Stochastic Analysis of Chloride Profiles in Concrete Structures

Abstract

Chloride attack has been identified as one of the most important factors initiating the corrosion of reinforcing bars. Chloride penetration into the concrete is generally recognized as a diffusion-controlled process. Material properties of the concrete structures are widely recognized that it is of intrinsic randomness, and it is an urgent task to develop a stochastic analysis method for chloride diffusion. Firstly, a deterministic boundary element method with new time marching scheme is developed and described. The diffusion coefficient is described as the random field and discretized based on the local average theory. In order to conduct a stochastic analysis, combining the Monte-Carlo simulation technique with the proposed BEM, the Monte Carlo stochastic BEM for stochastic analysis is proposed in this paper and used in numerical examples. Stochastic analysises presented in this paper provide basis for the reliability analysis of the durability of the concrete structures.

Simulation Approach for Random Diffusion of Chloride in Concrete under Sustained Load with Cellular Automata

“…In the erosion, the influence coefficient (α 0 and β k ) used to determine the cell evolution rule will not be a fixed number but a new variable (j 0 and l k ) determined by the random probability density distribution f (y). For this reason, the relationship between the influence coefficients before and after the improvement can be simply expressed by Equation (19).…”

“…Considering random-effect and time-varying characteristics of chloride ion diffusion in concrete, and the effect of stress level on chloride ion diffusion, the relationship between the improved influence coefficients and erosion time can be described by Equation (19).…”

“…The random variable y was calculated according to the triangular distribution function. The time step varies with time and random variable y, and its value was defined according to Equations ( 14) and (19). According to the above parameters, chloride ion concentration distribution in a concrete cross-section after 0.5, 1, 2, 3, 6, and 8 years was randomly predicted.…”

“…In addition to the above model, some service life models, such as the mathematical statistics method [10], meshless Galerkin method [11,12], finite element method [13,14], artificial neural network model [15,16], and Novel Machine Learning Techniques [17] have been used for the prediction of the distribution of chloride concentration in concrete. Simultaneously, to facilitate the description of the uncertainty and randomness of the chloride diffusion process in concrete, some scholars have proposed a series of probabilistic prediction models for describing the chloride diffusion process in concretes by taking the randomness of the main calculation parameters into consideration [18][19][20]. In addition, considering the stringent requirements for boundary conditions when solving based on Fick's second law, some scholars have proposed that Fick's second law is solved by a new numerical algorithm called cellular automata, which could be well implemented for the solution of chloride diffusion problems with complex boundary conditions.…”

Probabilistic chloride diffusion modelling in cracked concrete structures by transient BEM formulation

“…Nevertheless, the computational effectiveness of BEM enables a probabilistic description via MCS. Very few studies have adopted this coupling scheme [29], [30], [34], thus motivating this research, which deals with the probabilistic modelling of chloride ingress in concrete pores.…”