Efficient elastic buckling assessment algorithm for steel members with random non-uniform corrosion

Abstract

This paper developed an efficient algorithm to study the elastic buckling strengths of non-uniformly corroded steel members. The algorithm is derived based on the authors’ previous work on buckling strength assessment for the steel members by further introducing random non-uniform corrosion damage to the cross-section along member. In the algorithm, the random nature of the corrosion is captured by the Monte Carlo simulations. Moreover, the non-symmetric section effects, including the noncoincidence of section shear center and centroid and Wagner effects, caused by the non-uniform corrosion are considered in the algorithm. The proposed algorithm permits more than ten million times of Monte Carlo simulations being completed efficiently, which greatly reduces the computational expense compared with the existing finite element analysis algorithms. The effects of corrosion depth, corrosion ratio, and corrosion location on the flexural buckling (FB), lateral-torsional buckling (LTB), and axial-torsional buckling (ATB) strengths of the steel member are studied. It is found that corrosion has the most adverse effect on ATB than LTB and FB strengths of the member. When the corrosion distributes more compactly through the section, its adverse effect on the buckling strength of the member could be more significant. Even for the same corrosion ratio, corrosion depth, and total corrosion region number, the reduction factor of the buckling strength can be varied by up to more than 20% due to the difference of the corrosion locations. The reduction factor of the FB strength is generally linear against the corrosion depth. While for the LTB and ATB strength, the reduction factors show nonlinear as per the corrosion depth.