Engineering-Based Finite Element Approach to Appraise Massive Structures Affected by Alkali-Aggregate Reaction

Abstract

Alkali-aggregate reaction (AAR) is one of the most harmful distress mechanisms affecting the performance of aging concrete structures worldwide. In massive structures (e.g., dams), AAR has led to economic losses and negatively impacted structural safety. Some of the observed detrimental structural implications include loss of equipment alignment, excessive cracking leading to leakage, map cracking, concrete mechanical properties deterioration, introduction of excessive stresses on pre-stressing strands, inability to operate gates, and closure of expansion joints. Traditional structural design standards and hand-calculations are not able to accurately represent the effects of AAR on the performance of affected concrete structures due to its overall complexity and anisotropic induced expansion. Therefore, finite element (FE) modelling is one of the most reliable alternatives to evaluate AAR's structural implications. Several prediction models have been developed to assess the macroscopic consequences of AAR, ranging from simplified approaches that do not consider the implications of essential parameters to very complex micro-meso models that require extensive fitting. However, a thorough and comprehensive approach able to assess the current damage state (diagnosis) and to predict potential of further damage (prognosis), while (1) accounting for the most important parameters affecting the reaction, (2) accounting for the mechanical properties of deteriorated materials, and (3) not requiring fitting of parameters, is still lacking. In this context, Gorga (2018) proposed a practical, yet accurate engineering-based FE modelling framework for assessing AAR damage and predicting the future behaviour of affected slender structures, such as bridges. It is currently one of the most comprehensive approaches to assess AAR-affected slender structures, as it does not require fitting parameters. Even so, there is currently no macro-model capable of assessing AAR-affected massive structures (such as dams) that is able to achieve all the previously listed features. Dams are more complex systems for many reasons, including geometry and structural behaviour, unusual rehabilitation techniques (i.e., slot cutting), unique AAR kinetics, and thermal effects. Moreover, the economic impact and safety risks associated with the operation of massive structures are also much larger than those for slender structures, as they are often critical structures. Regarding rehabilitation of AAR-affected massive structures, often the only alternative is slot cutting, which consists of cutting vertical slots on the dam body to release the built-up compressive stresses and reduce permanent displacements. Even though there are several reports in the literature describing successful implementation of this rehabilitation technique, its efficacy remains the subject of ongoing debate among researchers. In this context, the main goals of the current thesis are: (1) to develop an updated all-encompassing modeling approach to assess AAR-affected concrete structures (with focus on massive structures) based on what had been proposed by Gorga (2018) and (2) to simulate and verify the effectiveness of slot cutting as a rehabilitation technique for AAR-affected massive structures. The first goal intends to ensure that all relevant phenomena are accounted for while keeping the same overall characteristics of the original approach (engineering-based assessment without the need for fitting). The commercially available software package Abaqus is used as the FE modelling platform. Key features in dam behaviour that are necessary to implement include: (i) updating the simulation to make it a fully integrated thermo-structural-AAR model, (ii) developing a new stress history-dependent creep subroutine, (iii) validating the thermal behaviour of mass concrete, (iv) validating the analysis in prestressed concrete structures, and (v) adopting an updated semi-empirical AAR model that considers the effect of leaching and alkali release from aggregates. All these steps are validated through data from laboratory tests or AAR-affected dams (Paulo Afonso IV and Bemposta dams). The second goal intends to settle the discussion within the engineering community regarding the effectiveness of slot cutting as a rehabilitation technique, clearly defining the mechanisms involved, their effect on the progression of the reaction, the influence of the most important variables on the slot cuts, and identifying in which situations the rehabilitation technique can be beneficial to the serviceability of the structure.