Sensitivity of key input parameters on chloride binding and distribution in concrete with cast-in chloride

This study addresses the complex challenge of modelling cast-in chloride and its binding and distribution in partially saturated concrete. An innovative approach, combining the theoretical model with a global sensitivity
analysis tool, is applied to determine the influential parameters. For the first time, this study evaluates the application of an extended Fourier Amplitude Sensitivity Test for parametric sensitivity to a concrete model with
cast-in chloride. Novelty stems from the ability to evaluate the influence of individual parameters and their interactions through main and total effect sensitivity indices. The rigorous approach considered four key input
variables for analysis: temperature, humidity, water-cement ratio, and aggregate fraction. An adaptive timesteppingmethod combined with a forward difference scheme was adopted to solve the governing equations of
chloride transport. The proposed theoretical model strongly correlates with published data, achieving a Pearson’s r value exceeding 0.90 and an error of less than 5 %. Unlike other sensitivity analysis tools, the present model effectively explained the interaction effect of assumed input variables. The findings reveal that when varied individually, the sensitivity index for humidity is 85 % higher than that for temperature, while on the interaction, the index is 58 % higher. The influence of temperature was limited during the initial period and diminished as time progressed. Further, temperature and humidity show strong higher-order interaction with a coefficient of 0.914. The proposed model shows moderate sensitivity to temperature and water-cement ratio when varied individually; however, their influence on interaction rises significantly. For a given mix design, the average daily variations in environment conditions over a year demonstrated that chloride diffusivity peaked between the peaks of temperature and humidity. The study serves as an important benchmark enabling the development of robust models for accurate predictions of concrete durability. These predictions will provide a real-world impact on implementing effective reinforced concrete infrastructure maintenance strategies.