Why These Cases Matter

Fastener failures rarely make headlines as fastener failures. The public record remembers a bridge collapse, a plane crash, or a sinking ship; the investigation afterward often finds a single pin, rivet, bolt, or anchor at the root of it. These cases are studied repeatedly in forensic engineering and failure analysis because the failure mode is almost never exotic. It is usually one of a small number of recurring themes: an unreviewed design change, a non-redundant load path, corrosion in a connection nobody could inspect, fatigue from cyclic loading that was not accounted for, or the wrong material specified for a sustained load. Reviewing the specifics case by case is more useful than reciting the theme in the abstract, so this report works through eight incidents in enough technical detail to draw real lessons from.

1. Hyatt Regency Walkway Collapse (Kansas City, 1981)

Two suspended walkways inside the hotel atrium, built one above the other, hung from the ceiling on steel rods. In the original design, a single continuous rod ran from the ceiling through both walkways, with each walkway’s box-beam supported by a nut threaded onto that one rod. During construction, the steel fabricator proposed a change for buildability: instead of one long threaded rod, use two shorter rods, with the upper walkway’s rod anchored to the ceiling and a second rod running from the upper walkway down to the lower one.

The change looked minor but was not. In the original design, the fourth-floor connection only had to carry its own walkway’s load, because the rod passed through it. In the revised design, the fourth-floor box-beam connection now had to carry the full combined weight of both walkways, since the lower walkway’s load was transferred through it rather than past it. The connection’s load effectively doubled, and the change was never adequately re-analyzed by a structural engineer before construction. The walkways collapsed during a crowded dance event, killing 114 people and injuring more than 200. The case is now a standard teaching example in engineering ethics courses on the responsibility to review, not rubber-stamp, fabricator-proposed changes to a structural connection.

2. Silver Bridge Collapse (Point Pleasant, West Virginia, 1967)

The Silver Bridge was an eyebar-chain suspension bridge, a design in which the main cables are replaced by chains of pin-connected steel eyebars. A single eyebar in the chain failed from a small stress corrosion crack, roughly a tenth of an inch deep, that had grown undetected at the pin connection over the bridge’s 40 years of service. Because the eyebar-chain design had essentially no redundancy, the failure of that one link caused the entire structure to collapse into the Ohio River during rush-hour traffic, killing 46 people.

The investigation highlighted two compounding issues: the susceptibility of the eyebar steel to stress corrosion cracking at the highly stressed pin connections, and the complete lack of structural redundancy in the load path, so a defect too small to have been reliably found by the inspection methods of the time was enough to bring down the whole bridge. The disaster directly led to the US National Bridge Inspection Standards and a much greater emphasis on redundancy in connection design for critical structures.

3. Mianus River Bridge Collapse (Greenwich, Connecticut, 1983)

A 100-foot section of the I-95 deck, connected to the rest of the span through a pin-and-hanger assembly, dropped into the Mianus River, killing three people. Pin-and-hanger connections are used to accommodate thermal expansion between bridge deck sections, with a steel pin passing through a hanger plate to let the joint pivot slightly.

The investigation found that a drainage design flaw let road runoff, carrying de-icing salt, pool directly onto the pin-and-hanger assembly instead of draining away from it. Years of corrosion caused the pin to migrate sideways within the hanger, gradually losing its bearing engagement until it walked out of the connection entirely under traffic loading. The failure mode was not overload; it was a slow, invisible loss of fastener retention from corrosion in a connection that was difficult to inspect and had gone without adequate maintenance. The case reinforced the importance of designing connections for drainage and inspectability, not just strength, and contributed to more rigorous bridge inspection protocols for pin-and-hanger details specifically.

4. Kemper Arena Roof Collapse (Kansas City, 1979)

The roof of this large-span arena, suspended below an exposed space-frame structure via hanger rod connections, collapsed during a severe thunderstorm. Fortunately the building was empty at the time. Investigators found that the bolted hanger connections supporting sections of the roof structure had been subjected to repeated cyclic loading from wind-induced movement of the roof over the structure’s service life, and that ponding rainwater on the roof added additional sustained load the connections had not been designed to carry cyclically. The repeated loading led to fatigue cracking and ultimate failure of the connections, bringing down a large section of roof.

The case is frequently cited alongside Hyatt Regency as a reminder that bolted connections in large structures need to be evaluated not just for peak static load, but for the cyclic and dynamic loading the structure will actually experience over its lifetime, including loads, like ponding water on a roof, that are easy to underestimate during design.

5. Big Dig Ceiling Collapse (Boston, 2006)

In Boston’s I-90 connector tunnel, a section of suspended concrete ceiling panel, along with the epoxy anchor bolts holding it, detached from the tunnel roof and fell onto a passing car, killing one occupant. The anchors were epoxy-set bolts: steel rods bonded into drilled holes in the concrete ceiling using an adhesive anchor system rather than a mechanical expansion anchor.

The National Transportation Safety Board investigation found that the epoxy formulation used was a fast-curing adhesive not rated for the kind of sustained, long-term tensile load the ceiling panels placed on it. Under sustained load, that epoxy was prone to creep, slowly deforming over time even well below its short-term rated strength, which let the anchors gradually pull free from the concrete over several years. The root cause was fundamentally a material selection failure: the wrong adhesive anchor system for a permanent, continuously loaded application. The case became a widely cited reference in the adhesive-anchor industry for distinguishing short-term-rated from sustained-load-rated epoxy products, and prompted stricter adhesive anchor qualification testing and installer certification requirements.

6. Alaska Airlines Flight 261 (2000)

This MD-83’s horizontal stabilizer is positioned by a jackscrew assembly, a large acme-threaded screw and matching nut that drives the stabilizer up and down for pitch trim. The investigation found that the acme nut’s threads had worn down nearly to failure because the required lubrication interval for the jackscrew assembly had been extended well beyond what the components could tolerate without adequate lubrication, and the end-play checks meant to catch excessive thread wear before it became critical had not caught the degradation in time.

The worn nut threads ultimately stripped completely in flight, causing a sudden and uncontrollable loss of pitch control. The aircraft crashed into the Pacific Ocean, killing all 88 people aboard. The case is a central example in maintenance engineering of how a mechanical fastener assembly’s safe life depends entirely on the lubrication and inspection program actually being followed as specified, and of the risk in extending maintenance intervals without re-validating the wear margins that interval was based on.

7. RMS Titanic (1912)

The sinking of the Titanic is usually remembered for the iceberg collision, but metallurgical analysis of recovered hull fragments decades later raised a fastener-specific theory about why the damage was as severe as it was. Much of the hull plating in the bow section was joined with wrought iron rivets rather than the stronger steel rivets used elsewhere on the ship, a common practice of the era for areas that were difficult for hydraulic riveting machines to reach. Metallurgical testing of recovered rivets found that some contained a higher-than-typical concentration of slag inclusions, which can make wrought iron more brittle, particularly in the near-freezing water temperatures of the night of the collision.

The prevailing theory among several researchers is that these rivets failed in a brittle manner on impact, popping their heads off and opening seams along the hull plating over a larger area than a single clean gash would have caused, accelerating flooding. This remains a debated area of maritime forensic history rather than a settled conclusion, but it is widely cited as an early illustration of how fastener material quality control, not just the design strength of the fastener itself, can determine whether a structure fails gracefully or catastrophically.

8. Aloha Airlines Flight 243 (1988)

A Boeing 737 lost a large section of its upper fuselage skin in flight, decompressing the cabin and killing one flight attendant who was swept out of the aircraft. The investigation traced the failure to the riveted lap joint running along the fuselage, where multiple rows of adjacent rivet holes had each developed small fatigue cracks after years of pressurization cycles.

Compounding the fatigue cracking, corrosion had degraded the cold-bond adhesive between the lap-jointed aluminum skin layers, a bonding method intended to help the joint share load more evenly across the rivets. As the adhesive bond deteriorated, the rivets alone had to carry more of the cyclic load than the joint was designed for, accelerating fatigue cracking around the rivet holes. Individually small cracks around adjacent rivets eventually linked up into one long crack, and the skin panel tore away. The case drove major changes in how aging aircraft are inspected for multiple-site fatigue damage, cracking distributed across many adjacent fastener holes rather than concentrated at one, which is far harder to catch with single-hole inspection methods.

Recurring Themes Across the Cases

Failure theme	Cases	Engineering takeaway
Unreviewed design change to a connection	Hyatt Regency	Any field or fabrication change to a load-bearing connection needs a full re-analysis, not a visual check that it looks similar
Non-redundant load path	Silver Bridge	A single fastener or pin should not be the sole load path in a structure where failure is catastrophic
Corrosion in an inaccessible or poorly drained connection	Mianus River Bridge, Silver Bridge	Connections need to be designed for drainage and inspectability, not just strength
Fatigue from cyclic or dynamic loading not fully accounted for	Kemper Arena, Aloha Airlines 243	Peak static load is not sufficient; the actual cyclic loading over service life has to be analyzed
Wrong fastening material or system for the sustained load case	Big Dig	Short-term rated strength is not the same as long-term sustained-load performance; adhesive and anchor systems must be qualified for the actual load duration
Maintenance program not matched to actual wear mechanism	Alaska Airlines 261	Extending inspection or lubrication intervals requires re-validating the margins the original interval was based on
Material quality control at the fastener level	Titanic	A structure’s nominal design strength assumes the fasteners actually meet their material specification; quality control at manufacture is part of the safety margin, not separate from it

Practical Takeaways

A few practices recur across the post-failure recommendations in nearly all of these investigations, and are worth treating as standing checklist items on any critical bolted, riveted, pinned, or anchored structure:

• Treat any fabricator- or contractor-proposed change to a connection detail as requiring the same engineering review as the original design, regardless of how minor it looks.
• Avoid single-point-of-failure fastening in structures where failure is catastrophic; build in redundancy at the connection level, not just at the member level.
• Design connections for inspectability and drainage, since a connection that cannot be seen or that collects water cannot be reliably maintained.
• Validate fastening and anchor systems against the actual duration and cycling of the in-service load, not just a short-term proof test.
• Treat maintenance intervals as calculated values tied to a specific wear or degradation model, not administrative defaults that can be extended without re-analysis.
• Maintain real material traceability and quality control on fasteners themselves; a correct design calculation is only as good as the fastener that was actually installed.