The structural design was completed by Gillum-Colaco (co-la-so) Inc out of St Louis Missouri. Jack D Gillum was the engineer of record, with Daniel M Duncan was the project engineer. What does this mean, you ask? Gillum’s engineering stamp was on the structural blueprints; he was the legal engineer of record. But Duncan completed the structural engineering work and managed the day to day design tasks. It’s fairly common for the owner or partner in an engineering firm to oversee and stamp major designs, while engineering staff complete the day to day work. Although there is typically significantly more oversight than what occurred here; but more on that later.
The main structure of each walkway consisted of two box beams running the length of the walkway. The box beams were made from two C channels welded together lengthwise to create a box shape, giving them the name box beam. Cross bracing between the box beams, q-deck laid on top with a layer of concrete created the walking surface. Each walkway weighed roughly 8,500 kilograms.
All three walkways were suspended from the atrium roof structure by hanger rods. Because the fourth and second floor walkways were stacked, the original design intent was to hang them from the same set of six rods suspended from above. The original design meant that one set of rods would support both walkways; commonly referred to as a single rod system. Notice I said original design intent, the single rod system is not what was built on site.
The connection details to support the walkway included the hanger rod passing through the box beam, with a threaded nut on the underside of the beam.
There are three types of connections, simple, complex and special.
Simple connections have no unusual loads or forces applied to them and can be designed using the American Institute of Steel Construction Manual of Steel Construction. As I understand it, these are standard details that can be re-used within the AISC manual parameters without requiring specific calculations to be completed.
Complex connections have extreme or unusual loads or loads from several directions and require engineering to design.
Special connections are a hybrid of both. When a simple connection has a concentrated load and/or the Manual of Steel Construction can no longer provide all of the necessary info to design, it becomes a special connection and required engineering design.
There are also redundant and non-redundant connections. If, when a connection fails, it would not result in failure of the structure, because there are other connections available to carry those loads, it is considered a redundant connection. But if when a connection fails, the structure collapses, that connection is non-redundant. The box beam, hanger rods, threaded nut connections at each walkway, were considered to be both special and non-redundant connections.
As I said earlier, during construction, the design was changed. The contractor, Havens Steel Company, had argued that if they used a single rod system, with one set of rods to support both walkways, the rods would be damaged when the fourth floor walkway was hoisted into place. Which I have to admit, seems to be a reasonable concern. You’d have to lift the fourth floor walkway into place very carefully in order to protect the rods. And, what is plan B if you took all possible precautions and the rods were still damaged?
The design change during construction resulted in two sets rods, or a double rod system. Engineers don’t like to get caught up in complicated names for things. Single rod systems have one rod, double rod systems have two rods. Easy peasy.
The double rod system resulted in one set of six upper rods from the structure to the fourth-floor walkway. And then a second set of six separate lower rods hung from the fourth-floor to support the second-floor walkway below. The upper and lower rods through the fourth floor beams were offset by 100mm.
In the original design scenario with a single rod system for both walkways, let’s consider the load applied to each walkway connection detail to be “X”.
However, in the re-design scenario, which is what was installed, with the second floor hung from the fourth floor, or a double rod system, the load applied to the fourth floor walkway was now 2 times “X”. The double rod system, doubled the load on the connection details at the fourth floor beams. This would have been fine, except the connection details were not re-engineered or even checked when the hanging method was switched to a double rod system. Not the beams, not the rods, not the threaded nuts. Not even when Haven Steel submitted shop drawings of the walkways and connection details to Gillum-Colaco (Co-la-so) for review; which they approved.
To recap, the design was revised to a double rod system, which doubled the load applied to the fourth floor walkway, and it was not revised or even checked during design or construction. But wait, there’s more. Because it never seems to be just one thing.
Remember the box beams, running lengthwise under each walkway, that were constructed from two c-channels welded together to create a box. Well the welded seam meant that of the four sides of the box beam, two of the opposite sides would have welded seams and the other two sides would be solid steel. Rather than running the hanger rods through the solid sides, they ran through welded sides. This is the weakest part of the beam. And this wasn’t just case on the second and fourth floor walkway; all three walkways were built like this. I`m no structural engineer, but even I can see this was a really bad idea. The welds were originally intended to be on the sides of the beams. But since Gillum-Colaco did not complete adequate field reviews of the walkways during construction, the beam orientation was not caught or corrected.
Following analysis of the failure, the welded seam on the bottom of the fourth-floor beams split, causing the threaded nuts to slip through the gap and the walkway to drop several inches. Then the upper seam of the beams failed as well causing the fourth-floor walkway to fall and since the second floor walkway was hung from it, they both collapsed and onto the floor below. Further analysis following the collapse suggested that the east line of hanger rods failed first, although it was determined that a failure was inevitable.
Had the support detail not changed and one set of rods supported both walkways, the third-floor walkway would have actually been the one with the highest probability of failure, as it was wider and could fit more people, while having similar design flaws with respect to the box beam construction. All of the walkways were under designed.
The calculated maximum loads acting on the connection details, as required by the Kansas City Building Code, were 181 kilonewtons at the fourth-floor walkway, and 90 kilonewtons at the second-floor walkway. Notice that the fourth floor load is double the second floor load.
The loads on the fourth and second floor walkways at the time of failure were estimated to be 95 and 51 kilonewtons respectively; roughly 50% of what was required. The connection details did not meet Kansas City Building Code or the American Institute of Steel Construction requirements. Any of the fourth-floor connections were candidates for failure; and with little capacity rating in each, collapse was imminent. The walkways were so under designed that they could barely support themselves, let alone any additional weight. It`s surprising they hung for a year before the collapse. And it’s too bad they hadn’t collapsed under their own weight when no one was in the atrium. But unfortunately, that wasn’t the case.
At the time of the collapse, it was believed that there were 40 people on the second-floor walkway and 16-20 people on the fourth.
When the walkways collapsed on July 17th, 1981, 114 people were killed and over 200 were injured. The collapse is still the deadliest non deliberate structural failure in US history, and the deadliest structural disaster until the collapse of the north and south world trade center towers on September 11th, 2001.
Volunteers quickly responded to the plea from the rescue teams and brought jacks, cranes, ropes, and saws to assist in the rescue mission. To mitigate the risk of fire, the power was been cut, which impacted rescue operations.
The main floor exhibition area was used as a temporary morgue and the lawn and driveway acted as a makeshift triage.
The last person was rescued at 430am following removal of walkway spans by heavy crane. Mark Williams had spent more than 9 hours trapped, had dislocated legs, and almost drowned by water from the sprinkler piping that burst when the walkways collapsed. I honestly couldn’t imagine what that was like. He was very lucky to survive the collapse, so many others were not as lucky.
An investigation into the cause of the collapse reported the following findings:
Following the atrium roof collapse during hotel construction, Gillum-Colaco (co-la-so) was given an additional fee to recheck the structural design. Gillum himself assured the owners that “he would personally look at every connection in the hotel”. But he didn’t, he only checked the atrium roof steel, not the walkways.
The steel fabricator Havens Steel Company had in house structural engineers that could have designed the connection details, had the drawings specified that they were to do so. But the drawings stated that the structural connection details had been designed by the engineer, and design by the fabricator was not required.
Special welds, stiffeners and bearing plates, as well as high strength steel hanger rods were required but not specified by the structural engineer.
The investigation did not find the necessary tests and calculations in Duncan’s records which would show that he had verified his connection detail design for a double rod system.
It also found that Duncan did not adequately review the shop drawing sent by Havens Steel. Nor did he review the connection detail in relation to the design specifications of the Kansas City Building Code or the project contract documents. Duncan argued during proceedings that his intention was for the fabricator to complete the calculations and therefore assumed that they had done so before the shop drawings were submitted. What he intended is not really relevant here. Only what was specified on the contract documents; which the investigation found had clearly stated the connection details were designed by the engineer, not the fabricator. Drawings and specifications have to be clear to avoid any finger pointing.
And lastly, the investigators found that that neither workmanship nor materials played a significant part in the failure. This meant that there was no one else to blame but the design and the people who were responsible for that design.
In the aftermath of the disaster the structural engineers were found culpable of gross negligence, misconduct and unprofessional conduct in the practice of engineering.
There were three main parties that received disciplinary action. The engineer of record (Jack D Gillum), the chief project engineer (Daniel M Duncan), and the engineering firm (Gillum-Colaco (co-la-so) Inc) lost their license to practice in Missouri, Kansas and Texas, and their membership with the American Society of Civil Engineers (ASCE). The ruling was later appealed and upheld.
Duncan was an engineer himself, and while he did not stamp the drawings, he was the project engineer and in charge of the design. The review commission found him guilty of gross negligence in preparation of the structural drawings and failure to properly review the steel and erection shop drawings. Although Duncan did not stamp the drawings, as an engineer himself, he had an obligation to public safety. He should have known better.
As the engineer of record, it was Jack D Gillum`s stamp on the structural plans. Gillum was found liable and responsible for the acts and omissions of Duncan, and Gillum was also negligent in failing to review or assure someone had reviewed the drawings before he stamped them. Gillum also failed to review the installation on site, even after requested by the architect.
Gillum-Colaco Inc was found liable and responsible for the acts and omissions of Duncan and Gillum.
So what can we learn from this failure? Drawings, or contract documents, must be clear. Always double check the details of critical components to your design. Even if they are designed by someone else; a second set of eyes can save alot of people. And lastly, engineers must review their design on site, during construction. The beam orientation could have been caught and corrected during construction, but it wasn’t.
Today, the lobby remains the same original layout and design, although the walkways were never re-built. The hotel was renamed the Hyatt Regency Crown Center in 1987 and has been called the Sheraton Kansas City Hotel at Crown Center since December 1st 2011.
A memorial was built in 2015 in Hospital Hill Park across the street from the hotel. It was dedicated on November 12th by the Skywalk Memorial Foundation which is a non-profit organization created for victims of the collapse. This included a $25,000 donation from Hallmark Cards.