impressive to most asset owners it usually disguises the lack of knowledge, responsibility and liability of the consultant involved. It is concluded that asset owners will be easy prey for consultants to play a lucrative numbers game eventually providing a desirable and realistic outcome, mostly already known beforehand.
Predicting residual service life of concrete infrastructure: a considerably controversial subject