MODELING OF AIR POLLUTANT EMISSION AND PREDICTION OF SPATIAL DISTRIBUTION BASED ON DEEP LEARNING METHOD

The development of human economy and society is deeply dependent on the consumption and combustion of fossil fuels, but the combustion of fossil fuels will inevitably produce a large amount of pollutants, which will lead to environmental pollution and damage to the living environment. The research, based on the deep learning theory, is going to study the spatial and temporal distribution characteristics of pollutants and analyzed the factors that affect the transport and distribution of pollutants. Also, this study attempts to control and reduce the impact of pollutants on the environment and human society in the transport link. The research constructed a mathematical model of atmospheric pollutant emissions and transport, and made a sensitivity analysis. The research found that the mathematical model constructed is extremely fast in operation and is suitable for the simulation analysis of pollutant diffusion behavior. The diffusion of pollutants is a gradual process, and the concentration of pollutants near the pollution source tends to increase slightly as the process of combustion continues. The final concentration of gaseous pollutants near the pollution source has risen to 0.21mg/m(3). The oxygen concentration will limit the full combustion of the fuel, so that the overall concentration of pollutants in the confined space is relatively reduced under the condition of high oxygen content. When the oxygen content was increased from 20% to 40%, the rate of pollutant diffusion increased by nearly 7.5. Combustibles are more likely to be fully burned at high temperatures, resulting in a faster diffusion rate of pollutants and a lower concentration of pollutants. This study can provide theoretical basis and support for exploring the emission and diffusion behavior of air pollutants.