DEVELOPMENT OF NONLINEAR FINITE DIFFERENCE MODEL FOR CHLORIDE DIFFUSION IN CONCRETE

Abstract

Owing to problems associated with chloride-induced corrosion, the reliable prediction of chloride ingress into concrete is one of the key elements of the durability design and redesign of concrete structures exposed to chloride environments. In this thesis, a nonlinear chloride penetration model based on the finite difference approach is developed that aims to predict chloride content at the reinforcement level after a certain period of exposure. The chloride ingress into concrete is modeled using a modified Fick’s second law of diffusion, where the chloride binding is dealt with as a separate term in the model, allowing the nonlinear binding to be accounted for in the modeling process. Two finite difference models for both simplified and modified Fick’s second law equations are derived and used to model chloride ingress into concrete. The influence of time and temperature on the effective chloride diffusion coefficient is accounted for in the model. The chloride parameters, which describe the predicted chloride profiles, are obtained by fitting to the experimental chloride profiles of acid-soluble chlorides the predicted chloride profiles resulting from the nonlinear model. Chloride parameters are obtained for three binding cases: no binding, linear binding, and nonlinear binding. The results of the nonlinear model are then compared to those of the error function solution and Life-365. A parametric study is also conducted to examine the sensitivity of the predicted service life against changes in the values of key parameters of interest used in the model.

In addition, the influence of the curing temperature on the chloride diffusion coefficient and age parameter (value of m) is experimentally investigated, and the influence of chloride binding on concrete porosity and pore volume at different exposure conditions is explored. Water-soluble and acid-soluble chloride profiles obtained at different exposure periods and temperatures are obtained and investigated. The error function solution and linear finite difference model are used to evaluate the chloride parameters of both types of chloride profiles and the results are compared and discussed. The results reveal that both the diffusion coefficient and the value of m are temperature-dependent. Furthermore, binding is found to have a significant influence on concrete porosity and pore volume, especially at 22.4o C or lower. The diffusion coefficients of water-soluble chloride are higher than those of acid-soluble chloride, although a slight difference is observed in the value of m for both types. Elevated temperature is also found to significantly influence the value of m, based on the results of acid-soluble chloride profiles.

The results of the nonlinear model show that binding strongly affects the value of the effective diffusion coefficient, while the nonlinear binding relation results in the lowest value of the effective diffusion. Binding is also found to have a strong influence on the shape of free, bound, and total chloride profiles. The predicted service life notably increases when nonlinear binding is considered and drops drastically if binding is ignored. The linear binding relation still provides a reasonable estimation of service life, but on the conservative side compared to the nonlinear relation. The sensitivity results show that the value of m, temperature, and concrete cover have the greatest influences on the predicted service life with respect to the base case concrete mixture used in this study. The main contributions of this thesis are determining the effective diffusion coefficient using the total chloride profiles, and modeling nonlinear binding with the time and temperature dependencies of the effective diffusion coefficient.