Crack growth resistance modeling and the fracture-mechanical fatigue limit
Thesis event information
Date and time of the thesis defence
Place of the thesis defence
OP auditorium (L10), Linnanmaa
Topic of the dissertation
Crack growth resistance modeling and the fracture-mechanical fatigue limit
Doctoral candidate
Master of Science (Technology) Kimmo Kärkkäinen
Faculty and unit
University of Oulu Graduate School, Faculty of Technology, University of Oulu Graduate School, Faculty of Technology, Materials and Mechanical Engineering
Subject of study
Engineering Mechanics
Opponent
Professor Yves Nadot, Pprime Institute
Second opponent
Professor Pavel Hutař, Czech Academy of Sciences
Custos
Docent Tero Frondelius, University of Oulu
The fatigue limit can be predicted by simulating crack growth
Metal fatigue is the single largest cause of failure in structures and machine components. With increasing use of high-strength steels and additive manufacturing, the share of fatigue failures can be expected to further increase. Over a hundred years of research have not produced sufficient solutions to the problem. The complexity and many dependencies complicate not only the transferability of fatigue test results but also understanding the phenomenon itself.
A physics-based approach to fatigue prediction may offer a solution to the transferability problem. Understanding, quantifying, and recognizing dependencies of the observation-generating process allow creating a bottom-up, physics-based prediction model. Such a model could yield accurate results even if initial parameters of the problem—such as material or loading—change.
A key to a physical prediction model may be a quantitative description of premature crack closure. The term refers to a phenomenon intrinsic to metals that resists crack growth and is used to explain many experimental observations in fatigue, such as effects of load history and environment. However, the microscale phenomenon is difficult to measure, limiting most considerations to a qualitative level. Crack closure has become a wildcard that is played whenever an experimental observation—such as crack arrest—cannot otherwise be explained. It is necessary to determine which observations—and to what degree—the phenomenon can really explain. With advances in simulation and computation, numerical modeling shows readiness for the task.
Simulation-based results of the dissertation show that crack closure is often the determining factor for the fatigue limit, and that the fatigue limit can be predicted through simulation. However, sensitivities of the prediction differ from those observed experimentally, which is explained by microstructural inhomogeneity and crack initiation resistance. A comprehensive description of crack growth resistance requires, in addition to these aspects, crack closure mechanisms caused by crack surface roughness and oxidation.
The dissertation provides significant new knowledge for its field on the physics of short cracks, which are essential to understanding fatigue. The research clarifies our conception of fatigue, forms a basis for its physical prediction, and identifies the next steps toward a complete prediction model. By incorporating the roles of crack initiation, growth, and microstructure, the research unifies the fields of materials science, fracture mechanics, and classical fatigue.
A physics-based approach to fatigue prediction may offer a solution to the transferability problem. Understanding, quantifying, and recognizing dependencies of the observation-generating process allow creating a bottom-up, physics-based prediction model. Such a model could yield accurate results even if initial parameters of the problem—such as material or loading—change.
A key to a physical prediction model may be a quantitative description of premature crack closure. The term refers to a phenomenon intrinsic to metals that resists crack growth and is used to explain many experimental observations in fatigue, such as effects of load history and environment. However, the microscale phenomenon is difficult to measure, limiting most considerations to a qualitative level. Crack closure has become a wildcard that is played whenever an experimental observation—such as crack arrest—cannot otherwise be explained. It is necessary to determine which observations—and to what degree—the phenomenon can really explain. With advances in simulation and computation, numerical modeling shows readiness for the task.
Simulation-based results of the dissertation show that crack closure is often the determining factor for the fatigue limit, and that the fatigue limit can be predicted through simulation. However, sensitivities of the prediction differ from those observed experimentally, which is explained by microstructural inhomogeneity and crack initiation resistance. A comprehensive description of crack growth resistance requires, in addition to these aspects, crack closure mechanisms caused by crack surface roughness and oxidation.
The dissertation provides significant new knowledge for its field on the physics of short cracks, which are essential to understanding fatigue. The research clarifies our conception of fatigue, forms a basis for its physical prediction, and identifies the next steps toward a complete prediction model. By incorporating the roles of crack initiation, growth, and microstructure, the research unifies the fields of materials science, fracture mechanics, and classical fatigue.
Last updated: 15.10.2025