Browsing by Author "Marquis, Gary, Prof., Aalto University, Finland"

Now showing 1 - 2 of 2

The influence of interacting small defects on the fatigue limit of steels - Vuorovaikuttavien pienten vikojen vaikutus terästen väsymisrajaan
(2020) Åman, Mari
School of Engineering | Doctoral dissertation (article-based)
Metallic engineering components possess numerous small natural defects, which can potentially become sites for fatigue crack initiation. Although the influence of a single small crack on the fatigue strength is well studied, knowledge of the behavior of interacting defects is more limited. Current design rules and standards for the interaction effect, such as British standard BS 7910, are based on material-independent criteria. Consequently, this thesis examines experimentally the effect of the material on defect interaction. In addition, this work introduces an efficient finite element-based method for stress intensity factor (SIF) evaluation of interacting arbitrarily shaped 3D cracks. In experimental investigations, fully-reversed tension-compression fatigue tests were performed using different steel grades. Small artificial defects were manufactured onto the specimen surface with an accurately determined distance between them. Fatigue limits were determined by the non-propagating condition of cracks that had emanated from artificial defects. The size of natural non-propagating cracks (NPC's) varies among different materials. Thus, the influence of the material's effect is determined on both NPC characteristics and defect interaction on the fatigue limit. In order to evaluate SIF's of arbitrarily shaped 3D cracks, this thesis develops the Stress Component Division Method (SCDM), which utilizes the theory of elasticity, the superposition principle, and stress component division. The key point of the SCDM is that it separates the total stress in an element into two components, singular and non-singular terms, the former of which is associated with the SIF. This study showed that the behavior of interacting defects varied greatly among different steels. In low strength steel, the defects coalesced at the fatigue limit regardless of the spacing between them, whereas defects never coalesced at the fatigue limit in high strength steel regardless of the spacing between the defects. In case of moderate strength steels, the interaction phenomena were more complicated and the coalescence at the fatigue limit depended on the materials hardness, the relative size of NPC's, microstructure and the spacing between original defects. The numerical SIF solutions for arbitrarily shaped 3D cracks computed by SCDM are in good agreement with the known numerical solutions even when a coarse mesh is employed. It was concluded that, by itself, the commonly used material-independent interaction criterion is insufficient in the evaluation of fatigue strength for interacting crack problems in different steel grades. The experimental results presented here provide new information and guidance to improve the interaction criteria in real engineering structures. It was also shown that in addition to the aspect ratio of the defects and the spacing between defects, the defect shape must also be considered in SIF evaluation of interacting defects.
Interaction of non-metallic inclusions, microstructure, and fatigue loading with small crack growth in high-strength steels
(2017) Roiko, Andrew
School of Engineering | Doctoral dissertation (article-based)
The increased demand and requirements for high-strength steels drives the need to better understand and predict the fatigue endurance and crack growth challenges related to their use in critical machine components. Non-metallic inclusions or defects in the steel become increasingly important as the hardness or strength of the steel increases. The distribution and the ability to predict the largest inclusion that causes failure is crucial for the proper and successful design and production of the components. The extreme value distribution is effective in predicting the maximum inclusion in a volume of steel. The proper prediction and use of inclusion data gathered from polished specimen as well as differences in anisotropy are important to consider when gathering data for use in design and prediction of fatigue life or failure. The difference in non-metallic inclusions and the forging direction affects the distribution of the size of the inclusions as well as the fatigue endurance limit and its scatter of the steel. The extreme value distributions combined with the Murakami-Endo model are used as a design approach for fatigue failure for components with ultra-long fatigue lives and step loading. This design approach uses the master curve for Optically Dark Area (ODA) growth obtained by Murakami et al. and combines it with the prediction of the largest non-metallic inclusion along with the estimate of the fatigue life of the component. The initiation and growth of small cracks from inclusions as well as small Focused Ion Beam (FIB) notches behave in a similar manner and show a strong tendency to follow the local microstructure. The effect of the local microstructure on the small fatigue crack growth is studied using FIB milling to create cross-sections of the microstructure. This showed that the microstructure is also linked to the formation of ODA around non-metallic inclusions in ultra-long fatigue. The behaviour of a small crack growing from notches in high cycle fatigue is studied by using high-speed microscopy and Rumul fatigue testing machines. The test results show that small cracks initiate and grow quickly in the beginning of the fatigue life after which they propagate slowly at a stress intensity range lower than the large crack growth threshold until it is reached. Comparing different data results for different R-ratios shows that the parameter ΔK+ works well to compare the crack growth rate of small cracks in the studied quenched and tempered steel. For crack arrest the ΔK+ or Kmax thresholds are lower for cracks with higher compressive loading. Also test results showed that increasing only the compressive portion of loading can reinitiate arrested small cracks. Finally, the ΔK+ or Kmax values for small cracks are lower for similar crack growth rates under larger compressive loads.