Recently, NKB was called by an existing client to assist them with a problem they began experiencing with ice falling from window sills several stories above the ground. In mid-January, Old Man Winter comes to Upstate New York and dropped 14 inches of snow in about 36 hours. Temperatures ranged from 15ºF to 25ºF and wind speeds were consistently between 5 and 15 mph with occasional gusts of up to 35-40mph. Over the two days which followed the completion of the storm there were reports of ice buildup and falling to grade below. The client called us in to examine what might be causing the ice build-up. What we found when we walked the perimeter of the building and looked out on various facades of the building from facing windows was very telling.
Each window sill showed some snow and ice still in place near the edge of the sill. The 6-8 inches nearest to the window was clear in all locations. Every – Single – Window. All 245 instances, showed the same clear area adjacent to the windows with ice buildup at the edge. Our client admitted that they did not recall if this was a problem prior to the facade renovation project. It had been several years since pedestrians were allowed anywhere near the building so there was no institutional memory of occurrences like this happening previously.
We began running through the list of things that could possibly cause this melting condition:
The windows are known to be headaches to the building management. Infiltration problems are known to exist so could we be getting warm air leakage at the windows causing snow melt? Well, warm air leakage would likely be a less uniform condition. Coupled with the fact that the facade renovation project that was just completed included the resealing of the entire perimeter of each of the windows, we decided that this was not a major contributor.
Since this is a horizontal surface, it could be that the sun was causing the snow melt, right? When we noticed that the condition existed at all faces, regardless of orientation (including the north facing facade which gets essentially zero sun at this time of year), we ruled out solar melting as a primary contributor to the problem.
Upon a closer inspection of the window sills and a reexamination of the original construction documents from the 1970’s, we realized that this pre-cast concrete window sill was likely acting as a thermal short. It extended, unbroken, from the exterior to the interior of the building.
A thermal short is a condition in which thermal energy is allowed to flow along a path of least resistance from a heat source to a colder region on the other side of the barrier. Consider an electrical short circuit. An electrical short is one in which the electricity is allowed to flow along a shorter path than intended and in doing so damages components and ceases to work properly (hopefully without killing anyone in the process). A thermal short is a similar phenomenon. In most homes the walls contain some form of insulation to provide greater resistance to heat loss in the winter or air conditioning in the summer. In a commercial building, a thermal short is most often exterior panels supported by metal clips and brackets which are tied to metal structural components. This string of metal tied to metal creates an easy path along which thermal energy can flow. This building did not have an unbroken string of metal from exterior to interior making identification of a thermal short in this case particularly difficult. However, the uninsulated concrete still served the function of a thermal short by acting as a path of least resistance for the thermal energy in the building to escape.
Most people consider concrete to be of sufficient thermal mass and to be a relatively capable thermal insulator. Compared to metal window frames and door frames or z-channel, this understanding is exactly right. Compared to air gaps and closed-cell insulation, this is an error in judgement. In an effort to better explain what was causing the observed conditions and to predict a likely solution we put together, a Finite Element Analysis (FEA) thermal model of the window sill conditions. We created 50 different runs with a varying set of boundary conditions and thermal loads. These runs examined the following conditions:
Insulation in PTAC cavity
Wind Speed Across Face of Building
What came out of the simulations was both expected and surprising at the same time. The results of the model with the existing conditions as inputs produced outputs which would predict the melting and refreezing conditions exactly as observed.
We then created a series of simulations in which we insulated the interior faces of the concrete so as to attempt to create the thermal break required to prevent the wholesale loss of building heat through the concrete sill. The results of these simulations predict that 1 inch of polyisocyanurate insulation on the interior face of the concrete sill will prevent melting on the sill in all conditions except those which also do not allow the snow melt to refreeze on the concrete.
Over the remaining winter months, we have found that this condition seemed to be limited to larger snow events. We had several smaller snow falls (less than 4 inches at a time) and it seems that there is a critical mass of snow which is required to create the pieces of ice that are large enough to separate from the building. This critical mass has not been reached since that snow storm in mid-January but considering that we, in Syracuse, routinely win the Golden Snowball award for largest snowfall total during the winter months, we can be assured that we will typically get at least one storm capable of creating the conditions to cause this occurrence each year. The implementation of the findings of this analysis is yet to be completed and will be the topic of ongoing discussion with the client.
So what is the moral of this story? Thermal energy is a slippery devil. It finds unexpected paths and causes unexpected consequences. Thermal shorts in all of their forms can be a significant building performance and energy consumption issue. These issues impact not only the long term performance of the building itself, but the people who use the building day-in-and-day-out. This case study serves to highlight the usefulness of a relatively straight-forward Finite Element Analysis steady state thermal model in helping to determine the causes of a given condition as well as helping to identify what recommendations we can make in order to best address them.