Steel Structure Factory Fasteners: High-Strength Bolts, Anchor Bolts, and Purlin Braces

1. 1Every crane pass is one load cycle — a typical fatigue analysis counts 100 passes/day × 300 days/year × 30 years ≈ 900,000 cycles, with an impact factor of 1.25 carried on the vertical wheel load entering the fatigue check (lateral force enters at 50% of full value)
2. 2Longitudinal creep of the rail relative to the girder stretches or shakes loose the hook and clip bolts — the AISC journal is explicit that these connections must be inspected and maintained on a regular basis
3. 3Bolts loosen or stretch → the rail fixing fails → the rail misaligns and starts gnawing — the field problem list also covers foundation settlement, misalignment from uneven wear, and environmental corrosion
4. 4Gnawing and misalignment worsen the wheel-load distribution → local overload → the feedback accelerates, feeding the damage back into the rail and its fixings
5. 5Crane girders are non-redundant load paths — a single fracture can collapse the runway; since the 1960s welded girders have failed in fatigue far more often than riveted ones, some after only 2-15 years of service, with cracks starting at the top-flange-to-web fillet welds and stiffener welds
6. 6Intervention point: put clip and hook bolts on a scheduled inspection-and-retighten list — the opposite of the rest of the frame, which is maintenance-free after final tightening, this is the one subsystem in the factory's fasteners that needs watching during service

The 900,000 cycles are a textbook-calculated order-of-magnitude estimate (100 passes/day × 300 days/year × 30 years); the 1.25 impact factor and the lateral force at 50% of full value follow the AISE TR13 fatigue-analysis basis.

The lateral force at about 15% of wheel load is an order-of-magnitude estimate; the per-clip lateral capacities of 21-250/30-300 kN are Gantrail full-series product data; 23G525 is the current standard atlas, replacing 05G525 and covering duty classes A1-A7 for cranes 5t-250t.

1. 1Assemblies are re-tested in batches on arrival — large hexagon heads: torque coefficient + wedge load + nut proof load; twist-off type: tightening axial force + wedge load + nut proof load; any failed batch is returned in full (a main-control item of GB 50205)
2. 2Tightening runs as initial / re-tightening / final — the initial and re-tightening torques equal 50% of the final torque; the erection preload is taken as Pc = 1.1P — the roughly 10% preload loss is compensated once, at erection, not by re-tightening during service
3. 3Most of the preload loss happens in the first hour after final tightening, stabilizes within a day or two, and stops after a month — which is why final-tightening quality checks are uniformly limited to the 1-48 h window
4. 4Inside the window: re-check the final torque and count no more than 5% of twist-off splines left intact; miss the window and the preload you measure is no longer the final-tightened state
5. 5Once accepted, main-frame joints are maintenance-free for life — the code system contains no in-service re-tightening clause: selling "maintenance-free" is correct, selling "re-tightenable" is wrong
6. 6The one exception: crane-runway clip and hook bolts are the system's only in-service exception — they go on a scheduled inspection-and-retighten list and get no maintenance-free pass

No order-of-magnitude estimates are used in this slot.

• Torque-controlled tightening converts preload as T_c = k·P_c·d — with a bad k, too low slips and too high drives into yield
• Re-test on arrival: torque coefficient mean 0.110-0.150 with standard deviation ≤0.0100 for high-strength large hexagon assemblies (JGJ 82-2011 Table 6.3.1) — any failed batch is returned in full
• Watch the metric change: GB/T 1231-2024 already switched the scatter indicator from "standard deviation" to "coefficient of variation (CV)" — accept new assemblies against the new standard, while the JGJ 82 values remain the current construction-code basis
• The three field traps: damaged or contaminated threads, privately added lubricant or anti-seize, and counterfeit parts entering the supply chain — all directly change k (field case: an A325 snapped at 1-2% under a calibrated Skidmore wrench, with over-torque / lubrication shifting k / counterfeits as the three suspected causes)
• Consequence chain: k scatter → insufficient preload → friction-surface slip → the joint switches to bearing-type action → hole-wall bearing and bolt shear → connection deformation; or preload too high → plastic elongation → preload decays or the bolt snaps outright

The 1-2% snap rate is an order-of-magnitude estimate, for trend judgment; the torque coefficient 0.110-0.150 with σ≤0.0100 and the full-batch rejection rule are standard acceptance values.

• Shear division of labor: column-base anchor bolts should not be relied on for the horizontal reaction at the column base — that force is carried by friction between the base plate and concrete (friction coefficient can be taken as 0.4) or by a shear key (GB 50017-2017 §12.7.4)
• Material and anchorage: non-load-bearing anchors should be Q235B; anchorage length should be no less than 20d; for d>40mm weld an anchor plate at the end and use an anchorage length of no less than 12d (note to GB 50017-2017 §12.7.6)
• Positioning tolerance: anchor center deviation ≤2.0mm, and no negative deviation on exposed length or thread length (GB 50205-2020 §10.2) — this is where nuts that cannot take full threads come from
• Secondary grouting must use cement-based grout (GB/T 50448-2015 Table 4.1.1): 28-day compressive strength ≥50MPa for Class I and ≥60MPa for Classes II-IV; zero bleeding; chloride <0.1%; vertical expansion ratio 0.1-0.35% at 3h
• Failure chain: shrinkage cracking of ordinary cement mortar → moisture ingress → anchor corrosion with reduced section → pullout under wind or seismic loads; placement is a three-way trade-off: embedded J-type bolts are cheap but position poorly, post-drilling positions well but may hit rebar, and a reserved pocket plus grout is economical but epoxy-grout filling is expensive

No order-of-magnitude estimates are used in this slot.

The listed standards are current public editions, cited by number and scope for procurement navigation only; acceptance values follow the current published texts. No order-of-magnitude estimates are used in this slot.