The synergistic combination of mechanical fatigue stresses and environmental agents acting together can be more detrimental than that of the summation of the contributions of each mechanism acting separately. Major attempts to understand the contribution of the different agents (microstructure, chemical composition of environment, temperature, loading conditions, etc.) have been reported in the literature. Nevertheless, current knowledge is insufficient to address life estimation with a sound physical basis from the initiation of localized corrosion (such as pitting) to the estimation of crack propagation. Major simplifications and assumptions have been required in the development of life prediction methodologies. This paper reviews recent effort made by the different interested parties, both in academia and industry, in the development of corrosion fatigue lifetime prediction procedures. The paper mainly focuses on methodologies proposed in the literature for O&G, nuclear, energy generation and aerospace applications, dealing with pitting corrosion-fatigue (CF) damage in aluminium alloys, carbon and stainless steels. The transition of a pit into a small crack and its growth is influenced by interaction of the pit stress/strain concentration and the local environmental conditions, making the modelling of this stages of the utmost complexity. A major trend in the models reviewed in this paper is to simplify the analysis by assuming the pit (a volumetric defect) as a sharp crack, decouple the CF problem and account for the mechanical and environmental contributions separately. These procedures heavily rely on fitting experimental data and exhibit low generality in terms of application to varying system conditions. There is a clear opportunity in this field to develop mechanistically based methodologies, considering the inherent dependence of the damage mechanism on the interaction of environmental, metallurgical and mechanical features, allowing more realistic lifetime estimates and defect tolerance arguments, where pit-to-crack transition and small crack initiation stages pose a significant challenge.

Here we analyse the relationship between the monotonic and cyclic behaviour of cylindrical AlSi10Mg (CL31 AL) samples fabricated by Selective Laser Melting (SLM) to the presence of manufacturing defects (pores, voids, oxides, etc.) and the beneficial effect of post-processing-T6 and hot isostatic pressing (HIP)-treatments. Correlative Computed Tomography (X-ray tomography, optical microscopy, electron backscatter diffraction, SEM and TEM) is used to characterise the microstructure and the three-dimensional (3D) structure of fatigue samples and to shed light on the role of defects on the experimental fatigue behaviour. Pancake-shaped pores are observed in the plane of the deposited layers having a 130% higher volume fraction for the vertical layering depostion (VL) than for horizontal layered (HL) orientations, and being larger and flatter. Further, while T6 treatment had relatively little effect on reducing porosity, the HIPping reduced the pore fraction by 44% and 65% for VL and HL samples, respectively. T6 and Hipping decreased the yield stress and the ultimate tensile strength considerabl y while increasing elongation and reduction of area accordingly. Although results are not conclusive and further work is required, our results suggest that the fatigue life seems to be dominated by the presence of these crack-like (pancake-like) defects perpendicular to the loading direction such that it is better to build samples transverse to the highest fatigue loads. Both T6 heat treatment and HIPping appear to reduce the fatigue streng of the material regardless of the AM deposition scheme as they tend to enlarge and collapse pores/voids to flat crack-like defects.

No abstract

Highlights  Stress-assisted localised corrosion is simulated using cellular automata finite element approach.  Localised corrosion component of the damage is modelled using cellular automata.  Stress concentration effect of pit geometry is analysed using finite element method.  A feedback loop between the cellular automaton and finite element models allows simulation of stress-assisted localised corrosion.  Comparison of simulation results with experimental measurements show good agreement.