Ryan G. Johnson, BAE/MAE, Penn State, 2009

Introduction Summary

The Hartford Civic Center Coliseum collapsed in the early morning hours (4:19 A.M.) on January 18, 1978. Hosting site for various concerts and hockey games, the building was filled with nearly five thousand spectators only hours before the collapse. It is not uncommon to see heavy snow storms in Hartford, Connecticut, but the winter on 1978 caused the largest snow storm in the five year life span of the arena. Loading from the snow caused the 300 ft. by 360 ft. space frame roof to deflect so much, the center fell 83 ft. in upon itself, as seen in Figure 1(ENR, Jan 1978).

Space frame construction consists of top and bottom square grids with nodal joints on center, connected by diagonal staggered bars. This type of construction resembles pyramid trusses of steel bars linked together into a uniform roof structure (Ramaswamy, 2002). The horizontal bars of the Hartford Civic Center connected at one node, while the diagonal members connected at another separate node below the first, creating a bending moment. Nodes being separated from one another was a byproduct of the use of steel angles configured into a cross shape. A cross shape offers significantly less efficiency in compression than a tube shape because it bends and twists under relatively low stresses. Lateral bracing of the top chords was met through diagonals in the interior of the frame, but along the edges there was no means to prevent out-of-plane bending (LZA, 1978).

Key Word:

Computer Modeling, Design Errors, Deflection, Faulty Welds, Lateral Bracing, Long Span, Radius of Gyration, Space Truss, Snow Loading, Torsional Buckling

Figure 1: Image courtesy of LZA investigation report

Events Leading up to the Failure

As many as five different subcontractors were involved in the construction, leading to confusion over who was primarily responsible for the entire project. Hints of danger from design errors were noticed during construction and just after the frame was completed in 1972, Design engineers dismissed any concerns, citing modeling discrepancies were to be expected. Construction photos taken during the lifting of the roof showed extensive deflection, as much as double the predicted amount (LZA, 1978). The large deflection had caused the roof fascia subcontractor to make on-site modifications to the attachment method of the panels, from using predefined bolt holes to welding the panels to support beams (Delatte, 2009). Even with these concerns brought to light, nothing was done and the building stood for five years, until the final collapse in 1978.

Design and Construction

Design of the roof space truss was done using a complicated computer program. Fraoli, Blum, and Yesselmann, the structural engineers of the arena, used the computer model to save time and money, an estimated $500,000. The new innovative model consisted of two main roof layers composed of steel members in a 30'x30' grid. Horizontal bars on the upper and lower grid were spaced 21' apart, while being diagonally braced at nodes and, therefore a third intermediate horizontal layer braced the main diagonals, as seen in Figure 2 below. Top layer horizontal members were also braced at their midpoints with intermediate diagonals. A total of 4,455 members made up the roof truss system, where any number of them could have failed due to the building symmetry (Goldberger, 1978). This pyramid design was not typical of space frames of the time, and differ in multiple ways:

Figure 2: Image drawn by Ryan Johnson with information from LZA, 1978

  1. Typical truss nodes normally the intersection point of member center lines. However in the case of the Hartford Civic Center's frame, the top horizontal bars intersected at one point and the diagonals at another. This design had caused bending stresses in the members. Refer to Figure 3 below for allowable loading differences.
    Figure 3: Images drawn by Ryan Johnson based on ENR, April 6, 1978

  2. A cross shape built of four steel angles was used for the frame's top horizontal members. The cross shape section has a much smaller radius of gyration and , therefore does not provide very good resistance to buckling and may may bend or twist under relatively low stresses. I-shapes or tube section configurations with the same amount of material provide better resistance to buckling.
    Figure 4: Images drawn by Ryan Johnson with information from LZA investigation report

  3. Vertical posts at nodes were added above the top horizontal layer to support the roof panels. The designer claimed variations in posts height would help with drainage and the top bars would not be subjected to bending stresses. This differs from a typical frame, where the top members act as a structural member and support the roof.
    Figure 5: Image courtesy of LZA investigation report
  4. Four large pylon legs set 45' inside of the edges of the frame supported the roof, rather than boundary columns or walls.
  5. A fast track method of construction was implemented on the Civic Center project. Work was divided to five different subcontractors, coordinated by the construction manager. The roof construction was one of the subcontracted pieces, where the subcontractor was in charge of assembly and erection of the steel. Space truss members were pieced together on the ground and raised into place in large sections atop temporary supports. Concrete pylons were then built under the truss sections. This method was employed so workers did not need to be suspended in the air for most of the construction, and to save time and money. Roof lifting took almost two weeks to complete and was completed on January 16, 1973. Once the roof truss was in place, the construction manager altered the roof material, increasing the dead load by 20% (Feld and Carper, 1997).

Design Flaws

  • Dead loads were underestimated by more than 20%. The actual frame weight was 23 lb/ft^2, but was estimated to only be 18 lb/ft^2. Under design led to the East and West face face exterior top layer compression members to be 852% overloaded, the North and South face exterior compression members to be 213% overloaded, and the interior East-West compression members to be 72% overloaded (ENR, April 1978).
  • The computer model assumed all of the top chords were laterally braced, but in fact only the interior frame met the criteria because of the diagonal bracing. The exterior edges of the frames diagonal bracing only offered in plane restraints, which allowed the top bars to buckle outward from the bracing.
  • The slenderness ratio of the built up diagonals violated section of the AISC code, which prohibits slenderness ratio of individual components of a built-up member to be larger than the slenderness ratio of the entire built-up member (LZA, 1978). Other violation were that bolt holes punching through sme members exceeded 85% of the gross area.

Construction Errors

  • Interior members were insufficiently braced at midpoints while exterior members were only braced at 30' rather than at 15' as specified in design. No midpoint braces were provided on the top layer members.
  • 3/8" steel cables were found in the wreckage strung throughout the truss. The location of the cables were in the most deficient parts of the frame, which should have raised questions of their presence. All of the contractors involved denied installing the cables, but it was speculated the cables were installed to force the truss back to shape during fascia installation (ENR, April, 1978).
  • Welding of filler plates located on several were found to considerable reduce the connection capacity.
  • Some of the diagonal members were misplaced and the wrong steel strength was used (Feld and Carper, 1997).

Conflicting Accounts of the Failure

Three different consulting engineers were hired to investigate the cause of the failure. Along with different causes of failure, there are discrepancies as to the location of the failure mechanism.

  • LZA believed the initial cause of the collapse was a design deficiency related directly to the inadequate bracing of all top chord compression members of the truss (LZA, 1978)
  • Loomis and Loomis Inc. believed the cause of the failure was due to torsional buckling of the compression members, and that members close to the middle of the truss were critically loaded even before live loads were added. This means of failure is usually overlooked as a cause of failure because it is so uncommon (ENR, June, 1979).
  • Hannsrarl Bandal believed that a faulty weld connecting the scoreboard to the roof was the primary cause of failure. A massive amount of energy would have been caused by the volatile weld release, causing the entire structure to collapse (Feld and Carper, 1997).

Could the Failure have been Prevented

LZA's conclusion specified that there was inadequate inspection and quality control. The city should have hired an independent structural engineering consultant, with full time involvement, to review the designs and be there during assembly and erection of the space frame. Hartford did require an independent peer review of technical designs on large projects for private developers, but since the Civic Center was owned by the city, there was no review performed.

Legal Repercussions

  • An out-of-court settlement between all parties involved was reached six years after the collapse. Hartford settled for $25 million, about half the original amount claimed.
  • Designers may be hired to preform traditional services, but courts may still find them responsible because they are licensed professionals who are liable for public safety.

Lessons Learned

A system of checks and balances needed to be preformed. The use of traditional design factors of safety, used with nontraditional methods is something that should never be overlooked. However, such a large error in design cannot be compensated by an arbitrary design factor of safety increase. The computer is simply an analytical tool and can never guarantee the correct solution. The operator should be experienced and competent about all information that is put into the model and fully understand all of the information given out. Assumptions should be checked and compared to the as-built conditions to verify that field measurements match those of the original design. In general terms, it was noted that there was confusion over a number of design and construction responsibility issues that contributed to the collapse but could have been avoided.

The Hartford Coliseum roof collapse did not put an end to space frame roof designs. Many similar truss structures are found all over the world: Standsed Airport in London,England, SkyDome in Toronto, Canada, and Collins Place in Melbourne, Australia. A notable structure that was of similar design and also collapsed was Kemper Arena, in Kansas City.


  • Delatte, Norbert J. Jr., Ph. D., P.E., 2009, Beyond Failure: Forensic Case Studies for Civil Engineers, ASCE, New York, NY, 174-184
    • Offers a wide variety of information dealing with design and construction, causes of failure, technical aspects, professional and procedural aspects, ethical aspects, and educational aspects.
  • Feld, Jacob, and Carper, Keneth L., 1997 2nd Edition, Construction Failures, Wiley and Sons, INc., New York, NY, 198
    • Discusses general stability problems in the structure from the time of construction until it eventually fell. Information on how the city of Hartford handled the event is also discussed.
  • Kaminetzky, Dov., 2001, Design and Construction Failures: Lessons from Forensic Investigations, McGraw Hill, New, NY, 220-225
  • Lev Zetlin Associates, Report of the Engineering Investigation Concerning the Causes of the Collapse of the Hartford Coliseum Space Truss Roof on January 18, 1978, June 12, 1978
  • Levy, Mathys, and Salvadori, 1992, Why Buildings Fall Down: How Structures Fail, Norton and Company, Inc., New York, NY, 68-75
    • A first hand narrative of an occupant from the hotel across the street discusses what he had seen/felt. Discusses concerns the public had with the building construction and eventual collapse and how the building was rebuilt.
  • Martin, Rachel, and Delatte, Norbert J., 1999, Another Look at Hartford Civic Center Coliseum Collapse, ASCE, New York, NY
    • In depth variation on Delatte’s book dealing with how the computer is only an analytical tool and results must be computed by the designer. Lessons for engineering students of the concerns that arise during design and construction of complex structures.
  • Petroski, Henry, 1985, "From Slide Rule To Computer", To Engineer is Human, St. Martin's Press, New York, NY, 198-200
    • Information on why the use of computers without a background on the technical information the program uses leads to concerns. A computer is frequently used for optimization of design through simplification but engineer should be able to take what information a computer gives and partially replicate it by hand calculations.
  • Ramaswamy, G. S., Eekhout, M., Suresh, G. R., 2002, Analysis, Design and Construction of Steel Space Frames, Thomas Telford Publishing, Heron Quay, London, 2-4
    • An introduction into advantages of space frames and the components of it are discussed. The information presented is up to date with modern techniques, which may not have been available during the design in 1978. If the information was available, the outcome may have been prevented
  • "Space frame roofs collapse following heavy snowfalls", ENR, January 26, 1978
  • "Design flaws collapsed steel space frame roof", ENR, April 6, 1978
  • "Collapsed space truss roof had a combination of flaws", ENR, June 22, 1978
  • "New theory on why Hartford roof fell", ENR, June 14, 1979
  • Fellows, L., "Coliseum Roof Collapse at Hartford Civic Center", Special to the New York Times, January 19, 1978
  • Goldberger, P., "Hartford Cave-in Still a Mystery", New York Times, January 20, 1978
  • News Dispatches, "Hartford Civic Center Roof Falls", The Washington Post, January 19, 1978