Dynamic simulation of smoke diffusion and gas pollution

Abstract

Gas pollution is a common phenomenon in natural life. Aiming at the low efficiency of simulating smoke diffusion based on physical model and the poor effect of drawing gas pollution based on empirical model, a hybrid model method is proposed to draw the dynamic gas pollution caused by smoke diffusion. Firstly, the semi Lagrangian method is used to model the smoke, and the k-d tree is introduced to improve the computational efficiency; secondly, to solve the problem of insufficient details in smoke simulation, the fluctuating wind field based on linear filter method is introduced into the external force term to optimize the trajectory of smoke particles; the bidirectional shot function combined with the real smoke texture is selected for rendering to avoid the problem of obvious particle sense and significantly improve the smoke diffusion details; then, the optimized Gaussian plume model is introduced to establish the relationship between the physical model and the empirical model, and the pollution attenuation formula and the optimized Perlin noise are used to improve the lack of details of global gas pollution and increase the realism of gas pollution changes; by improving the time axis algorithm, the problem of fixed gas pollution color is solved, and the dynamic gradual gas pollution is obtained. Finally, several groups of analysis and comparison experiments are designed. The results show that this method can draw a realistic dynamic gas pollution scene in real-time.