Prediction of pollutant concentration variation inside a turbulent dispersing plume using PDF and Gaussian models

摘要

In order to evaluate the impact of emission of pollutants on the environment, it has become increasingly important that the dispersion of pollutants be predicted accurately. Recently, USEPA has proposed stringent guidelines for regulating the diesel exhaust emissions, specifically, NO{sub}x, CO{sub}x, SO{sub}x, and particulate matter (PM) due to greenhouse effect, and ozone depletion. Modelling pollutant transport in the atmospheric environment is complicated by the fact that there are many turbulent mixing time scales and spatial scales present which directly influence the dispersion of the plume. The traditional approach to predicting pollutant dispersion in the atmosphere is the use of Gaussian plume models. The Gaussian models are based on a steady-state assumption, and they require the flow to be in a homogeneous and stationary turbulence state. The dispersion correlations in these models need to be modified for applications where inhomogeneous effects, for example dissipating turbulent eddies near the physical obstacles are present. The current research is focused on predicting such dispersion correlations from the detailed fundamental computational fluid dynamics (CFD) solution of the equations of conservation of mass, momentum and energy. The CFD model is modified in the current study to truly reflect actual conditions experienced by vehicles, and hence it is useful in predicting dispersion of emissions accurately. Commercially available CFD software was applied on a heavy- duty truck's plume operating inside a wind tunnel to solve the species concentration using a probability density function (PDF) mixture fraction formulation. An excellent agreement with experimental data on CO{sub}2 concentrations in a turbulent plume was observed by using the PDF formulation and the modified Gaussian model.