The Construction Phase

Problem Solving

• Problems will Develop. How problems are managed often determines the success or failure of a project.

When Communication Fails:

The Hyatt Regency Hotel Disaster

The lobby of the Kansas City Hyatt Regency Hotel and

the skybridges which connect the upper

levels of the hotel to the convention center

•The Hyatt Regency Hotel in Kansas City, Missouri included a large lobby area that was used for entertainment events.

•The lobby featured several ‘skywalks’ which connected different functions on the upper floors of the public areas by way of bridges which passed through the lobby space at the second, third and fourth floors.

•The skywalks on floors two and four were stacked on top of one another, the bridge on the third floor was alone.

•The skywalks were suspended by steel rods from a roof truss system which spanned the entire lobby area.

The Skywalk Original Design called for a

single rod, supporting the weight of each walkway held up by

the bolts.

The General Contractor requested that the design be modified because of constructability issues.•The rods were designed by the engineer to be threaded the length of the rod, which seemed impractical to the Contractor•The contractor suggested that the rods be split, to make threading the nuts easier.

•The proposed substitution: Two rods rather than one.

•The Contractor suggested this alternative in the Shop Drawings and submitted them to the Architect and to the Engineer for approval.

•The Contractor received the Shop Drawings from both the Architect and the Engineer with their stamps to proceed with the substitution.

The two alternatives appear to be similar. Each relies upon a suspending rod, or rods. The welded box beam configured from

two steel channels is identical in both versions. The steel beams connecting the deck to the box beams are also identical

in both options.

The Contractor went ahead with the substitution as indicated on the approved Shop Drawings and installed the skywalks with the method indicated on the right.

•The building had been open for barely a month, when on July 17, 1981, a party was held accommodating over 1600 people in the lobby space and adjacent public areas.

During the festivities, the fourth floor bridge collapsed, and the two bridges, loaded with people, fell to the ground

The two skywalks pancaked on top of one another on the lobby floor.

•What caused this disaster which claimed 114 lives and injured over 200 others?

•Resonance, said some. The bridge was loaded down with hundreds of people, DANCING. They overloaded the bridge because the were jumping up and down on the bridge in unison.

•Quality of Materials, said some. The Contractor probably used inferior materials which gave way under loading conditions. Contractors are often caught only after a disaster exposes their errors.

•Poor installation, said others. Work is only as good as the workers who installed the work, and clearly they must have been in a rush to get it done and didn’t provide quality welds on the steel.

Catastrophic structural failure often occurs without warning, and is over in seconds.

Forensic analysis of the debris often provides an insight into what actually happened.

Were any of the rumors being passed around true?What caused this disaster to happen?

The investigation immediately began to focus on the box beams, and the nuts on the threaded rods.

There was deformation that clearly was caused prior to the structure hitting the ground. The box beams were deformed in ways that indicated they tore through the nuts on the rods.

The smoking gun.

The bolt is still in position on the threaded rod

So What Caused This Failure?

Recall the proposed substitution: Two rods

rather than one.

•This solution places the loading from the lower walkway onto the upper box beam rather than the rod.

•The ‘new’ design was included on the Shop Drawings from the fabricator for approval by the Engineer

•The Shop Drawings were stamped by the Architects ‘Approved’ and by the Structural Engineers ‘Reviewed’.

•Calculations done after the accident confirmed that the actual Dead Load for the bridges was actually 8 percent higher than the design loads used because of a concrete deck included in the specifications but not taken into account in the original engineering.

•The Building Code required a Live Load be assumed of 72,000 lbs for EACH walkway.

•A video taken on the evening of the disaster indicated 63 people on the upper bridge, which would be a live load of only 9,450 lbs.

•To determine the actual cause of the failure, mock ups were made of the elements and subjected to real loading.

•The six (6) rods carried a load of 24,000 lbs. each.

•The nut on the rod first bent the lower flange of the box beam, then the upper, then pulled clean through.

•The midspan rods failed first, then the others.

The final investigation concluded the following:•The walkways collapsed under loading substantially less than loads specified by the building code.

•Any of the box beams at the fourth floor could have failed based upon the design in the Shop Drawings

•Neither the box beams nor the hanging rods, even in the original design, met the minimum code requirements. The original design may have been strong enough to avoid the collapse, but was not sufficient to ever pass a design review.

•Neither the quality of the materials nor the quality of the workmanship contributed to the collapse.

•The Missouri court system identified the Engineers as responsible parties in the collapse for failing to recognize the fundamental difference between the original design and the substitution request.

“It wasn’t a matter of doing something wrong, they just never did it at all. Nobody ever did any

calculations to figure out whether or not the particular connection that held up the skywalks up would work. It got built without anybody ever figuring out if it would be strong enough. It just slipped through the cracks.”

-Patrick McLarney, attorney representing the state licensing board

What Lessons have been learned here?•BE EXPLICIT. If you have questions about the safety of a design element, do not keep it to yourself. Ask questions, and demand backup. Ask to see the calculations.

•When proposing substitutions, make sure the element you are modifying is not a structural element. If it is, be explicit. Request that the substitution be checked by the engineer, and that the calculations be provided to you for your records.

•If a substitution does involve structure, do not include it with a bunch of other information (in a set of Shop Drawings, for example), but instead pull it out and make it a unique piece of correspondence.

•Track each element all the way through to installation. If what is installed looks different from your last memory of events, ask questions. Look it up in the drawings, and follow through. Don’t just ‘assume’ that someone else has checked it.