Abstract
A wind tunnel study was used to investigate flow and the dispersion of elevated emissions over and in the wake of both generic building groups and a real industrial site. Much research has been previously carried out on low level releases passing through extensive building arrays, but very little has treated elevated releases. The objective of the present research was to investigate the requirements of 'practical' models to predict dispersion from elevated releases above an industrial complex. The experiments were carried out in the EnFlo wind tunnel at the University of Surrey in simulated neutral and stable boundary layers at 1:500 scale. The behaviour of the mean flow and turbulence in array wakes was compared with predictions from an analytical, three dimensional, eddy viscosity model. The model was successful when arrays were aligned normal to the approach flow but, as it did not treat the secondary flows associated with roof vortex systems, it was not adequate for other wind directions. The measurements showed that when roof vortex systems were prominent they created strong mean streamline deflections downwards over the wake, leading to velocity excess rather than deficit within the central part of the wake. Diagonal wind directions produced the higher ground level concentrations, or the greater reduction in effective stack height, for all the building groups studied. This correlated well with the measured mean streamline deflections. Dispersion measurements in the stable boundary layer for releases below 1.5 building heights were very similar to those made in the neutral boundary layer, but the building effect appeared larger in the stable boundary layer for the higher releases. This latter feature probably resulted from interactions with the boundary layer edge in the relatively shallow stable layer. Analysis of extensive measurements in the neutral boundary layer showed that a group of buildings could be adequately represented by a single, effective building of appropriate dimensions as far as the maximum ground level concentrations were concerned. However, the size of the effective building was a function of array orientation relative to the wind. Vertical plume spread in elevated releases was unaffected by the presence of the buildings. Streamline deflections or downwash, which caused the mean height of a plume to reduce, were the predominant cause of increased ground level concentrations in the array wakes. In situations where the plume was significantly entrained into the wake, the lateral plume spread was enhanced. A modified Gaussian plume model including downwash and enhanced lateral spread represented the resulting concentration field quite well.