Hartford Civic Center Roof Collapse: Causes, Lawsuits, and Lessons

The Hartford Civic Center roof collapse occurred at roughly 4:19 a.m. on January 18, 1978, when the coliseum’s 300-by-360-foot steel space-frame roof dropped 83 feet onto 10,000 empty seats, about six hours after a UConn men’s basketball game had emptied the arena.¹ No one was killed or hurt. Investigators concluded the heavy snow on the roof that morning was only the trigger. The design itself had been failing from the day it was built.

What Happened That Morning

The night before, nearly 5,000 fans watched UConn beat UMass 56–49 inside the arena. The winter of 1978 had already brought the worst weather the five-year-old building had seen, and a fresh storm the previous evening added to roughly ten days of accumulated snow and ice sitting on the roof.² In the early morning hours, with the building empty, the center of the roof caved in on itself and destroyed the seating bowl below.³ The timing is what kept it from becoming a mass-casualty disaster. A collapse during the game, or during cleanup, would have been catastrophic.⁴

The Roof That Failed

The coliseum had opened in January 1975 as Hartford’s main entertainment venue and the home of the New England Whalers.⁵ Vincent G. Kling and Partners of Philadelphia designed the building, with structural engineering by Fraioli-Blum-Yesselman.⁶

The signature feature was the roof: a 1,400-ton steel space frame made up of 4,455 members arranged in a pyramidal truss pattern and resting on four concrete columns.³ It was one of the first large-span roofs designed and analyzed entirely by computer, with the goal of saving several hundred thousand dollars by trimming material.⁵ The frame was assembled on the ground, lifted into place, and the arena was built around it.²

The project used a fast-track construction method with as many as five subcontractors under a construction manager. Hartford required independent peer reviews for private development, but no outside review was done here because the city owned the building.³

Why It Collapsed

Three forensic engineering firms examined the wreckage: Lev Zetlin Associates (LZA), hired by the city and reporting in June 1978; Smith and Epstein; and Loomis and Loomis. The U.S. Army’s Cold Regions Research and Engineering Laboratory separately studied the snow load on the roof at the time of failure.³⁷ All of them agreed the structural design was fundamentally flawed. The city investigation concluded the roof had been undergoing “progressive failure as soon as it had been installed.”¹

Compression Members Too Weak for the Loads They Carried

The top chords of the space frame were built from four steel angles arranged in a cross-shaped (cruciform) section. That shape buckles more easily than a tube or an I-section.⁴ The diagonal bracing members all sat in the same inclined plane, so the top chords were braced in only one direction. The design assumed an unbraced length of about 4.57 meters, but the chords were effectively unbraced over 9.14 meters, roughly double what the calculations relied on.⁸ LZA found that exterior compression members on the east and west faces of the frame were overloaded by as much as 852 percent.

Joints That Did Not Meet Where the Design Assumed

The top horizontal bars and diagonal bars were supposed to meet at a single node. In the built structure they met at different points. That joint eccentricity introduced bending forces in members that the computer model had never considered.⁴ The frame also lacked camber, the upward curve engineers build in to offset expected sagging, and it had no structural redundancy. Once a member failed, nothing was in place to take the load, and failures cascaded.

Overreliance on a Computer Model

The analysis software used a simplified model that did not capture the frame’s real three-dimensional behavior and missed local weaknesses.⁹⁴¹⁰

Disagreement on the Precise Trigger

The investigators agreed on the design flaws but split on the exact failure mechanism. LZA pointed to inadequate bracing of the top-chord compression members. Smith and Epstein described compression members, braced in only one plane, buckling out of plane and shedding load to neighbors until the frame gave way. Loomis and Loomis argued torsional buckling, not lateral buckling, started the collapse, and their analysis found members near the support columns were close to torsional capacity even before the storm; as little as 12 to 15 pounds per square foot of added load would have been enough.⁴ A separate investigator, Hannskarl Bandel, blamed a faulty weld connecting the scoreboard to the roof.³

Lawsuits and Settlement

The City of Hartford and The Travelers Insurance Company sued the parties responsible for the design and construction. In March 1984, six years after the collapse, the Hartford City Council unanimously accepted an out-of-court settlement under which the city stood to receive $1.8 million and Travelers more than $10.6 million.¹¹ A separate accounting puts the total settlement across all parties at roughly $25 million, about half the amount originally claimed. The number of subcontractors involved in the fast-track project made it hard to pin down primary responsibility, though the investigating firms noted that licensed design professionals remain accountable for public safety regardless of how a project is organized.³

What Engineers Took From It

The Hartford collapse became one of the most studied failures in American structural engineering. A 2001 paper in the American Society of Civil Engineers’ Journal of Performance of Constructed Facilities by Rachel Martin and Norbert J. Delatte presented it as a lesson in the “difficult technical, professional, procedural, and ethical issues that may arise during the design and construction of a complex, high-occupancy structure.”¹⁰

The central takeaway concerned the limits of computer analysis. Software is a tool, not a substitute for judgment, and computed results have to be checked against what the structure is actually doing. Deflections at twice the predicted values during construction were a warning that went unanswered. The Institution of Structural Engineers has pointed to Hartford as a historical example of expertise bias and overreliance on analysis software.¹² The absence of an independent peer review on a city-owned project also shaped how jurisdictions came to think about outside oversight of government construction.

The arena itself was rebuilt with a new roof design compared to a railroad truss bridge, the columns raised 12 feet to add about 4,000 seats, and a protocol requiring deflection measurements at the four corners of the roof at least twice a year. The rebuilt building reopened in February 1980.²¹³

1
Connecticut History. Almost a Tragedy: The Collapse of the Hartford Civic Center
2
Connecticut History. Civic Center Roof Collapses
3
Penn State Engineering. Hartford Civic Center
4
SE Licensure. Structural Failures: Hartford Coliseum
5
Today in Connecticut History. Hartford Civic Center Roof Collapses
6
The New York Times. Fears of a Roof Defect Preceded Collapse
7
ERDC Library. Snow Loads on the Roof of the Hartford Coliseum
8
Brady Heywood. Hartford Civic Center Collapse Analysis
9
ASCE. Engineers Are Not Infallible
10
ASCE Library. Another Look at Hartford Civic Center Coliseum Collapse
11
The New York Times. Hartford Settles Roof-Collapse Suit
12
IStructE. Lessons From Failure
13
ICC. Renovations Improved Construction, Hockey and Music Saved the PeoplesBank Arena