Mechanical parts fatigue testing evaluates how components behave under repeated cyclic loading, generating critical data on fatigue life, endurance limits, and crack growth behavior. By following internationally recognized standards such as ISO 1099, ASTM E466, and GB/T 3075, accredited third-party laboratories help manufacturers validate design reliability and prevent unexpected in-service failures. This article explains the principles of fatigue testing, typical sample types, test items, test methods, and common industrial applications.
Technical Overview
Fatigue is the progressive, localized structural damage that occurs when a mechanical part is subjected to repeated cyclic loading. Crucially, fatigue failure can occur at stress levels far below the ultimate tensile strength, and often below the yield strength, of the material. It is widely estimated that the majority of mechanical service failures in engineering components involve fatigue as a contributing mechanism.
Fatigue failure typically develops in three stages: crack initiation at a stress concentration site (such as a notch, fillet, keyway, weld toe, machining mark, or material defect), stable crack propagation under continued cycling, and final rapid fracture when the remaining cross-section can no longer sustain the applied load. Fatigue testing is designed to quantify each of these stages in a controlled, repeatable manner.
The central output of fatigue testing is the S-N curve (also known as the Wöhler curve), which plots applied stress amplitude against the number of cycles to failure. For many steels, the curve approaches a horizontal asymptote known as the fatigue limit or endurance limit, below which the part may theoretically survive an infinite number of cycles. Aluminum alloys and many welded structures do not exhibit a true fatigue limit, so a finite-cycle fatigue strength is specified instead.
Because fatigue data exhibit significant statistical scatter, even for nominally identical specimens, valid programs require multiple test points and statistical evaluation, typically in accordance with ISO 12107 or ASTM E739. Key factors influencing fatigue performance include surface finish, notch geometry, residual stress, heat treatment, microstructural cleanliness, mean stress, loading mode, and service environment.
Testing Samples
Standard Laboratory Specimens
Machined coupons such as hourglass-shaped round bars (typically 5–12 mm in gauge diameter) or flat plate specimens are used to characterize the intrinsic fatigue behavior of a material. Critical surfaces are finely polished in the axial direction to remove machining marks, since surface condition dominates fatigue crack initiation. Specimen geometry and preparation follow standards such as ASTM E466, ISO 1099, and GB/T 3075.
Actual Components and Sub-assemblies
Finished or near-net-shape parts—including transmission shafts, gears, coil and leaf springs, fasteners, bearings, welded joints, and brackets—are tested in their as-manufactured condition. Component-level testing captures the combined effects of manufacturing processes, surface treatment (shot peening, nitriding, carburizing), geometry, and assembly loads, and therefore correlates more directly with field performance.
Coupons Removed from Production Parts
In quality control and failure investigations, small specimens can be excised from delivered components or failed parts to verify material condition, detect process drift, or benchmark batch-to-batch consistency. Additively manufactured coupons are increasingly tested in multiple build orientations to qualify 3D-printed parts for cyclic service.
Testing Items
High-Cycle Fatigue (HCF)
Stress-controlled testing in the regime of roughly 10,000 to 10 million cycles or more, where stresses remain primarily elastic. HCF testing determines fatigue strength at specified cycle counts and the fatigue limit.
Low-Cycle Fatigue (LCF)
Strain-controlled testing at higher load amplitudes, typically below 10,000 cycles, where significant plastic strain occurs each cycle. LCF data, including cyclic stress-strain curves, are essential for components experiencing thermal transients or overload events, such as turbine disks and pressure components.
Fatigue Limit and Endurance Limit Determination
Staircase or modified staircase methods are used to statistically establish the stress level at which specimens survive a defined runout, commonly 10⁷ cycles for steels.
S-N Curve (Wöhler Curve) Generation
Full S-N curves are developed across a range of stress amplitudes, with statistical analysis to define characteristic curves at specified survival probabilities (for example, P₅₀ and P₉₇.₅).
Fatigue Crack Growth Rate Testing
Measures crack propagation rate (da/dN) as a function of stress intensity factor range (ΔK), yielding the Paris-law constants used in damage-tolerance analyses and remaining-life assessments.
Specialized and Environmental Fatigue Testing
Elevated-temperature fatigue, thermomechanical fatigue (TMF), corrosion fatigue in aqueous or saline media, fretting fatigue, and multiaxial (combined bending-torsion) fatigue address demanding service environments.
Testing Methods
Fatigue tests are performed on servo-hydraulic, electrodynamic, or resonant test systems equipped with calibrated load cells, extensometers, and environmental chambers as required. Common loading configurations include:
- Axial (tension-compression) load-controlled fatigue per ASTM E466, ISO 1099, or GB/T 3075, typically at stress ratios (R) such as -1, 0.1, or 0.5 and frequencies from a few hertz up to several hundred hertz.
- Rotating bending fatigue per ISO 1143 or GB/T 4337, in which a rotating specimen under a stationary bending moment creates fully reversed cyclic stresses—historically the classic method for determining steel fatigue limits.
- Torsional and combined bending-torsion fatigue to simulate multiaxial stress states in shafts and axles.
- Strain-controlled low-cycle fatigue per ASTM E606, ISO 12106, or GB/T 15248, with radial or axial extensometry and hysteresis loop recording.
- Fatigue crack growth testing per ASTM E647 or ISO 12108 using compact tension (CT) or middle-crack-tension specimens, with crack length monitored by compliance, optical, or potential-drop methods.
- Resonant high-frequency fatigue testing, which enables runout cycles of 10⁷ to 10⁸ to be reached efficiently for long-life qualification.
Test planning and data evaluation follow ISO 12107 and ASTM E739 for statistical design, and ASTM E1049 for cycle counting when complex service load spectra must be reduced into equivalent blocks. After testing, fractographic examination—often using scanning electron microscopy—confirms crack initiation sites and failure modes, distinguishing fatigue from overload, corrosion-assisted, or manufacturing-defect-driven failures.
A typical test report from an accredited third-party laboratory includes material certification and heat-treat condition, specimen geometry and surface preparation, stress amplitude and ratio, test frequency, environment, cycles to failure or runout, the resulting S-N data, and statistical analysis with survival probabilities.
Applications
Fatigue testing underpins design validation, durability sign-off, material selection, process optimization, and failure analysis across virtually every engineering industry:
- Automotive: crankshafts, connecting rods, suspension springs, wheels, steering knuckles, and welded body structures validated against target service lives.
- Aerospace: airframe structures, landing gear, fastener systems, and engine disks evaluated using damage-tolerance and safe-life methodologies.
- Rail transportation: axles, wheels, bogie frames, and rail welds subjected to high-cycle and multiaxial fatigue programs.
- Wind energy: blade root sections, main shafts, gearbox components, and welded tower connections qualified for 20-plus-year cyclic service.
- Oil, gas, and marine: drilling tools, risers, mooring hardware, and pressure equipment assessed for corrosion fatigue and variable-amplitude loading.
- Heavy machinery and manufacturing: gears, shafts, springs, dies, and welded fabrications in cranes, presses, and mining equipment.
- Additive manufacturing qualification: build-direction and orientation-dependent fatigue properties for 3D-printed metal parts.
FAQ
What is the difference between high-cycle and low-cycle fatigue testing?
High-cycle fatigue testing is stress-controlled and involves primarily elastic deformation at long lives (typically above 10⁴ cycles). Low-cycle fatigue testing is strain-controlled, involves significant plastic strain per cycle, and covers short lives of a few hundred to about 10⁴ cycles.
How many specimens are required to build an S-N curve?
A preliminary S-N curve may be generated with 8–15 specimens, while statistically validated curves with defined survival probabilities generally require 15–30 or more specimens distributed across stress levels, as guided by ISO 12107 and ASTM E739.
What cycle count defines a runout?
For steels, runout is commonly set at 10⁷ cycles, although 10⁸ cycles may be specified for very-high-cycle fatigue applications such as springs, bearings, and rail components. Aluminum alloys and welded joints usually require a specified finite fatigue strength rather than an endurance limit.
Can fatigue tests be performed under non-ambient conditions?
Yes. Accredited third-party laboratories routinely conduct fatigue testing at elevated or sub-zero temperatures, in corrosive media, under thermal cycling, and with combined mechanical loading to replicate real service environments.
Which standards govern mechanical parts fatigue testing?
Widely applied standards include ISO 1099, ISO 1143, ISO 12106, ISO 12107, ISO 12108, ASTM E466, ASTM E606, ASTM E647, ASTM E739, ASTM E1049, GB/T 3075, GB/T 4337, and GB/T 15248, along with sector-specific specifications for automotive, aerospace, and rail components.
How long does a fatigue test take?
Duration depends on frequency and target life. Servo-hydraulic tests at 20–100 Hz reach 10⁷ cycles in roughly one to six days per specimen, while resonant systems operating near 100–300 Hz or ultrasonic systems can reduce long-life testing to hours. Full statistical programs spanning multiple stress levels typically run from several weeks to a few months.