Prepared by: W. Carter Sinclair Jr.
On September 11, 2008, I was contacted by Lieutenant Jason Zaccarine of the Chesterfield County Police Department, and asked to participate in a review of the collapse of the Fountain Square Condominiums. The Fountain Square Condominiums had collapsed on September 10, at approximately 1200 hours and caused the death of one construction worker. I arrived on site at approximately 1200 on September 11.
I met with William Dupler, the Building Official of the County of Chesterfield, Captain Robert Skowron, Lieutenant Jason Zaccarine, both with the Chester County Police Department and James Frye with the Bureau of Capital Outlay Management for the Commonwealth of Virginia. The purpose of the meeting was for a briefing and to determine what resources I might require for the investigation. I reviewed the project plans that the contractor provided Lieutenant Zaccarine, in addition to final drawings and roof truss shop drawings provided by the Building Official. I learned in the meeting that framing inspections and roof truss shop drawings provided by the Building Department officials and Special Inspections were required for the project. I requested a copy of the Special Inspections forms and some additional shop drawings that included the floor trusses. The plans were discussed as well as the level of inspections completed for the structure. The structural supports and lateral requirements were discussed. It was noted that the plans went through several reviews prior to approval. The complete set of plans was prepared by Keith Engineering, Inc.
The meeting was followed with an inspection of the collapsed structure. The following were present for the inspection of the building, Jason Zaccarine, William Dupler, James Frye, the police photographer, the Fire Chief and myself. At various times during the inspection we were joined by members of the Search and Rescue team. For clarification the descriptions of direction are as follows; the bottom of the plan is referred to as the front of the structure, the top of the plan is referred to as the rear of the structure, the right side of the plan is referred to as the right end of the structure and the left side of the plan is referred to as the left end of the building. For unit numbering purposes the numbers start at the left end of the building and increase to the right end of the building. The units on the second level are numbered 1 to 9.
The following is review of observation from my first inspection of the collapsed structure. At the front of the structure, there were collapsed balconies with cracked concrete topping on the ground. Approximately two thirds of the front of the structure was on the ground. The front of the structure at each was still vertical. We first entered into the building on the left end of the building. There was cracking in the brick veneer on the exterior of the building. The exterior wall and first interior bearing wall was still vertical, but there was some buckling of the metal studs at the front section of the interior bearing wall. The first interior bearing wall had several openings in the wall, with 2 x 8 gang studs supporting LVL beams above. The next interior bearing wall was either collapsed or buckled. The collapse was complete in the front section and as the wall neared the rear of the structure, the studs were semi-collapsed. The next area entered was on the second floor at the rear of the structure. The site was sloped such that the second level was near ground level on the rear of the structure. The rear wall across much of the building was out of plumb, however it had not collapsed. This was due to the stabilizing effect concrete wall beneath the exterior wall at the second level. There were 8 x 8 wooden posts at the rear of the structure supporting an overhanging canopy. These posts were also out of plumb. The unit on the second floor at the left end of the building, unit 1, was relatively intact. Drywall sheathing was installed in the unit, however there was cracking in the drywall caused by the lateral movement of the structure. It was noted along the rear of the building that there was separation of the floor system on the second level from the concrete slab that made up the entry porch into the rear of the building. The separation gap was most severe in the center of the building and there was no visible attachment of the floor system to the concrete wall and slab. The next place of entry was the last unit on the right side of the building, unit 9, at the second level. The wall studs were covered in drywall and the resilient channel was installed in parts of the roof trusses. It was noted that there were some field truss modifications to the ends of the trusses on exterior wall of the unit. The wall sheathing in this unit did not show any cracking. The next area of inspection was in the concrete tunnel on the first level at the rear of the structure. The interior bearing wall on the first level at the left end of the structure, the studs adjacent to the concrete wall were buckled, but not completely collapsed. Moving away from the rear wall the studs were completely collapsed towards the front of the building. We moved to the front of the building and reviewed the support of the 8 x 8 posts, it appeared that one of the posts was not completely on the concrete support, it was discovered that is was shifted one inch off the support, however it appeared to be built that way. I entered the front of the building where the veneer collapsed onto the victim. The second level was still relatively intact, however it was resting at ground level with the first floor completely collapsed. The roof mansard was also relatively intact. The Search and Rescue team had installed shoring to inspect the remainder of the structure for any remaining people. I then went up the Fire ladder over the top of the roof. There was no indication of any blocked drains prior to the collapse. There were the remnants of a pool of water that had formed from rain the previous evening. The pool was pumped out the previous evening.
The most important result of the inspection was the lack of any type of lateral bracing for the studs on the interior bearing walls on the structure with the exception of the interior wall at the right end of the structure. There was no drywall installed, or any type of temporary or permanent steel or wood bracing.
The inspection was followed with a meeting with the contractor, Barney McLaughlin and a representative from Keith Engineering, Jeff Keith. The contractor was told that he would have access to the building the next day for his insurance engineers. I did not have any questions for the contractor, but I did ask Jeff Keith about his level of Special Inspections. He told me that he was only involved in the inspections of the foundations and not in any of the framing. I asked him to send me a copy of his inspections. It was stated by the building official that they had not received any of the Special Inspections reports; however this is normal not to have the reports until the end of the Special Inspections period.
The Search and Rescue portion of the operation was concluded the evening of September 11 and the building was then released to the Police department. I met with Lieutenant Zaccarine to determine a plan for the following day. Chesterfield County hired a demolition firm to begin a controlled demolition of the structure. This operation was important for its impact on determining how many pieces of drywall were stacked and where were the stacks located. It was determined that most of the recently delivered drywall was stacked in unit 7 and 8. The initial plan was to begin a controlled demolition, beginning with removing a section of parapet at the rear of the building for safety. The section of parapet at the rear of the building had a brick veneer and it cantilevered over the roof, risking further collapse. The next planned step was to remove the roof over units 7 and 8. The units 7 and 8 were collapsed and posed a safety risk for entry. The plan was for a review of the structure in the morning for the contractor and insurance representatives and controlled demolition to begin around 1200. I was scheduled to return to the site at the beginning of controlled demolition.
On September 12, I returned to the site around 1200. The demolition was delayed due to a request by the contractor to save the right end units of the building. In order for that to occur, the roof diaphragm had to be cut. The safest way to access the roof was with the use of a man lift to hover above the roof and have a man use a chop saw to cut the roof. While waiting for the lift, it was recommended to the demolition contractor to remove the parapet at the rear of the building. The removal began and in conference with the demolition contractor, it was determined that the parapet was fairly stable and would not pose a risk to the roof removal, since that would take place in front of the structure. The equipment moved to the front of the structure. Demolition began on the front of the structure, and it was determined that the front of the structure at unit 7 and 8 was not very stable and continued demolition could cause the exterior wall at the first level to collapse. The demolition plans changed at this point in time, to do a review of unit 7 and 8 from a man lift, looking in through the windows and doors of the second level. A man lift was delivered to the site, and I did a visual review of the drywall stacks in units 7 and 8. A police forensic photographer took still shots and videos of the second level in unit 7 and 8. At this point in time, my review of the site was complete.
The following section of the report is a description of the structural components based upon the plans and what was actually in the field.
In a review of the plans it shows a two story structure. The roof structure was shown on plan as wood trusses with a required design by the truss supplier. The roof sheathing was shown on plan as 5/8 sheathing supported by the wood trusses. A double 2 x 4 stud wall with a spacing of 16 inches on center was shown supporting the roof trusses. There were locations shown at ther ear of certain units that had LVL beams in place to form an opening. The LVL was supported by gang studs at each end. The second floor was shown as 18 inch deep open web trusses at 16 inches on center. The required deflection of the trusses was a maximum of L/480 deflection with final design the responsibility of the truss manufacturer. The plans show 8 inch 20 gage studs on the first floor bearing walls supporting the second floor trusses, it is noted that an earlier set of plans shows a 6 inch 20 gage stud as the bearing wall. In certain locations 3 studs were shown on the plan to support a point load from the roof loading above. The steel studs were shown on plan to be supported on two courses of eight inch CMU and supported on a concrete spread footing. The exterior walls for the structure are shown on plan as 2 x 4 wood studs at 16 inches on center. The exterior wall on the rear of the structure at the first floor level was shown as either 8 inch CMU or concrete, forming the interior wall of the rear tunnel. The exterior wall of the rear tunnel was shown as a concrete wall to be designed by the concrete contractor. One item of note is there were no details found for the connection of the second floor diaphragm to the rear concrete tunnel. The balconies on the front of the structure were supported by wood framing, attaching to a ledger at the building and spanning to a wood beam supported on 8 x 8 wood posts. The wood framing was shown with a concrete topping. The posts were supported by concrete spread footings. There were no connection details for the posts to the foundation or beam above. The canopy on the rear of the building appears to be shown on plans without any column support required. The canopy is shown bolted to the exterior wall of the structure, without provisions for column support at the outside edge of the canopy. The plans showed a 4 foot section of drywall panel at each end of the bearing walls for the provision of lateral support of the structure in the front to back direction. The plans called for OSB sheathing nailed to the exterior wall studs for lateral bracing in the left to right direction. All of the design loading shown on the plans met the requirements of IBC (International Building Code).
The structure as built appeared to follow the building plans. The first floor unit at the left end of the building had multiple penetrations in the bearing wall that were not shown on the original plan; however subsequent plans were submitted to the County for approved modifications. There were 8 x 8 columns added to the rear of the structure which were not evident on the plans as being load bearing.
Based upon an inspection of the collapse and a review of the plans it appears that the collapse was the result of overloading of the bearing wall studs on the first level. The critical part is to determine what caused the overloading and where the overloading occurred. There can be overloading of a vertical structural support in two directions, vertical and lateral. That means that loads can be applied by gravity forces or by lateral forces. The collapse of this building is similar for example, to a collapse experienced in an earthquake, due to a soft first story. The soft first story is a building with poor lateral support on the first floor. However in the case of this building collapse there was no indication of any type of lateral loading on the building from wind or earthquake loading. There was a report of the drywall delivery truck boom pulling on the building; however there was no indication of any type of damage around the window where the boom truck was loading the drywall. With the elimination of the lateral loading, that leaves gravity loading as the only source of overloading of a vertical element. Therefore the overloading had to be the result of overloading in the vertical direction at the beginning of the collapse.
This leads to what caused the critical vertical loading. Almost all of the dead load was in place for the structure on the second level; the dead load consists of all the construction materials that will be in place for the duration of the structure, such as framing, plywood, siding, roofing, flooring, etc. The exception was in units 7 and 8 where the drywall was not installed. The installation of drywall had begun in unit 7, but had not been started in unit 8. There was a large amount of drywall stacked in units 7 and 8 and it resulted in a substantial loading of the first level steel studs. Based upon the visual review of the drywall in addition to a sketch provided by the Police department, obtained by a subcontractor, a vertical load of 1889 pounds was calculated. The calculations are included in an appendix at the rear of the report. The worst case scenario was the combination of several vertical stacks of drywall with two horizontal stacks of drywall along with several buckets of drywall compound. These loads combined into loading onto a single stud. Based upon the allowable loading calculation supplied by Clark Weston; the allowable vertical load on an 8 inch 20 gage stud unbraced in the weak direction at a height of 9’-2" is 848 pounds. That means an overstress of the stud by a factor of 2.22. The stress maximum should be a factor or 1.0. The engineer at Clark Weston stated that there is a safety factor of 1.6 in the material, therefore global buckling (the entire height of the stud) leading to localized buckling would occur at any point past the 1.6 level. Additional calculations were performed on stud loading for the units that were already sheathed in drywall. Two cases were evaluated, one with a dead load of 15 psf for the floor and roof system and a second case of 10 psf for the floor and roof system. The first case resulted in a stud loading of 1099 pounds per stud and second case resulted in a stud loading of 733 pounds per stud. The first case results in a 1.29 stress level or overstress and second case results in a .86 stress level. As the adjacent studs tried to pick up the additional load released by the collapsing stud, they became overstressed and then transferred the loads to their adjacent studs. There was no redundancy along the bearing wall to pick up a single overload condition. When the floor and roof began to move, dynamic loading was introduced into the structure. Dynamic loading is a load caused by movement. When the second floor began to come down it exerted a greater force on the structure, than if the structure was in a static (motionless) condition. In addition lateral forces were introduced into the building structure. As the floor came down, tensile forces along the plywood floor became lateral forces, this is similar to the loading of a trampoline; as trampoline has a load applied the strings around the outside start to stretch, and the more motion on the load the more the spring stretches. One indication of the tension forces in the collapse was a tearing of the floor sheathing at the initial collapse spot in unit 8. The question remains as to why the exterior wall did not collapse as well as each end of the building. It appears that the bay windows in unit 7 and 8 were sheathed at the time of the collapse. This probably provided enough lateral support in the exterior wall to prevent collapse. In addition unit 8 was adjacent to the fully sheathed bearing wall on the first level between units 8 and 9. It appears that the left end of the building remained standing due to the use of wood studs along the bearing wall. The wood studs are much stronger in the weak axis than the metal studs prior to sheathing. Therefore the bearing wall between units 1 and 2 on the first level is able to carry more load than the remaining walls. However it was noted that the front section of the metal stud walls showed signs of buckling. In summation the collapse mechanism probably occurred as follows, there was a single point overstress of the stud, as it buckled it transferred loads to adjacent studs and they buckled, as the line collapsed lateral loads were put into the floor system and the structure was unstable laterally, the adjacent bearing walls were overstressed and the second level started the downward motion putting dynamic and lateral loads on the structure.
The problem of the collapse appears to be centered upon the unbraced 8 inch 20 gage stud. The question is why the stud was unbraced at this point in the construction sequence. It is reasonable to assume that as soon as the studs are erected, there is no lateral bracing; however at this point in the construction sequence the studs should have had lateral bracing. Included in the appendix is the bracing requirement for the studs, which is a maximum of 48 inches on center. None of the studs in the collapsed area was braced in the weak direction. Although the drywall does provide lateral bracing, separate steel bridging was required, and it was not used on the project. All means and methods, which include temporary and permanent bracing, are the responsibility of the contractor. There is also a problem with the design of the steel studs. This is based upon the American Iron and Steel Institute. The requirement is based upon the web to material thickness ratio. The ratio is the depth of the stud to the material thickness which is h/t or 8/.0346 which is equal to 231. The maximum allowable is 200 unless web stiffeners are provided. The designer did not require the submittal of a designed metal stud system, therefore he is responsible for either calling out the web stiffeners or increasing the gage of the metal studs to meet the required ratio. This might not have prevented the collapse; however it could have limited the amount of the collapse.
In conclusion the problem rests with the use of the 8 inch 20 gage studs for the bearing wall and their being in an unbraced state at the time of collapse. At one point in time the plans called for a 6 inch 20 gage stud for the bearing walls. This would not have prevented the collapse, but the 6 inch studs did not require web stiffeners. There was a bearing issue with the 6 inch studs, so 8 inch studs were substituted. I assume the reasoning was that if 6 inch was good then 8 inch must be better; but this is not the case with metal studs in their unbraced condition. It did not appear that the contractor was cognizant of the results of lateral loads on a structure. There was no lateral bracing in place along the lower level, not even temporary bracing. The question becomes should the building official be responsible to insure that the lateral bracing is installed. The building official is responsible for all permanent bracing; however he is not responsible for temporary bracing during construction. This is a gray area with metal studs as the bridging can be installed prior to the sheathing being installed, however it is still the ultimate responsibility of the contractor. In addition to bracing in the studs, there was no lateral bracing evident in the roof trusses. This just paints a picture of a contractor who is not diligent about installing all required lateral bracing. The drywall stacks caused the start of the collapse, however if the studs were properly braced the studs would have carried the load.
W. Carter Sinclair Jr., P.E.
