Steel+Structural+System+Collapses+&+Failures+During+Construction

Collapses of Steel Structures During Construction //Brandon Rupert, EIT // //BAE Penn State 2013 //

[[image:failures/Penn state steel building under construction.JPG width="350" height="263" align="left" caption="Photo of steel erection Photo Credit: M. Kevin Parfitt"]]Introduction
toc According to STRUCTURE magazine, between 1990 & 2008 the Occupational Safety and Health Administration (OSHA) investigated 96 structural collapses during construction that involved fatalities and injuries (Ayub 2010 p. 12). Of these 96 incidents, 60 involved the collapse of various types of steel structures whether temporary or permanent. Every engineer's goal is to design an economical structure that meets code and operates in a way that best suits the owner, but a designer's main goal is the safety of the occupants once the building is finished. The construction process has many risks including erection of the steel frame, ability of any temporary structures to support day-to-day operation, and the stability of any structures on site. The Union Carbide Building in Toronto, Husky Stadium in Washington, and the Big Blue Crane Collapse at Miller Park in Milwaukee are all examples of how stability of steel structures during construction must be a major concern for design engineers, construction managers, owners, and all other people involved in any construction project. A structural engineer's job is to design a building that can stand up to the elements, natural or man-made, for a long period of time. But often it is assumed a building under construction can sufficiently support loads it will see during construction because they are generally not as high as what the building is designed for. Although this may be true, many times the normal assumptions we make during design are not always sufficient for what a structure may see during construction. It is often necessary to look into construction loading and the overall construction process on top of the codes and design loads used to design a building to also ensure the safety of the people working on the project as well as the occupants once construction is finished. Although design engineers are generally not responsible for means and methods of construction, each case detailed throughout this article will demonstrate that many incidents can often be avoided if design engineers are more involved in the construction process or if engineers are retained by contractors to specifically monitor construction.

Whether a collapse during construction is due to an error in the process or an unforeseen load being to an unfinished steel frame, the uncertainty of a structure's stability is a major concern during construction. With construction being as dangerous as it already is, engineers and contractors must work together where ever possible to try to predict some of the uncertainties that come with the construction process. In 1987 when the UW Husky's stadium was being expanded the collapse could have been avoided had the guy wires not been cut out of order during the demolition of the roof structure. The Union Carbide Building collapse in 1958 was not necessarily as avoidable due to the unpredictability of Mother Nature, but had the construction process been different and the concrete spandrel beams added strategically with the rest of the structure perhaps the collapse may never have happened. In other cases though, Mother Nature can be a sign something may be unsafe like in the case of the Big Blue Crane at Miller Park. Big Blue collapsed due to the perfect storm of problems, some natural and some that could have been avoided had there been more supervision. As anyone in the industry knows, construction is a very expensive process and getting a project done quickly often means a cheaper or more profitable project in the end. But with quickness many problems can be overlooked, accidentally caused, or exacerbated, and no amount of money saved is worth risking the lives of the many men and women that go to work on these sites every day.

Throughout the following case studies one will notice a common factor. Lateral stability seems to consistently be a factor in the collapses summarized below and many other structures. Structures while being built can be very unstable just by nature. Structural engineers design a building to be laterally stable considering the entire building structure and its components. During construction this may not be the case. ASCE 7-05 simply states, "Temporary bracing should be provided to resist wind loading on structural components and structural assemblages during erection and construction phases," but in no way says how to design for wind loads during construction. ASCE 37, a standard that focuses on structural loads during construction, only vaguely states what a designer should do to ensure lateral stability. The OSHA standard, 29 CFR 1926, [|1926 Subpart R-Steel Erection] also touches on temporary bracing but much like the standards above does not include much on how to design temporary bracing on a structure not yet completed. The following case studies will prove more focus is needed on how to assemble and design temporary bracing to ensure structures are stable throughout the entire construction process not just once the building is occupied.

Union Carbide Building: September 6, 1958
======

According to Jack McCormac, author of the 4th edition of Structural Steel Design, in the past 150 years there have been many failures due to wind alone, and a large percentage of these building failures occurred while the steel frame was being erected (McCormac 2008 p. 45). One major case is the Union Carbide Building in Toronto, Canada. The Union Carbide Building, designed by Shore and Moffat, was to be a state of the art building that would house the management of Union Carbide's Canadian operations and some of its various subsidiaries (Bradburn 2011). The building was a 180,000 square foot, 215' x 65' wide steel framed structure with columns at 20' centers (Feld 1996 p.147-150). One feature of this building was the lack of interior columns to allow for maximum space.

Erection of the steel frame began in mid-June of 1958 at 123 Eglington Avenue in Toronto. On Friday September 5, 1958 all of the connections were welded completely up to the 9th floor by the end of the work day Friday. To stabilize the top two floors of the building temporary bracing was put in place, and the top two floors would be welded at the beginning of the following week (Feld 1996 p. 147-150). But the events of September 6, 1958 would halt construction and turn a quiet street in Toronto upside down. At approximately 6:20 PM a lightning storm hit Toronto with wind gusts reportedly reaching speeds greater than 90 kph (Bradburn 2011). With the strong winds and a possible lightning strike, the southwest corner of the steel frame began to sway and then the entire 1850 tons of structural steel came crashing to the ground, as scene in Fig. 1, in a scene described as, "a falling house of sticks and a folding accordion," (Bradburn 2011).

Investigations done by the city, insurance companies, and consultants proved that the original building design was sufficient (Feld 1996 p. 147-150). The investigations proved the temporary bracing used on the upper two floors was not sufficient enough to resist the extreme conditions of the September 6th storm (Bradburn 2011). Once the temporary bracing failed the southwest corner of the building collapsed down onto the lower nine floors and the weight plus the force of the impact exceeded the rest of the frame's load capacity, resulting in the pancaking of the entire structure. The original design called for girder to column moment connections with deep concrete spandrel beams in the longitudinal direction, but erection of the concrete beams was not started by the time of the building collapse (Feld 1996 p. 147-150). Without the concrete beams, the only bracing the 11-story columns had were the very light longitudinal tie beams at each floor in the face of the walls that did not provide nearly enough rigidity to brace the structure against the wind forces that evening (Feld 1996 p.147-150).

Thankfully the events of September 6th were not nearly as devastating as the could been. The only person that was on site that evening, because it was a weekend, was a security guard who happened to be on the other end of the construction site at the time of the collapse (Bradburn 2011). On Eglington Avenue, the heroic and skillful actions of a bus driver saved the lives of himself and several others. The driver heard the loud crash of the structure and swerved out of harms way in just enough time to avoid the bus being crushed by the crumbling steel (Bradburn 2011). The damage caused by the Union Carbide Building collapse was luckily only materials, and a stoppage in construction. Several parked cars were crushed, and none of the original steel was salvageable even though only one of the several hundred welds failed (Feld 1996 p.147-150).

After the investigations were completed as mentioned previously, the building design was deemed sufficient, but to ensure another collapse would not happen again the consultants recommended the addition of deep horizontal trusses between columns the columns of each floor to add more lateral stability during construction and occupancy (Bradburn 2011). The designers took these recommendations into consideration, and rebuilt the structure with the newly added trusses, and the Union Carbide Building was completed and opened by July 1960. After opening in 1960, the Union Carbide operated without incident until 1990 when it was demolished to make room for new construction as seen in Fig. 2 (Bradburn 2011).

Husky Stadium: February 25, 1987
In 1987 the University of Washington hoped to add to their football team's home, Husky Stadium. This addition was to be a 17,000 seat addition to the North side of their stadium completed in time for the next season's opening game against Stanford on September 5th, but the events of February 25, 1987 would make that goal seem very unreasonable (Griffin). That morning was a sunny clear day at the University of Washington, but something unusual was noticed by some of the construction workers on site, a large crack in the steel frame had begun to form. At 10:07 AM a loud bang was heard all across the UW campus followed by the entire collapse of the North Stands Addition (Griffin). The entire 12 second collapse was documented in the time lapse photos shown below in Fig. 2.

After an extensive investigation, the overall design of the structure was deemed sufficient and the collapse was caused by insufficient lateral bracing during the construction process. The roof structure included guy wires which kept the roof structure from experiencing unusual torsional forces on the structure, and the demolition process required cutting of these guy wires. This had to be done in a very controlled and methodical sequence. Unfortunately some of these wires were cut out of order. This resulted in the twisting of the structure which caused the original crack in the roof truss support beams mentioned earlier (Griffin).

One contributing factor to the cause of the collapse was how the tubular guy wires were originally designed. Generally when tubular sections are designed it is assumed the section stays circular when subjected to deformations (Chen 1988 p. 1088). A report done on local and post-buckling behavior of tubular beam-columns shows that tubular sections with walls significantly thinner than the diameter of the section subjected to significant deformations can experience local buckling and distortion of the cross section. The local buckling and distortion of the cross section causes a reduction in the load carrying capacity and the energy absorption section of the guy wire (Chen 1988 p.1088). When the workers cut the guy wires out of order, they subjected the remaining guy wires to loads well above the original design loads. This could have caused large deformations which reduced the roof structure’s ability to resist the wind loads experienced on site.

Thankfully the only damage caused by this incident was economic. All 20-40 of the ironworkers on site were not hurt or injured because of a construction supervisor who halted construction and forced everyone off site once the crack in the roof structure was noticed (Griffin). Original estimates say the damage cost between $500,000 and $1,000,000 (Lange 2001). Although the collapse was a major setback in the construction schedule, construction was completed prior to September 5, 1987, the Huskies first home football game of the ’87 football season.

Big Blue Crane Collapse at Miller Park: July 14, 1990
media type="youtube" key="p8fiixoGtM0" height="247" width="328" align="left" With faster construction scheduling the need for faster steel erection has become a must, and this has resulted in the need for larger booms on mobile cranes (Feld 1996 p. 228). This was the case at the Milwaukee Brewers' new stadium. The Brewers' stadium was three years in the making, and involved a state of the are multi-panel retractable roof that was to be supported by curved truss assemblies. With the size of the site and the size of the roof panels, weighing between 200-450 tons and the largest being 176' x 200' x 16', a large crane was a must. The crane, referred to as Big Blue, had a 340' lattice boom w/ a 200' jib, 2 crawler tracks, and had a 2400 kip counter weight, as seen in Fig. 3 (NIOSH 1999). Big Blue had a maximum capacity of 1040 kips, well under the weight of the final roof section to be hoisted into place (NIOSH 1999).

[[image:failures/_bbcrane.jpg width="276" height="253" align="right" caption="Fig. 4: Big Blue Crane Structure Photo Credit: NIOSH FACE Report 99-11"]]
(Video 1: Big Blue Crane Collapse Video Credit: YouTube)

On July 14, 1999 the construction of the Milwaukee Brewers would be halted in a very disastrous series of events. On this day, there would be between 400-700 workers working on various parts of the project. Five of these workers were part of the lift crew which was under the direction of one lift supervisor. Like all of the roof sections, the section to be lifted on July 14th was constructed on the ground to limit the risk of fall hazards and to speed of the time of construction. Each lift began at a pick point to the North of the site, where the section was hoisted and checked to ensure it was properly suspended. Once that was done the section would then be swung to the west over the south end of the stadium and moved to the set point at the south end of the stadium where the section would then be lowered into place, as seen in Fig. 4 below (NIOSH 1999).

When hoisting a section this large, it requires frequent checks of the attitude of the member being hoisted. The attitude at which an object hangs depends on the location of the center of gravity in relation to the attachment points and the lengths of the lines suspending the section (NIOSH 1999). Once the section was hoisted it was noticed that the connections at the south end of the section were higher than the north end connections and this had to be fixed to ensure the section could be properly placed (NIOSH 1999). There are several ways to adjust this; move the attachment points, shortening the suspension lines, or add weight to one side to lower it. It was decided that with the time constraints and the size of the section, the best route was to add a 6500 lb. concrete block to the south end to adjust the height of the connections (NIOSH 1999). This addition made an already difficult lift even more difficult, and the perfect storm of events that would take the lives of three construction workers had just begun.

[[image:failures/crane_tip_over_NIOSH(2).jpg width="181" height="281" align="left" caption="Fig. 5: Critical Crane Locations Photo Credit: NIOSH FACE Report 99-11"]]
Once the 6500 lbs. was added, the lift supervisor noticed the crawlers of the front transport sinking into the ground. To do a lift like this the crane must be level, but swinging the section to the west would put the crane out of level (NIOSH 1999). The decision was then made to move the crane to more competent ground to ensure the crane stayed level. By moving the crane to a new location, the crew was then forced to change the lift plan. This new lift plan called for the section to be hoisted over the north end of the stadium rather than the south end and swung to the east to be lowered into place (NIOSH 1999). As has been explained in previous sections of this case study, wind forces can cause many significant accidents, and that was not any different today. Both the steel erection and the roof construction contractors had a strict policy that prohibited and lift operation to be done if the wind was greater than 20 mph at the top of the crane (NIOSH 1999). Once the section reached its highest point in the lift, 300 feet, the lift supervisor called for regular updates on the wind conditions. At approximately 2:30 PM a mast-mounted anemometer read wind conditions between 17-20 mph, but the lift was not halted (NIOSH 1999).

The reasons for the lift operation not being halted are left up to much speculation, but more information can be found at the Big Blue Crane Collapse case study also found on the failures wiki website. Big Blue traveled approximately 500 feet to reach the set point and came to a stop, and the section would be stabilized while it was lowered (NIOSH 1999). The section needed to be stabilized because once the crane came to a stop the section began moving in a pendulum motion, swinging Northward. The plan was to stabilize the section on the next swing when it was going southward, but the pendulum motion along with the high winds caused the crane to become unstable (NIOSH 1999). With the crane becoming unstable it further increased the swinging of the roof section and the crane then tipped north and slightly eastward. The counterweight was now not enough to bring the swinging section, the tipping crane, and the added weight used to adjust the attitude back down to stable ground, and the whole crane tipped and crashed into the stadium wall parapet which snapped the craned at about the midpoint of the boom.

With the sudden nature of the collapse, taking less than 40 seconds, as seen in Video 1 above, none of the lift operators observing the lift in hoisted platforms were able to move in time. The roof section swung into a platform hoisted on another part of the site with three workers on it, killing all three of the workers when they fell the 300' to the ground below (NIOSH 1999). An investigation was done on the crane collapse, and showed that there were several contributing factors to the crane collapse. The combination of the side loads from the wind on both the crane and the roof section, the out of level ground conditions, and the pendulum motion of the section all contributed to the perfect storm of conditions resulting in the death of three workers (NIOSH 1999).

Many accidents are completely unavoidable, but actions must be taken to minimize the risk. It seems that some of the risk factors explained above were well covered and policies were in place to avoid them, but those policies were not completely followed during this particular lift. On top of that it seems some factors may not have been very well analyzed. Whether it was the ground conditions, the wind conditions, or overall lift plan many things could have been done to help avoid this accident. Many of these will be outlined in the Lessons Learned section of this case study below.

Lessons Learned
There is never only one culprit to blame in any failure case. Very often there are several different causes of a collapse. In the Big Blue crane collapse it was the combination of wind loads, uneven ground, and pendulum motion of the hoisted load. The Union Carbide Building collapsed because of weather and the use of insufficient temporary bracing. The University of Washington's Husky Stadium collapsed because of how the otherwise normal wind loads caused the structure to react after guy wires were cut mistakenly. One thing each of these cases has in common is, they incidents could have possibly been avoided had more care been taken.

Structural engineers are very rarely consulted when it comes to the means and methods of construction until after something goes wrong, and it is virtually impossible for an engineer to account for everything that may or may not happen in the future. However, it is often helpful to consult an engineer or for the engineer to have specific specifications that pertain to a structure's construction. In the case of the Union Carbide Building a specification stating an engineer must be consulted when any temporary bracing is installed may have caused the workers on site to contact an engineer to see if the temporary bracing was sufficient. The storm that occurred in Toronto that day had gusts up to 90 kph (approx. 56 mph), which is not necessarily abnormal for the area, and a simple check of the structure's stability may have called for more bracing.

The Husky Stadium collapse was in large part due to poor sequencing. The guy wires were cut out of order by mistake, and that is completely avoidable had the workers paid more attention or the supervisor in charge been more cognizant of how the work was being done. Perhaps less emphasis was put on the sequence of demolition because it is only demolition, but had the construction supervisor not noticed a buckle in the roof truss there would have been several injuries or fatalities. This case was completely avoidable, as explained above, had the work been done correctly. Many accidents are unavoidable, so anyone on a construction site must do everything possible to avoid the accidents that can be avoided.

The tragedy in Milwaukee was caused by several different things which can be seen above, and sadly that accident may have been avoided had more been done. The Fatality Assessment and Control Evaluation Report done by NIOSH pertaining to this case includes many conclusions they have drawn to prevent similar incidents. Unfortunately, these conclusions were made with the benefit of hindsight, and the reader is invited to read the NIOSH In-house FACE Report 99-11 to see all of these recommendations so an accident such as this does not happen again. Some of the recommendations made included: implementing specifically engineered lift plans for any critical lift operations, ensuring any cranes and work areas had equipment capable of measuring wind velocity during the hoist operation, and finding alternate methods of observing a lift that does not involve putting multiple people in harms way. On all jobs there is risk and having engineers and designers constantly evaluating the risk every day can ultimately save lives. At the Brewers' stadium the wind velocity seemed to be ignored even though both the erection and roof contractors had strict policies restricting lifting operations if wind speeds are above 20 mph. Had the contractors paid more attention to the instruments that were placed properly at the elevation of the lift the operation may have been halted and the three men that died could have gone home to their families that evening. One lesson anyone should take from any tragedy such as this, is no amount of money is worth a human life. It will always be the responsibility of everyone involved to remember that and do whatever it takes to keep everyone safe.

As has been mentioned, it often is not possible for a structural engineer to be involved in means and methods during construction so it is pertinent to touch on the people on site being more knowledgeable of how structures act in certain situations. ACCE (2000) guidelines state,

"The Constructor must have an understanding of the contribution of the design disciplines' processes. The Constructor must be able to communicate with the design professionals, and should be capable of participating during the planning phase of design-build projects. Construction sciences and architectural or engineering design topics selected to facilitate communications with the design disciplines and to solve practical construction problems are to be considered in this category." (ACCE p. 10)

Because of these guidelines any construction curriculum should include some level of structural design and analysis. In the three cases listed above, and many other cases many structures collapse during construction due to lack of lateral stability. Understandably so, it is the structural engineers responsibility to design for lateral forces, but construction personnel must also have knowledge of a structure's lateral system to ensure a safe and stable sequence to the construction of the steel frame. Most of the construction curricula do not cover lateral stability of buildings; as a result, students are not aware of its importance (Gujar 2006 p. 3). In cases such as the Husky Stadium roof collapse, perhaps if the construction superintendents and workers had a better knowledge of lateral stability the guy wires may not have been cut incorrectly. In an ideal world, engineers, architects, contractors, and owners work together in all aspects and are involved at all levels to ensure a safe working environment and safe, economical, and innovative final product. Considering the world is not always as ideal as one would prefer, everyone on any project must be educated in all areas of the project, even if in the smallest of capacities, to ensure everything comes together in a safe manner.

Technical Information & Codes
There are several different codes and regulations that govern the design of structures and dictate the types of loads structural engineers must design for. When designing a building a structural engineer will use which ever version of ASCE 7 governs at the time to determine the loads the building will be designed for. During construction though, it is often assumed the loads used as per ASCE 7 are sufficient enough to ensure a safe structure during construction, but sometimes it is necessary to look at other codes as well. As has been discussed throughout this case study the construction process can have many unforeseeable issues, whether they are loads from natural events, poorly designed structures, or construction materials being laid out on an unfinished structure, often it is impossible to design for all of them.

ASCE 37 is a standard that expands on ASCE 7 to include design loads on structures during construction. In some areas ASCE 37 can be very vague, whether by design or to leave the engineer with more options, it can create some issues. This may be one of the reasons it is not always talked about in engineering classes and offices. That being said, ASCE 37 does contain a lot of pertinent information on designing a structure to resist certain forces caused by varying construction loads when combined with other forces such as wind. Therefore, if a structural engineer has at least a basic knowledge of what is contained in ASCE 37 some issues may be avoided.

Some of the case studies listed above, such as the Union Carbide Building, do not fall under such design codes necessarily because of the time at which it was constructed, and it the case of the Union Carbide Building, its location outside of the United States. Nevertheless, perhaps a better knowledge of the possible forces a structure may see, and the unique way in which a structure not completed can possibly distribute those forces could have very well resulted in stronger temporary bracing being used which would have saved the owner and designers a large sum of money in the long run.

Conclusion
In an industry that is moving in the direction of a more integrated process between owners, designers, and builders with methods such as building information modeling and integrated project delivery, we hope to eliminate any problems a project may have. With a more integrated process engineers, architects, and contractors must also work together to try to create a safer environment before, during, and after construction. Although the only accident discussed in this case study involving casualties was the Big Blue crane collapse, the other two could have been just as deadly if not more had it not been for a combination of lucky timing and heads up supervision. Everyone on a project hopes all of their colleagues go home to their families each and every night and sometimes there are accidents that can not be avoided. That is why all involved in the construction industry must learn from the accidents that have occurred, take away the lessons they have taught us, and work closer together with other disciplines to avoid accidents at all costs. Structural engineers are not always involved in the means and methods of construction, but sometimes talking with contractors and working with them can go a long way and save lives.

Additional Resources

 * American Society of Civil Engineers (ASCE) (2002). “ASCE Standard: Design Loads on Structures During Construction.” ASCE, Reston, VA.**
 * Gives standards for design loads on structures during construction, including structures that are not yet enclosed and may have different loads on them than once the building is occupied caused by different distribution of forces along the structure.


 * McDonald, Brian, Ross, Bernard, and Saraf, Vijay (2007). “Big Blue Goes Down. The Miller Park Crane Accident.” //Engineering Failure Analysis,// 14(6), 942-961.**
 * This is a journal report on the Crane collapse at Miller Park, it describes the engineering analyses used to disprove some of the theories as to the cause of the failure.toc