Despite adequate structural design and a less-than-design snow event, a scribed log- and sawn-lumber-framed commercial lodge structure experienced a catastrophic collapse less than two years after it was built. The subsequent investigation involved a piece-by-piece selective dismantling of the debris pile to determine the cause of the devastating failure.
The recreational ranch facility where the subject building was located is a vacation-type resort situated on leased federal land in a high alpine environment of the Rocky Mountains. The summer activities of the ranch cater to horseback riding, hiking, bike riding, and exploring the nearby national park. During the months of October to April, the area where the ranch is located can receive upwards of 20 feet or more of snow, making wintertime access by tourists impractical for some and infeasible for others.
Under the terms of the ranch owner’s long-term land lease with the United States National Forest Service, the ranch was permitted to construct permanent structures on the property. As part of a broader facility expansion, the ranch owner began procuring plans for an approximate 3,100-square-foot, one-story fourplex log building that included a wrap-around exterior wooden deck. The plans for the log building were developed by a professional engineer who was licensed in the state where the structure was to be built. The snow loads specified in the construction plans were appropriate for the proposed location of the log building.
The ranch owners/managers were acting as the general contractors for the project. The onsite log erection/assembly, roof covering installation, and rough framing of some of the interior walls took place in early summer. The interior finishes, including interior drywall, insulation, floor coverings, electrical wiring, plumbing, and HVAC equipment, were not installed before the onset of the first winter. Due to project financing issues, all construction progress stopped, and the project remained incomplete for the first winter, the following spring, summer, and fall. The status of the project remained nearly unchanged for the second winter season.
On May 1, the building collapsed catastrophically (Figure 1). The damage was significant and global. It was believed that only the concrete foundation could be reused if the structure was to be rebuilt. Information provided by witnesses who discovered the collapse indicated that 10 or more feet of snow and a thick layer of solid ice had accumulated on the roof at the time of the collapse.
Unified Building Sciences and Engineering, Inc. (UBSE) was hired by an insurance company to determine the cause of the building collapse. As part of the investigation, USBE conducted a site visit once the roads became navigable, approximately one month after the collapse. Based on a detailed review of the weather records, the maximum roof load during the first winter after the initial construction was determined to be 20 to 30 pounds per square foot. The maximum roof snow load on May 1st at the end of the second winter was determined to be 60 to 70 pounds per square foot.
USBE visually and tactiley inspected the logs, framing, connectors, and foundation as part of the investigation. The structural members were still in a near-new state. No corrosion, decay, rot, or deterioration was observed that could have caused or contributed to the building collapse. Aside from the stresses imparted by the collapse, the logs were in good condition and well-suited for their intended function. Similarly, conditions were not observed that would suggest the logs or dimensional framing were improperly fabricated, shipped, or assembled onsite.
The configuration of the building debris after the structural collapse indicated that the failure originated near the center of the building. USBE analyzed the design that was set forth in the construction documents and determined, before the inspection, that the beams, headers, and roof trusses were appropriately designed for the anticipated loads and would have been expected to endure the snow loads that were presented in the weather records.
The approved construction plans specified that the roof structure was to be vertically supported by two triangular-shaped log frames that would be vertically supported along the bottom chord by two 12 foot tall first-floor 2×6 stud-framed interior load-bearing walls. The roof ridge was also to be supported in the transverse direction with two back-to-back 2×4 continuous walls that were to serve as the fire assembly party walls.
The examination of the physical evidence located within the building collapse revealed the following:
- The two 2×6 first-floor interior load-bearing walls were installed; however, they were not sheathed with a gypsum wallboard panel product as specified in the approved construction plans.
- The lack of wall sheathing violated the National Design Specification for Wood Construction (NDS) slenderness ratio for solid columns and rendered the studs vulnerable to weak axis buckling.
- Without sheathing, the code-permissible (allowable) load-bearing capacity of the 2×6 wall was effectively 0 (zero) pounds per square foot; however, the calculations indicate the ultimate (failure load) capacity would have been reached with approximately 55 to 65 pounds per square foot of snow on the roof.
- Blocking had not been installed between the studs that comprised the central 2×6 load-bearing wall.
- The two back-to-back 2×4 continuous first-floor party walls specified on the plans were not installed. The failure to install the specified walls resulted in the roof ridge being vertically unsupported; however, at a laterally unsupported height of 18 feet, these 2×4 walls would have had an ultimate capacity (failure load) of less than 5 percent of the applied load.
- The 2×6 studs that comprised the center load-bearing walls buckled and collapsed under heavy snow loads in early May. The compression buckling of the load-bearing walls caused the log roof frames to fail and the roof structure to come crashing down through the first-floor framing into the crawl space.
Discussion
The conventional 2×4 and 2×6 sawn lumber wall framing was sufficiently tall that Euler buckling action controlled the allowable design load. Had the 2×6 stud framing been sheathed/covered with gypsum wallboard, the calculated allowable axial load for Douglas-Fir No. 2 wooden studs would have been approximately 5,000 pounds per stud with the expected mode of failure being in the direction of the strong axis. The ultimate (failure load) would have been approximately 14,000 pounds per stud in the direction of the strong axis. Without the gypsum wallboard sheathing, the allowable axial load of a Douglas-Fir No. 2 wood stud would have been approximately 400 pounds per stud due to weak axis buckling, and the ultimate (failure load) would have been approximately 1,120 pounds per stud.
The NDS limits the slenderness ratio of axially loaded members to 50 or less for in-service loads and 75 or less for construction loads. A 12-foot tall 2×6 stud without blocking has a calculated slenderness ratio of 96, which means the stud wall framing, as it existed at the time of the collapse, was not suitable to receive service or construction loads.
Conclusion
Based upon the forensic structural engineering evaluation of the building collapse, the following conclusions were reached:
- The structure was appropriately designed for the local design ground snow load of 175 pounds per square foot.
- The fourplex log building collapsed due to the general contractor’s failure to install the interior load-bearing walls in conformance with the plans.
- The interior load-bearing stud walls were not covered with gypsum wallboard sheathing at the time of the collapse. Without the gypsum wallboard sheathing attached to the narrow face of the studs, the wood studs were vulnerable to buckling along the narrow/weak axis.
- Had the gypsum wallboard been installed, the axial capacity of the wood wall studs would have been approximately 12 times greater.
- The general contractor did not realize that the interior finishes, specifically the gypsum wallboard, were necessary for the load-bearing wall to support the anticipated loads.
- The structural engineer did not anticipate that the construction might stop for two winters, leaving the load-bearing wall without the sheathing that was needed to prevent weak axis compression buckling of the wall studs.
- When the load-bearing interior stud walls were subjected to moderate loads due to an accumulation of ice and snow on the roof, the individual 2×6 studs within the two primary interior support walls buckled. The buckling of the load-bearing walls caused the supported roof structure to collapse downward. The loss of integrity of the roof framing placed outward forces on the exterior walls, which caused the perimeter log walls to lean outward, leading to a large-scale catastrophic collapse (Figure 2).
- The collapse could have been avoided had the structural engineer of record been notified of the work stoppage and the incomplete status of the load-bearing wall construction. Similarly, had the plans and notes indicated that some finishes, such as the drywall, were needed to support the loads that were likely to occur, the general contractor may have been alerted to the need to sheath the walls before stopping work.
The structural members in this building were appropriately sized for final finished conditions. However, proper considerations were not in place to account for loads encountered during the construction period. Although the construction documents communicated that the interior load-bearing walls were to receive sheathing, it was not understood by the builder or the property owner that the specified sheathing needed to be in place before stopping the work for the duration of the winter. Had the structural engineer of record been informed of the state of construction and the desire to stop work, the collapse could have likely been avoided. ■
About the Author
Peter Marxhausen, PE, is a full-time forensic structural engineer with Unified Building Sciences and Engineering, Inc. (UBSE) and a part-time non-tenure professor of civil engineering at the University of Colorado Denver. (pmarxhausen@ubse.com)