steel-structure-fatigue-assessment
Steel Structure Fatigue Assessment: Stress Range & Crack Detection
Interior fatigue-inspection scene in an industrial workshop. An inspector wearing headphones runs an ultrasonic probe along the lower-flange weld toe of an overhead crane girder, one hand steadying the flaw detector whose screen shows an A-scan waveform; in the background, bridge cranes and steel columns frame a skylit roof.
Fatigue does not care about your static design check. A crane girder that passed every load case can still crack at the weld toe after 20 years of tens of thousands of lifts—because fatigue is about repetition, not peak load. Steel structure fatigue assessment estimates how many cycles an existing member has left: measure the real stress range, find the cracks early, and accumulate damage with Miner's rule—especially on crane girders.
This guide covers the stress range and S-N curve logic, the non-destructive crack-finding toolkit, Miner's linear accumulation for remaining life, and the crane girder hotspots where fatigue actually starts. Fatigue design of new members is covered in our steel structure fatigue design article; this one is about assessing a frame that has been working for decades.
Why Assess Fatigue on an Existing Frame?
Steel fails in fatigue at stresses far below yield—that is the trap. After enough repeated cycles (often millions), a tiny defect at a stress concentration grows into a propagating crack, even though no single load ever stressed the material past its elastic limit. Cracks almost always start where the section changes: weld toes, hole edges, stiffener ends, and cope cuts. The members that see the most cycles are the obvious casualties—crane girders, suspended crane runways, railroad bridges, and structures under repeated wind-induced vibration.
An existing frame earns an assessment for three reasons. It was designed to an older code that may have underestimated cycle count or live load. Its use has changed—heavier cranes, faster lifting, longer shifts—that quietly slashes the remaining life. Or the owner simply needs an answer: how much longer can this run, and do we replace it now? The difference from new design is decisive: when you design a new member you choose a detail category and an allowable stress range; when you assess an existing member you measure the stress range and hunt for cracks already started. Crane-served buildings are the natural focus—see overhead crane steel building—and the same logic transfers to bridge girders under repeated wheel load; see steel bridge design fabrication.
Stress Range & S-N Curves
In a steel structure fatigue assessment, the variable that controls fatigue is not the peak stress but the stress range Δσ = σmax − σmin. Two lifts that both peak at the same high stress are not equal: one that swings from −10 to +50 ksi (Δσ = 60 ksi) consumes far more life than one that sits at +40 to +50 ksi (Δσ = 10 ksi), even though the maximum is identical. You obtain the actual stress range either by live strain-gauge measurement (most reliable) or by finite-element analysis, then read it against an S-N curve—a plot of stress range Δσ against allowable cycles N, drawn as a straight line on log-log paper. Move Δσ up by a modest amount and the allowable cycle count drops by orders of magnitude.
Where the detail sits on that curve is set by its fatigue detail category. A smooth, ground weld toe and an as-welded, unfilled undercut toe are the same steel at the same nominal stress but live dramatically different lives. AISC 360 and AASHTO assign categories by detail type; an un-ground weld toe or an undercut drops the category and shortens life, and the stress-concentration factor at the detail must be included. Connection detailing is therefore central to fatigue life—see steel structure connection design—and the weld quality on site sets the real category achieved; see steel building welding process.
Mean stress shifts the result too: a tensile mean stress worsens fatigue (a bad combination of steady tension plus cycling), while a compressive mean stress helps close the crack. Deflection limits interact with the whole picture because excessive live-load deflection is itself a symptom of repeated demand—see steel structure deflection control. S-N curves and detail categories are tabulated in the fatigue appendix of AISC 360 Specification.
| Typical Fatigue Detail | Category (typical) | Allowable Stress Range at 2×10⁶ cycles | Notes |
|---|---|---|---|
| Plain members, no holes / welds | A | ~24 ksi (165 MPa) | Airy surface; benchmark |
| Built-up plate, continuous weld | B | ~16 ksi (110 MPa) | Standard girder body |
| Welded girder web-to-flange, ground toe | B' | ~12 ksi (83 MPa) | Grinding pays off |
| Welded stiffener end / cope | C | ~10 ksi (69 MPa) | Common crack origin |
| Un-ground transverse weld toe, undercut | E | ~4.5 ksi (31 MPa) | Worst; detail degradation |
Illustrative categories and ranges; exact values depend on the governing code and detail geometry. consult our engineers for your actual S-N data.
Crack Detection & NDT
Fatigue cracks do not announce themselves with noise until late, so inspection is scheduled rather than reactive. The practical rule is a first full inspection after roughly 10–15 years of crane service, then on a fixed cycle; before any crane upgrade (heavier or faster); and immediately if a member develops unusual noise or a sudden change in vibration.
Three non-destructive methods do the work. Magnetic particle testing (MT) magnetizes a ferromagnetic part and shows surface and near-surface cracks at the weld toe—it is the workhorse for carbon-steel crane girders. Dye penetrant testing (PT) bleeds a colored dye out of an open surface crack; it is cheaper and portable but reads only surface-open flaws. Ultrasonic testing (UT) sends sound through the metal and returns the depth and length of internal or subsurface cracks—essential for thick webs and for measuring how deep a found crack has grown. Visual plus light tap testing is the initial screen; members designated fracture-critical (their failure would be catastrophic) get closer intervals and UT coverage. Standards for these methods are gathered under ASTM Non-Destructive Testing Standards.
When a crack is found, the response is ranked by severity: drill a stop-hole and grind the toe, locally re-weld, or replace the damaged segment—and re-run the remaining-life calculation. A member with a propagating crack must not continue to be loaded at full crane capacity. Factory and third-party NDT follow the same methods; see steel structure quality inspection and steel building third-party inspection. For running girders, continuous strain or acoustic monitoring between inspections is increasingly used—see steel structure IoT monitoring.
| NDT Method | What It Detects | Depth Range | Cost Level | Notes |
|---|---|---|---|---|
| Visual + tap | Surface condition, loose parts | Surface only | Low | Initial screen |
| Dye penetrant (PT) | Surface-open cracks | Surface only | Low | Portable; any material |
| Magnetic particle (MT) | Surface & near-surface | ~0–0.1 in (0–3 mm) | Medium | Ferrous steel only; demagnetize after |
| Ultrasonic (UT) | Internal cracks & depth | Deep, thick plate | Medium–high | Measures crack depth; needs coupling |
Method chosen by flaw type: surface crack on steel weld → MT; internal depth → UT; quick surface screen → PT. Fracture-critical members get dual MT + UT.
Wondering How Many Lifts Your Crane Girders Have Left?
A girder that passed its load test can still have a crack starting at the weld toe. Tell us your crane capacity, lifts per day and years in service, and our engineers will estimate the stress range, plan the NDT, and calculate the remaining fatigue cycles.
Miner's Rule & Remaining Life
Real loading is not one stress range but a mixture—a few heavy lifts, many light ones. Miner's rule (linear cumulative damage) sums the fraction of life consumed at each level:
Σ (nᵢ / Nᵢ) ≤ 1.0
where nᵢ is the actual number of cycles experienced at stress range level i, and Nᵢ is the number of cycles the S-N curve allows at that level. Each level consumes a slice of the total; when the sum reaches about 1.0, failure is expected. In practice engineers do not wait for 1.0—they set an alarm threshold around 0.5–0.7, because real scatter, corrosion, and occasional over-lifts make 1.0 an optimistic number.
To apply it you build the load spectrum: distribution of lift weights, lifts per day, and working days per year—data that comes from crane logs and the crane manufacturer. Feeding the measured stress ranges and the spectrum into Miner's rule is the core calculation of any steel structure fatigue assessment, and it yields remaining years. Two caveats sharpen the answer. Measured stress ranges beat computed ones; strain gauges under a live lift are the gold standard. And corrosion accelerates fatigue—corrosion fatigue lowers the allowable N at every stress range—so a corroded girder consumes life faster than the clean S-N curve assumes. Corrosion and fatigue are tracked together in steel structure corrosion maintenance schedule. After an over-lift or mishap, re-assess promptly—see steel building post-disaster assessment—and keep the inspection cadence tied to steel building maintenance lifecycle.
| Stress Range Level | Actual Cycles (nᵢ) | Allowable Cycles (Nᵢ) | Damage nᵢ / Nᵢ |
|---|---|---|---|
| High (full load) | 40,000 | 200,000 | 0.20 |
| Medium | 300,000 | 1,000,000 | 0.30 |
| Low (light lifts) | 1,500,000 | 5,000,000 | 0.30 |
| Total | — | — | 0.80 |
Illustrative spectrum: cumulative damage ≈ 0.80, above the 0.5–0.7 warning threshold—remaining life is short and action (grind toes / reduce lifts / replace) is due. Values are demonstrative; your numbers come from strain measurement and crane logs.
Crane Girder Fatigue Case
The overhead crane girder is the textbook fatigue member because every wheel pass is a cycle, repeated day after day. The braking and trolley horizontal forces add an extra stress range at the connections on top of the vertical wheel load. Cracks start at a predictable list of hotspots: the end of a web stiffener (where a fatigue-prone weld terminates), the flange-to-web weld toe, flange splice plates, and the lower flange directly under the rail.
Changing the crane is the hidden life-killer. Doubling the tonnage roughly doubles the stress range, which—on the steep S-N slope—cuts life by nearly an order of magnitude. Adding a shift doubles the cycle count and halves the remaining years. The rule is simple: re-assess the girders before, not after, you change the crane. Vibration behavior also shifts with a heavier crane, so see steel structure vibration control when symptoms appear. When a girder must be replaced, the new section follows normal material-substitution discipline—see steel material substitution.
Life-extension measures are modest but real: grinding the weld toe smooth raises the effective detail category; restricting crane duty cycle and eliminating over-lifts lengthens life; and locally replacing a damaged web segment avoids full girder replacement. The building reference for crane-served structures is overhead crane steel building.
| Crane Girder Hotspot | Cause | Inspection Focus |
|---|---|---|
| Web stiffener end | Weld termination stress concentration | UT / MT at stiffener toe |
| Web-to-flange weld toe | Repeated wheel-load bending | Grind and MT; measure depth if cracked |
| Flange splice plate ends | Bolt-hole / weld-edge concentration | MT at fastener holes |
| Lower flange under rail | Wheel load, rail wear, impact | UT for flange thinning & cracks |
Hotspots ranked by frequency; fracture-critical joints get closer inspection. Always re-check before a crane capacity upgrade.
Cost & Decision
Numbers are planning ranges. A fatigue assessment including NDT runs about $2,000–$10,000 depending on the number of members. Live strain-gauge measurement runs roughly $5,000–$20,000. Weld-toe grinding for life extension is priced by linear meter; full girder replacement is priced by tonnage and section weight. The owner's decision is then three-way: continue as-is (low damage, short cycle), derate or restrict lifting (moderate damage, control demand), or replace (high damage or a propagating crack). Insurance often asks for the same documentation—see steel building insurance—and the assessment should be written to a clear engineering standard; see steel structure technical specification. Tie the recurring cadence back to steel building maintenance lifecycle.
Conclusion
Steel structure fatigue assessment is measured stress range plus crack detection plus Miner accumulation, and the crane girder is where it matters most. Remember three things: fatigue is driven by the stress range, not the peak stress; upgrading the crane forces an immediate re-assessment because life falls disproportionately; and a found crack must be acted on, not left under load. Treat Miner's sum as a warning at 0.5–0.7, never wait for 1.0. Resolve steel structure fatigue assessment with real strain data and a scheduled NDT plan before the next lift adds the cycle that starts the crack.
Need the Real Fatigue Life of Your Crane Girders?
We assess existing steel frames around real operation—measure the stress range, find the cracks with NDT, and run Miner's rule against your actual lift spectrum. Tell us your crane capacity and years in service.
🏭 Explore: Steel Factory · Steel Workshop
Case Example
A 25-year-old overhead-crane girder system in a Midwest U.S. plant, 12,000 m2 (about 129,000 ft2) served by two 30 t (33 tonne) cranes, needed an answer after the owner upgraded crane capacity. The team strain-gauged the real stress range, ran magnetic-particle NDT at weld toes and stiffener ends, and accumulated damage by Miner's rule against the S-N detail category. Measured stress range came to 75 MPa (10.9 ksi) against a 90 MPa allowable, no propagating cracks were found, and the predicted remaining life was about 18 years at the current cycle count. Weld-toe grinding extended life by roughly 40% for about 18,000 USD, far below girder replacement. Repetition, not peak load, controls; see overhead crane steel buildings and steel structure fatigue design for the detail categories behind the assessment.
Reference Links
- AISC 360 Specification for Structural Steel Buildings
- ASCE 7 Minimum Design Loads and Associated Criteria for Buildings and Other Structures
- ISO 12944 Corrosion protection of steel structures by protective paint systems
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 the difference between fatigue design and fatigue assessment?
Fatigue design picks a detail category and an allowable stress range for a new member. Fatigue assessment evaluates an existing member that has already seen millions of cycles—measuring the real stress range, inspecting for cracks, and using Miner's rule to estimate remaining life.
What controls fatigue—the peak stress or the stress range?
The stress range Δσ = σmax − σmin, not the peak stress. On an S-N plot, a small increase in stress range reduces the number of cycles to failure by orders of magnitude. Mean tensile stress worsens it; compressive mean stress helps.
How do you find fatigue cracks early?
Use MT/PT for surface cracks and ultrasonic testing (UT) for internal depth, plus visual and tap testing. Fracture-critical members get closer inspection. The first full inspection is typically after 10–15 years of crane service.
What is Miner's rule?
Miner's rule accumulates damage linearly: Σ (nᵢ / Nᵢ) ≤ 1.0, where nᵢ is actual cycles at a stress level and Nᵢ is the S-N allowable cycles at that level. Treat 0.5–0.7 as a warning threshold rather than waiting for 1.0.
What happens if I increase my crane capacity?
Raising crane tonnage roughly raises the stress range, which reduces fatigue life disproportionately—often by an order of magnitude. Adding shifts doubles the cycle count and halves life. Always re-assess the girders before upgrading the crane.
steel-structure-corrosion-inspection
steel-building-seismic-retrofit