Dozens stood on the walkways. At 7:05 PM, the walkways on the second, third, and fourth floor were packed with visitors as they watched over the active lobby, which was also full of people. The fourth floor bridge was suspended directly over the second floor bridge, with the third floor walkway set off to the side several meters away from the other two. Construction issues led to a subtle but flawed design change that doubled the load on the connection between the fourth floor walkway support beams and the tie rods carrying the weight of the second floor walkway.
This new design could barely handle the dead load weight of the structure itself, much less the weight of the spectators standing on it. The connection failed and both walkways crashed one on top of the other and then into the lobby below, killing 114 people and injuring more than 200 others. Mainbody FIG-1 FIG-2 Originally, the 2nd and 4th floor walkways were to be suspended from the same rod (as shown in fig-1) and held in place by nuts.
The preliminary design sketches contained a note specifying a strength of 413 MPa for the hanger rods which was omitted on the final structural drawings. Following the general notes in the absence of a specification on the drawing, the contractor used hanger rods with only 248 MPa of strength. This original design, however, was highly impractical because it called for a nut 6. 1 meters up the hanger rod and did not use sleeve nuts. The contractor modified this detail to use 2 hanger rods instead of one (as shown in fig-2) and the engineer approved the design change without checking it.
This design change doubled the stress exerted on the nut under the fourth floor beam. Now this nut supported the weight of 2 walkways instead of just one Conclution FIG-3 FIG-4 Neither the original nor the as-built design for the hanger rod satisfied the Kansas City building code making the connection failure inevitable. If, however, the building design had contained more redundancy this failure may not have resulted in the complete collapse of the walkway.
Kaminetzky (1991) suggests two much stronger design alternatives for the connectors. The toe-to-toe channels used in the Hyatt Regency provided for weak welding which allowed the nut to pull through the channel/box beam assembly initiating the collapse. A back-to-back channel design using web stiffeners when necessary (fig-3) or the use of bearing crossplates in conjunction with the toe-to-toe channels (fig-4) would have made the connection much stronger making it much more difficult for the nut to pull through