Roof load predictions using artificial snow

A series of companion field and laboratory studies have shown that roof snow loads can be efficiently and effectively modeled in a cold room. Water and air are precooled by a heat exchange unit before passing through mixing nozzles in a cold room where the temperature is controlled to meet the requirements of the experiment. The artificial snow has small crystals with specific gravity from 0.13 to 0.26. Small-scale gable and shed roofs, equipped with load cells and thermocouples, are placed in the cold room, and snow is deposited until the required depth is reached. The room temperature is regulated to yield the design weather pattern. Loads and temperatures are monitored periodically by a data logger and the information is transmitted to a central computer. Probabilistic models were developed and simulation used to predict 50-year mri roof loads. These loads were comparable to those predicted by a field-derived model.