Analytical and experimental studies of vehicle pollution dispersion
Using a dual approach of wind tunnel experimentation and the development of an analytical model, the dispersion of pollutants from road vehicles has been investigated. A series of novel tests were conducted in an environmental wind tunnel which looked at the dispersion of propane in the wake of a 1=50th scale model of a lorry which was fired across the wind tunnel using a special rig. Time histories of concentration and air flow were taken as the lorry sped across the tunnel. Two experimental scenarios were investigated. The first, a simulation of a typical rural boundary layer, confirmed the existence of a wake behind the moving lorry. Concentration measurements revealed that the dispersion was largely Gaussian in nature and that at low cross wind speeds the vehicle-induced turbulence was the dominant mixing effect. Measurements were taken as the lorry passed along a model of an idealized urban street canyon. Time histories for individual firings exhibited two peaks as the propane was swept around the canyon in the resident vortex. Ensemble averages of several firings allowed a quantitative assessment of the rate of dispersal from the canyon to be made. A computer model has been developed which predicts the dispersion of vehicular pollutants in both the rural and the urban street canyon environments. The model, based around the Gaussian Puff Method, extends the range of applicability of earlier models in several areas. It is a transient model which enables the investigation of traffic congestion and non-steady above canyon wind fields. It is also the first model to include individual vehicles as sources of both pollutant and turbulence. A detailed sensitivity study is presented, followed by an application of the model which attempts to predict probability distributions of pollutant in a street canyon. Finally, a comparison between the analytical model and the experimental program is presented which demonstrates that the model is capable of modelling a real situation to a good degree of accuracy but also demonstrates that further validation is required.