Similitude considerations for roof snow loads

Abstract

Scale-model gabled roofs were tested in a cold room laboratory using artificial snow to obtain conditional probabilities of sliding for various roof configurations. A companion field study was conducted on prototype roofs for the same purpose. All of the structures were unheated, gabled roofs with slippery metal cladding. The similitude relationships were established using dimensional analysis. During sliding, the weight of the snow mass overcomes the compressive, frictional, adhesive, and tensile resisting forces. Ten parameters were judged to be significant in causing snow to slide from the roofs. Using the Buckingham Pi theorem, we developed five relationships between model and prototype. Laboratory results were related to field information by comparing degree-hours, snow density, depth of roof snowpack, diameter of snow crystals, and weight of roof snowpack. In some cases the roofs did not meet the similitude requirements of depth or weight. This slight distortion did not significantly affect the results. Our study considered neither ice dams nor slides cause by addition of load from precipitation. Comparison of prototype and laboratory data reveals that model results produce reasonable predictions of roof snow-sliding.