S-N curve fatigue analysis is the fundamental stress-life method used to determine how metallic and non-metallic materials withstand cyclic loading before failure. By plotting stress amplitude against the number of cycles to failure, engineers derive the fatigue strength and endurance limits required for safe, long-life structural design. This article explains the technical principles of S-N curves, specimen types, test items, standard testing methods, industrial applications, and frequently asked questions.
Technical Overview
S-N curve fatigue analysis, also called stress-life fatigue analysis, is the most widely used experimental method for characterizing how materials and structures fail under repeated cyclic loading. Fatigue is the progressive and localized accumulation of damage that occurs when a component is subjected to fluctuating stresses, often at levels far below the ultimate tensile strength of the material. The S-N curve, historically known as the Wöhler curve, graphs the applied stress amplitude or stress range on the vertical axis against the number of cycles to failure on the logarithmic horizontal axis.
The method addresses high-cycle fatigue, where stresses remain nominally elastic and failure occurs beyond roughly 10,000 cycles. For many ferrous alloys, the curve flattens into a fatigue limit or endurance limit, below which the material can theoretically withstand an unlimited number of cycles. In laboratory practice, a runout condition is defined, commonly between 10 million and 100 million cycles depending on the material and specification. The finite-life region of the curve is typically described by the Basquin power-law relationship, in which stress amplitude equals the fatigue strength coefficient multiplied by twice the number of reversals raised to the fatigue strength exponent.
Because fatigue data exhibit considerable scatter, a single deterministic S-N curve is rarely sufficient for critical designs. Statistical analysis produces P-S-N curves at specified survival probabilities, while mean stress effects introduced by non-zero load ratios are corrected using models such as the Goodman, Gerber, Soderberg, or Walker relations before the data are applied to real service loading spectra.
Testing Samples
Standard Metallic Fatigue Coupons
Machined round or flat specimens with hourglass, parallel-gauge, or waisted profiles are the reference geometry for baseline material characterization. Surfaces are fine-turned, ground, and often mechanically or electrolytically polished to remove machining marks, since surface finish strongly influences measured fatigue life.
Welded Joint Specimens
Butt-weld, cross-joint, and T-joint specimens are tested to capture the reduced fatigue performance caused by weld geometry, residual stresses, and micro-defects. Data for welded details are usually presented as nominal or structural hot-spot stress ranges against life, in line with recommendations of the International Institute of Welding.
Fasteners and Mechanical Joints
Bolts, rivets, screws, pins, and lugs are tested under representative loading conditions to establish fatigue ratings, quantify the benefit of thread rolling or shot peening, and validate preload recommendations.
Additive Manufactured Specimens
Samples produced by laser powder bed fusion, directed energy deposition, or electron beam melting require as-built versus machined and heat-treated comparisons, because porosity, surface roughness, and build orientation dominate fatigue behavior.
Polymers, Composites, and Elastomers
Viscoelastic materials show strong frequency and temperature sensitivity, and fiber-reinforced composites degrade through matrix cracking, delamination, and fiber breakage rather than a single dominant crack, requiring adapted geometries and careful data interpretation.
Full-Scale Components
Structural subassemblies such as axles, crankshafts, suspension arms, and implant devices are tested directly to verify fatigue performance with all real manufacturing effects included, which coupon tests cannot reproduce.
Testing Items
A complete S-N fatigue test program typically evaluates and reports the following items:
- Stress amplitude, stress range, maximum stress, and stress ratio R for every test.
- Cycles to failure or runout at the specified cutoff life.
- Fatigue strength at defined lives such as 100,000, 1 million, and 10 million cycles.
- Fatigue limit or endurance limit for materials that exhibit a knee point.
- Fitted Basquin parameters, including the fatigue strength coefficient and exponent.
- Mean stress sensitivity and construction of Haigh or Goodman diagrams.
- Statistical scatter bands and P-S-N curves at 50, 90, and 95 percent survival probabilities.
- Fractographic examination of initiation sites, crack propagation regions, and final fast fracture areas.
Testing Methods
Axial Stress-Controlled Fatigue Testing
Servo-hydraulic or electrodynamic machines apply fully reversed or offset tension-compression loading to axially aligned coupons, following ASTM E466 and ISO 1099 for metallic materials. Load frequencies typically range from 1 Hz to 100 Hz, and machine alignment is verified per ASTM E1012 to limit unintended bending strains.
Rotating Beam Bending Fatigue Testing
The classical Wöhler method rotates a cantilever or four-point-bending specimen under a stationary load, producing a fully reversed bending stress at frequencies up to about 100 Hz. ISO 1143 and equivalent national standards govern this configuration, which remains an economical route for establishing the endurance limits of steels.
Plane Bending Fatigue Testing
Sheet, plate, and thin welded sections are cycled in load- or displacement-controlled bending to evaluate forming, surface treatment, and joining effects under deformation modes representative of actual components.
Ultrasonic Very High Cycle Fatigue Testing
Piezoelectric systems excite specimens at approximately 20 kHz, allowing gigacycle testing within days instead of months. This method is essential for investigating interior-initiated failures that occur beyond 10 million cycles in high-strength steels and turbine alloys.
Data Evaluation and Reporting
Constant-amplitude results are reported in accordance with ASTM E468 and analyzed statistically using ASTM E739 to define confidence intervals, compare material conditions, and convert raw data into design curves suitable for durability assessment and finite element based fatigue analysis.
Applications
S-N curve fatigue analysis underpins durability qualification and certification across nearly every engineering sector:
- Aerospace airframes, landing gear, and engine disks, where safe-life and damage-tolerant philosophies rely on statistically substantiated fatigue data.
- Automotive chassis, powertrain, and suspension components validated against measured road-load spectra.
- Railway axles, wheels, and rails, continuing the original application of Wöhler's research.
- Wind turbine blades, hubs, and towers qualified for twenty-year service lives.
- Offshore platforms, risers, and ship structures subjected to wave-induced and vortex-induced cyclic loading.
- Steel bridges and civil structures designed against traffic and wind fatigue.
- Medical implants such as hip stems, spinal devices, and dental abutments certified under international fatigue standards.
- Pressure vessels, piping, and rotating machinery subject to thermal and mechanical cycling.
FAQ
What is the difference between S-N and strain-life fatigue analysis?
S-N stress-life methods address high-cycle fatigue under nominally elastic conditions, typically for lives above roughly 10,000 cycles, whereas strain-life methods characterize low-cycle fatigue involving cyclic plasticity and are governed by standards such as ASTM E606.
How many specimens are required to generate an S-N curve?
A preliminary curve can be produced with eight to twelve specimens, but statistically reliable design curves generally require fifteen to thirty tests distributed across several stress levels plus runout points, following the sample planning guidance in ASTM E739.
What is a runout in fatigue testing?
A runout is a specimen that survives the specified maximum number of cycles without failure, commonly 10 million cycles for steels and up to 100 million cycles for very high cycle applications. Runouts are treated as suspended data in statistical analysis.
Do all materials have a fatigue limit?
No. Most carbon and low-alloy steels exhibit a horizontal endurance limit after roughly 1 million cycles, whereas aluminum alloys, titanium alloys, and many high-strength steels continue to fail at progressively lower stresses even beyond 100 million cycles, which is why gigacycle testing is increasingly performed.
How is mean stress handled in S-N analysis?
Fully reversed data are corrected for tensile or compressive mean stress using equations such as Goodman, Gerber, Soderberg, or Walker, often visualized on a Haigh diagram, or the effect is measured directly by testing at several stress ratios R.
Which factors most strongly affect S-N results?
Surface finish, residual stress, material defects, specimen geometry, loading mode, mean stress, and environmental conditions such as temperature and corrosive media are the dominant influences, which is why test conditions must replicate service conditions as closely as possible.