What Are the Fatigue Properties of Steel and Why Do They Matter?
The fatigue properties of steel matter any time a part works under repeated loading, even when one load does not look serious on the drawing. If you specify shafts, springs, bridge details, pipeline steel, fasteners, welded frames, or stainless parts, yield strength and tensile strength do not tell the whole story. You also need to know how steel behaves after thousands, millions, or hundreds of millions of cycles. For more background on related material behavior, see the Properties section.
Steel is used in large volumes, so fatigue is a normal engineering issue, not a rare topic. The World Steel Association reported in World Steel in Figures 2025 that world crude steel production reached about 1,885 million tonnes in 2024, with apparent steel use of 214.7 kg per person. With that much steel in service, a small design detail, a rough surface, or a poor weld toe can turn into a field failure.
What Do the Fatigue Properties of Steel Really Mean?
Fatigue is the way steel loses strength under repeated stress. It often starts at a surface mark, notch, inclusion, weld toe, thread root, corrosion pit, or small crack. A part may pass a static load check and still fail later because cyclic stress keeps moving a crack forward. That is why fatigue has to be checked with fatigue data, not only static strength.
Cyclic Stress and Stress Range
The main value is often stress range, not just the highest stress number. A rotating shaft, for example, may see tension on one side and compression on the other side during each turn.
When the range is larger, damage builds faster. Mean stress also matters, because tensile mean stress usually makes crack growth worse.
Fatigue Strength and Endurance Limit
Fatigue strength is the stress a steel specimen can carry for a stated number of cycles, such as 106 or 107. Many steels show an apparent endurance limit in standard tests, but it should not be used as a fixed safe line for every job.
Surface condition, size, corrosion, temperature, and loading history can all lower the real limit. This is why test data should be read together with the service condition, not away from it.
Crack Initiation and Crack Growth
Fatigue life has two main stages: crack initiation and crack growth. Smooth polished specimens may spend a large part of life just forming the first crack.
In welded or notched steel, the crack-like detail may already be there before the part goes into service. In that case, the work shifts from simple material selection to damage-tolerant design.
How Are the Fatigue Properties of Steel Tested?
Useful fatigue data comes from controlled testing, not from one tensile certificate. Standards help you compare one steel with another, and they also remind you that test conditions must be close to service conditions. A polished lab bar can give a clean reference value, but it is not the same as a welded bracket working in road salt.
ASTM E466 Axial Fatigue Testing
ASTM E466-21 covers force-controlled, constant-amplitude axial fatigue tests for metallic materials, mainly for cases where strain stays mostly elastic. ASTM says this test can study effects from material, geometry, surface condition, and stress.
The same standard also notes that the data fits design work only when specimen conditions reasonably match service. That point is easy to overlook when a buyer only asks for a fatigue number.
ISO 12107 Statistical Planning
Fatigue scatter is normal in steel testing. Two specimens from the same steel can fail at different cycle counts, even when the test setup is well controlled.
ISO 12107:2012, confirmed in 2019, gives methods for fatigue test planning and statistical analysis. Its purpose is to determine fatigue properties with high confidence while still using a practical number of specimens.
ASTM E647 Crack Growth Measurement
ASTM E647-24 deals with fatigue crack growth rate data. It notes that physically small cracks can be less than 1 mm, and that near-threshold data from long cracks may give non-conservative life estimates for small cracks.
That warning is short, but it matters in safety-critical parts. Small cracks do not always behave like long cracks, so the wrong data can make remaining life look better than it is.
Which Factors Change Steel Fatigue Performance Most?
You rarely buy fatigue life from the steel grade alone. The same alloy can behave very differently after machining, welding, heat treatment, coating, shot peening, or exposure to hydrogen. In shop terms, fatigue often starts where the surface or detail is least forgiving.
Surface Finish and Residual Stress
Polished steel often lasts longer than rough-machined steel in high-cycle fatigue. Compressive residual stress can help because it works against crack opening, while tensile residual stress can make fatigue worse.
NIST reported in 2006 that a pipeline steel study examined compressive residual stress at the outer pipe surface. This shows why surface condition belongs in the fatigue discussion, not only in final inspection.
Welds, Notches, and Sharp Geometry
Sharp corners raise local stress. Weld toes, undercuts, holes, keyways, threads, and stamp marks can do the same thing.
In many fabricated steel structures, detail category can matter more than base metal strength. That is why bridge, crane, and pressure equipment rules spend so much time on geometry and weld quality.
Microstructure, Heat Treatment, and Cleanliness
Grain size, inclusions, hardness, retained austenite, decarburization, and case depth can shift fatigue results. These details may look like mill or heat-treatment issues, but they can decide how a cyclic part performs.
ASTM E466 specifically lists variables such as hardness, cleanliness, grain size, composition, directionality, surface residual stress, and finish as factors to control when comparable data is needed. The list is long because fatigue is sensitive to small changes.
How Do Different Steel Applications Use Fatigue Data?
Steel fatigue decisions are not the same for a bridge girder, a gas pipeline, and a stainless rotating component. The basic science is related, but the design checks are different. Start with the load history, then choose the right data type: S-N fatigue strength, strain-life data, crack growth rate, or a code-based detail category.
Bridge Details and Stress Cycles
The FHWA Design and Evaluation of Steel Bridges for Fatigue and Fracture manual explains AASHTO fatigue detail categories, finite life, infinite life, stress range, and remaining life checks. This is practical bridge work because truck passages keep adding stress cycles year after year.
In one FHWA steel girder example, a 75-year fatigue calculation with 3,000 single-lane ADTT and 1.5 cycles per truck passage gave 123,187,500 cycles. A number like that shows why a small detail at the start can become important later. See also: Application.
Pipeline Steels and Crack Growth
A NIST 2006 study tested six pipeline steels from Grade B to X100. It found similar fatigue crack growth rate behavior across most of the stable crack growth regime, with only minor threshold differences.
Larger differences appeared near final crack growth and failure. The useful lesson is simple: grade strength alone does not tell the whole fatigue story.
Stainless Steel and Surface Engineering
ASM International published a 2017 example on 316 stainless steel round bar tested in tension-compression cycling. Low-temperature carburized specimens showed about two orders of magnitude longer fatigue life at the same maximum stress.
The stress at 107 cycles rose from about 200 MPa to 350 MPa. For some stainless parts, surface engineering can be a real fatigue tool, not just a finishing step.
How Should You Select Steel for Better Fatigue Life?
Better fatigue life starts with matching steel, process, geometry, and inspection plan to the real load case. A higher tensile grade can help in some machined parts, but it will not fix a poor weld detail or a corrosion pit. This sounds basic, yet it still comes up in purchasing and supplier review meetings.
Match the Steel to the Load History
For high-cycle elastic loading, use S-N data and mean-stress correction methods accepted in your design code. This fits parts where the stress cycles repeat many times and local plastic strain is not the main issue.
For low-cycle loading with local plastic strain, strain-life data may be a better fit. For known flaws, crack growth analysis is often more useful than a plain endurance limit value.
Control the Manufacturing Route
Ask for the process details that affect fatigue: heat treatment, hardness range, surface finish, grinding direction, weld procedure, peening, coating, and inspection. These items are not just paperwork because they change the surface and the stress state.
If a supplier changes one of these, the fatigue result can change even when the chemical composition still meets the grade. This is why fatigue-critical orders need process control, not only grade control.
Design Away from Crack Starters
Use generous radii, smoother transitions, cleaner weld profiles, proper bolt preload, and corrosion control. These choices reduce the places where a crack can start under repeated loading.
If you must keep a notch or hole, treat it as a fatigue feature, not a drafting detail. A small burr at a keyway can matter more than a few extra megapascals on a datasheet.
What Common Mistakes Should You Avoid?
The biggest fatigue mistakes are usually basic ones. People use tensile strength as a shortcut, copy lab data into dirty service conditions, or ignore small cracks because the part still looks strong. Steel may give warning signs if you inspect the right areas, but fatigue often leaves little time near the end.
Using Tensile Strength as Fatigue Strength
Ultimate tensile strength is measured under a steadily increasing load. Fatigue strength is measured under repeated loading.
They are related in some steels, but they are not the same property. Treating them as interchangeable can make the spreadsheet look clean while the part is still at risk.
Ignoring Environment and Hydrogen
Environment can change crack growth quickly. NIST reported in 2013 that API 5L X52 and X100 pipeline steels had significantly higher fatigue crack growth rates in pressurized hydrogen gas than in air, with tests at 1.7, 7, 21, and 48 MPa hydrogen pressure.
If hydrogen, seawater, or corrosion is present, use data that matches that condition. Air test data is not enough for those service cases.
Forgetting Inspection and Scatter
Fatigue data has scatter, and real parts carry defects. ISO 12107 exists partly because statistical planning matters.
For critical parts, pair material choice with inspection methods, acceptance limits, and replacement rules. A design that depends on perfect steel forever is not really a design.
FAQ
Q1: What Are the Most Important Fatigue Properties of Steel? A: The most useful values are fatigue strength at a stated cycle count, endurance limit when it applies, S-N curve shape, crack growth rate, threshold stress intensity range, and sensitivity to mean stress, surface finish, notches, welds, and environment.
Q2: Does Steel Always Have an Endurance Limit? A: Many steels show an apparent endurance limit in standard high-cycle fatigue tests, but service conditions can reduce it. Corrosion, size, rough surfaces, welds, tensile residual stress, and very long cycle life can make a simple endurance-limit assumption unsafe.
Q3: Is Higher Strength Steel Better for Fatigue? A: Sometimes, but not always. Higher strength may help smooth machined parts, yet it can be more notch sensitive and less forgiving around defects. In welded structures, detail geometry and weld quality often control fatigue life more than base metal strength.
Q4: Why Do Welded Steel Parts Often Fail in Fatigue? A: Weld toes, undercuts, residual stress, lack of fusion, and abrupt profile changes create local stress concentration. Repeated loading can start cracks at these locations even when the nominal stress in the member looks acceptable.
Q5: What Data Should You Ask a Steel Supplier for? A: Ask for grade, heat treatment, tensile data, hardness, cleanliness if relevant, surface condition, fatigue test method, S-N curve or fatigue strength data, crack growth data for critical applications, and any limits tied to temperature, corrosion, hydrogen, or welding.
