Nonlinear regression using Gaussian-Lorentzian functions to empirical modeling of convective-diffusive chloride transport in concrete

• A quali-quantitative chloride profile modelling using Gaussian-Lorentzian functions were done in this work. • For chlorides profiles that presente smooth peaks, M1 functions presented better performance, whereas, when sharp peaks are observed M2 function presented better performance. • M2 function is more influenced by the convection zone than M1 function. Reinforcement corrosion due to chlorides action is the main problem regarding the durability of reinforced concrete structures present in the marine environment. Fick's second law of diffusion is used to model chloride concentration profiles in concrete for service life analysis, however, the material's properties, allied to environmental characteristics, create, close to the concrete surface, a zone where diffusion is not the chlorides transport mechanism, but also convection. Fick's second law of diffusion neglects the convective effect resulting in a large discrepancy between modeled and measured data, affecting the service life modeling of field marine structures. This study proposes an empirical chloride transport model, contemplating the convection-diffusion zones, through two Gaussian–Lorentzian functions using nonlinear regression in chloride profiles obtained from field structures located in different marine aggressive zones exposed to the marine environment for more than 40 years. The parameters of the two functions were interpreted in order to represent the physical perspective of the convective/diffusive process of chlorides penetration in reinforced concrete field structures for realistic and assertive modeling. The results showed that chosen functions can qualitatively and quantitatively describe the chloride transport in filed concrete structures, presenting better results than methodology based on Fick's second law of diffusion.