I-35W Mississippi River Bridge — Half-Inch Gusset Plates That Buckled Under Added Weight

Summary

On 1 August 2007, at 6:05 p.m., the eight-lane steel deck-truss bridge carrying Interstate 35W over the Mississippi River in Minneapolis, Minnesota collapsed in seconds during the evening rush, dropping a 456-foot main span and its approaches into the river and onto the banks. Thirteen people died and 145 were injured; 111 vehicles were on the failed deck. The National Transportation Safety Board (NTSB), the federal investigating body, traced the collapse to a single class of component: the gusset plates that connected the steel members at the bridge's main-truss joints. At the joints designated U10, those plates were roughly half the thickness they should have been — a design error baked into the structure when it was built in the 1960s.

The mechanism was instability, not rupture from corrosion or fatigue. A gusset plate is the steel sheet that ties a truss's diagonal, vertical, and chord members together at a joint, and it must be thick enough to carry the combined forces without buckling. The U10 plates were 0.5 inches thick where the design demanded roughly twice that, and for four decades they carried traffic because the loads stayed within the margin even an undersized plate retained. The NTSB found two slow, additive overloads erased that margin: concrete resurfacing raised the permanent dead load about 20 percent, and on the afternoon of the collapse a resurfacing project parked an estimated 578,000 pounds of equipment, sand, and aggregate over the weakest joints.

When the demand on the U10 plates finally exceeded their buckling capacity, the plates failed by lateral instability — they folded. The deck truss was non-redundant and fracture-critical: with a main-truss connection gone, there was no alternate load path, and the loss of one set of joints unzipped the center span. The collapse propagated across the full 1,907-foot bridge in roughly four seconds.

What makes I-35W a permanent case is that the fatal flaw was not wear, weather, or neglect, but an original calculation that was never done. The plates were sized below the loads they would carry, the error survived design review, and decades of added dead load plus one day of stacked construction material brought the demand to the point the deficient plates could not hold. The bridge did not fail because it grew old; it failed because it was never strong enough at one joint, and no one ever checked.

Timeline

1964-1967

Design and construction

Sverdrup & Parcel and Associates designs Bridge 9340 to the 1961 AASHTO specifications as a continuous steel deck truss carrying I-35W over the Mississippi. The main truss spans 456 feet over the navigation channel; total length is 1,907 feet. The structure is fracture-critical — it lacks redundant members, so the failure of certain components means loss of the span.

1967

Bridge opens

The eight-lane bridge opens to traffic in November 1967. The undersized U10 gusset plates — 0.5 inches thick — are in the structure from day one, carrying loads below their already-deficient buckling capacity.

1977

First overlay added

A two-inch concrete overlay is placed on the deck, raising the permanent dead load the truss must carry for the rest of its life.

1998

Second overlay and widening

Further deck and median modifications add weight. Cumulatively, resurfacing and modifications raise the bridge's dead load by roughly 20 percent over the original design.

1990s-2000s

Repeated "structurally deficient" ratings

Federal inspections rate the bridge "structurally deficient" for years, citing fatigue and corrosion in the deck truss. Inspectors check for cracks and section loss, not for whether the plates were correctly sized in the first place.

Early 2007

Fatigue study weighs retrofit vs. inspection

Concern over fatigue cracking leads to a study; the state opts for intensified inspection over a costly retrofit. The gusset-plate deficiency stays undetected because no one re-runs the original capacity calculations.

Summer 2007

Resurfacing project begins

A contractor mobilizes to resurface the deck. Lane closures concentrate traffic, and construction equipment and materials are staged on the bridge.

1 August 2007, afternoon

Construction loads stacked over U10

An estimated 578,000 pounds of equipment, sand, and aggregate is positioned on the deck near the center span, directly over the U10 joints — a heavy, concentrated live load atop the already-elevated dead load.

1 August 2007, 6:05 p.m.

The U10 plates buckle

During the evening rush, demand on the U10 gusset plates exceeds their buckling capacity and the plates fail by lateral instability. With a main-truss connection lost, the deck truss has no alternate load path.

1 August 2007, 6:05 p.m.

Progressive collapse

The center span drops into the river; the loss propagates through the continuous truss and the bridge collapses across its full length in about four seconds. 111 vehicles fall with the deck; 13 people are killed and 145 injured.

August 2007

NTSB opens federal investigation

Investigators recover the wreckage, reconstruct the joints, identify fractured and buckled gusset plates at eight locations in the center span, and recompute the design loads the plates should have been sized for.

14 November 2008

NTSB report adopted

The Board issues report HAR-08/03. Probable cause: inadequate load capacity of the U10 gusset plates, due to a design error by Sverdrup & Parcel, which failed under the combined dead-load increases and concentrated construction and traffic loads.

The Deck Truss and the Joints That Held It Together

Bridge 9340 was a continuous steel deck truss: the roadway rode on top of a deep triangulated steel frame, the trusses hanging beneath the deck rather than beside it. A truss carries load by resolving it into tension and compression along straight members — the diagonals, verticals, and top and bottom chords. At every joint where those members met, a pair of flat steel sheets called gusset plates riveted the members together and transferred force from one to the next. The gusset plate is the truss's connection, where the load path turns a corner: if a member is the road the force travels, the gusset is the intersection.

A gusset plate carries force in shear and bearing, but its governing limit is often stability. A flat plate in compression does not need to rupture to fail; it can buckle — fold sideways out of plane — at a stress well below the steel's strength. The thicker the plate, the higher the load it carries before becoming unstable, so sizing a gusset is a buckling calculation, not merely a strength check. At the U10 joints, that calculation produced plates 0.5 inches thick where roughly one inch was required — about half the thickness the loads demanded. They were not corroded down to that thickness over time; they were fabricated that way. The deficiency was geometric and permanent, sitting in the structure's most heavily loaded main-truss joints from the first day of traffic.

How Slow Weight Met an Undersized Plate

An undersized plate is not the same as an immediately failing one. The U10 gussets carried four decades of traffic because, even at half thickness, they retained some reserve above the original 1960s loads. The disaster was the closing of that reserve from two directions.

The first was dead load that crept upward and never came back down. The two-inch concrete overlay placed in 1977, plus later deck and median modifications, added permanent weight the truss bore every second of every day — an increase of about 20 percent over the original design dead load. Unlike traffic, dead load does not pass; it accumulates and stays, and each overlay pushed the steady-state demand on the U10 plates closer to their buckling threshold, narrowing a margin that had never been generous.

The second was a concentrated live load at the worst place. The 1 August resurfacing operation staged roughly 578,000 pounds of equipment, sand, and aggregate on the deck near the center span — directly over the U10 joints. Stacked atop the elevated dead load and rush-hour traffic, that concentrated weight pushed the demand on the half-thickness plates past their capacity to resist instability, and the NTSB found the U10 plates buckled. The bridge had absorbed forty years of growing load with no visible distress because the plates were merely overstressed, not yet unstable. The stockpile supplied the final increment that turned an undersized connection into a buckled one — and the structure had no second path to carry the load the failed joint dropped.

The Reckoning: A Calculation Never Performed

The NTSB's conclusion was unusually clean for a structure rated "structurally deficient" for years and widely assumed to have died of old age and fatigue. The Board found that corrosion and fatigue cracking, though present, did not cause the collapse. The cause was an original design error: the U10 gusset plates were sized below the loads they had to carry, and that under-sizing escaped the original design review. The Board named contributing factors with equal precision — the designer's lack of quality-control procedures to ensure the main-truss gusset-plate calculations were actually performed, and inadequate review by the firm and by the federal and state agencies who later inspected the bridge without ever questioning whether the plates had been correctly sized.

This is the defining feature of the case. The bridge had been inspected for decades, with inspectors crawling the truss for cracks, corrosion, and section loss — damage to a presumed-adequate structure. No inspection regime in that era checked whether the gusset plates had been designed thick enough to begin with, because design adequacy was assumed settled in 1965. The error was not introduced by time; it was present at completion and invisible to every inspection that searched only for deterioration. There was no villain, no fraud, no defective steel — only an undersized plate, a verification never done, and a slow accumulation of weight that eventually found the limit no one knew was there.

Contributing Factors

01

The governing gusset-plate limit was buckling, and the plates were sized below it

A gusset plate in compression fails by lateral instability long before it ruptures, so its required thickness is set by a buckling calculation, not a strength check. The U10 plates were about half the thickness their loads demanded. When the governing failure mode of a connection is stability, under-thickness is the whole margin. Size connection plates for buckling, and treat thickness as the safety variable it is.

02

Dead load crept upward and was never re-checked against capacity

Two inches of added concrete overlay and later modifications raised the permanent dead load roughly 20 percent over the original design. Dead load, unlike traffic, accumulates and never leaves; each addition was approved as routine resurfacing, but their sum permanently raised the demand on every truss joint. Any addition to permanent weight is a load increase that must be checked against the as-designed capacity of the governing members.

03

Construction staging concentrated a heavy live load over the weakest joints

The resurfacing project parked an estimated 578,000 pounds of equipment and material near the center span, directly above the U10 joints — the most heavily loaded connections on the bridge. Concentrated construction loads on an in-service structure must be analyzed against its actual reserve and located away from its critical members. A stockpile is a load case; where it sits decides what it threatens.

04

The structure was fracture-critical, so one joint's failure was total

The deck truss had no redundant members and no alternate load path. The loss of the U10 connection did not cause local distress that could be caught and shored; it removed a link in the only chain carrying the span, and the collapse propagated across 1,907 feet in seconds. In a non-redundant structure, every governing connection must be designed and verified as if its failure ends the bridge — because it does.

05

Design adequacy was assumed settled, so no inspection ever re-verified it

Decades of inspections searched for corrosion and fatigue cracks — deterioration of a structure presumed correctly designed. None re-ran the gusset-plate capacity calculations, because the original design was treated as a closed question, and an error present at completion is invisible to a regime that only looks for new damage. Periodic re-rating must ask whether the structure was ever adequate, not merely whether it has degraded.

Aftermath

The I-35W collapse killed 13 people and injured 145, and it broke the assumption that a bridge inspected for cracks and corrosion is a bridge whose strength is known. The NTSB's finding — that an original gusset-plate design error, not deterioration, brought the structure down — forced a national reckoning over a component that bridge inspection had largely ignored. The Federal Highway Administration issued advisories directing owners to load-rate the gusset plates of steel-truss bridges, not just the members, and to inspect them for distortion; agencies recomputed capacities for hundreds of similar deck-truss bridges nationwide that had never had their gusset plates independently rated. The collapsed structure was replaced by the I-35W Saint Anthony Falls Bridge, opened on 18 September 2008, designed with redundancy and instrumented with sensors. Beyond the codes, "I-35W" became shorthand in the profession for the danger of treating original design as settled, and for the way slow, additive dead load can quietly consume the margin of an undersized connection until a single day of concentrated weight finds the limit no one knew was there.

Lessons

Size connection plates for their governing failure mode — for a gusset in compression that is buckling, not strength — and treat thickness as the safety margin it physically is.
Re-check capacity every time you add permanent dead load: an overlay or a widening is a structural change, and the demand it raises never goes away.
Locate and limit construction staging by the structure's actual reserve, and keep concentrated loads off the critical, most heavily loaded joints of an in-service bridge.
Design and verify every governing connection in a non-redundant, fracture-critical structure as though its failure ends the span — because without an alternate load path, it does.
Re-rate for original adequacy, not only for deterioration: an inspection that searches only for new cracks and corrosion will never catch a member that was undersized on the day it was built.