C.W. Post College Auditorium Collapse
Nate Babyak, BAE/MAE, Penn State 2011

Permission pending for illustrations

Introduction

In the early hours (between 2:00 and 3:00 A.M.) of January 21, 1978, the C.W. Post Dome Auditorium in Brookville, NY collapsed. The building had survived seven years of wind, rain, and snow, but the winter of 1978 proved to be too much for the dome to overcome. Fortunately, for those at C.W. Post University, the 3,500 person auditorium was empty at the time of the collapse. While it was drifting snow and ice that ultimately caused the collapse of the 170 foot span shallow dome to collapse, an investigation into the collapse discovered that a faulty design theory left the dome vulnerable to collapse.
Babyak_C.W._Post_Location.JPG
Figure 1. Location of Brookville, NY (Bing.com)


Keywords

C.W. Post University, Long Span Structure, Design Errors, Snow Caused Collapse, Membrane Theory, Buckling

Dome Design

Babyak_CW_Post_Dome_Design2.JPG
Figure 2. Diagram of Dome Design (Adapted from illustrations by Kevin Woest in Why Buildings Fall Down)



In 1970, the Dome Auditorium was built to provide a 3,500 person theatre for C.W. Post University in Brookville, NY. The shallow dome spanned 170 feet and was supported by 20 exterior steel columns. The roof rose 24 feet from the column tops and stood 43 feet above the arena floor (ENR, January 1978). The dome itself was constructed of 40 steel pipe trusses that served as meridians while steel channels circled the dome creating hoops. To further strengthen the dome, a steel compression ring was at the top of the dome while a canopy surrounding the bottom of the dome acted as a tension ring. In order to brace the dome, two steel pipes served as cross bracing between the meridian trusses, but they were only placed at alternating sections between adjacent trusses (Salvadori 2002). The Dome, designed, built, and erected by Butler Manufacturing Company, stood in compliance with local building codes and passed multiple inspections prior to its collapse in 1978.

Collapse


Seven years after the Dome Auditorium's completion, the dome collapsed following a round of strong snowstorms. Just months earlier, an inspector from the Brookville building department found the auditorium in satifactory condition and issued a permit of occupancy during an annual inspectation of the facility. The permit of occupancy indicated that the building's drawings had been checked by the New York State Code Bureau and nothing was found to be structurally deficient. So, if the building was determined to be code compliant, what caused the collapse? Did the Butler Manufacturing Company make a mistake in the structural design? Was there an error in the construction of the dome? An investigation by Nichoals W. Koziakin, a structural engineer with Mueser, Rutledge, Johnston, and DeSimone in New York City, concluded that the simplified theory used in the dome's design was inapplicable to the reticulated dome's structure (Salvadori 2002).

As Koziakin became more familiar with the collapse of the dome, he focused on how the dome was designed. He found that the dome had been designed using a simplified membrane theory that did not fit the dome's actual construction. The membrane theory has several key assumptions that must be met in order for the theory to be applicable to a domed structure. Unfortunately for C.W. Post University, the Dome Auditorium did not meet these criteria. The membrane theory was derived to evaluate the structural behavior of perfectly spherical or rotational dome shapes which are:
1.) made out of materials with identical properties at each point, and in any direction on the dome surface,
2.) under gravity loads perfectly symmetrical about the vertical axis of the dome or
3.) in the case of winds, under horizontal loads equally pushing on the wind side and pulling on the leeward side of the dome (Salvadori 2002).
The wind assumption has long been disproved by further computer analysis and we now understand that the wind actually causes suction on the top of a shallow dome. This theory was often applied to classical concrete or masonry domes like the Pantheon or Brunelleschi's Dome.


Babyak_CW_Post_Dome_Collapse_Diagram2.jpg
Figure 3. Diagram of Dome Collapse (Adapted from Illustrations by Kevin Woest in Why Buildings Fall Down)
The first assumption that the dome acted with isotropic properties can be easily deemed incorrect when you consider that the dome was made up of steel trusses, cross bracing, and wood fiber roofing. The steel only exists where the members are placed and no where inside the roof mesh itself. For this reason, the dome should be considered nonisotropic. The second assumption that the roof would only face uniform loads shows that the original design did not take snow drift and ice build up into consideration placing the dome in a very vulnerable position during winter time. Figure 2, at left, illustrates how snow and ice built up on the leeward side of the roof and placed greater stresses on that section of the dome. The snow drift problem exacerbated the potential of buckling in the dome. The diagonal cross bracing was not designed to take substantial compressive forces but only to prevent the trusses from twisting (Salvadori 2002). As the snow built up on the dome, these members eventually were compressed passed their limit causing buckling and the dome to snap through.

In addition to the design theory assumptions, there were other errors that may have also factored into the collapse. The dead load had been undervalued by 17 percent, the compressive force in the meridional trusses had been alloted equally to their upper and lower chords (despite bending in the trusses), and an incorrect safety factor had been made in the evaluation of the dome's stability against snap through (Salvadori 2002). All of these design errors came to a front during the winter of 1978. The result was the collapse of the $1.5 million C.W. Post Dome Auditorium leaving C.W. Post University without a large auditorium that it had long desired.

Industry Reaction

Two major discussions raged following the collapse of the C.W. Post Dome Auditorium. One discussion related to how the building codes dealt with snow loads. Over the winters of the late 1970s, hundreds of roofs collapsed across the United States due to snow loads. These collapses caused a public outcry for improvements to local and national building codes, but engineers, architects, and contractors had to discover what the root cause for these collapses were. Following these discussions and analyses, designers were prompted to revise ANSI A58.1 (the snow code at the time which has developed into ASCE 7) in 1980. The second discussion dealt with the safety of long span buildings. In addition to the snow caused collapses, a rash of long span structures in the United States collapsed in the late 1970s (See Hartford Civic Center, Kemper Arena, and Rosemont Horizon Arena). These collapses prompted the American Institute of Architects (AIA) to convene a panel of prominent architects, engineers, and contractors to review the procedures for design and construction of long span structures (Kliment 1981). The panel investigated and presented findings on the unique issues of long span structures, controlling the process for building long span structures, design codes and materials among other issues. These discussions have ultimately led to safer building designs.

Reaction to Snow Code

Due to the hundreds of collapses that occured during the winters of the late 1970s, it was clear that something must be done to help minimize the risks caused by snow loads. Engineers realized that it was impossible to prevent all roof collapses, but there were some things that they could do to allow structures to better perform under heavy snow conditions. The first problem in the code they sought to rectify dealt with snow drift. According to insurance statistics, 75% of snow-related collapses are due to snow drifts on lower level roofs (Seltz-Petrash 1979). Up to that point in time, most structures had been designed strictly for a uniform snow load, often 25-30 psf depending on location. Charles Thorton, president of Lev Zetlin Associates at the time, remarked, "there are an awful lot of roofs designed to 30 psf that have never had a problem" (Business Week 1978). With a typical factor of safety of about 2, roofs would only be able to safely hold 50 psf of snow on the roof. However, it was becoming clear that this safety factor was not enough. In 1972, ANSI A58.1 added snow drift recommendations and calculations. This occured after the design of the Dome Auditorium illustrating that the designers of the dome may not have fully understood the consequences of snow drift. This doesn't mean that the designers should get a free pass. The Douglas Fir Use Book, a document from 1958, includes recommendations for calculating unbalanced loading by loading half the roof with the snow load and putting no load on the other half. The designers of the Dome Auditorium clearly didn't understand the potential hazards of snow drift. Following the late 1970s roof collapses, the ANSI A58.1 code recommendations were altered.

There were several key improvements to ANSI A58.1 in 1980. One improvement was to create better ground snow load maps by gathering evidence from more stations and using longer records than had previously been possible. While this allowed the code to improve, the ground snow load maps remain a hot spot for code controversy. Many municipalities including State College, PA, have adopted local amendments to the ground snow load. Drift requirements were added to deal with pitched or curved roofs, valleys formed in multiple series roofs, roof areas adjacent to sloping roofs, and lower level roofs adjacent to higher roofs (Seltz-Petrash 1979). The revised code also began to consider roof exposure and thermal properties of a flat roof and how those properties affected the roof snow load. Today, you can still see the results of the changes to snow codes in 1980. Snow drift is a major factor in almost every building design in a snow laden area and calculations can be derived straight from ASCE 07-10. We suffered millions of dollars of damage due to snow-related collapses, but we took the collapses as an opportunity to learn more about how snow affected structures and, today, we apply what we have learned to design safer stronger structures.

Safer Long-Span Structures

Babyak_CW_Post_Collapsed_Dome.JPG
Figure 4. Aerial View of the Dome Collapse (Towards Safer Long Span Buildings 1981)

Following the collapses of several prominent long span structures in the 1970s, the AIA produced a guide, Towards Safer Long Span Buildings, based on a panel discussion amongst architects, engineers, and contractors that sought to develop recommendations and guidelines for the design and construction of long span buildings. The guide discusses long span buildings comprehensively focusing on design, engineering, and construction aspects. The guide raises awareness to issues that plague long span designs and helps focus designers on ways to improve the safety of long span structures. The panel investigated and presented findings on the unique issues of long span structures, controlling the process for building long span structures, design codes and materials among other issues.

The discussion first focused on the uniqueness of long span structures. One area of discussion was the lack of redundancy in long span buildings. While shorter smaller structures typically have many ways of alternative support, long span structures have relatively few major structural elements, and fewer alternative paths of load resistance (Kliment 1981). Lack of redundancy makes each support and each connection that much more important to the structure as a whole. It is imperative for designers to allow for a greater safety factor in order to prevent collapse. Long span buildings also present a great danger to people in the event of collapse due to the vast amount of people that they typically house. This adds another factor of safety that designers must account for in order to prevent loss of life. Fortunately, none of the mentioned long span structures were full of people when they collapsed .

The panel then focused on the roles and responsibilities of each design team member and recommended ways to control the entire process of completing a long span building. The uniqueness of a long span structure makes it imperative that each team member know their role and be willing to work together to create a safe structure that the owner desires. Today, with the increasing advances in building information modeling, the collaboration amongst all team members has made it easier to work together, share ideas, and diffuse issues before they exist in a built environement. In the design process, it becomes increasingly important for the contractor to be involved early in order to suggest ways of constructing such large structures and allow the designers to see the consequences of certain design decisions. Engineers and contractors must also work closely to ensure that the structure is built as designed and that the building is properly inspected throughout the construction process. The panel emphasized the importance of the role and responsibility of the owner, in addition to the design team. The owner must understand the consequences of different project delivery methods and how they want to assemble the design team when deciding to build a long span structure. A successful team can help produce a safe long span structure that the owner can enjoy.

The final discussion among the panel related to design codes and materials. With the collapse of the C.W. Post College Auditorium and the Hartford Civic Center occuring during the winter of 1978, it was no suprise that the engineers felt that greater importance be placed on the effects of snow, especially drifts. As discussed earlier, there are now much improved code regulations focused on snow loading. Secondary effects like creep, shrinkage, and temperature become much more important in long span structures due to fewer expansion joints in most long span structures (Kliment 1981). It is important for engineers to not neglect these effects like they may in certain smaller structures. In some instances, the secondary effects may control causing the designers to increase the factor of safety. Materials can have a major impact on the severity of secondary effects. It is important to consider the shape and form of the structure when choosing materials because some materials will better suit the needs of the design and the stresses applied to the material. The panel concluded that more careful selection and tighter quality control of materials become more important as span increases (Kliment 1981).

The overall impression the panel sought to deliver was that it isn't possible to just scale up ideas that worked for smaller structures and apply them to long span structures. A long span structure demands greater knowledge of structural loads and materials. Details that may not have been important in smaller structures can suddenly have great importance in a long span structure. It is the responsibility of the architect, engineer, and contractor to confront the design challenges that a long span structure provides and deal with them in a controlled, collaborative manner to provide the long span structure that the owner desires.

Lessons Learned

Babyak_C.W._Post_Today.JPG
Figure 5. Aerial View of C.W. Post Dome Auditorium Today (Bing.com)
Long span structures present many critical issues that typical structures do not face. A few of these issues can be solved through compliance with building codes, but others require solid engineering analysis and judgement. Today's building codes have changed to be more critical of snow drifting and wind load effects than they were in the 1970s, but it's still the responsibility of the designer to ultimately understand what he creates. The code can't make up for poor decisions. It's critical to understand the assumptions of design theories and latest computer modeling techniques to ensure that the actual design fits the engineering assumptions made in the process of design. Engineers must have a feel for the behavior of long span buildings in addition to the ability to use the computer as a tool for analysis (Kliment 1981). The collapses of the C.W. Post Dome Auditorium and other long span structures forced architects, engineers, and contractors to better understand how long span structures work in order to better preserve public safety.

The collapse of the Dome Auditorium left C.W. Post without the largest assembly hall on its campus and the university had to decide what to do with what was left of the Dome Auditorium (Murphy 1982). The university decided to construct a new assembly hall inside the walls of the old one. This time instead of constructing a shallow dome, the designers created a stepped roof plan. Large trusses were placed where the steps occur in the roof and structural joists were placed in between the trusses to create a flat roof. The new auditorium completed in 1982 still stands proving that engineers have learned from their previous mistakes and succeeded in overcoming previous shortcomings.

Conclusion

The C.W. Post Dome Auditorium collapse could have been prevented. If the engineers had fully understood the membrane theory that they had used to design the dome, they more than likely would have chosen a more accurate method to model and design the dome's structure. Drifting snow and ice may have been what ultimately caused the collapse, but it was really the faulty design assumptions that left the dome vulnerable to drifting conditions. Fortunately, no lives were lost the night the dome collapsed.

Engineers, Architects, and Contractors should also be commended for building safer long span structures. It does seem that the 1970s incidents really pushed the industry to further understand these structures and find ways to better design them. The advances of computer modeling analysis and building information modeling has helped designers make safer structures. It is now much easier to analyze complex structures using computer modeling programs rather than relying on simplified theories. This doesn't mean that designers can freely rely on the computer program. The designer is still responsible for the input and for correctly reading the output. The ability to see complex long span structures in three dimensions before they are constructed allows designers to fully visualize the design and alleviate potential problems before they occur.

References


  • Business Week (February 6, 1978). "An Inquest into Why All The Roofs Fell In." 46.
    • News Article:The article briefly discusses the investigations into multiple roof collapses due to a round of snowstorms. It includes quotes from investigative engineers and roof design loads.

  • ENR (January 26, 1978). "Space Frame Roofs Collapse Following Heavy Snowfalls." 8-9, 64.
    • News Article:This article was one of the first reports on the collapses of multiple space frame roofs due to snow.

  • Kliment, Stephen. (1981). Towards Safer Long Span Buildings. American Institute of Architects.
    • Report: This report focuses on the architectural reaction to long span building collapses. It summarizes a discussion by a panel of architects, engineers, and contractors to revisit the planning, design, and construction of long span buildings.

  • Murphy, J. (August 1982). "Out of Round." Progressive Architecture, 49-54.
    • Journal Article: This article discusses the new design of the auditorium at C.W. Post University that replaced the collapsed dome.

  • Salvadori, Mario and Matthys Levy. (2002). Why Buildings Fall Down: How Structures Fail. New York: W.W. Norton & Company.
    • Book: Salvadori and Levy include an in-depth case study of why the C.W. Post Dome collapsed and include illustrations and diagrams of the failure mechanism. They focus on how the design theory did not fit the actual design of the dome.

  • Seltz-Petrash, Ann. (December 1979) "Winter Roof Collapses: Bad Luck or Bad Design?" Civil Engineering-ASCE, 42-45.
    • Journal Article:The article revisits the numerous collapses that occurred due to a round of snowstorms. Seltz-Petrash includes industry reaction to the collapses and ways to prevent further collapses.

  • West Coast Lumberman's Association. (1958) Douglas Fir Use Book. USA: Daily Journal of Commerce.
    • Book: This book contains contains structural data and design tables that are related to the construction and design of wood structures.

Additional Resources

  • Carper, Kenneth, and Jacob Feld. (1997). Construction Failure. New York: John Wiley.
    • Book: Carper and Feld provide a brief description of the C.W. Post Dome collapse among case studies into the failures of the Hartford Civic Center and Kemper Arena.

  • ENR (February 9, 1978). "Flat Roofs Suffer Most Collapses in Heavy Snowfalls." 9-10.
    • News Article: The article illustrates the vast amount of failures that occurred due to the same snowstorms. There is also a brief interlude that discusses what can go wrong with space frame roofs.

  • Geis, James, Kristen Strobel, and Abbie Liel. (2010). Snow-Induced Building Failures. University of Colorado College of Engineering.
    • Report:This report investigates building failures due to snow and hopes to establish any building failure trends that are present in snow caused collapses.

  • Minimum Design Loads for Buildings and Other Structures (7-10). (2010). American Society of Civil Engineers.
    • Book: This standard provides requirements for general structural design and is a revision of the 7-05 version.

Figure Permissions
  • Figure 4 reproduced with permission of The American Institute of Architects, 1735 New York Avenue, NW, Washington, DC 20006.