When Fatigue Assessment Is Required

Fatigue is the progressive and localised structural damage that occurs when a material is subjected to repeated loading. In structural steelwork, fatigue manifests as crack initiation and growth at stress concentrations — typically at welds, bolt holes, and sharp changes in geometry.

Per AS 4100 Clause 13.1, fatigue assessment is required when:

1. Crane runway girders and supporting structures — the dominant fatigue-governed application in Australian steel design
2. High-cycle machinery supports — vibrating screens, crushers, fans, compressors (> 10,000 cycles design life)
3. Bridges and gantries — road and rail bridges (though AS 5100 is the primary standard for bridges, AS 4100 Clause 13 is referenced)
4. Wind-sensitive slender structures — stacks, masts, and towers subject to vortex shedding (typically > 1 million cycles per year)
5. Seismic applications — low-cycle fatigue in ductile links (not covered by Clause 13; see AS 4100 Clause 10 for seismic design)

Fatigue assessment is NOT required when:

• The number of design loading cycles N_sc < 20,000 (deemed a static problem — AS 4100 Clause 13.2)
• The stress range f* < 27 MPa for any detail category (the infinite-life threshold for welded details)
• The detail has a fatigue category of 160 or higher (parent metal and smooth transitions — effectively infinite life for normal service loading)

AS 4100 Stress Range Method — Clause 13.3

The stress range method is the core of fatigue assessment. It is based on the principle that fatigue damage is proportional to the applied stress range (f* = f_max - f_min), not the mean stress.

S-N Curves and Detail Categories

AS 4100 Table 13.3 provides detail categories (DC) numbered by their fatigue strength in MPa at 2 million cycles (the constant-amplitude fatigue limit — CAFL). Higher categories = better fatigue performance.

The Design S-N Curve

For each detail category, the fatigue life (N_sc, number of cycles to failure) is related to the nominal stress range f by:

For N_sc <= 5 x 10^6 (slope m = 3): f^3 x N_sc = (DC)^3 x 2 x 10^6

For 5 x 10^6 < N_sc <= 10^8 (slope m = 5, constant amplitude fatigue limit exceeded): f^5 x N_sc = (0.730 x DC)^5 x 5 x 10^6

Cut-off limit (infinite life): f_cutoff = 0.405 x DC (at 10^8 cycles — stress ranges below this do not cause fatigue damage)

Design Check per Clause 13.7

The fatigue check for a single stress range level: f* <= phi x f_c / gamma_f

Where:

• f* = design stress range from factored load (AS 1170.0 fatigue combination)
• f_c = nominal fatigue strength from the S-N curve for the required design life (N_sc)
• phi = 0.70 (capacity factor for fatigue — lower than static phi due to the brittle nature of fatigue failure and the statistical scatter in fatigue data)
• gamma_f = 1.0 (fatigue already accounts for variability in the S-N curves)

For constant amplitude loading (most crane applications): f_c = DC x (2 x 10^6 / N_sc)^(1/3) for N_sc <= 5 x 10^6

Palmgren-Miner Cumulative Damage Rule

For variable amplitude loading (multiple stress range levels), the Palmgren-Miner linear damage accumulation rule is used per AS 4100 Clause 13.8.2:

SUM (n_i / N_i) <= 1.0

Where:

• n_i = number of applied cycles at stress range f_i
• N_i = number of cycles to failure at stress range f_i (from S-N curve)

Equivalent constant amplitude stress range (Miner's sum format):

f_eq = [SUM (n_i x f_i^3) / SUM (n_i)]^(1/3)

The equivalent stress range f_eq is checked against the detail category for the total number of cycles N_total = SUM(n_i).

Crane Runway Girder — Cumulative Damage Example

A typical crane runway girder experiences a spectrum of load cycles:

• 20% of lifts at full capacity (f_full)
• 50% at half capacity (0.5 x f_full)
• 30% at quarter capacity (0.25 x f_full)

The equivalent stress range f_eq is:

f_eq = f_full x [0.20 x 1^3 + 0.50 x 0.5^3 + 0.30 x 0.25^3]^(1/3) = f_full x [0.20 + 0.50 x 0.125 + 0.30 x 0.01563]^(1/3) = f_full x [0.20 + 0.0625 + 0.00469]^(1/3) = f_full x [0.267]^(1/3) = f_full x 0.644

This means that despite the crane operating at partial load 80% of the time, the equivalent fatigue damage is still 64% of the full-capacity damage — because of the cubic relationship (small reductions in stress range have disproportionately large effects on fatigue life).

Crane Runway Girder — Worked Example

Problem: A 25 t overhead travelling crane operates on a simply supported runway girder spanning 8.0 m. The crane applies a factored vertical wheel load of P* = 185 kN per wheel (including dynamic allowance per AS 1418.1). Wheel spacing a = 3.6 m. Girder: 610UB101 (Grade 300). The girder has welded stiffener-to-web and stiffener-to-flange fillet welds at 1.5 m centres. Design life = 2 million crane cycles (50 years at 40 lifts/day x 5 days/week x 50 weeks/year x 50 years = 500,000 cycles — but with multiple wheels per lift, total stress cycles = 2 x 500,000 = 1,000,000 cycles. Adding an allowance for light lifts gives N_sc = 2 x 10^6 cycles).

Step 1 — Determine the maximum moment and stress range:

For the worst wheel position (both wheels on span, one wheel near mid-span): R_A = P x (L - a/2 + a/2) / L ... this is a moving load problem.

For a simply supported beam with two moving loads at spacing a, the maximum bending moment when a < L occurs when one wheel is at 0.5a from mid-span:

M_max = 2 x P x L / 4 x (1 - a/(2L))^2 = 2 x 185 x 8.0 / 4 x (1 - 3.6/(2 x 8.0))^2 = 740 x (1 - 0.225)^2 = 740 x 0.601 = 445 kN.m

Step 2 — Stress range at the stiffener-to-flange fillet weld (bottom flange, tension):

Section modulus of 610UB101: Z_x = 2,530 x 10^3 mm^3 (elastic). Nominal stress range: f* = M_max / Z_x = 445 x 10^6 / 2,530 x 10^3 = 175.9 MPa.

Step 3 — Determine the fatigue detail category:

The stiffener-to-flange fillet weld is a transverse fillet weld at an attachment. The stiffener plate is welded transversely to the beam flange. Per AS 4100 Table 13.3, this detail corresponds to:

• Detail Category 56 for a transverse stiffener welded to the flange with full-length fillet welds
• The stiffener ends at the flange, creating a stress concentration at the weld toe

Use Detail Category 56 (conservative — the stiffener is not loaded axially, but the weld transverse to the tension flange creates a fatigue-sensitive detail).

Step 4 — Determine allowable stress range for N_sc = 2 x 10^6 cycles:

At exactly 2 x 10^6 cycles, f_c = DC = 56 MPa (by definition of the detail category — the CAFL).

Design check: f* = 175.9 MPa >> phi x f_c = 0.70 x 56 = 39.2 MPa. FAIL — fatigue governs.

Step 5 — Redesign options:

The stress range of 175.9 MPa greatly exceeds the allowable 39.2 MPa for Detail Category 56. Possible solutions:

(a) Improve the detail category: By removing the stiffener-to-flange weld and using a stiffener that stops clear of the flange (web-only stiffener), the flange detail category improves to at least 90 (transverse butt weld or clean parent metal). The stiffener must then resist shear buckling through the web connection only — this requires a web depth check.

(b) Reduce the stress range: Increase the girder size. For a 610UB125 (Z_x = 3,200 x 10^3 mm^3): f* = 445 x 10^6 / 3,200 x 10^3 = 139.1 MPa — still too high for DC 56.

For a 760UB147 (Z_x = 5,000 x 10^3 mm^3): f* = 445 x 10^6 / 5,000 x 10^3 = 89.0 MPa — still exceeds DC 56 limit.

(c) Use a rolled section without transverse attachments: If stiffeners are placed only on the web (clear of the flange), and no attachments are welded to the bottom tension flange, the detail category improves to 160 (parent metal). Then:

phi x f_c = 0.70 x 160 = 112 MPa at 2 x 10^6 cycles > 89.0 MPa (for the 760UB147). OK.

OR keep the 610UB101 but eliminate ALL transverse attachments from the tension flange: f* = 175.9 MPa >> 112 MPa — still fails, the beam itself is too light.

Step 6 — Final selection:

Use a 760UB147 Grade 300 with web-only stiffeners (no stiffener-to-flange weld on the tension flange). Detail category = 160 (parent metal or continuous longitudinal fillet weld). phi x f_c = 112 MPa > f* = 89.0 MPa. OK.

Crane classification per AS 1418.1: Class C4 (heavy duty) with 2 x 10^6 cycles design life. Stiffener spacing = 1.5 m. Stiffeners fillet welded to the WEB ONLY (leave a 40 mm gap between stiffener and tension flange to avoid the transverse weld stress concentration). Web stiffener connection designed for shear buckling per Clause 5.10.

Local Stress Effects — Clauses 13.9 and 13.10

Stress Concentration Factors (SCF)

For complex details not covered by the standard categories in Table 13.3, AS 4100 permits the use of stress concentration factors determined from finite element analysis (FEA) or published literature (e.g., IIW recommendations). The hot-spot stress method:

f_hotspot = SCF x f_nominal

Where SCF accounts for the local geometry (e.g., cope holes, gusset plate terminations, cope hole profiles). Typical SCF = 1.5-3.0 for common details; values above 3.0 indicate a detail that should be redesigned.

Thickness Correction

For plates thicker than 25 mm (t > 25 mm), the fatigue strength is reduced by the factor:

k_t = (25/t)^0.25 (for butt welds and transverse fillet welds — AS 4100 Clause 13.3.2)

For a 40 mm plate: k_t = (25/40)^0.25 = 0.889 — a 11% reduction in fatigue strength.

This thickness correction accounts for the statistical size effect — thicker plates have larger heat-affected zones, higher residual stresses, and a higher probability of containing a critical defect.

Bolted Connections — Fatigue Considerations

Bolted connections generally have better fatigue performance than welded connections because:

• Bolts transfer load through bearing, avoiding the metallurgical discontinuities of welds
• The hole is typically drill-finished (better surface than flame-cut or sheared edges)
• Bolt pretension creates compressive clamping that reduces the stress range at the faying surface

Bolt Fatigue Categories

Per AS 4100 Table 13.3:

• Pretensioned bolts (slip-critical): Detail Category 100 for bolts in shear, category 71 for bolts in tension (due to the lower fatigue strength of bolt threads)
• Bearing bolts (snug-tight): Detail Category 80 for bolts in shear (slightly lower due to fretting at the bolt-hole interface)
• Bolt holes in tension members: Detail Category 63 for net section through holes (accounts for the hole stress concentration)

Crane Girder Splice — Bolted vs Welded

For crane runway girder splices, bolted connections are strongly preferred over welded splices because:

1. Bolted splices are Detail Category 100 (pretensioned) vs 56 (transverse fillet weld) — nearly double the fatigue strength
2. Bolted splices can be inspected and retightened during service
3. Fatigue cracks in bolted connections are visible (loose bolts, rust staining) before catastrophic failure; fatigue cracks in welds can propagate undetected

Fatigue Design Checklist for Crane Runway Girders

1. Determine crane classification per AS 1418.1 (Class A = light to Class F = very heavy)
2. Calculate the total number of stress cycles N_sc over the design life
3. Identify the governing fatigue detail at each critical location (bottom flange, web stiffener terminations, splices)
4. Calculate the stress range f* at each critical location for the maximum and frequently-occurring load levels
5. For variable loading, compute the equivalent stress range f_eq using Palmgren-Miner
6. Check f* <= phi x f_c for each detail category
7. Document fatigue-critical details on the fabrication drawings with special welding and inspection requirements

Frequently Asked Questions

At what cycle count does fatigue assessment become mandatory per AS 4100? AS 4100 Clause 13.2 defines the threshold at N_sc = 20,000 cycles. Below this, fatigue is deemed negligible and standard static design governs. For 20,000 to 2 million cycles, the constant amplitude fatigue limit (CAFL) applies with m = 3 slope. Above 2 million cycles, the variable amplitude threshold applies with m = 5 slope (reduced slope, more damage per additional cycle).

Why is the fatigue capacity factor phi = 0.70 lower than the static phi values? The lower phi reflects the brittle nature of fatigue failure (no plastic redistribution), the significant statistical scatter in fatigue test data (coefficient of variation typically 30-50%), and the severe consequence of fatigue failure in critical members (crane girders, bridges). The static phi of 0.90 assumes ductile behaviour with redistribution capacity — fatigue has none of this.

How do you assess an existing crane runway girder for increased load capacity? For assessment of existing crane girders: (a) Determine the accumulated damage from the historic crane usage records (if available — often they are not), (b) Assume a conservative historic usage (Class C or D for typical industrial cranes), (c) Calculate remaining fatigue life using Miner's rule, (d) Perform non-destructive testing (NDT) — magnetic particle or ultrasonic testing — of fatigue-critical details (stiffener-to-flange welds, splices, cope holes), (e) If cracks are found, assess whether crack arrest methods (stop holes, grinding, replacement) can extend the life, and (f) Implement an inspection schedule (6-monthly or 12-monthly visual plus annual NDT) if continued operation is permitted.

What is the most fatigue-critical detail in steel construction and how can it be avoided? The most fatigue-critical common detail is a transverse fillet weld on a tension flange (Detail Category 56 or lower). This covers stiffener-to-flange welds, gusset plate terminations, and cover plate end welds. It can be avoided by: (a) terminating stiffeners short of the tension flange and welding to the web only, (b) using extended stiffeners with a tapered transition (1:4 slope min) and ground-weld profile (may improve to DC 71 or 80), (c) specifying flange plates with a transition radius ground smooth (improves to DC 80-90), or (d) relocating the fatigue-critical area to the compression flange where possible (stress range is in compression, and fatigue crack propagation in compression is much slower).

This page is for educational reference. Fatigue assessment per AS 4100:2020 Clause 13. All structural designs must be independently verified by a licensed Professional Engineer or Structural Engineer registered with Engineers Australia or the relevant state registration board. Results are PRELIMINARY — NOT FOR CONSTRUCTION.

Disclaimer: This content is for educational purposes only. Results must be verified by a licensed professional engineer. Steel Calculator provides preliminary design tools — NOT a substitute for professional engineering judgment.