Numerical Framework for Simulation of Anisotropic Expansion due to Alkali-Aggregate Reactions in Concrete Members

Alkali-Aggregate Reaction (AAR) is a type of destructive and time-dependent chemical process in concrete that occurs between alkali ions in the cement and reactive minerals in certain types of aggregates. First recognized in 1930s, AAR is divided into two major categories of Alkali-Silica Reaction (ASR) and Alkali-Carbonate Reaction (ACR), both of which produce an expansive gel in the concrete which expands as a result of water absorption. The expansion of the AAR gel exerts significant internal pressure in the concrete, which may lead to internal and external cracks. With the occurrence of such cracks, many parameters affecting the stiffness and strength of the structure, such as compressive strength, tensile strength, and modulus of elasticity are diminished. As a result, the safety and serviceability of the structure may be seriously impacted. While advances in concrete materials science have led to means to prevent AAR in new construction, numerous existing structures worldwide, such as dams, power plants, and bridges, are affected by these reactions, the replacement of which may be impractical, or in some cases, impossible. As a result, it is crucial to simulate the behavior of such structures for reliable estimation of their safety and providing rehabilitation measures as necessary. One of the major indicators of AAR is the anisotropic expansion it generates inside the concrete member, which changes drastically based on the boundary conditions and internal and external restraint imposed on the expansions. As a result, the prediction of the anisotropic expansions is of utmost importance in successful simulation of AAR-affected reinforced concrete structures. This paper presents a practical simulation methodology for estimating the directional distribution of AAR expansions. The methodology makes use of the user subroutine capability in the finite element software Abaqus. A mathematical model is used to simulation AAR-related expansion based on the stress tensor, whereas concrete damage is simulated using the concrete damage plasticity model. The model is used to simulate a variety of AAR-affected reinforced and plain concrete cube and beam specimens for which the directional expansion data have been reported in the literature. Comparison between numerical and experimental results shows that the proposed methodology is capable of reliably simulating the AAR-induced expansions and the interaction between AAR expansions and the ensuing damage for a variety of reinforcement configurations. The model showed that the yield strength of reinforcing bars plays a major role in the directional distribution of expansion. However, changes in the mechanical properties of concrete were found to be inconsequential in the distribution of the expansions. Moreover, changes in distance between reinforcing bars and the reinforcement ratio in each direction were observed to affect the accuracy of the model. However, the model was found to be successful in reasonably capturing the trends in all case studies investigated. The results of this study are of great value to the simulation of AAR-affected reinforced concrete structures.