Parametric Study of Pitting Corroded Steel Bars for Reinforcement Concrete Beams using Statistical Analysis

Corrosion of steel reinforcement has a complex process which leaded the reduction of crass section bars and degradation of concrete structure. the corroded reinforcement bars were the most important issue of the concrete permanence in the marine structures. The important effects of corrosion are in the damages made by corroded concrete structure which include all kinds of structural and non-structural damages. The first one is more important because of the reduction of the safety factor of the structure against the applied external loads. These failures include the reduction of cross-section bars and the changes in the steel mechanical behavior. Corrosion includes two processes: (1) corrosion initiation and corrosion propagation. The initiation time is when the corrosive ions receive on the surface bars and lead to activation of the steel. (2) The propagation time is when the structure loses its capability subject to the loss of the cross-sectional area of reinforcing steel bars, reduction of bond and crack initiation and propagation. The predicative models of life-service of a reinforced concrete structures should be included the two processes of corrosion. Because for new structures, the initiation time of corrosion and insurance from the long-time of corrosion initiation time, and for the existing structures, controlling the corrosion propagation is more important.
In this paper, statistical characteristics of the chloride diffusion coefficient, corrosion initiation time and corrosion rate including the best probability distribution function and its parameters are investigated based on Mont Carlo simulation of pitting corrosion data. The distribution function parameters of the corrosion variables i.e. the chloride diffusion coefficient, corrosion initiation time and corrosion rate were calculated using the Maximum likelihood method based on mathematical pitting corroded model that corrosion initiation and corrosion propagation processes are considered in this model. the probability density functions such as: Gamma, Gumbel, Lognormal, Normal, and Weibull were used in the statistical analyses of corroded pitting parameters. The best probability distribution function was selected using chi-square statistic. the Lognormal distribution function was obtained the best probability function for the threshold chloride concentration, the corrosion initiation time and the corrosion rate. The corrosion initiation time depend on four basic random variables such as: the concrete compressive resistance, the concrete cover, the threshold chloride concentration and the surface chloride concentration. Thus, their statistical effects of these random variables on corrosion initiation time are parametrically investigated using 10000 Mont Carlo simulations. It is obvious that increasing the concrete resistance leads to increasing the corrosion initiation time and standard deviation of the density function. The concrete physical and mechanical characteristics are effective variables on the corrosion initiation time but the threshold chloride concentration and the surface chloride concentration are insensitive variables on the mean of corrosion time but lead to significant changes in standard deviation of the corrosion time. Finally, Various diameters bars such as: 8, 12, 16 bars and 20 were investigated in time-depended area of the corroded steel of concrete beams. Results illustrated that and the cover depth is important variable in corroded crass section of bars. Also, increasing the bar diameter and decreasing the corrosion time period were leaded to reduce the rate of the crass section bars.