Snow Load Assessment Using the Finite Area Element Method: A Parametric Study to Validate the Feasibility of Using Large-Eddy Simulations as the Wind Velocity Input

In structural snow load modeling, physical wind tunnel testing of scale models has been a widely accepted method for gathering bulk flow data over building roofs. While this approach is well-established and reliable, its ability to provide high-resolution data can be limited by the size and spacing requirements of instrumentation and the building size and geometry at scale. Recently, computational fluid dynamics (CFD) simulations using Reynolds-averaged Navier–Stokes (RANS) turbulence models have gained traction for predicting wind speeds on and around buildings at planes of interest, such as roof surfaces. The computational approach has demonstrated some success in estimating bulk wind speeds used to derive snow-induced structural loads. However, RANS models exhibit limitations in wind speed predictions, particularly in areas where they struggle to resolve recirculating and turbulent flows, often leading to overestimated snow loads in those zones. Building on previous research that employed RANS for generating high-resolution input flow fields for snow loading software, this study explores the use of large-eddy simulation (LES) to potentially improve the flow field accuracy at low-wind regions and enhance the understanding of the relationship between building aerodynamics and snow accumulation. A generic building step model with adjacent upper and lower roofs, associated with classic recirculation phenomena and low-wind zones, was selected for the analysis. The results of the comparative analysis between the LES, the RANS, and the wind tunnel approaches indicate that by capturing detailed, time-dependent turbulent flow, LES offers a more nuanced depiction of wind recirculation zones that RANS models generally fail to capture, resulting in better agreement with the baseline wind tunnel results.