Understanding the Current and Forecasting the Future Behaviour of DEF and ASR-Affected Concrete Beams

Abstract

Delayed ettringite formation (DEF) and alkali–silica reaction (ASR) are amongst the most harmful damage mechanisms affecting critical concrete infrastructure worldwide. Very often, these mechanisms are seen combined in the field due to physicochemical synergy. Among the factors influencing the induced development of DEF and ASR, moisture gradients and reinforcement-induced restraint are the most significant, since they modify the magnitude and distribution of induced expansion and damage. The impact of moisture gradients and reinforcement-induced restraint has been investigated individually in recent research. Yet the combined influence of these factors on the reaction kinetics and amplitude of damage over time remains insufficiently quantified at the material scale.

This PhD thesis focuses on a detailed multi-level investigation aiming to better understand how moisture gradients and reinforcement configurations modify the development of microscopic deterioration, the associated mechanical response, and the potential for further expansion in DEF- and DEF+ASR-affected concrete members. An experimental program was conducted on concrete beams affected by DEF of various damage potential (i.e., moderate and high), and by DEF+ASR, subjected to controlled moisture gradients representative of partial wetting and late water supply (rewetting). Beams included non-reinforced members and members with distinct reinforcement-induced restraint levels (i.e., moderately, and heavily). Cores were extracted along selected heights and orientations (i.e., transverse, longitudinal, and vertical) to enable direction- and depth-dependent assessment. Current condition was comprehensively assessed through a multi-level protocol integrating microscopic analysis using the Damage Rating Index (DRI) and mechanical testing using the Stiffness Damage Test (SDT). The potential for further expansion was appraised through residual expansion testing under favourable moisture and temperature conditions. A semi-empirical moisture-transfer model was further developed to support exposure-history interpretation, and a Bayesian formulation of residual expansion was implemented to produce probabilistic projection intervals, adding uncertainty to the analysis.

Results demonstrate that moisture history plays a major role in height-dependent deterioration and that reinforcement-induced restraint redistributes induced damage through direction-dependent constraints. Microscopy confirmed that, in combined mechanisms, DEF-induced damage features are predominant in high water availability locations, whereas ASR-induced damage features are highly present in location under less favourable moisture. A DRI-based partitioning approach enabled interpretation of mechanism contributions to the "overall damage" under moisture gradients. The diagnosis-to-prognosis framework revealed that residual expansion potential depends on both exposure history and the current damage extent, while the Bayesian formulation provided bounded uncertainty envelopes suitable for conservative screening.

Finally, the findings provided a better understand of the impact of the moisture gradients and reinforcement-induced restraint on the reaction kinetics and damage development over time, which in turn can improve protocols for conditions assessment of affected concrete.