Longitudinal Variability of cross-sectional areas of corroded prestressing wires

Corrosion of prestressing wires may lead to premature failure of prestressed concrete structures. Time-dependent reliability of corroded prestressed concrete structures strongly depends on longitudinal variability of... [ view full abstract ]

Corrosion of prestressing wires may lead to premature failure of prestressed concrete structures. Time-dependent reliability of corroded prestressed concrete structures strongly depends on longitudinal variability of cross-sectional areas of corroded prestressed wires. The correlation of average corrosion degrees of the central wire and outside wires, which are twisted to be a prestressing tendon, and the longitudinal variation of cross-sectional areas are experimentally investigated. An indicator, k, which is the coefficient to describe the relationship between the average corrosion degree of the center wire and the prestressing tendon, is introduced to describe the corrosion correlation of center wire and the whole prestressing tendon. An indicator, R, which is the ratio between the average cross-sectional area and the minimum cross-sectional area of a steel wire with a certain length, is introduced to characterize its longitudinal variation of cross-sectional areas. Statistical analysis shows that k is a stochastic variable for prestressing wires of a certain type, and it is independent on the average corrosion degree of prestressing wires. That means, the average corrosion degree of central wire and outside wires can be estimated once the average corrosion degree of prestressing wires is obtained. The indicator, R, is statistically analyzed with different element lengths. The results show that the indicator R of corroded prestressing wires can be fitted well by the Gumbel distribution, and the distribution parameters increase linearly with increases in corrosion degree. With increases in element length, the location parameter increases significantly. Because the increased element length of a corroded prestressing wire will lead to the increased likelihood of a prestressing wire coming into contact with a more adverse environment in the concrete, which can further result in the development of a larger amount and deeper penetration of corrosion along the surface of the prestressing wire. While the scale parameter is independent on the element length. The relationship between the location parameter and the element length can be characterized by a logarithmic linear correlation. Then a probability distribution model for indicator R is established. Based on this probability distribution model, static tensile load-deformation curves of corroded prestressing wires are analyzed through numerical simulation and the results agrees well with previous experimental results.