steel-structure-fatigue-design
Steel Structure Fatigue Design: Stress Range, S-N Curves & Detail Categories
Technical close-up of the end of a steel overhead crane girder—microscopic fatigue crack visible at the weld toe where a web stiffener meets the flange, developer dye-penetrant pink revealing the crack line, cold industrial lighting, crane runway receding in the background.
A steel beam can be strong enough for the biggest single load it will ever see—and still break years later from a million small ones. That slow, repeat-loading failure is fatigue, and it is the silent enemy of crane girders, railway bridges, and any steel structure that lives under cyclic load. The steel itself rarely yields; it simply cracks one load cycle at a time until a detail can no longer hold.
Steel structure fatigue design is not about how strong the steel is. It is about the stress range of every load cycle and the detail category of every weld and bolt hole. Get those two right and a heavy industrial frame runs for decades; get them wrong and you are looking at an 8-year-old crane girder with a crack at every stiffener end.
This guide walks through the mechanism, the stress range, the S-N curve tool, the fatigue detail categories, and the two classic fatigue-critical applications—crane girders and steel bridges. Note that this is about repeated load over decades, not the rare large event. For that, see our steel building seismic design guide.
What Is Fatigue and When Does It Matter?
A static failure happens once: one overload exceeds the material strength and the member breaks or yields. A fatigue failure never happens at one load. It happens after thousands to millions of cycles at stresses well below the yield strength—often below 30–40 % of yield. The process has three ingredients: a repeated cyclic stress, a stress concentration (a hole, weld toe, or section change), and enough repetitions to grow a crack from microscopic to critical. Most codes trigger explicit fatigue checks above roughly 10,000 cycles over the design life. That threshold is the gate at which steel structure fatigue design begins to matter.
Which structures actually need it? A standard warehouse column under dead, live, and wind load cycles almost never—and is designed by strength, not fatigue. But put a 20-tonne overhead crane in the same building and every crane travel becomes a stress cycle; run it two shifts a day and you have millions of cycles in a decade. Highway and railway bridges live under repeated wheel loads. Structures exposed to wind-induced vibration or wave loading also enter the fatigue regime.
| Structure Type | Typical Cyclic Exposure | Fatigue Check Needed? | Typical Driver |
|---|---|---|---|
| Low-rise warehouse / office | Wind gusts, a few lifts per week | No | Static strength governs |
| Heavy-duty crane building | Thousands of runway passes per day | Yes, mandatory | Crane duty class A6–A8 |
| Road / rail steel bridge | Millions of wheel passages | Yes, mandatory | Traffic loading |
| Structure under wind vibration | Large cyclic wind stress | Case by case | Slender / resonant members |
| Marine / offshore structure | Waves, vessel berthing | Yes | Wave cycles over life |
The distinction matters for scoping. Seismic design (see the steel building seismic design article) handles a rare, once-in-decades event and relies on ductility. Fatigue design handles small loads repeated many times. A building can need one, the other, or both.
Stress Range: The Real Load Variable
Here is the counter-intuitive core of steel fatigue design: the life is not controlled by the maximum stress, but by the stress range, Δσ = σmax − σmin—the difference between the most tensile and least tensile stress in one cycle. A high mean stress with a tiny range may be harmless; a modest maximum stress with a large range can be deadly.
For a crane girder, every lift produces one cycle. A full load, a half load, and an empty hook give three different stress ranges that mix into a load spectrum. Designers classify the crane by duty (CMAA Class A–F, or FEM 1Am–5m) to estimate both the number of cycles and the stress-range mix. Because a real structure sees variable ranges, cumulative damage is summed with the Miner linear damage rule: D = Σ (nᵢ / Nᵢ) ≤ 1, where nᵢ is cycles spent at range i and Nᵢ is the cycles to failure at that range. Controlling this range is the single highest-leverage move in steel structure fatigue design.
Stress concentration is where fatigue actually starts. At a bolt-hole edge, a weld toe, or a sudden section change, the local stress range can be magnified 2–3 times above the nominal value—that is the crack origin. Good fatigue design therefore avoids slots, mis-drilled holes, and abrupt flange/web transitions, and it treats every stiffener termination as a potential hot spot. For how crane loading integrates into the frame, see overhead crane steel building.
S-N Curves: The Fatigue Backbone
The S-N curve is the working tool of every fatigue engineer. On a plot of stress range S (vertical, logarithmic) versus cycles to failure N (horizontal, logarithmic), each structural detail has a near-straight line. The relationship is Δσ = C / Nᵐ, so doubling the number of cycles cuts the allowable stress range by a fixed fraction. Two things follow directly:
- Longer life means lower range. To design for 10× the cycles, you must accept a noticeably smaller stress range on the same detail.
- Each detail has its own curve. A rolled-edge member, a drilled-hole connection, and a partial-joint-penetration weld do not share one S-N line—they sit on different curves ordered by detail category.
For welded steel, there is usually no true infinite-life threshold; design is anchored to a reference life of about 2 × 10⁶ cycles. Non-welded, machined, or rolled surfaces do show an endurance limit at high cycle counts, below which life is effectively infinite. Using the curve is straightforward: fix the design life, read the allowable stress range [Δσ] for that detail category, and confirm that your factored range stays under it. The table below is an illustrative ordering; actual values come from code tables.
| Cycles to Failure N | Relative Allowable Stress Range (typical) | What It Means |
|---|---|---|
| 10⁴ – 10⁵ | High (≈ 1.0× reference) | Few cycles; strength check dominates |
| 2 × 10⁶ | Reference design point | Standard welded detail benchmark |
| 10⁷ | Lower than reference | Longer life demands smaller range |
| 10⁸+ | Lowest (welded) / flattens out (unwelded) | Unwelded members approach endurance limit |
Illustrative relative ordering only. Always read absolute allowable ranges from the AISC 360 Specification (Appendix 3, Fatigue) or EN 1993-1-9 tables for your project.
Detail Categories: Where Design Wins or Loses
The single biggest lever in steel structure fatigue design is not the steel grade—it is the detail category. Codes classify details by construction and inspection quality: AISC uses Categories A through F; European practice uses FAT classes. Higher category means a larger allowable stress range on the S-N curve—and that translates directly to more life at the same stress.
The rankings are intuitive once you think about crack initiation. Parent metal on a rolled, flame-edge-finished section (Category A) is the most fatigue-resistant. Drilled (not punched) bolt holes and smoothly ground welds stay high. Crosses of longitudinal and transverse fillet welds, partial-joint-penetration terminations, and rough thermal cuts drop to the lowest categories. The same beam in the same grade can last several times longer if the weld toe is ground and the stiffener ends run out smoothly instead of stopping on a hard edge.
For complex multi-brace intersections where welded gussets cannot cleanly frame three or more hollow sections, a cast steel node pours the geometry as one piece with smooth internal fillets that drop the stress concentration factor from 2.0–3.5 (welded plates) to 1.2–1.8; our steel cast steel node design guide covers G20Mn5 material, sand-casting process, NDT acceptance, and FAT 80–125 fatigue detail classes for cast joints.
Welding quality and inspection decide where a detail lands. Automatic submerged-arc welds with ground toes outperform field hand welds; full-penetration groove joints beat fillet welds in the same geometry. Punched holes introduce cold-work and microcracks; drilled holes do not. For the shop processes behind these decisions, see our steel building welding process guide, and for the joint logic itself, bolted vs welded steel connection. The AWS (American Welding Society) standards govern the weld quality that sets the category.
Even with the best detail category and weld quality, fatigue cracks initiate at weld toes and propagate over millions of load cycles—often before visual inspection catches them. For crane girders and other cyclic-loaded details, installing crack gauges and strain gauges at weld toes provides early warning: a 0.01 mm crack elongation triggers a warning inspection, and modal frequency drops flag stiffness loss at connection nodes. Our guide to fatigue crack detection via SHM covers how structural health monitoring complements fatigue design by tracking crack growth and stress ratios in real time.
| Detail Type (Illustrative) | Relative Fatigue Category | Key Do / Don't |
|---|---|---|
| Rolled parent metal, finished edge | Highest (A) | Good; avoid notches |
| Drilled bolt holes, smooth surface | High | Do drill, don't punch |
| Automatic groove weld, ground toe | Medium-high | Grind toe; NDT inspect |
| Transverse fillet weld termination | Medium | Run out smoothly, no hard stop |
| Partial-penetration weld end / rough cut | Lowest | Avoid; redesign or inspect |
Illustrative relative ranking; absolute categories and allowable ranges per AISC 360 Appendix 3 / EN 1993-1-9.
Typical Allowable Stress Ranges at the 2×10⁶ Cycle Reference Point
The table below gives illustrative AISC 360 Appendix 3 allowable stress ranges at the welded-design reference life of N = 2 × 10⁶ cycles. These are code-look-up values; always confirm against the current AISC 360 table and your project's actual cycle count and stress ratio.
| AISC Detail Category | Typical Construction | Allowable Δσ at 2×10⁶ cycles (MPa) | Allowable Δσ (ksi) |
|---|---|---|---|
| A | Rolled parent metal, clean flame edge | ~165 MPa | ~24 ksi |
| B | Drilled (not punched) bolt holes; smooth mill finish | ~110 MPa | ~16 ksi |
| B' | Automatic groove weld, weld toe ground flush | ~97 MPa | ~14 ksi |
| C | Transverse fillet weld termination, well-run-out | ~69 MPa | ~10 ksi |
| E | Partial-joint-penetration weld end / rough thermal cut | ~31 MPa | ~4.5 ksi |
Two companion benchmarks round out a fatigue scope. A heavy-duty crane (CMAA Class D–F / FEM 4m–5m) accumulates roughly 500,000–2,000,000 stress cycles per year, which is why the fatigue check is mandatory from the first design meeting. Stress concentrations at welded gusset joints typically raise the local stress range to 2.0–3.5× the nominal value, versus 1.2–1.8× for a smooth cast-steel node—an order-of-magnitude detail choice that often matters more than beam section size (see steel cast steel node design).
For the lowest of these welded details in particular, our guide on steel structure fatigue detail Category E works through partial-joint-penetration terminations, rough thermal cuts and stiffener-end hot spots, and the grinding, drilling and run-out details that lift a Category E connection toward Category C.
Adding a Crane or Designing a Fatigue-Critical Span?
Fatigue is won at the detail: the weld, the hole, the stiffener you almost forgot. Our engineers specify fatigue-rated details and check the stress range against the right S-N curve before fabrication. Tell us your crane duty cycle or vehicle loading.
Crane Girders and Bridges: The Two Classics
The two structures that dominate real-world fatigue practice are the crane girder and the steel bridge.
Crane girders are the canonical fatigue members of industrial buildings. Every trolley passage cycles the top flange over the rail, and the hot spots are well known: rail joints (uneven joints amplify impact and range), the connection to the brake girder, and the terminations of web stiffeners. A heavy-duty crane (CMAA Class D–F / A6–A8) demands an explicit fatigue check—it cannot be sized by static bending alone. A real-world lesson: a 20 t heavy-duty (A6) girder that cracked after roughly 8 years was re-rated not by upsizing the beam, but by grinding the stiffener weld toes and redesigning the termination to lower the hot-spot stress range. Detail, not section size, controlled the life.
Steel bridges live under wheel loads. Floor beams, stringer splices, and orthotropic deck welds are the traditional hot spots; the fatigue of orthotropic deck-to-rib welds is a modern design focus. For the broader bridge context, see steel bridge design & fabrication.
Fatigue cannot be "fixed later" the way strength can. Once a crack initiates, repairs are expensive and temporary. The countermeasures must be designed in: control the stress range, select high-detail-category details, eliminate stress concentrations, and plan periodic ultrasonic or magnetic-particle inspection during service.
| Fatigue-Critical Member | Typical Hot Spot | Countermeasure |
|---|---|---|
| Crane girder web / stiffener | Weld toe at stiffener end | Ground toe, smooth termination, fatigue-rated detail |
| Crane rail seat | Rail joint impact | Ground joints, welded rail, impact reduction |
| Bridge stringer splice | Bolt-hole / weld termination | Drilled holes, Category-E detail, NDT |
| Orthotropic deck | Rib-to-deck weld | Improved weld detail, grinding, inspection plan |
Practical Checklist
When scoping a project, answer four questions before drawings are finalized:
- Does the structure see enough cycles? Heavy-duty cranes, bridges, wave- or wind-pulsed structures, and pedestrian-induced vibration on long spans trigger fatigue checks. A normal warehouse or office frame does not.
- What is the stress range? Reduce the peak of the load spectrum wherever possible—smaller range buys disproportionately longer life.
- What is the detail category? Specify ground toes, drilled holes, smooth terminations, and full-penetration welds at every hot spot.
- Is it written on the drawings? Put the fatigue category, design cycle count, and allowable stress range into the design notes so the fabricator does not "simplify" a detail into a lower category.
The rule of thumb: do not buy fatigue life by buying stronger steel. Welded fatigue strength is governed by the detail and the range, not by Fy. Spend the effort on details and load control. That is the whole of steel structure fatigue design in one sentence.
Conclusion
Steel structure fatigue design is the discipline of surviving a million small loads, not one big one. The outcome is set by three quantities: the stress range, the number of cycles, and the fatigue detail category—read against an S-N curve. Crane girders and steel bridges are the two proving grounds, and the winner is the designer who controls the hot-spot range and specifies welds and holes the right way. Fatigue is never solved by thicker steel; it is solved at the weld toe, the drilled hole, and the smooth termination. Specify those correctly, check the range against code, and the frame outlasts its first decade comfortably. Where an existing frame is already in service, the same stress-range and detail-category questions flip into an inspection-led exercise—see our guide to steel structure fatigue assessment for hot-spot NDT, remaining-life estimation, and the decision logic between grinding-out cracks, retrofitting details, and retiring the member.
Design for a Million Cycles, Not Just One Big Load
We engineer fatigue-rated steel structures—crane girders, industrial frames, and heavy-duty systems—checking stress ranges against code S-N curves and specifying high-quality details. Tell us your loading cycles and duty class.
🏭 Explore our products: Steel Workshop · Steel Factory
Reference Links
- AISC 360 Specification for Structural Steel Buildings
- Eurocode 3 Design of steel structures (EN 1993)
- ASCE 7 Minimum Design Loads and Associated Criteria for Buildings and Other Structures
About the Author
Senior Structural Engineer
With over 20 years of hands-on experience in steel structure design and prefabricated building engineering, our in-house senior structural engineer has personally contributed to more than 500 steel building projects—including warehouses, industrial factories, aircraft hangars, agricultural buildings, and commercial structures. The focus is on translating design codes such as AISC 360, ASCE 7, and Eurocode 3 into buildable, cost-effective steel solutions that balance structural performance, fabrication efficiency, and total project cost.
Learn more about our engineering team
Frequently Asked Questions
What is fatigue in steel structures?
Fatigue is cracking caused by repeated cyclic loading at stresses well below the steel's yield strength. A beam survives a single huge load but fails after millions of smaller load cycles. It depends on three things: the stress range (Δσ = σmax − σmin), the number of cycles, and the detail category of the weld or hole where the crack initiates.
How is fatigue design different from seismic design?
Seismic design handles a rare, large, once-in-decades event and relies on ductility and energy dissipation. Fatigue design handles small loads repeated millions of times over the structure's life. They are separate checks: a warehouse may need seismic design but no fatigue check, while a heavy-duty crane girder needs fatigue design regardless of seismicity.
What is an S-N curve?
An S-N curve plots the allowable stress range (S) against the number of cycles to failure (N), on log-log axes. For a given detail category, it tells you the maximum stress range the steel can survive for a chosen number of cycles. Choose the allowed N (e.g., 2 × 10⁶), read off the allowable stress range, and check that your calculated range stays below it.
Why are crane girders and bridges fatigue-critical?
They see the most cycles: a heavy-duty crane travels thousands of times a day, and every wheel on a bridge passes a million times. Hot spots are stiffener ends, welded cover-plate terminations, and rail joints. That is why heavy-duty (A6–A8) crane girders and steel bridges require explicit fatigue checks—not just strength selection.
Does using a stronger steel prevent fatigue?
Usually no. For welded details, fatigue strength is largely governed by the detail category and stress range, not the steel's yield strength—using a higher-grade bar rarely helps. The real fixes are lowering the stress range, choosing better details (ground weld toes, drilled not punched holes, smooth terminations), and avoiding stress concentrations.
steel-structure-stability-design
steel-logistics-distribution-center