S-N curve determination testing is the fundamental fatigue experiment used to establish the relationship between cyclic stress amplitude and the number of cycles to failure for engineering materials. This article provides a comprehensive overview of S-N curve determination experiments, covering testing principles, specimen preparation, experimental methods such as the group method and staircase method, applicable standards like ASTM E466 and ISO 1099, and typical industrial applications in aerospace, automotive, and structural engineering.
Technical Overview
The S-N curve, also known as the Wöhler curve, is a graphical representation of the relationship between the applied cyclic stress amplitude (S) and the number of cycles to failure (N) of a material under fatigue loading. S-N curve determination experiments are the cornerstone of high-cycle fatigue (HCF) evaluation, in which materials endure more than approximately 10,000 to 100,000 cycles while the nominal stress remains within the elastic range.
In a typical experiment, identical specimens are tested at several stress levels until failure occurs or a predetermined runout cycle count is reached. The resulting data points are plotted on semi-logarithmic or log-log coordinates, and the finite-life region is commonly fitted with the Basquin power-law relationship, expressed as S = Sf × (2N)^b, where Sf is the fatigue strength coefficient and b is the fatigue strength exponent. For many steels, the curve becomes asymptotic beyond roughly 10^6 cycles, defining a fatigue limit below which failure is not expected. Non-ferrous alloys such as aluminum generally show no true fatigue limit, so a fatigue strength at a specified number of cycles, such as 5×10^8, is reported instead.
Key parameters controlled during S-N curve determination include the loading mode (axial, rotating bending, plane bending, or torsion), the stress ratio R (the ratio of minimum to maximum stress, with R = -1 representing fully reversed loading), loading frequency, waveform, temperature, and environmental conditions. The experiments are governed by international standards, including ASTM E466 for axial force-controlled fatigue of metallic materials, ISO 1099 for axial testing, ISO 12107 for statistical planning and analysis, and ASTM E2948 for fatigue testing general principles.
Testing Samples
Metallic Materials
S-N curve determination is most widely performed on metallic engineering materials, including carbon and alloy steels, stainless steels, aluminum alloys, titanium alloys, nickel-based superalloys, and cast irons. Both wrought products and additively manufactured metals can be evaluated, provided the material is sufficiently uniform and representative of the intended application.
Specimen Geometry
Common specimen configurations include smooth hourglass specimens with a continuous radius in the gauge section, cylindrical specimens with a parallel gauge length, and flat or sheet-type specimens for thin materials. Notched specimens with a defined elastic stress concentration factor (Kt) may also be tested to evaluate notch sensitivity. All dimensions must comply with the relevant standard, and the minimum cross-section must be calculated accurately because it directly determines the applied stress.
Surface Preparation
Because fatigue cracks almost always initiate at the surface, specimen surface condition is critical. Machining must be followed by progressive grinding and longitudinal polishing, typically achieving a surface roughness of Ra 0.2 μm or better. Careless preparation can introduce circumferential scratches, machining-induced residual stresses, or overheating, all of which can scatter or bias the results. Specimens should be free of corrosion, burrs, and handling damage, and should be stored carefully before testing.
Non-Metallic Materials
Polymers, composites, and biomedical materials can also be characterized with S-N curves, although testing parameters such as frequency must be controlled carefully to avoid hysteretic self-heating in polymers, and composite failure criteria must be clearly defined.
Testing Items
S-N curve determination experiments typically measure and report the following items:
- Cycles to failure (N) at each controlled stress amplitude or maximum stress level
- Fatigue strength at specified lives, such as 10^5, 10^6, and 10^7 cycles
- Fatigue limit or endurance limit determination, where applicable
- Complete S-N curve and fitted Basquin equation parameters (Sf and b)
- Statistical S-N characteristics, including median, mean, and P-S-N curves at defined survival probabilities such as 50%, 90%, or 95%
- Effects of stress ratio R, mean stress, and loading mode on fatigue performance
- Effects of notch, surface finish, temperature, and corrosive environment on fatigue strength
- Runout confirmation at the specified cycle limit without detectable cracking
Testing Methods
Group Method (Multiple-Specimen Method)
The group method is the most rigorous approach for S-N curve determination. Specimens are divided into groups of three to five (or more) identical test pieces, and each group is tested at a different constant stress level. Typically four to seven stress levels covering lives from about 10^4 to beyond 10^7 cycles are used. At each level, the logarithms of the failure lives are treated statistically to obtain mean values and standard deviations, producing a statistically characterized S-N curve as described in ISO 12107. A total of 12 to 20 or more specimens is common for a reliable full curve.
Staircase (Up-and-Down) Method
The staircase method is an efficient technique for estimating the fatigue limit or fatigue strength at a specific long life, usually near 10^6 to 10^7 cycles. The stress applied to the next specimen depends on the outcome of the previous test: if the specimen fails, the stress is decreased by a fixed increment; if it survives (runout), the stress is increased. After approximately 15 to 30 specimens, statistical analysis of the response data yields the mean fatigue limit and its standard deviation.
Small-Sample Method
For preliminary screening or material comparisons with limited specimens, a small number of pieces, often three to eight, are tested at progressively reduced stress levels to outline an approximate S-N curve. This approach is economical but carries higher statistical uncertainty and should not be used for design allowables without further verification.
Runout Criteria and Validation
A runout is declared when a specimen completes the specified cycle count, commonly 10^7 cycles for steels and up to 10^8 or 10^9 cycles for non-ferrous alloys or very-high-cycle fatigue studies, without visible cracking. Runout data are censored rather than treated as failures in regression analysis. After testing, fractured specimens should be examined visually or with microscopy to confirm that failure initiated in the gauge section rather than at the grips or fillets; invalid failures must be excluded from the data set.
Equipment and Data Processing
Testing is conducted on servo-hydraulic axial fatigue machines (typically 5 to 300 kN capacity, frequencies of 1 to 100 Hz), electromagnetic resonance systems (up to approximately 300 Hz for efficient long-life testing), or rotating bending and plane bending fatigue rigs. Loads are calibrated according to ASTM E4 or ASTM E467. Data analysis involves plotting stress amplitude against log N, applying linear regression to the finite-life region in log-log space, and constructing P-S-N curves for reliability-based design.
Applications
S-N curve determination experiments provide the essential fatigue design data used across virtually every engineering industry:
- Aerospace: airframe structures, landing gear, fan and turbine disks, and fasteners evaluated for safe-life and damage-tolerance certification
- Automotive: chassis components, suspension springs, crankshafts, connecting rods, and welded body structures subjected to road-load spectra
- Railway: axles, wheels, and rails where very-high-cycle fatigue integrity is safety-critical
- Energy and pressure equipment: pressure vessels, piping, drill pipes, turbine blades, and wind turbine components designed for decades of cyclic operation
- Medical devices: orthopedic implants such as hip stems, spinal fixation devices, and stents tested under physiological cyclic loading
- Civil and offshore structures: bridge details, welded joints, mooring chains, and risers assessed against fatigue limit states
- Materials development: ranking alloys, heat treatments, surface treatments such as shot peening and nitriding, and additive manufacturing process parameters
FAQ
What is the difference between an S-N curve and an ε-N curve?
An S-N curve is a stress-based representation used for high-cycle fatigue where the nominal strains remain essentially elastic. An ε-N (strain-life) curve is used for low-cycle fatigue, where significant plastic strain occurs and the total strain range is controlled, typically following the Coffin-Manson relationship.
How many specimens are needed to determine an S-N curve?
A statistically valid S-N curve following the group method generally requires 12 to 20 or more specimens, distributed across at least four stress levels with three to five replicates each. A staircase fatigue limit determination typically requires 15 to 30 additional specimens, although fewer may be acceptable for exploratory studies.
Why do runout specimens matter?
Runouts demonstrate survival at the specified limit life and are essential for establishing the fatigue limit or the long-life portion of the curve. They are treated statistically as suspended data, not as failures, when fitting the S-N relationship.
What stress ratio should be selected?
The choice depends on the service condition being simulated. Fully reversed loading (R = -1) is the traditional baseline for material characterization, while R = 0.1 is common for tension-tension loading, and actual service spectra may require mean-stress corrections such as the Goodman or Walker models.
Does loading frequency affect the results?
For most steels in benign, ambient environments at moderate temperatures, frequencies below a few hundred Hz have negligible influence on fatigue life, allowing resonance machines to accelerate testing. However, frequency effects must be considered for polymers, materials susceptible to corrosion fatigue, or elevated-temperature conditions.