Why Do Bolts Loosen: Four Mechanisms of Vibrational Loosening and Matching Anti-loosening Solutions
Executive Summary
Bolt loosening is not simply "the thread turned itself loose." Transverse slip, rotational loosening, preload decay and fatigue fracture are four different mechanisms, calling for entirely different anti-loosening solutions. This article breaks down each mechanism and its criteria, and gives the selection logic among friction, mechanical, adhesive and structural anti-loosening methods.
Table of Contents
- On the Vibration Table, the "Tightest" Bolt Loosened First
- Mechanism 1: Rotational Loosening from Transverse Slip
- Mechanism 2: Rotational Loosening from Axial Load
- Mechanism 3: Preload Decay (Embedding & Relaxation)
- Mechanism 4: Fatigue Fracture
- Selection Logic Across Four Anti-loosening Categories
- Mechanism-to-Solution Reference Table
- Conclusion
On the Vibration Table, the "Tightest" Bolt Loosened First
A motor bracket underwent vibration testing with three options compared: group A used plain bolts tightened to the specified torque, group B added spring washers, and group C used nylon-insert lock nuts. After 30 minutes, residual preload was checked: group A decayed 12%, group B decayed 15%, group C decayed 4%.—the group with spring washers was actually the worst.
This is not an anomaly; it follows from the mechanism. The split of a spring washer creates an off-center load when compressed, and its rebound force is far too weak to resist transverse slip. To understand this result, you must first distinguish the four entirely different mechanisms of bolt loosening.
Mechanism 1: Rotational Loosening from Transverse Slip
This is the most common form of loosening and the main cause of preload decay under vibration.
How it happens
When a joint sustains transverse (perpendicular to the bolt axis) alternating load, micrometer-level relative slip occurs between the clamped parts. Slip makes the thread-flank contact state change repeatedly; under load the flanks elastically rebound, and each cycle contributes a tiny rotation.
Key quantitative relationship: transverse load causes loosening orders of magnitude faster than axial load. At the same amplitude, transverse load can cause loosening dozens of times faster than axial load.
This is why transverse vibration tests (such as testers based on the Junker principle) can visibly loosen a bolt within a few hundred cycles, while pure axial vibration hardly loosens it at all.
Diagnostic features
- Residual preload declines continuously and rapidly, with the decline rate correlated to the number of transverse load cycles;
- Annular wear marks on the bearing surface, fretting debris on the hole wall or mating face (reddish-brown iron-oxide powder);
- The relative rotation direction of bolt and nut is opposite to the tightening direction.
Mechanism 2: Rotational Loosening from Axial Load
When a joint sustains axial alternating load large enough to cause an open-close cycle of the joint faces, the bolt is stretched and rebounds in each cycle. At the instant of unloading, the pressure on the thread flanks approaches zero, friction resistance is lost synchronously, and the nut slowly rotates as it is pushed by the lifting flanks.
Diagnostic features
- Rust-stain seepage or oil traces at the joint faces, indicating repeated opening;
- Fracture often occurs at the first loaded thread tooth, with clear fatigue features;
- Residual preload declines relatively slowly, but joint stiffness shows a sudden change.
The root cause of this loosening is insufficient preload—with enough preload the joint faces do not open, and the whole mechanism never starts. The design should ensure FM > (1 − Φ) × FA,max, with a 1.2~1.5× safety margin.
Mechanism 3: Preload Decay (Embedding & Relaxation)
The defining feature of this "loosening" is: the bolt did not rotate, yet preload truly declined.
Embedding loss
Micro-asperities on the thread flanks and bearing surface are flattened under pressure, reducing the bolt's elastic elongation. This is unrelated to vibration; most of it completes within tens of hours after assembly.
| Contact Surface Condition | Typical Embedding | Preload Loss |
|---|---|---|
| Precision-machined surface, single joint face | 5~10 μm | 3%~8% |
| Common machined surface | 10~20 μm | 8%~15% |
| Multiple joint faces or rough surfaces | 20~40 μm | 15%~25% |
| Coating layer or soft shim at the interface | 40 μm+ | Up to 30% |
Stress relaxation
Under sustained load and temperature, material creeps and preload decays slowly. Temperature is the dominant accelerating factor; near or above the material's creep threshold, relaxation becomes the leading mechanism.
Mechanism 4: Fatigue Fracture
Strictly speaking, fatigue fracture is not "loosening," but it is often mistaken on site for loosening—the bolt broke, the joint came loose. The causality is often reversed: fatigue fracture is the long-term result of separation or slip.
Judgment points:
- The fracture shows clear fatigue marks (clam-shell propagation zone + rough final-fracture zone);
- The fracture concentrates at the first loaded thread tooth or the head transition;
- The bolt shows no sign of rotation, and residual preload may still be high (because it has already broken).
The direction of action is entirely different: loosening requires anti-loosening measures; fatigue requires lowering the stress amplitude, lowering the load distribution factor, and improving the thread-root transition.
Selection Logic Across Four Anti-loosening Categories
| Category | Typical Solution | Effective Mechanism | Applicability & Notes |
|---|---|---|---|
| Friction anti-loosening | Spring washer, serrated washer, Belleville washer | Increase bearing-surface friction, maintain preload | Spring washers have limited effect on transverse slip; Belleville washers compensate for embedding loss |
| Mechanical anti-loosening | Nylon-insert lock nut, all-metal lock nut, cotter pin, tab washer, wire locking | Directly restrict relative rotation of the thread pair | Most reliable, but mostly not reusable; wire locking suits bolts in the same group |
| Adhesive anti-loosening | Pre-applied micro-encapsulated adhesive, anaerobic adhesive | Fill thread gaps; cured bond forms | Also seals; removal needs heat or higher torque; mind shelf life |
| Structural anti-loosening | Raise preload, add dowel/spigot, enlarge bearing surface, use conical fit | Remove the cause of slip and separation | Treats the root cause; should be considered first, not last |
Engineering consensus on selection order: first use structural means to remove the loosening mechanism, then use mechanical or adhesive measures as a backstop; do not use friction-type measures to solve a slip problem.
Mechanism-to-Solution Reference Table
| Dominant Mechanism | Field Features | First Choice | Second Choice |
|---|---|---|---|
| Transverse slip | Bearing-surface wear, fretting debris, rapid decay | Raise preload + add spigot or dowel | All-metal lock nut, serrated washer |
| Axial opening | Leakage traces at joint faces, fatigue fracture | Raise preload to meet separation-resistance criterion | Increase clamped-part stiffness to lower Φ |
| Embedding loss | Short-term decay after assembly, no rotation | Reduce joint faces, raise bearing-surface hardness | Belleville washer compensation, delayed re-tightening after assembly |
| Stress relaxation | High-temperature service, long-term slow decay | Creep-resistant materials and high preload margin | Periodic re-tightening schedule |
| Fatigue fracture | Fatigue fracture surface, no rotation | Lower Φ and stress amplitude | Rolled threads + thread-root strengthening |
Conclusion
The question "why do bolts loosen" can only be answered by splitting it into four mechanisms. Transverse slip causes rapid rotational loosening; axial opening causes slow rotation rooted in insufficient preload; embedding and relaxation are "non-rotating loosening," against which lock nuts are useless; fatigue fracture is the long-term result of repeated separation and slip.
Do one thing before choosing an anti-loosening solution: judge whether the nut actually rotated. If it rotated, use mechanical or adhesive means to restrict it; if it did not, go back to preload, joint-face stiffness and surface state to find the cause. Get this step right, and you will not choose the wrong solution.
