Failure Modes and Fracture Identification
Failure analysis begins at the fracture surface, because the surface records the loading history. Fatigue fracture is the most frequent mode on cyclically loaded joints. It shows a smooth region with beach marks where the crack grew and a rough final rupture zone, and the origin is usually at the head-to-shank fillet or at the first engaged thread root, where stress concentration is highest. Overload or tensile fracture is different: the surface is flat and dull for a high-strength steel, sometimes with a cup-and-cone profile on a ductile material, and there is no progression pattern. Hydrogen embrittlement produces a brittle, intergranular fracture at the thread root and typically appears within hours or days of electroplating, before the joint has ever seen a full service load. Corrosion fatigue combines both signals: pitting at the origin plus fatigue progression, common on coastal and tunnel sections.
Root Cause Analysis Workflow
Once the mode is known, the analysis chain is short. Record the failed position and the tightening history of the fastener. Check whether the measured preload was above or below the design value, then check the joint geometry for eccentric loading and the coating process for a hydrogen source. Fracture surfaces are examined visually, then with magnetic particle testing to ASTM E709 on the adjacent fasteners to find partially cracked bolts before they fail, and with liquid penetrant testing to ASTM E1417 where the geometry or material makes magnetic methods unsuitable. Hardness and tensile results are compared with the property class declared for the fastener, and the nominal tensile strength and proof-load values for that class are those listed in ISO 898-1, with GB/T 3098.1 giving the equivalent requirements for the metric fastener classes used on Chinese railways.
Prevention Measures Matched to the Failure Mode
For fatigue, the practical levers are load and geometry rather than material alone: increase the shank or thread section, reduce eccentricity so the bolt sees true axial load, increase the fillet radius at the head transition, and raise the initial preload so that the cyclic component applied to the bolt is only a small fraction of the clamp force. For overload, the correction is usually a lower tightening torque or a higher property class, applied after confirming that the joint does not separate. For hydrogen embrittlement, control is entirely process-based: bake plated high-strength fasteners within a short time of plating, keep preload in the range of 60 to 70 percent of yield strength, and prefer zinc-flake coatings that involve no electrolytic hydrogen for property classes 10.9 and above. For corrosion, the fix is coating selection and sealing of the engaged threads.
| Failure mode | Primary evidence | Preventive action |
|---|---|---|
| Fatigue fracture | Beach marks, origin at fillet or first thread | Larger section, larger fillet radius, higher preload, reduced eccentricity |
| Overload fracture | Flat or cup-and-cone surface, no progression | Correct tightening torque, higher property class, verify joint separation load |
| Hydrogen embrittlement | Brittle intergranular fracture at thread root soon after plating | Bake after plating, preload 60 to 70 percent of yield, zinc-flake coating |
| Corrosion fatigue | Pitting at origin plus fatigue progression | Higher coating mass per ISO 4042, sealing compound, salt spray verification to ISO 9227 |
Selection by Track Scenario
On straight sections of ordinary line, where vibration is mild, a spring washer with scheduled torque checks is usually sufficient and costs least. On high-speed and heavy-haul alignments the dynamic input is severe, and an all-metal prevailing-torque thread form or a threadlocker should be used, because both survive continuous vibration without consumable parts. Turnouts and switch components warrant double-nut locking so that the joint remains restrained even if one nut rotates. In wet, coastal or tunnel sections the anti-loosening measure should be combined with a corrosion-resistant coating rather than applied alone. Joints that will never be disassembled, such as permanently fixed track fittings, can be locked by welding.
Frequently Asked Questions
Q: How can fatigue fracture be distinguished from overload fracture in the field?
A: Fatigue surfaces show beach marks and a rough final rupture zone, and the origin sits at the head fillet or the first engaged thread. Overload surfaces are flat or cup-and-cone with no progression pattern.
Q: What preload level is recommended for high-strength fasteners?
A: A common design range is 60 to 70 percent of yield strength. Above that range the risk of hydrogen-assisted cracking and of embedment-driven preload loss increases, and below it the bolt absorbs a larger share of the cyclic load.
Q: Why do plated grade 10.9 bolts crack soon after installation?
A: Electroplating introduces hydrogen into the steel, and a highly stressed thread root can then crack in a brittle manner. Baking after plating and switching to a zinc-flake coating remove the hydrogen source.
Q: Which inspection method finds partially cracked bolts still in service?
A: Magnetic particle testing to ASTM E709 on the adjacent fasteners is effective on ferritic parts. Liquid penetrant testing to ASTM E1417 is used where the geometry or material rules out magnetic methods.
Q: Does increasing bolt diameter always solve a loosening problem?
A: No. Loosening is usually a preload and friction issue rather than a strength issue, so the first correction is a better locking method and a verified tightening specification, not a larger bolt.
Q: How is coating performance compared for corrosive corridors?
A: Coating mass and thickness requirements are set by ISO 4042 for electroplated fasteners, and comparative performance is demonstrated by neutral salt spray testing to ISO 9227 at the thickness class specified for the corridor.