Discretizing Three-Dimensional Wind Field For Falling Snow Simulation Based On D3bQ15 Lattice

The simulation of the natural phenomena has always been one of the most challenging problems. The simulation of snow in the natural landscape could greatly enhance the realistic effect of the virtual scene. The irregular shape of snowflakes, the uncertain movement and the regularity of the snow by the environmental factors determined that its modeling approach and the description of its movement was more difficult. To make snow particles immersed into the wind field is the premise of simulating the realistic falling snow scene. Therefore, this paper focused on the following areas.Firstly, the D3bQ15 lattice to improve the model of wind field was used. For the LBM, the numerical accuracy of the wind field was lower when the quantity of the discrete velocity was fewer, but the computational efficiency was slower when the quantity of the discrete velocity was more. A new method was developed. The D3bQ15 of Body-Centre Cubic lattice was used to discrete three-dimensional wind field in view of the limitations of the Cartesian lattice. D3bQ15 lattice has better isotropic and the fewer samples in specific space. It can reconcile the contradictions between the accuracy and the efficiency caused by the quantity of discrete velocity. The new developed method can better meet the demand in the authenticity and the real-time for the falling snow scene than the Cartesian lattice.Secondly, the movement of the snow particles was analyzed. For the movement of the snow particles must be along the lattice lines when based on the lattice of D3bQ15 in simulating the falling snow scene. But the locations of the snow particles in the next time were not exactly corresponded to the three -dimensional lattice nodes. We solve this problem with the approach based on the probability. This method makes sure that the average position of the snow particles change with the expected speed flow, while movement along the lattice lines. In this way, we not only simulate the randomness of falling snow, but also guarantee the macro level of physical accuracy of the falling snow particles.Finally, to validate the effectiveness of the D3bQ15, a falling snow simulation system was designed and realized. The results in the experiments indicated that the new developed method can better meet the demand in the authenticity and the real-time for the falling snow scene than the Cartesian lattice. This study is a useful exploration for falling snow simulation.