PSI - Issue 62

Simone Celati et al. / Procedia Structural Integrity 62 (2024) 361–368

362

2

Simone Celati et alii/ Structural Integrity Procedia 00 (2019) 000 – 000

1. Introduction Assessing existing prestressed concrete (PC) bridges, particularly those with post-tensioned tendons (PT), is a critical challenge in structural engineering. PT structures face increased uncertainty, and conventional visual inspections often fall short in reliably evaluating their degradation. New inspection methodologies, such as those outlined by Mazzatura et al. (2023), are under development. Evaluating the remaining lifetime of these structures involves modelling the degradation process, often focusing on cable corrosion, see for example Guo et al. (2016) or Nguyen et al. (2013). Since these structures are exposed to chlorides from de-icing salt or a marine environment, chloride-induced corrosion is recognized as a major factor in degradation. However, recent research indicates that chloride concentration alone is insufficient to initiate long-term corrosion in sound concrete due to its high pH (Melchers and Chaves (2018)). Furthermore, in-situ tests are performed to accurately assess PT structures, and, to the best of the authors' knowledge, there is a lack of studies on the sensitivity of variables in modelling PT structures under corrosion degradation, crucial for accurate assessment and resource optimization. The proposed method for assessing the reliability of PT girders over time incorporates both a corrosion model based on empirical evidence that corrosion only initiates after chlorides reach the steel surface and alkalinity is reduced (Melchers and Li (2006)) and a sensitivity analysis that offers insights into variable importance at different degradation stages. The method enables estimating the remaining service life of these structures and includes (i) the probabilistic characterization of severe corrosion progression in Section 2, (ii) a structural probabilistic approach considering extreme loads and capacity influenced by severe corrosion (Section 3), (iii) and a sensitivity analysis to assess the significance of input variables over time and explore opportunities for simplifying the complexity of the model (Section 4). 2. Corrosion process in post-tensioned concrete structures The corrosion process of steel in concrete is generally divided into two sequential phases: the initiation period and the propagation period. The former is defined as the time required for aggressive ions to break down the naturally formed passive layer on steel in an alkaline environment, whereas the latter is characterised by active corrosion during which the steel area reduces, and rust develops. The corrosion initiation of the PT system can occur through two distinct mechanisms. The first mechanism, as illustrated in Guo et al. (2016), involves the penetration of aggressive agents from the surrounding environment through the cover and then through the metallic ducts, to finally diffuse into the grout towards the wire surfaces. The second mechanism, as described by Nguyen et al. (2013), considers the presence of voids inside the ducts contaminated by corrosive agents. In this case, corrosion occurs after the aggressive species diffuse through the thin grout cover of the wires. The initiation phase is often studied through the diffusion process of chloride ions or carbon dioxide, which can lead to chloride-induced corrosion or carbonation (Vereecken et al. (2021)). Despite this approach being broadly employed in literature, where authors typically focus on one of the two causes of passive film breakdown, there is evidence, as reported in Melchers and Chaves (2018) that chloride diffusion alone cannot trigger long-term corrosion until the pH level drops below approximately 9, as the required thermodynamic conditions are not met. Considering the phenomenological model described by Melchers and Chun Qing Li (2006), active corrosion can occur only after both chloride diffusion and pH reduction. Therefore, the time to active corrosion is herein modelled as the longer of the two periods needed by chlorides to reach the steel surface and carbonation develop. In other words, the initiation phase is logically modelled as a two-component parallel system, consisting of chloride diffusion and carbonation. This definition is founded on the assumption that the two processes develop independently, and the pH decreases solely due to the carbonation process. 2.1.1. Chlorides diffusion model The most broadly used chloride diffusion models rely on Fick’s diffusion theory. The model selected in this paper is based on Fick’s second law and is reported in Schießl et al. (2006). 2.1. Initiation phase modelling