Time-dependent reliability of deteriorating RC structures: experimental and numerical investigations

Abstract

Due to environmental effects (e.g., chloride penetration, concrete carbonation) and extreme hazards (e.g., seismic, hurricane), the performance of reinforced concrete (RC) would deteriorate, which could significantly affect structural performance and safety, causing economic losses and even posing considerable challenges to life-cycle design and maintenance. Traditional life-cycle assessment and management are based on deterministic or semi-probabilistic approaches and models. Due to the randomness of material properties and changing environment, the performance of deteriorating RC structures is highly stochastic so that deterministic and semi-probabilistic methods may not be appropriate. Thus, it is of great importance to propose a probabilistic performance evaluation framework for deteriorating RC structures. Regarding the deterioration model of RC structures, previous studies have mainly applied simple analytical models to evaluate the performance of RC structures, lacking effective experimental validation, which may misestimate the life-cycle performance of RC structures in practical engineering. In this dissertation, a series of evaluation models associated with their durability performances and mechanical behaviours are developed and experimentally validated accordingly. Existing studies proved that chloride-induced reinforcement corrosion is one of the most important threats to RC structures. Thus, the time-dependent performance of RC structures under marine atmospheric environment is investigated. To overcome the drawbacks of constant environment parameters and Fick-law-based chloride diffusion models in previous studies, a time-dependent environmental model and a two-dimensional chloride transport model are developed to investigate the changing climate and convection effect in chloride ingress, and calibrated by the collected in-situ experimental data. Besides, considering the corrosion non-uniformity and its spatial effects on RC structures, a series of mechanical analysis models, including analytical, semi-analytical, and finite element method (FEM) models, are proposed and validated by experimental results. Existing studies related to RC structures subject to environmental effects only focused on their durability and mechanical behaviour, while economic and social impacts may need to be considered in practical engineering. Therefore, a performance-based probabilistic analysis method related to the durability of RC structures is proposed to address the uncertainty of environmental, material, and external load parameters. Furthermore, the proposed method could successfully assess the long-term economic and social impacts resulting from structural deterioration. In addition, a probabilistic analysis framework based on polynomial chaos expansion (PCE) and FEM modelling is proposed to perform global sensitivity analysis and time-dependent reliability analysis. Such a framework can effectively reduce the computational burden of probabilistic analysis in the application of complicated FEM models. On the other hand, conventional reliability analysis methods of deteriorating RC structures focus on statistic moments of the target variables, such as the first-order second-moment method (FOSM) and Monte Carlo simulation (MCS), which have limitations in terms of computational cost and application areas. This dissertation proposes a series of probability density function-informed methods (PDFM) accounting for different application scenarios. Considering the continuous deterioration and sudden damage of engineering structures during their service life, a two-step translation method-based PDFM is proposed and illustrated by numerical cases related to the time-dependent reliability of corroded RC structures. In addition, a reliability analysis framework based on point evolution kernel density estimation is proposed, considering maintenance behaviours, including preventive and essential maintenance actions, and their effects on the reliability analysis of ageing engineering structures. Furthermore, to improve the accuracy and efficiency of the proposed reliability analysis method in rare event estimation scenarios, a novel adaptive small failure probability estimation method is developed from the ideas of polynomial chaos kriging (PCK), subset simulation (SS), and PDFM. A series of benchmark cases validate such a novel approach. Finally, conclusions are summarized, and the future work is highlighted.