Multiaxial Fatigue Testing Methods: Standards, Specimens, and Evaluation Techniques

Multiaxial fatigue testing evaluates how materials and components withstand combined cyclic loading—such as simultaneous tension, torsion, bending, and internal pressure—that mirrors real-world service conditions far more accurately than conventional uniaxial testing. Governed by internationally recognized standards including ASTM E2207, ASTM E606, ISO 1099, and ISO 12106, these methods are essential for industries such as aerospace, automotive, energy, and oil and gas. This article provides a comprehensive technical overview of multiaxial fatigue testing, covering specimen types, testing items, evaluation methodologies, and typical applications.

Technical Overview

Multiaxial fatigue refers to the degradation and eventual failure of a material subjected to cyclic loading in which more than one stress or strain component acts simultaneously. Unlike uniaxial fatigue, where a single normal stress governs crack behavior, multiaxial conditions involve combined normal and shear stresses acting on multiple planes. This produces complex stress states at critical locations, changes the dominant crack plane, and can significantly reduce fatigue life compared with predictions based on uniaxial data alone.

A fundamental distinction in multiaxial fatigue is the difference between proportional and non-proportional loading. In proportional loading, the principal stress and strain directions remain fixed and the components vary in phase with one another. In non-proportional loading, the components vary out of phase or follow complex load paths, causing the principal axes to rotate continuously during each cycle. Non-proportional loading is particularly damaging because it induces additional cyclic hardening in many ductile metals, lowers effective fatigue strength, and promotes early crack initiation along multiple planes.

Multiaxial fatigue testing is typically performed by accredited third-party laboratories using servo-hydraulic axial-torsional machines, planar biaxial frames, or combined axial-torsion-pressure systems. The governing international standards include ASTM E2207 for strain-controlled axial-torsional fatigue testing with hollow specimens, ASTM E466 and ASTM E606/E606M for axial force- and strain-controlled fatigue, ISO 1099 and ISO 12106 for metallic materials under axial force and strain control, ISO 12111 for strain-controlled uniaxial testing, ISO 1352 for torque-controlled fatigue, and ISO 12107 for the statistical planning and analysis of fatigue test data.

Testing Samples

Solid Cylindrical Specimens

Solid round specimens are used for force-controlled axial-torsional fatigue testing, particularly in the high-cycle fatigue regime where strains remain largely elastic. They are economical to machine and well suited to generating comparative S-N data under combined tension and torsion, although the stress gradient through the cross section must be considered during data reduction.

Thin-Walled Tubular Specimens

Thin-walled hollow specimens are the preferred geometry for strain-controlled axial-torsional testing in accordance with ASTM E2207 and for combined axial-torsion-internal pressure testing. The reduced wall thickness ensures a nearly uniform stress and strain distribution through the thickness, which permits accurate local strain measurement and makes these specimens ideal for low-cycle fatigue research and for the calibration of multiaxial constitutive and life models. Typical recommendations limit the ratio of wall thickness to outer diameter to keep shear stress variation across the wall below acceptable tolerance.

Cruciform Specimens

Flat cruciform (cross-shaped) specimens are employed in planar biaxial testing frames to apply independent cyclic loads along two perpendicular in-plane axes. This geometry is used to study biaxial tension-tension and tension-compression behavior, thin-sheet alloys, and welded or adhesively bonded joints under plane stress conditions representative of pressure vessel skins, panels, and shell structures.

Component-Level and Notched Samples

Actual components or notched specimens—including shafts, fillets, weldments, and fastener holes—are tested to capture the combined effects of multiaxial nominal loading, stress concentration, surface condition, residual stress, and size. Results from these tests support design validation, fatigue strength verification, and remaining life assessment of real engineering parts.

Specimen Preparation Requirements

• Machining should be performed in stages with progressively finer cuts to avoid introducing cold work or residual tensile stresses at the gage surface.
• Gage sections are typically mechanically polished to a surface roughness of 0.2 μm Ra or finer, with the final polishing performed in the axial direction to eliminate circumferential scratches that act as crack initiators.
• Dimensional verification of wall thickness, concentricity, and straightness is mandatory for hollow specimens, since wall-thickness variation produces parasitic bending and non-uniform shear.
• Specimens should be stored and handled carefully to prevent corrosion pitting, nicks, and fingerprint contamination prior to testing.

Testing Items

Multiaxial fatigue test programs can be structured to quantify a wide range of material and component behaviors:

• Proportional (in-phase) axial-torsional fatigue testing at various normal-to-shear stress or strain ratios.
• Non-proportional (out-of-phase) fatigue testing at defined phase angles, typically 30°, 60°, and 90°, to characterize additional cyclic hardening and life reduction.
• Planar biaxial tension-tension and tension-compression fatigue testing on cruciform specimens.
• Combined axial load, torque, and internal or external pressure testing on tubular specimens to simulate pipeline and pressure vessel stress states.
• Strain-controlled low-cycle multiaxial fatigue (LCF) with cyclic stabilization and hysteresis loop recording.
• Force-controlled high-cycle multiaxial fatigue (HCF) to generate S-N curves and establish fatigue limits under combined loading.
• Mean stress and mean strain sensitivity studies under multiaxial conditions.
• Crack initiation life, crack plane orientation, and early crack growth measurement.
• Variable-amplitude and spectrum multiaxial loading tests for service-load simulation.
• Runout verification at specified cycle counts, commonly 10^6 to 10^7 cycles depending on the application.

Testing Methods

Test Equipment and Control

Servo-hydraulic axial-torsional testing systems apply independent cyclic force and torque through a biaxial load frame and a combined actuator, while planar biaxial frames use four independent actuators acting on the arms of a cruciform specimen. Closed-loop digital controllers coordinate the commanded waveforms, hold the prescribed phase relationship, and monitor peak loads continuously. Deformation is measured with biaxial extensometers that capture axial displacement and twist angle simultaneously, or with non-contact digital image correlation (DIC) for full-field strain mapping on complex geometries.

Test Procedure

• Verify machine force and torque calibration, extensometer calibration, and system alignment in accordance with the applicable standard before testing.
• Perform an alignment check using a strain-gaged calibration specimen to confirm that bending and parasitic shear remain within the limits specified by the standard.
• Mount the specimen with grips that provide rigid torsional restraint while minimizing fixture-induced bending.
• Apply the prescribed cyclic waveforms—sinusoidal or triangular—under force, torque, strain, or twist control, maintaining the specified ratio of shear to normal amplitude and the specified phase angle.
• Select test frequency to avoid self-heating of the specimen; strain-controlled low-cycle tests are commonly run below 0.5 Hz, while force-controlled high-cycle tests may run at several hertz.
• Monitor hysteresis loops, peak load or strain response, and specimen surface condition throughout the test to detect cyclic softening or hardening.
• Define failure by a specified drop in stabilized load or strain response (for example, 5% or 10%), by a defined crack length, or by complete separation; record runouts that survive the specified cycle count.
• Analyze the data statistically in accordance with ISO 12107 to derive mean S-N or ε-N curves and design allowable values.

Data Evaluation and Life Prediction Models

Because no single damage parameter is universally valid, several analysis approaches are applied to multiaxial fatigue data:

• Equivalent stress-strain approaches extend uniaxial criteria—such as von Mises or Tresca equivalence—to correlate multiaxial states with uniaxial fatigue data. These methods work reasonably well for proportional loading but tend to be non-conservative for non-proportional conditions.
• Critical plane approaches assume that cracks initiate and grow on specific planes experiencing the greatest damage. Widely used models include the Fatemi-Socie model for shear-dominated failure, the Brown-Miller model combining shear strain and normal strain on the critical plane, and the Smith-Watson-Topper (SWT) parameter for tension-dominated failure. These models naturally account for phase angle effects and crack orientation.
• Energy-based approaches use the plastic strain energy density or total strain energy density per cycle as the damage parameter and are effective where both normal and shear deformation contribute significantly.
• Fractographic examination of failed specimens confirms the predicted crack initiation site, crack plane angle, and failure mechanism, closing the loop between model and observation.

Reporting of Results

A complete multiaxial fatigue test report should include the material certificate and heat treatment condition, specimen geometry and surface preparation, the applied load path with amplitude ratio and phase angle, control mode, waveform, frequency, environment, stabilized stress-strain hysteresis data, cycles to crack initiation and to failure, observed crack orientation, and the statistical analysis method applied.

Applications

Multiaxial fatigue testing supports design, qualification, and failure analysis across many safety-critical industries:

• Automotive engineering: crankshafts, drive shafts, wheel hubs, suspension arms, and steering knuckles experience combined bending and torsion throughout their service life.
• Aerospace engineering: turbine discs and blades, landing gear, wing-fuselage attachments, and fastened joints are qualified under combined thermal, centrifugal, and vibratory loads.
• Power generation: gas turbine rotors, steam turbine shafts, wind turbine main shafts and pitch bearings, and nuclear piping subjected to thermal cycling and seismic loading.
• Oil and gas: drill string components, risers, wellhead equipment, and subsea connectors undergo combined tension, torsion, bending, and pressure cycling.
• Rail transportation: wheelset axles experience combined rotating bending and press-fit fretting at wheel seats, making multiaxial qualification essential.
• Medical devices: orthopedic implants such as hip stems are loaded in combined bending and torsion during gait cycles.
• General machinery: gear shafts, couplings, springs, and welded structures benefit from multiaxial fatigue data for lightweighting and warranty prediction.

FAQ

What is multiaxial fatigue testing?

Multiaxial fatigue testing is a laboratory method in which a material or component is subjected to cyclic loading along two or more axes—such as combined axial force and torque, biaxial tension, or axial-torsion-internal pressure—to reproduce the complex stress states that occur in actual service and to measure fatigue life under those conditions.

What is the difference between proportional and non-proportional loading?

In proportional loading, all stress or strain components vary in phase and the principal directions remain fixed. In non-proportional loading, the components vary out of phase or along complex paths, causing continuous rotation of the principal axes and inducing additional cyclic hardening, which typically shortens fatigue life markedly.

Which standards govern multiaxial fatigue testing?

The principal standards include ASTM E2207 for strain-controlled axial-torsional fatigue testing with hollow specimens, ASTM E466 and ASTM E606/E606M for axial fatigue under force and strain control, ISO 1099 and ISO 12106 for metallic materials under axial force and strain control, ISO 12111 for strain-controlled uniaxial fatigue, ISO 1352 for torque-controlled fatigue testing, and ISO 12107 for statistical analysis of fatigue data.

Why are thin-walled tubular specimens preferred for research-grade multiaxial testing?

Thin-walled tubes provide an almost uniform distribution of stress and strain through the wall thickness, allowing accurate local strain measurement, straightforward separation of normal and shear response, and direct comparison of test data with critical plane and energy-based life models.

How is fatigue failure defined in a multiaxial test?

Failure is typically defined as the point at which the stabilized response changes by a specified amount—for example, a 5% or 10% drop in peak load or strain—or when a surface crack of a defined length is detected. Complete separation is also used as a criterion, and specimens surviving a specified cycle count without failure are recorded as runouts.

Can uniaxial fatigue data be used to predict multiaxial fatigue life?

Uniaxial data can be extended to multiaxial conditions only through validated models. Equivalent stress-strain methods are acceptable for proportional loading, while critical plane approaches such as Fatemi-Socie, Brown-Miller, and SWT, or energy-based parameters, are recommended for non-proportional loading because they capture the effect of phase angle and rotating principal stresses.