The collapse of the Hyatt Regency Hotel: what lessons must be learnt?
In 1981, at the opening of the magnificent Hyatt Regency Hotel in Kansas City, a tragic accidentoccurred in which many people were killed and maimed. Several investigations into thedevastating accident were carried out and it appears to be the case that an elementary error instructural engineering design was to blame. However, much of the data also suggested that 'amore fundamental cause was a series of communication failures between the structural engineerand the steel fabricator' (Banset & Parsons, 1989: p.273).
The background to the incident is as follows. The hotel and restaurant were crowded and therewas a dancing competition in the atrium. Above the atrium, on the second and fourth floors,were walkways (Fig. 2). As people crossed the walkways, they stopped to watch the dancing,and began tapping their feet in time to the music. Suddenly the two walkways began to fall. Theycrashed one on top of the other and then into the atrium below: 114 people were killed and morethan 200 others were injured.
Several studies seem to agree that the people who were tapping to the music created resonance inthe walkway structure. Once resonance starts, it continues to increase in intensity until the causeis removed or the structure breaks up. The central hanger rod in the fourth floor walkway (seeFig. 2) tore through its box beam. The remaining rods could not support the extra load, so theytore through the beams one by one, leaving the rods hanging from the roof trusses while thewalkway fell loose. Because the second-floor walkway was suspended from the fourth-floorwalkway, the two fell together.
However, resonance aside, there was probably a much simpler explanation of why the structurefailed. An investigation found that the builders had changed the design specifications duringconstruction. In order to hang the walkways, they had used two shorter rods instead of onecontinuous threaded rod through both walkways - from the roof to the fourth floor, and from thefourth floor to the second floor (Fig. 4). This decision was due to the cost of producing acontinuous threaded rod of this length, which then meant that the top box beam-rod connectionhad to carry the weight of both walkways, rather than just the top walkway as it was originallydesigned to. In hindsight, it would seem obvious that this design change was bound to lead tofailure, yet no one detected it. As Taylor and Moncarz observed, the collapse was the result of 'anill-considered change of an ill-defined structural detail' (2000: p.46).