Posted on July 30, 2012
The New England area experienced an accumulation of ice and snow between December 26, 2010 and February 8, 2011 that has not been seen in more than 15 years. It may have been a windfall for roofing contractors and skiers this winter; but, it frayed a lot of nerves of property owners across the region.
Cissell Investigative Engineering (CIE) is often asked to address property defects and causes of collapses. The first part of the analysis requires assessing snow accumulation by taking field measurements at several locations near weather stations. It's a simple test, collect an area of snow, measure the depth, and weigh the accumulated amount. Then, we compare this data to weather station records to find a correlation factor. Later, when the snow melts, we have a way to assess the amount of accumulated snow and ice, as well as the weight it imposed on a structure. The next step is to assess if there are any special considerations, such as, the potential for drifting, the manner in which snow was removed, and any special loading situations.
CIE typically has found a range of snow and ice accumulation loads ranging from 20 psf (pounds per square foot) to 30.5 psf. Normally, this loading would cause bending but typically does not break anything. Most design codes in the region require at least a 30 psf ground snow load. The design code tolerance allows for a specified amount of bending at the design load. Typically, a 1" in 20' is the tolerance for most roof structures.
The structural frame has to bend a little to absorb increased loads; this is normal. As previously mentioned, a certain amount of bending is expected and tolerated in the codes. The tolerances are based more on prevention of finished materials cracking than on collapse. If you can prevent the finishing materials from cracking, such as drywall, then you can be assured that the structure is not going to collapse unless there is some major connection failure or defect. More on that later.
Let's take a simple rafter framed sloped roof structure. The snow load pushes down on the rafter and the rafter will bend some to absorb the load. The rafter essentially pushes up to resist the load. Where does the load go from here? The vertical load that equals the snow and structure weight pushes down on the support wall, however, because the rafter is sloping, there is a horizontal load component. Engineers thrive on geometry, but your eyes may start to glaze over if I explain this in detail. The greater the slope, the more the horizontal load and forces that could push against the top of a support wall. The horizontal load is commonly restrained by ceiling joists and collar ties.
Many of the rafter problems observed this winter were due to the lack of lateral restraint. That is, the horizontal load from the rafter pushed on the top of the wall and the wall leaned out. The visual evidence of this is to look at the ridge line and see a sag in structure. Cathedral ceilings were a common source of these failures; there are special construction details needed to properly support a cathedral ceiling which are often overlooked. Even conventional ceilings were found with defective lateral restraint; typically, the rafters weren't connected to the ceiling joists, or ceiling joists ran the wrong way. We found a lot of "wrong ways" while investigating structural defects and collapes.
A number of the total collapses we observed had components of decay in the rafters. If you take a rafter that is undersized to start with and decay begins to chew on the fibers, the member gets overloaded pretty quickly. The snow events are a convenient way to demolish a delapitated structure. It is my opinion that the proximate cause of many of these collapses is not the result of ice and snow weight, but the strength reduction afforded by the neglect and decay. Decay is an inherent vice that progressively weakens the wood member.
Our team has been surprised by the number of trusses that failed. As previously mentioned, the snow loads approached the design load but rarely exceeded it. Even with the loads exceeded, the margins of safety would have prevented a collapse or member failure. Ask a fireman what they think of trusses, while they are an efficient use of materials, they rely on the integrity of each piece.
CIE investigations found manufacturing defects, design defects and installation defects. The most common failure involved metal plates with wooden trusses. We commonly found plates pulled off the members. Once you lose one member, the truss starts to tear apart leading to failure. Once one truss fails, there is a domino affect and more start to fail until there is a catestrophic collapse. The manufacturing defects found were typically poor quality control on the placement and embedment of the plates. Design defects included undersizing members and not accommodating extra loading needs. Installation failures included the lack of bracing per manufacturers specifications. The problem with trusses is that they can partially fail and remain hidden for years. CIE was called onto a claim where two trusses had punched through a ceiling. After climbing the roof, we did not find any unusual loading scenario; it was a long, wide slightly pitched roof. The attic revealed dozens of trusses of the same construction and after pulling back insulation we found an additional six (6) trusses that had failed connection plates. Disaster averted! What surprised me was that there were engineered repairs made previously to a few trusses on the other slope. CIE recommended the trusses be measured and independently analysed since the design loads were not met and the failure occurred due connection failures and a lack of bracing.
We have seen some proposed contractor solutions that often exceed the demands and needs of the loss. As an example, countless times contractors have stated that the whole roof needs to come off and be replaced. Fortunately, for some of our clients, we have the resources from a couple of structural contractors who are pretty clever in fixing these problems at small fraction of the replacement costs. Here are a couple of the simpler solutions that can be effected:
Unfortunately, many of the problems have nott been identified or recognized. This past winter collapsed a lot of the already weakened structures, it may take another dozen years before the structures weakened by this past snow and ice buildup cause a claim. We have also visited a number of structures where there was great concern on the part of the property owner, most of the time we found just excessive deflection cracks in the drywall and found a simple solution to prevent a reoccurrence. It is just as rewarding to provide peace of mind!