Mathematical Simulation of 50-Year Snow Drift Loads

Abstract

Snow loading on building roofs is an important consideration in the design process in many regions across the United States. It is especially important for buildings that have roof geometries that promote drifting of snow. When snow is transported across the roof of a building to a drift accumulation area, it causes a large concentrated load. Building roofs are more likely to suffer structural damage, or failure, from this large concentrated load rather than from a uniform load acting over the entire roof. Current building code provisions for roof snow drifts have typically been based on reviews of insurance case histories. This methodology considers a drift on a particular building, during the particular year of the insurance claim. As a result, there is no easily determined return period associated with the case history. Hence, the return period for the relationship developed from the case histories is unknown. By using a mathematical procedure to simulate roofsnow drift loads, the maximum annual drift load can be computed for each year over a period of years. From these annual maximums, the return period for a design drift can be estimated. The mathematical procedure is based on the physics of drift formation which involves a transport rate and trapping efficiency. The procedure used to simulate the maximum annual drift for a particular roof during a particular winter involves: (1) determining the amount of driftable snow in the roofs snow source area, (2) determining the amount of snow transported from the roof's snow source area based on the winter's wind speed record and transport rate, (3) and finally applying the trapping efficiency to determine the amount of transported snow that is in the drift. The procedure outlined above is applied to each winter's weather data for a period of 19 years at each of 46 weather stations across the United States. The maximum annual drifts from each station are then fit with a Gumbel Distribution to determine the 50- year drift load from each station for both a gable and lee-step roof geometry. These 50- year drift loads are than used to find the 50- year drift ratios, which are compared with the drift ratios for a leeward step and gable roof geometry using the methodology in the ASCE- 7 Load Standard. From this comparison conclusions are made on the accuracy of the current method for calculating drift loads.