LAUSR

Abstract a real snow drifting phenomenon may last from hours to several days, therefore the wind-snow phase boundary is dynamic, and can affect the local flow field when enough changes occur in the snow boundary. To simulate the dynamic process of snow boundary, this paper proposes an adaptive-mesh method using radial basis function (RBF) interpolation, which realizes the change of phase boundary in both two-dimensional and three-dimensional snow drifting. A time-marching method based on steady-state algorithm is used in this paper, which can effectively reduce the huge cost of transient computation. According to the method, the wind duration determined by meteorological data or wind tunnel tests is divided into several stages, and the boundary of snowpack is updated based on the results of the previous stage using RBF interpolation. Meanwhile, based on the relative residual of the average mass transport rate of snow phase, the optimum iteration method is adopted to determine the optimal subsection of duration. By employing the adaptive-mesh method proposed in this paper, the snow redistribution on stepped flat roofs is predicted and the process of dynamic change of three-dimensional snow boundary is revealed in detail. The numerical results are in good consistent with the field measurements and the wind tunnel tests. Finally, the differences between two-dimensional and three-dimensional numerical simulation are compared, showing that three-dimensional calculation can provide certain reference to the prediction of the distribution of snow loads on the typical roofs.