Roof Collapse
In a perfect world of consulting engineering, every investigation would result in a single, concise cause of damage to the product, structure, or system under investigation. In the real world, this rarely happens. For example, a portion of a roof structure over a two story office building fell overnight. The roof structure consisted of bar joists that spanned 40 feet front-to-rear, with the joists supported on the front and rear exterior walls. The joists were 20 inches deep, and were spaced at 4.0 feet center-to-center. Measurements showed that the roof had no slope.
There were two drains on the main portion of the roof, with three inch diameter pipes attached to the lower flange of the drains. The roofing, from top to bottom, consisted of a layer of ballast rock, a layer of spray foam roofing, and a layer of built-up roofing. The installation of the foam material over the built-up roofing and the installation of the ballast material indicates the building was re-roofed at some time. A one square foot area was marked on the roof in three locations, with the ballast rock from each area collected for analysis, along with a portion of the roofing material. The samples of the ballast rock were dried and weighed in EDT’s laboratory. Sample 1 had a total weight of 17 pounds, Sample 2 had a total weight of 10.5 pounds, and Sample 3 had a total weight of 9.5 pounds. One square foot of the roofing material was also weighed and had a total weight of 4.0 pounds.
Weather records from the National Weather Service were reviewed for the time period when the roof fell. The records showed that the rainfall in the area around the facility was less than the code required design level for rainfall. The capacity of the roof drains was analyzed per the requirements of the International Plumbing Code (IPC). The analysis showed that the two drains on the main portion of the roof did not meet the size requirements of the International Plumbing Code.
An analysis was done to determine if the bar joists used at the facility were sufficient to carry the applied loads. The Steel Joist Institute was provided the dimensions of the joists, and the institute determined that the joists at the facility were either 20J6 joists or 20H6 joists. Neither joist was capable of supporting the weight of the roofing system and the weight of the ballast, per the capacities listed by the Institute. The calculations also showed that the deflection of the joists at the center of the building, when the joists were supporting only the weight of the roof system, with no other loads included, was 3.04 inches.
The Manual of Steel Construction states that the depth of beams and girders supporting flat roofs shall be not less than the strength of the steel divided by 600,000, multiplied by their span length. Based on the properties of the members used at the facility and the Manual of Steel Construction requirements, the minimum depth of the joists was required to be 29.04 inches. The joists used at the facility were 20 inches deep. The use of joists with a depth of 20 inches, and the use of joists that were not capable of supporting the loads required by the building code are design defects.
Analysis showed that the bar joists were not capable of supporting the loads required by the building code. However, the building was 48 years old. The joists supported the roof for 48 years before the damage occurred, showing that the capacities as provided by the Steel Joints Institute include a safety factor. The fact that the joists supported the roof for a minimum of 48 years before the damage occurred shows that the loads applied to the joists increased at some point. The re-roofing work, which included the addition of up to 17 pounds per square foot of load on the roof, increased the load that the joists were supporting by up to 50 percent.
The weather records showed that heavy rainfall was recorded in the area around the time of the damage to the building. Had the roof been properly designed, with the joists capable of supporting the code required loads, and using joists with an adequate depth, the deflection of the joists would have been limited. The steel manual states that the minimum depth requirements are in place to “minimize the effect of “ponding,” wherein the deflection of supporting beams results in the retention of rain water which, in turn, causes additional deflection.” As shown above, the calculations based on the manual requirements showed that the joists should have been a minimum of 29 inches deep. Therefore, the root cause of the damage to the facility was the inadequate design of the roof structural system.
The investigation showed that the root cause of the damage to the facility was the inadequate design of the roof structural system. However, the addition of the ballast during the reroofing, the inadequate size of the roof drains, and the ponding of water on the roof were all contributing factors. Thus, instead of the preferred single cause of damage, there were multiple factors at work.
About the Author
Charles E. Whitley, P.E. is a consulting engineer with our Birmingham Office. Mr. Whitley provides technical consultation related to the design and construction evaluation of industrial, commercial, and residential structures; failure analysis; damage assessment; and expert testimony. You may contact Charles for your forensic engineering needs at cwhitley@edtengineers.com or (205) 838-1040.
Learn about how EDT Forensic Engineering & Consulting approaches construction assessments, scope of damage, and forensic engineering by assigning a file today.