Fatigue mechanics of materials from S-N curves to crack growth
Why fatigue mechanics matters
Fatigue mechanics of materials examines how repeated or fluctuating loads create damage that may not appear in a single tensile test. A component can carry a static load safely and still fail after thousands, millions, or even billions of cycles if local stress, defects, surface condition, or the operating environment supports crack initiation and growth. For engineers comparing materials properties, fatigue data are not a single value. They form a framework that connects stress amplitude, strain response, crack size, microstructure, manufacturing quality, inspection intervals, and acceptable risk.
The practical value of fatigue mechanics is that it turns the broad question, “Will this material last?”, into a set of more useful questions. What loading spectrum will the part see? Is the response mainly elastic or plastic? Are defects already present? How fast would a crack grow if it formed? How much scatter should be expected from the test method and material condition?
The fatigue process is a sequence, not one event
Fatigue failure is often described as sudden because final fracture can occur quickly. Mechanically, however, fatigue usually develops in stages. Localized cyclic slip, stress concentration, inclusions, pores, notches, machining marks, corrosion pits, or other imperfections can create favorable locations for damage. A small crack may then initiate and grow a little with each load cycle until the remaining cross-section can no longer sustain the applied load.
This sequence matters because each stage is controlled by different variables. Crack initiation is strongly affected by surface finish, residual stress, local plasticity, defect population, and microstructure. Long-crack growth is often treated through fracture mechanics parameters such as the stress-intensity factor range, ΔK, when the assumptions of linear elastic fracture mechanics are reasonable. Final fracture depends on the remaining ligament, toughness, geometry, and loading condition.
Because the dominant mechanism changes as damage evolves, fatigue results should always be read with their test conditions. Data from polished laboratory specimens under constant-amplitude loading may be useful for ranking materials, but they do not automatically represent a welded joint, a rough machined surface, a corrosive environment, or an additively manufactured part with internal lack-of-fusion defects.
Three common ways to read fatigue behavior
Fatigue mechanics is not a single calculation route. In practice, engineers move among stress-life, strain-life, and fracture mechanics approaches depending on whether the design question is controlled mainly by nominal stress, cyclic plasticity, or crack growth.
| Approach | Main output | Typical use | Important limitation |
|---|---|---|---|
| Stress-life, or S-N | Cycles to failure at a given stress amplitude | High-cycle fatigue where elastic response dominates | Often specimen- and condition-specific; less direct when large plastic strain occurs |
| Strain-life, or ε-N | Cycles to failure linked to total, elastic, and plastic strain | Low-cycle fatigue, notches, thermal cycling, and local plasticity | Requires careful strain control and material cyclic stress-strain characterization |
| Crack-growth mechanics | Crack growth rate, often da/dN versus ΔK | Damage-tolerant design, inspection planning, and remaining-life assessment | Depends on crack size measurement, stress analysis, geometry, environment, and transferability of data |
Stress-life data is useful but conditional
An S-N curve plots cyclic stress against cycles to failure. It is widely used because it is intuitive and fits many high-cycle fatigue situations where macroscopic plastic deformation is limited. ASTM E466-21 frames force-controlled, constant-amplitude axial fatigue testing for metallic specimens in air at room temperature, with the stated purpose of evaluating effects such as material, geometry, surface condition, and stress under repeated direct loading. (store.astm.org)
The condition-specific nature of S-N data is not a weakness; it is a warning label. Mean stress, stress ratio, specimen preparation, frequency, temperature, humidity, and surface condition can all shift results. When a data sheet gives an endurance strength or fatigue strength without enough context, it may not be suitable for direct design use.
Strain-life methods address cyclic plasticity
When local strain rather than nominal stress controls damage, strain-life methods become more relevant. ASTM E606/E606M-19 covers strain-controlled fatigue testing and emphasizes measurement of cyclic total strain and determination of cyclic plastic strain, making it especially important for low-cycle fatigue and cyclic stress-strain behavior. (store.astm.org)
This situation is common near notches, in thermal-mechanical loading, during start-stop service, or in components that experience overloads. A material that performs well in high-cycle elastic loading may behave differently when cyclic plasticity drives crack initiation.
Crack-growth data supports damage-tolerant decisions
Fracture mechanics asks a different question: if a crack exists, how fast will it grow under a defined cyclic load? ASTM E647-24 covers measurement of fatigue crack growth rates from near-threshold behavior to instability, with results expressed in terms of crack-tip stress-intensity factor range, ΔK. The standard also warns that residual stress and shielding effects such as crack closure can influence interpretation, especially at low ΔK and low force ratios. (store.astm.org)
This approach is central to damage-tolerant design because it connects material behavior with inspection. If the smallest reliably detectable crack is known, and if crack-growth behavior under representative conditions is characterized, engineers can estimate inspection intervals with a clearer basis than a simple pass-fail fatigue life value.
Variables that change fatigue performance
Fatigue is highly sensitive to detail because it often begins at local extremes rather than at average material conditions. The following variables deserve attention whenever fatigue data are compared across materials, processes, or suppliers.
- Stress amplitude and mean stress: Higher cyclic stress generally accelerates damage, but mean stress and stress ratio also influence crack opening and closure.
- Surface condition: Scratches, machining marks, roughness, fretting, and corrosion pits can raise local stress and promote crack initiation.
- Geometry and notches: Holes, threads, fillets, weld toes, sharp transitions, and contact edges create stress concentrations that may dominate over bulk material strength.
- Microstructure: Grain size, phases, inclusions, texture, heat treatment, and anisotropy can affect both initiation and crack-growth resistance.
- Residual stress: Compressive residual stress near a surface can be beneficial in many cases, while tensile residual stress can reduce fatigue resistance.
- Environment: Moisture, salt, hydrogen, oxidation, and elevated temperature can change crack initiation and growth mechanisms.
- Loading spectrum: Real service loads are rarely perfectly constant amplitude; overloads, underloads, dwell times, vibration, and variable frequency can change accumulated damage.
The practical conclusion is straightforward: fatigue mechanics is not just a property of a material grade. It is a property of a material in a defined condition, geometry, environment, and loading history.
How standards and handbooks help prevent misuse
Standards are valuable because they define how data is generated and reported. They do not remove the need for engineering judgment. ASTM E466-21 is useful for force-controlled axial fatigue of metallic specimens where strains are predominantly elastic. ASTM E606/E606M-19 is better aligned with strain-controlled fatigue and cyclic plasticity. ASTM E647-24 focuses on fatigue crack growth rate measurement and its fracture mechanics interpretation. Together, these documents show why one fatigue test cannot answer every fatigue design question. (store.astm.org) See also: Application.
NASA fracture-control guidance also shows the broader design context. NASA-HDBK-5010 addresses fracture control implementation for spaceflight hardware, including material selection, flaw screening, traceability, verification, and assessment of fracture-critical hardware. Spaceflight requirements are more stringent than many industrial applications, but the underlying lesson is broadly relevant: fatigue and fracture risk are managed through a combination of material data, flaw assumptions, inspection capability, stress analysis, and defined acceptance logic. (standards.nasa.gov)
For non-aerospace applications, the same reasoning can be scaled to the consequence of failure. A consumer bracket, a medical implant, an aircraft fitting, a pressure component, and a rotating shaft do not need the same control system, but none should be evaluated by an isolated tensile strength value when cyclic loading is the governing service condition.
Why new materials make fatigue interpretation harder
New material systems and processing routes can improve strength, weight efficiency, corrosion resistance, or design freedom. They can also make fatigue assessment more complex. Additive manufacturing is a clear example. NIST’s work on additively manufactured metallic components highlights fatigue-relevant issues such as defects, residual stresses, microstructure characterization, process condition, and damage-tolerant assessment. NIST/ASTM workshop findings also identified needs including better processing-structure-property understanding, process and material models, in-process monitoring, non-destructive inspection, predictive design tools, and qualification frameworks. (nist.gov)
The same caution applies to advanced coatings, surface treatments, composites, and hybrid structures. A process may improve one fatigue stage while weakening another. Shot peening, for example, can introduce beneficial compressive surface stress, but the result depends on process control and surface integrity. A coating may improve corrosion resistance while introducing defects, stiffness mismatch, or residual stress. A composite laminate may not behave like an isotropic metal because matrix cracking, delamination, fiber failure, and interface damage can interact.
For readers evaluating new materials, the most useful question is not “Is the fatigue strength high?” A better question is “Which fatigue mechanism controls the application, and has the test method reproduced it?”
A practical checklist for reading fatigue data
Before using fatigue data in a material comparison or design discussion, check whether the data answers the actual service question. This checklist can help prevent common interpretation errors.
- Identify the loading mode. Axial, bending, torsion, contact fatigue, rolling fatigue, thermal fatigue, and multiaxial loading are not interchangeable.
- Check stress ratio and mean stress. Data from fully reversed loading may not represent a tensile mean stress condition.
- Confirm specimen condition. Polished, machined, cast, forged, welded, printed, heat-treated, and coated surfaces can produce different fatigue behavior.
- Look for scatter and runout policy. Fatigue data usually show significant variability, so a single curve should not be treated as a universal boundary.
- Ask whether initiation or propagation dominates. Long-life smooth specimens may be initiation-controlled; flawed or inspected structures may be crack-growth-controlled.
- Match environment and temperature. Laboratory air at room temperature may not represent corrosive, humid, high-temperature, or vacuum service.
- Separate ranking from design approval. Screening data can help compare candidates, but design allowables and inspection intervals require more rigorous context.
This checklist is especially important for industry content, supplier literature, and comparative material tables. A concise fatigue value can be useful for early screening, but a fatigue mechanics interpretation needs the test boundary conditions.
Frequently asked questions
What is fatigue mechanics of materials?
It is the study of how materials respond to repeated loading, including crack initiation, cyclic plasticity, crack propagation, and final fracture. It connects material properties with geometry, stress history, defects, environment, and inspection assumptions.
Is fatigue strength the same as tensile strength?
No. Tensile strength describes behavior under monotonic loading, while fatigue strength describes resistance to repeated loading under specified test conditions. A material with high tensile strength may still be fatigue-sensitive if it has defects, poor surface finish, unfavorable residual stress, or low crack-growth resistance.
When should crack-growth data be used instead of an S-N curve?
Crack-growth data is most useful when a flaw may already exist or when inspection and remaining-life assessment are part of the design strategy. S-N data is often more useful for initiation-focused comparisons under defined specimen and loading conditions.
Why do fatigue test results vary so much?
Fatigue is sensitive to local features. Small differences in surface finish, inclusions, porosity, residual stress, grain structure, alignment, environment, and loading history can change where damage starts and how quickly it grows.
Can one fatigue test qualify a new material?
Usually not for critical applications. A single test can provide useful screening information, but qualification normally requires data that reflects the intended material condition, manufacturing route, geometry, environment, loading spectrum, and acceptable risk level.
