ABSTRACT
The purpose of this study is to establish a probabilistic model to describe the distribution of residual cross-sectional area of corroded steel bars along the length. Steel bars are embedded in concrete blocks and then subjected to accelerated corrosion test, and the area loss ranges from 0% to 30%. After corrosion test, the corroded steel bars are taken out of the concrete blocks, and then cleaned with a sand blaster. The surface morphology of the cleaned steel bar is scanned with a 3D laser scanner, and the distribution of the residual cross-sectional area is then determined with an image processing software. Statistical analysis is used to determine the probability distribution function (PDF) of the residual cross-sectional area along the length of corroded steel bars. The PDF used to fit the histogram is mixed mode normal distribution function. The relationship between the parameters of the PDFs and the average corrosion-induced mass loss is established, and an empirical PDF is proposed to quantify the distribution of the residual cross-sectional area of steel bars along the length with different corrosion levels.
