PSI - Issue 84

Kelly Camarena et al. / Procedia Structural Integrity 84 (2026) 999–1006

1001

Code for Service Life Design, have established probabilistic frameworks for chloride induced corrosion, treating key parameters as random variables. Within this context, DuraCrete proposes a predictive model to estimate chloride concentration as a function of exposure time and location. = C(x, t) = (1−erf( 2√ . )) (1) Where, C(x, t) : content of chlorides in the concrete at a depth x and at time t , : design value of critical chloride content, : design value of the chloride surface concentration, x : penetration depth, D : apparent coefficient of chloride diffusion, t : time, erf : error function. According to Chen et al. (2021), carbonation-induced corrosion is a major deterioration mechanisms in RC structures, especially those not directly exposed to chlorides. The advance of the carbonation front is commonly modelled using diffusion-based approaches derived from Fick’s laws, where carbonation depth increases with the square root of time. Performance-based standards, such as the fib Model Code for Service Life Design, employ advanced probabilistic models to account for the main influencing factors and improve service life predictions. ( ) = √2. . . ( . − 1 + ). . √ . ( ) (2) Where, ( ) : carbonation depth at the time t , t : time of exposure, : environmental function, : execution transfer parameter, : Regression parameter, − 1 : inverse effective carbonation resistance of concrete, : error term, : CO 2 concentration, W(t): Weather function. 3. Methodology Climate data: temperature and CO 2 concentration were selected as the main climatic parameters. Annual temperature data were obtained from the Clima ARPA Veneto, which provides high resolution historical records for 1950 to 2005 and future projections up to 2100 under RCP 2.6, 4.5, and 8.5. Municipality- scale CO₂ emission data was obtained from ARPAV and INEMAR Veneto (2021) and a background concentration was included as recommended by the fib Model Code. In the absence of regional emission projections, a sensitivity analysis was performed considering three scenarios: current emissions, +50%, and +100%. Deterioration models: chloride penetration was modelled using the probabilistic frameworks of DuraCrete and the fib Model Code, treating temperature, cover depth, and diffusion coefficient as random variables. To account for different chloride sources, the study area was divided into coastal, mountain, and inland zones representing marine aerosols and de-icing salt exposure, see Figure 1a. Carbonation depth was estimated using the DuraCrete approach, with CO₂ concentration, cover depth, and concrete carbonation resistance considered as random variables. Unlike the chloride model, zoning was not required because CO₂ data were available for al l 563 municipalities, see Figure 1b. Both models determine the time to corrosion initiation in years.

(a) (b) Fig. 1. (a) Zonification of Veneto region for chlorides diffusion modelling; (b) CO 2 concentration in Veneto. Source: ARPAV and INEMAR Veneto (2021). Structural assumptions: a representative simply supported RC bridge deck was selected as the reference structural typology, reflecting typical regional infrastructure. The assumptions included a concrete cover depth of structural