A wind turbine bolt that was correctly torqued at installation should not loosen in service — but it frequently does. The causes are specific and well-understood; the fix depends on correctly diagnosing which mechanism is operating, because different causes require different solutions.
Tower flange bolts loosen through three distinct mechanisms: embedding relaxation (mating surface roughness flattens under preload, reducing clamping force within the first 72 hours), vibration loosening (cyclic transverse loads reduce thread engagement over months to years), and corrosion creep (rust jack effect under thread engagement adds uncontrolled preload loss). Each mechanism requires a different response: re-torque at 6 months addresses embedding; Nord-Lock washers or anaerobic adhesive address vibration; material upgrade and coating improvement address corrosion creep.
- Best for
- O&M engineers diagnosing the mechanism of bolt loosening on a specific tower section and selecting the appropriate corrective action
- Not suitable for
- Re-torquing corroded bolts as the sole remediation — if corrosion is the mechanism, re-torquing without material upgrade will produce the same result within the next inspection interval
- Diagnosis steps
- 1 — Check timing of loosening (days/months/years post-installation) → 2 — Inspect thread and contact surface condition → 3 — Identify mechanism (embedding/vibration/corrosion) → 4 — Select corrective action for that specific mechanism → 5 — Document in O&M record
- RFQ information
- Bolt size and grade, tower section, loosening mechanism identified, replacement specification required, anti-loosening hardware needed, quantity
§ 01 The five root causes
Embedment relaxation
Thread and bearing-face asperities bed in under load within the first weeks of service. Typical preload loss: 5–15%. Occurs once; corrected by the initial re-torque check.
Vibration-induced loosening
Transverse vibration causes the nut to back off against friction. The Junker test (DIN 65151) quantifies this. Most prevalent at rotor-frequency harmonics in nacelle and hub connections.
Fatigue preload loss
Cyclic bending at tower flanges causes micro-slip at the joint interface under repeated loading. Distinct from vibration loosening — the nut does not rotate, but the effective clamp force decays.
Thermal cycling
Differential thermal expansion between the bolt and the flanged joint relaxes preload during temperature swings. More significant in black steel towers exposed to solar gain than in grouted foundations.
Joint disturbance
Any event that moves the joint — overload, impact, adjacent bolt re-torque, foundation settlement — disturbs preload in nearby bolts. Common source of puzzling "spot" loosening on otherwise stable arrays.
Wrong diagnosis
Repeatedly re-torquing a bolt that is loosening due to vibration — without adding a locking element — will not solve the problem. Match the remedy to the mechanism.
§ 02 Vibration loosening in detail
Vibration loosening — sometimes called self-loosening — occurs when transverse (shear) loads applied to the joint cause the clamped parts to slip relative to each other. This relative slip generates a small rotational motion at the nut face that, over many cycles, causes the nut to back off. The key word is transverse: purely axial vibration does not cause self-loosening under normal conditions.
In wind turbines, the main sources of transverse vibration that drive self-loosening are:
- Rotor imbalance and aerodynamic excitation at blade-passing frequency — directly transmitted to nacelle, hub, and blade root bolts.
- Tower bending modes excited by wind turbulence and wake effects — primarily affects upper tower flange bolts.
- Drivetrain torsional excitation — affects gearbox and generator mounting bolts.
The standard test for resistance to vibration loosening is the Junker test (DIN 65151 / ISO 16130), which applies controlled transverse displacement cycles to a bolted joint and measures residual clamp force. Wedge-lock washers (Nord-Lock type) retain significantly more preload than standard washers under Junker test conditions.
§ 03 Fatigue preload loss at tower flanges
Tower flange connections experience large bending moment cycles as the turbine operates. Each cycle slightly changes the stress distribution across the bolt circle. Under sustained cyclic loading, micro-slip occurs at the flange faces — especially if the flange surface finish is rougher than specified or if the bolt circle is under-torqued initially.
Unlike vibration loosening, fatigue preload loss does not involve nut rotation. The marking paint stripe across the nut may remain intact while the actual clamping force has dropped below the minimum required. This is why torque checks alone are insufficient as the sole inspection method for high-cycle locations — ultrasonic bolt elongation measurement or hydraulic re-tensioning provides a more reliable assessment of actual preload.
§ 04 Prevention methods matched to cause
| Cause | Prevention method | Notes |
|---|---|---|
| Embedment relaxation | Scheduled initial re-torque within first 3 months | One-time; part of commissioning procedure |
| Vibration loosening | Wedge-lock washers (Nord-Lock, Heico-Lock) or prevailing-torque nuts | Physical locking element required; re-torque alone is not sufficient |
| Fatigue preload loss | Correct initial preload; periodic re-tensioning; ultrasonic monitoring | Flange surface finish and flatness also critical |
| Thermal cycling | Increased inspection frequency; consider disc spring washers (Belleville) | Less common in grouted foundations; more relevant in bolted steel-to-steel joints |
| Joint disturbance | Re-check adjacent bolts after any re-torque or repair event | Often overlooked — document which bolts were disturbed and re-check the neighbours |
§ 05 Inspection and remedial action
When you find a loose bolt in the field, the sequence matters:
- Don't just re-torque and move on. A single loose bolt is a signal — inspect the full bolt circle before deciding the problem is isolated.
- Check the marking paint. A broken or displaced stripe indicates the nut has rotated. If the stripe is intact but the bolt feels loose, the joint interface has likely compacted (embedment) or fatigue preload loss has occurred without nut rotation.
- Count consecutive loose bolts. If three or more adjacent bolts are loose, a flange geometry or torquing procedure issue is more likely than individual bolt defects.
- Record everything. Log bolt position, amount of rotation found, torque applied at re-check, and date. Trend data across multiple inspections is far more valuable than any single snapshot.
For the correct torquing procedure to use when re-tightening, see How to torque wind turbine foundation bolts. For the locking elements that address vibration loosening specifically, see the forthcoming article on anti-loosening methods.
Related: Bolt Preload Loss in Wind Turbines: Causes and Troubleshooting
Evidence and decision boundary
- Direct evidence
- Fastener-property and torque/clamp-force standards define controlled characteristics and tests; they do not identify the cause of a specific tower-joint preload loss.
- Engineering inference
- Diagnosis should separate embedment, relaxation, friction scatter, insufficient initial preload, joint slip, fatigue damage, coating changes, temperature and installation error using records and inspection evidence.
- Typical or indicative value
- Retorque intervals and loss thresholds are planning values unless stated by the turbine OEM or project maintenance procedure.
Primary sources checked
- ISO 898-1:2013 - mechanical properties of carbon and alloy steel fasteners
- ISO 16047:2005 - torque/clamp-force testing
- ASTM F3125/F3125M-19 - high-strength structural bolt assemblies
- ISO/IEC 17025:2017 - competence of testing and calibration laboratories
Related commercial route: Compare wind-turbine clamp systems and RFQ inputs.