Effect of fatigue-initiating defect area measurement on defect size distributions and fatigue assessment of ductile cast iron
Year: 2026
Type: article
Abstract: Numerous methods have been suggested to quantify fatigue-initiating defect size on fracture surfaces, the most prevalent are based on the Murakami-Endo a r e a -parameter. However, there is an ambiguity in how to systemically determine defect areas. For instance, in literature on high-cycle fatigue of ductile cast irons the several different methods have been suggested: i) the traced contour, ii) the convex hull, iii) the minimum circumscribed circle, and iv) the minimum bounding rectangle. This work focuses on comparing and evaluating these methods by assessing the fatigue-initiating defect area distributions, and the influence on fatigue assessment using the a r e a -parameter. To this end, very high cycle fatigue data on ductile cast irons with different microstructures is used, where complex shaped defects are the root-cause for fatigue failures. It is shown that there is a significant difference in the area distributions, originating from the applied area measurement method. In addition, to enable and include fatigue assessment of high strength ausferritic ductile irons, two improved Murakami-Endo type models are proposed, which show satisfactory prediction capabilities over a wide range of ductile cast iron microstructures. To further evaluate the different area measurement methods, the suggested models are validated against ductile cast iron high-cycle fatigue data from literature having artificial defects and notches. Finally, it is concluded that the traced contour defect measurement method yields the best agreement between artificial and natural defects, and overall, the least prediction errors. • Several defect area measurement methods compared. • Extended fatigue strength prediction models for ductile cast iron. • Evaluation of influence on area distributions and fatigue assessment.
Source: International Journal of Fatigue
Authors: Max Ahlqvist, Viktor Norman, Daniel Leidermark
Institution: Linköping University
Language: English
Subfield: Mechanics of Materials
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,940
Funders: VINNOVA, Energimyndigheten
OpenAlex ID: W7124805458