Abstract

Fatigue of structural elements has been a major cause of many catastrophic failures of steel bridges. Corrosion is considered to be an influential factor that significantly contributes to the reduction of the fatigue life of steel structures. The objective of this research is to develop a numerical tool that is capable of predicting the fatigue life of steel members and connections while accounting for the effects of corrosion. To reach this goal, the research is divided into three phases.
In the first phase, a stress-life damage accumulation model (called the VTLC, for Virtual Target Life Curves) is proposed and validated using the experimental work available in the literature. The model has the advantage of including the effect of overloading. The VTLC model was verified using a case study of a riveted railway bridge.
In the second phase, a numerical tool is developed (named CorrFLP for Corrosion Fatigue Life Predictor). It acts as an add-on and uses available FEM packages as solvers. This tool is useful in predicting the fatigue life of railway or roadway steel bridges with riveted or bolted connections. CorrFLP uses the strain-life approach along with the theory of critical distances. CorrFLP is validated using several test results in the literature with and without the effect of corrosion.
To account for corrosion, a new fatigue strain-life model based on the Smith-Watson-Topper model is proposed in the third phase of this research. The model takes into account the corrosivity of the environment, the stress level, and the corrosive behaviour of the material used. A method is proposed to standardize a testing protocol to evaluate some new material constants that describe its behaviour in corrosive environments. The proposed strain-life method is implemented in CorrFLP and the resulting fatigue life predictions matched well with reported experimental results of twenty-four steel beams subjected to various fatigue and weathering conditions.
It is seen that the developed numerical tool, along with the corrosion-fatigue strain-life based model, would help structural engineers by generating several stress-life design charts to predict the fatigue life of several commonly-used structural components while accounting for different levels of corrosion.