Fastener failure analysis is a structured investigation that preserves the failed part, reconstructs the assembly and service history, identifies the physical failure mode, and tests the most likely cause before corrective action. A broken screw or loose nut is a symptom, not a root cause. The fastener may have failed by overload, fatigue, thread stripping, galling, corrosion, loosening, or incorrect installation. Replacing it with a stronger part before the load path and mating components are checked can move the failure elsewhere without preventing recurrence.
Preserve the failed fastener, mating nut or tapped part, washers, inserts, fragments, and nearby components. Photograph the assembly in place before disassembly where practical. Label orientation and location, and keep each part with its matching joint. Avoid grinding, aggressive cleaning, tightening, or destructive removal until the investigation plan is set, because those actions can erase fracture, wear, corrosion, or assembly evidence.
Record the assembly date, supplier and lot, part numbers, material and finish, tool and torque trace, lubricant, installation sequence, service loads, cycles, temperature, environment, maintenance history, and reported symptom. Ask whether the failure occurred during installation, initial operation, after a service interval, or during removal. Timing helps distinguish installation damage from progressive service failure.
Begin with what failed and where. A fractured bolt may show a fatigue origin, overload, corrosion attack, or a defect. A stripped thread can be in the fastener or the internal thread. A loose joint can result from rotation, settlement, gasket relaxation, thermal change, or inadequate initial clamp force. Galling can seize stainless threads before the assembly reaches its intended preload.
| Symptom | Possible failure mode | Evidence to preserve | First checks |
|---|---|---|---|
| Fracture near a thread root | Fatigue, overload, corrosion-assisted cracking, or material issue | Fracture faces, thread condition, nearby corrosion, matching fasteners | Load spectrum, preload, thread runout, bending and surface condition |
| Threads pulled or torn out | Internal or external thread stripping, cross-threading, wrong engagement | Loose material, lead threads, tapped part and insert | Thread size and pitch, engagement, material, hole depth and installation trace |
| Fastener seized during installation or removal | Galling, contamination, damaged threads, or coating interference | Uncleaned thread pair, debris, lubricant and coating records | Material pairing, lubrication, speed, alignment and reuse history |
| Joint separated or hardware rotated | Preload loss, slip, vibration loosening, relaxation, or wrong lock feature | Witness marks, washer faces, joint surfaces and torque records | Clamp-load retention, joint stiffness, transverse motion and service cycles |
| Repeated failures across a lot or station | Process variation, mixed parts, tool, feeder, or supplier change | Traceability records and unaffected comparison parts | Lot identity, dimensions, coating, tool program and inspection method |
The table is a starting diagnostic map, not a definitive failure-code system. Several mechanisms can contribute to one event.
Separate observations from hypotheses. For example, a fractured thread root is an observation; fatigue due to preload loss is a hypothesis until the fracture, load history, and joint condition support it. List alternative causes and the evidence that would distinguish them. Where possible, compare failed parts with unused parts from the same lot and with parts from a known-good assembly.
Analyze the complete load path. Check fastener size, property class, thread form, engagement, grip, head seating, washer, hole geometry, material, finish, joint stiffness, preload method, and external load direction. A bolt can be strong enough in simple tension yet fail because of bending, eccentricity, cyclic load, corrosion, or partial engagement. The NASA Fastener Design Manual provides engineering background on materials, threads, preload, fatigue, corrosion, and failure mechanisms.
Use suitable inspection or laboratory analysis when visual checks cannot distinguish causes. This may include dimensional measurement, hardness or material verification, microscopy, fracture-surface examination, coating analysis, or load testing. Select destructive tests only after preserving the original surfaces and documenting the sampled location.
A practical sequence begins with a timeline. Identify when the part was manufactured, coated, installed, inspected, loaded, and removed. Next compare the failed part with a same-lot unused part and a known-good assembly. Map the first visible damage and its direction relative to applied tension, shear, bending, and rotation. Then test the leading hypotheses using the least destructive method that can distinguish them.
For example, if the fracture originates at a thread root, inspect thread form, runout, engagement, and cyclic stress before concluding that the material was defective. If the nut seized, preserve the mating thread and identify transferred metal, lubricant, and coating condition. If a joint came apart, check whether the fastener rotated, the member relaxed, or the internal thread stripped. A root-cause statement should explain both how the failure occurred and why the production or service controls did not prevent or detect it.
Fatigue often leaves a progressive fracture pattern and starts at a high-stress area, but confirmation should come from qualified failure analysis. Review the dedicated fastener fatigue guide for cyclic loads, preload, and inspection. Galling can cause adhesive transfer and seizure; the stainless fastener galling article explains material pairing and lubrication controls.
Overload may follow an excessive external force, misassembly, over-tightening, or a load path different from the design assumption. Corrosion or environmental attack can reduce section or initiate cracking. Cross-threading and bottoming can create damage during installation. The thread engagement guide explains how internal material and usable thread depth affect joint capacity.
Standards define specific mechanical requirements but do not diagnose every application failure. ISO 898-1 specifies properties for covered carbon- and alloy-steel bolts, screws, and studs, and explicitly does not specify fatigue resistance. The product standard and drawing should be checked along with service loads and observed evidence.
This TU Delft lecture introduces failure modes in mechanically fastened joints. Its examples provide a framework for load-path thinking, while the actual failed assembly still needs its own investigation.
Corrective action should target a demonstrated cause. If preload was too low, determine why the installation method missed the required range. If the internal thread stripped, assess material, engagement, hole depth, and alignment. If fatigue is confirmed, review alternating stress, separation, bending, and surface quality. If galling occurred, evaluate material pair, lubricant, assembly speed, and reuse.
A stronger grade or a locking feature may help only when it addresses the verified failure mechanism and remains compatible with the joint. A repair or redesign should be checked for a new weak link in the bolt, nut, insert, sheet, gasket, coating, or adjoining component. The FAA AC 43.13-1B gives general accepted aircraft maintenance practices; approved maintenance data and the responsible design authority govern a specific aircraft or product.
Verify the corrective action with the same failure-relevant loads, environment, assembly process, and inspection method that exposed the problem. Define acceptance limits before testing. Update drawings, work instructions, incoming inspection, supplier controls, and traceability records so the correction survives into production.
Separate correction from corrective action. Replacing a damaged batch restores affected units; changing the drawing tolerance, fixture alignment, lubricant control, or inspection plan may prevent recurrence. Confirm that the proposed action does not create a new failure path, such as a stronger bolt overloading a thin insert or a locking compound obstructing a required electrical contact.
For an OEM supplier investigation, assign each test to a named part lot and maintain chain-of-custody for returned samples. Record the test setup, calibration status, photographs, operator, and raw measurements. Close the investigation only after the cause is supported by evidence, containment is verified, and production controls are updated.
Maintain a record linking the failure to part number, revision, lot, certificate, coating batch, assembly station, tool, installation program, and service history. For a custom fastener, the TNHO hex and square-head bolt family shows relevant product geometries; the quotation and approved drawing define delivered properties and dimensions. The hex bolt grades guide can help identify designation differences, but it does not replace material certification or failure analysis.
For an OEM investigation, request the original failed parts, mating parts, photos before disassembly, installation data, load and environmental history, quantity affected, and any known substitutions. A concise chronology and traceable parts prevent the investigation from mixing similar-looking but unrelated fasteners.
Only if insufficient fastener strength is the demonstrated cause and the rest of the joint can carry the increased load. A stronger bolt may move failure to the internal thread, sheet, gasket, or connected component.
Preserve it as received and document its condition first. Cleaning or handling can remove debris, corrosion products, lubricant, and fracture evidence. Follow the laboratory’s instructions before any preparation.
Visual appearance alone may not be conclusive. Examine the fracture surface and loading history, compare with unused fasteners, and use qualified materials analysis when needed. Fatigue usually involves progressive crack growth under cycles; overload is a final failure under a high load, but both can coexist.
Keep the drawing and revision, fastener lot and certificate, coating and lubricant, tool calibration and installation trace, joint stack, service loads and cycles, temperature, environment, maintenance history, and photographs before disassembly.
TNHO 
