By Tara Hoke
Situation
A few hours before dawn on January 18, 1978, the roof of the three year- old Hartford Civic Center (now the XL Center), in Connecticut, collapsed under a moderately heavy snow, bringing some 1,400 tons of steel and rubble down on the unoccupied arena below. Though no one was harmed in the collapse, the escape from tragedy was a narrow one; a college basketball game hours before had drawn some 5,000 spectators to the venue.
The roof had featured an innovative space-frame design, a design that offered the ability to span large areas with minimal need for interior columns, allowing for better visibility in all areas of the arena. The design was significantly cheaper than more conventional roof designs, making it appealing to the project's decision-makers from an economic as well as aesthetic point of view. Given the complexity of the roof's structure and the number of cost-saving measures built into the design, the engineers turned to what was, at the time, state-ofthe- art structural analysis software to test the roof's design and ensure that it would offer the desired factor of safety.
Yet almost immediately after construction began on the roof, signs emerged suggesting that its performance did not align with the computer model's predicted results. The roof frame was initially assembled at ground level, and an inspection agency reviewing the assembled structure found greater-than-expected deflections at some of the nodes. This observation was reported to the design engineers, but no further action was taken.
Next, the frame was lifted into position atop its supporting pylons, and deflections were measured once more. This time, the maximum deflection was found to be twice what the design software had predicted. Again, the design engineers were consulted, and again they were untroubled by the discrepancy. Noting that the computer's calculations had included some simplifying assumptions, the engineers advised that this discrepancy between theoretical and actual values was within the range of difference that might be expected in such cases.
Still later in the construction, a subcontractor complained that support brackets meant to attach fascia panels at the ends of the roof frame would not fit because of the frame's excessive deflection. A project manager instructed the subcontractor to handle the problem himself; he complied by modifying or refabricating the brackets as needed, and construction ultimately proceeded to its conclusion.
Less than a year after the arena opened to the general public, a concerned citizen contacted the City of Hartford to report the downward deflections he had observed in the stadium's roof, deflections he felt indicated the roof was unsafe. Alarmed, city officials contacted the design engineers, but they again expressed confidence in the design; the officials' concerns were allayed, and thus the last clear opportunity to avert the roof's collapse was lost.
After the collapse, an independent investigation revealed several major flaws in the roof's design, as a result of which some compression members in the top layer of the frame were overloaded by more than 800 percent. These overstressed members buckled under the weight of the wet snow, overloading other members, which in turn buckled in a progressive failure. The investigation also highlighted inadequacies in the modeling software used by the design engineers to validate their work, noting that the analysis considered the effects of roof loads only on a simplified model of the complex space frame and thus failed to detect weaknesses in certain areas of the frame.
Question
What ethical lessons can be learned from the Hartford Civic Center roof collapse?
Discussion
In the four decades since the roof collapse, the case has been featured in innumerable news reports, professional seminars, and scholarly works that have examined the technical and ethical causes of engineering failures. While it would not be possible in a column of this length to give a complete picture of the circumstances leading to this collapse, even this brief chronology of events offers some clear ethical lessons.
Fundamental Canon 2 of the ASCE Code of Ethics states: "Engineers shall perform services only in areas of their competence." While the language of this canon might seem to focus only on the need for adequate technical knowledge, the canon should perhaps be more broadly understood as requiring engineers to "perform services competently," meaning that engineers must not only possess the necessary knowledge, but they must also apply that knowledge to their work in an effective and diligent manner. The facts of this case suggest a failure to deliver competent service in both respects, as the design engineers relied on a computer program to answer critical questions about the roof's safety without sufficient knowledge or understanding to ensure the suitability of the program or the validity of its results; they also failed to take additional steps to procure that verification, either through their own calculations or by another professional with expertise in the needs or performance of such a structure.
Furthermore, it is almost impossible to discuss Canon 2 without recognizing how closely it intertwines with Fundamental Canon 1: "Engineers shall hold paramount the safety, health, and welfare of the public and shall strive to comply with the principles of sustainable development in the performance of their professional duties." More than simply a matter of meeting client expectations or avoiding contract liability, competence primarily serves the aim of protecting the public from the dangers of unsafe or unreliable engineering works. Conversely, engineers can only truly meet their ethical obligation to "hold paramount the safety, health, and welfare" when they have the competence to make informed judgments about risks or limitations.
Despite the early morning timing of the Hartford Civic Center roof collapse, the indications of a pending failure did not literally occur overnight, but rather they arose over an extended period of time in which red flags were raised and opportunities were missed. Even in view of their obvious faith in their design and the computer analysis that supported it, it is still incredible that the engineers in this case failed to heed signs of the structure's deficiency and take steps to investigate them--incredible, perhaps, until one considers that a similar string of disregarded warnings can be found in many of history's most catastrophic engineering failures, from the Quebec Bridge to the Hyatt Regency walkway in Kansas City, Missouri. Collectively, these cases stand as a reminder of the extent to which a personal bias--confidence, obstinance, or the unwillingness to recognize one's own fallibility-may hinder an engineer's ability to meet his or her ethical obligation to hold paramount the public safety, health, and welfare.
The Hartford Civic Center roof collapse is often cited as the archetypal case of overreliance on computer technology. In using computer models- or any other engineering tools-engineers must be mindful of the strictures of guideline a under Canon 1, which requires that engineers "recognize that the lives, safety, health, and welfare of the general public are dependent upon engineering judgments, decisions, and practice." The key phrase in this statement is "engineering judgments, decisions, and practice." Even the most sophisticated computer system is not an engineer, and thus even the most exhaustive computer analysis is no substitute for sound and ethical judgment by the human engineer who determines how to apply that analysis in practice.
Tara Hoke is ASCE’s general counsel and a contributing editor to Civil Engineering.
© ASCE, ASCE News, March,2020
