Crack propagation characterization and statistical evaluation of fatigue life for locally corroded bridge steel based on metal magnetic memory method

Abstract

Corrosion has a great endangerment for the fatigue performance of bridge steel members. As a newly developed self-magnetic leakage (SMFL) nondestructive testing (NDT) technique, metal magnetic memory (MMM) is regarded as a potential and feasible approach for examining fatigue damage of ferromagnetic material. Thus, the tension-tension high cycle fatigue (HCF) tests for the Q345qD bridge steel sheet specimens with local corrosion were carried out in this paper, and the normal component HSF(y) of SMFL field on the surface of specimens were investigated innovatively via MMM method. The experimental results show that the mutational crest (or trough) and the crest-trough (or trough-crest) behavior on the curves of HSF(y) signal and its gradient K can accurately indicate the range of corroded region and locate fatigue crack position. Then, the significant experimental correlations between magnetic characteristic parameters and crack size were demonstrated. The maximum value (or the minimum value) Km and the crest area Ka of the gradient curve can be utilized for the comprehensive characterization of crack initiation and propagation in the corroded region. Comparing with the experimental correlations proved that the established improved magnetic dipole models were reliable to simulate the variation of HSF(y) signal at the corroded region before and after cracking. Finally, based on the naive Bayesian model, the statistical evaluation of normalized fatigue life for corroded bridge steel was realized by classification via magnetic characteristic parameter. This research will serve as a base for future studies of fatigue crack detection and fatigue life estimation on the corroded bridge steel component based on MMM method.