Skip to content
Rotational slope failure along a highway embankment with a scarp and cracked shoulder.

CEGR 492 · Learning Tools

Failure Investigations

Embankment slope failure — the consequence when driving moment exceeds available shear resistance.

Failure Investigations

Timeline, failure sequence, structural analysis, redesign, and lessons — with proposal ideas.

Hyatt Regency Walkway Collapse

1981 · Steel connections / design review114 killed, 216 injured
Timeline
  • 1978 — Design: fourth-floor walkway hung from continuous rods through the second-floor box beams.
  • 1979 — Shop drawing change: single continuous rod replaced by two offset rods.
  • 1979 — Change accepted without independent structural review.
  • 17 Jul 1981 — Tea dance; walkways loaded with spectators; connection fails.
Failure sequence
  • Load from the fourth-floor walkway transferred into the second-floor box beam connection.
  • The revised detail doubled the load on the upper box beam washer/nut bearing.
  • The box beam welded seam tore; the nut pulled through the beam flange.
  • Fourth-floor walkway fell onto second-floor walkway; progressive collapse to the lobby.
Structural analysis
  • Original detail: box beam carries load of one walkway ≈ P.
  • Revised detail: upper box beam carries its own walkway plus the hanger reaction from below ≈ 2P.
  • Capacity of the welded box beam seam ≈ 0.6 × required for the original load — roughly 30% of the revised demand.
  • Even the original detail did not meet the Kansas City code; the change made it catastrophic.
Redesign
  • Use a continuous rod so each box beam carries only its own walkway load.
  • Provide a bearing plate/stiffened seat sized for the full hanger reaction, not just washer bearing.
  • Check local flange bending, weld strength, and block-type tearout at the rod penetration.
  • Require engineer-of-record review and stamped calculations for any connection change.
Lessons learned
  • Shop-drawing changes are design changes.
  • Load path must be re-traced after every detail modification.
  • Connection design is not a delegated afterthought.
Proposal ideas
  • Digital shop-drawing change-tracking workflow with automatic load-path re-check
  • Survey of connection review practice in regional fabrication shops

FIU Pedestrian Bridge Collapse

2018 · Concrete truss / construction staging6 killed, 10 injured
Timeline
  • 10 Mar 2018 — Span moved into place using self-propelled modular transporters.
  • 13 Mar 2018 — Cracking observed and photographed at node 11/12.
  • 15 Mar 2018 — Post-tensioning in diagonal 11 re-tensioned; span collapses onto traffic.
Failure sequence
  • Nodal region 11/12 had insufficient interface shear capacity between diagonal and deck.
  • Cracks widened after the span was moved and the temporary supports removed.
  • Re-tensioning increased demand on the already-cracked node.
  • Interface shear failure → the diagonal punched through the deck → total collapse.
Structural analysis
  • Interface shear demand at node 11/12 exceeded ACI 318 shear-friction capacity by roughly a factor of 2.
  • Shear-friction: Vn = μ Avf fy; the roughened area and crossing reinforcement provided were inadequate.
  • Redundancy was zero — a single concrete truss with no alternate load path.
  • Structure was not closed to traffic despite visible structural cracking.
Redesign
  • Increase interface reinforcement across the node and provide an intentionally roughened, keyed interface.
  • Provide redundancy (two trusses or a supplemental support system) for the erection condition.
  • Independent peer review of every construction stage, not just the final condition.
  • Mandatory road closure protocol when structural cracking is observed.
Lessons learned
  • Cracks are data, not cosmetics.
  • The erection condition can govern.
  • Non-redundant structures need higher review standards.
Proposal ideas
  • Construction-stage structural monitoring protocol for accelerated bridge construction
  • Shear-friction capacity study of node regions in concrete trusses

I-35W Mississippi River Bridge

2007 · Steel truss / gusset plates13 killed, 145 injured
Timeline
  • 1967 — Built; U10 gusset plates detailed at 1/2 in thickness.
  • 1977–1998 — Deck widened and re-decked, adding dead load.
  • 1 Aug 2007 — Construction materials staged over U10; bridge collapses during rush hour.
Failure sequence
  • Undersized U10 gusset plates carried increased dead load.
  • Construction staging added ~578 kips over the critical node.
  • Gusset plate buckled in shear/compression.
  • Non-redundant deck truss collapsed completely.
Structural analysis
  • Required gusset thickness ≈ 1 in; provided 0.5 in — a design-office error carried for 40 years.
  • Demand-to-capacity at U10 ≈ 2.0 at the time of collapse.
  • Gusset plates were not checked in routine inspections — inspections looked for corrosion and fatigue cracks.
  • Fracture-critical, non-redundant configuration meant no alternate load path.
Redesign
  • Thicken gusset plates and check Whitmore section, buckling, block shear, and net section rupture.
  • Add redundancy or classify and inspect as fracture-critical with gusset-specific checks.
  • Control construction staging loads with an engineered load-management plan.
Lessons learned
  • Original design errors can hide for decades.
  • Inspection scope must match the actual failure modes.
  • Staged construction loads need engineering review.
Proposal ideas
  • Gusset plate capacity screening of state-owned deck truss bridges
  • Load-management protocol for construction staging on non-redundant bridges

Champlain Towers South (Surfside)

2021 · RC flat plate / durability98 killed
Timeline
  • 1981 — Constructed with a pool deck flat slab over the garage.
  • 2018 — Engineering report documents major structural damage and failed waterproofing.
  • 2021 — Repairs not yet executed; the building collapses at 1:22 a.m.
Failure sequence
  • Chronic water intrusion through failed pool-deck waterproofing.
  • Corrosion of slab and column reinforcement, section loss at the slab-column interface.
  • Punching shear failure at pool-deck/column connections.
  • Progressive collapse propagated into the tower.
Structural analysis
  • Pool deck slab reportedly lacked required punching-shear capacity even as-built (low slab thickness, misplaced top steel).
  • Punching: vu at the critical perimeter d/2 exceeded φvc after corrosion-related section loss.
  • No structural redundancy in a flat-plate system once punching initiates — failure is brittle and propagates.
  • Deferred maintenance turned a durability problem into a strength problem.
Redesign
  • Provide shear studs/stirrups or drop panels at slab-column connections.
  • Design waterproofing as a structural durability element with a maintenance plan.
  • Increase cover and use corrosion-resistant reinforcement in wet exposure classes.
  • Mandatory structural recertification with enforcement teeth.
Lessons learned
  • Durability failures become strength failures.
  • Flat plates are punching-shear critical and non-redundant.
  • Reports without executed repairs save no one.
Proposal ideas
  • Punching-shear vulnerability screening of aging coastal flat-plate garages
  • Chloride-ingress service-life modeling for pool decks over occupied space

Tacoma Narrows Bridge

1940 · Aeroelasticity / suspension bridgesNo human deaths (one dog)
Timeline
  • 1940 Jul — Opened; immediate vertical oscillation observed ('Galloping Gertie').
  • 7 Nov 1940 — 40 mph wind induces torsional flutter; deck fails after ~70 minutes.
Failure sequence
  • Shallow, solid plate girder deck with very low torsional stiffness.
  • Wind produced self-excited aeroelastic flutter (not simple resonance).
  • Torsional amplitude grew until the deck and hangers failed.
Structural analysis
  • Deck depth/span ratio 1:350 versus contemporary practice ~1:100 — extremely flexible.
  • Solid girders shed vortices and prevented pressure equalization; an open truss would have vented.
  • Flutter is a negative-damping instability: energy input from wind exceeds structural damping.
Redesign
  • Open stiffening truss deck with wind grating for pressure equalization.
  • Increase torsional stiffness (closed box or truss).
  • Wind tunnel testing of section models — now standard practice.
Lessons learned
  • Slenderness driven by aesthetics/economy can create dynamic instability.
  • Wind tunnel testing became mandatory for long spans.
  • Flutter ≠ resonance.
Proposal ideas
  • Section-model wind study for a proposed long-span pedestrian bridge

Silver Bridge

1967 · Eyebar chain / fracture46 killed
Timeline
  • 1928 — Built as an eyebar-chain suspension bridge.
  • 15 Dec 1967 — Rush-hour traffic; a single eyebar fails; the bridge drops in under a minute.
Failure sequence
  • Stress corrosion cracking / corrosion fatigue in eyebar 330 at the pin hole.
  • A 0.1-in defect grew to critical size.
  • Non-redundant two-eyebar chain: one failure released the entire chain.
Structural analysis
  • Fracture mechanics: Kc reached at a crack depth of roughly 0.12 in for the eyebar steel and stress level.
  • The crack was internal to the pin joint and invisible without disassembly.
  • Two-bar chain means zero redundancy — the definition of fracture critical.
Redesign
  • Use multi-bar chains or cable systems with redundancy.
  • Design for inspectability; avoid concealed fracture-critical details.
  • Select tougher steels and control stress ranges.
Lessons learned
  • Created the U.S. National Bridge Inspection Standards (1971).
  • Redundancy and inspectability are design requirements.
Proposal ideas
  • Inspectability scoring method for fracture-critical members in a state inventory

Mianus River Bridge

1983 · Pin-and-hanger assemblies / drainage3 killed, 3 injured
Timeline
  • 1958 — Built with pin-and-hanger suspended spans.
  • 1970s — Deck drains paved over; water discharged onto the hanger assemblies.
  • 28 Jun 1983 — Suspended span drops into the river.
Failure sequence
  • Corrosion product buildup in the pin-and-hanger assembly forced the hanger outward.
  • The outside hanger slipped off the pin.
  • The remaining connection was overloaded and fractured; the span fell.
Structural analysis
  • Corrosion 'rust packing' generated enormous lateral force on the hanger.
  • Assembly was non-redundant and difficult to inspect.
  • Blocked drainage was the initiating maintenance failure.
Redesign
  • Eliminate pin-and-hanger details or provide catcher beams.
  • Design and maintain deck drainage away from all steel details.
  • Provide inspection access to every load-carrying connection.
Lessons learned
  • Maintenance decisions are structural decisions.
  • Detail selection determines lifetime inspection cost.
Proposal ideas
  • Retrofit catcher-beam design for remaining pin-and-hanger bridges in the region
© 2026 Civil Engineering Capstone Studio. All rights reserved.