Using LES and wind tunnel data to investigate peak-to-mean concentration in an urban environment

Abstract

The impact assessment of odorant sources, flammable gases and accidental or intentional release of toxic gases requires the estimation of short term averaged maximum concentration. The well-known power law function is frequently used to calculate the peak-to-mean concentration ratio. However, the use of this function for urban environments is still a challenge. The aim of this work is therefore, first, to examine the maximum mean (peak) concentration value considering different averaging times within a larger sampling time interval using wind tunnel (WT) data obtained in a model urban environment. The second aim is to investigate the potential of numerical modelling of atmospheric dispersion using Large Eddy Simulation (LES) to assess the influence of turbulent eddies on averaged concentration over short time intervals on dispersion influenced by a building array. Although the LES technique predicted smaller relative maximum instantaneous concentrations than those measured in the WT, relative concentration fluctuation (concentration variance divided by its mean) agreed quite well with wind tunnel data, especially for the sensors behind the buildings. LES and WT spectra essentially collapsed over a non-dimensional frequency range of two to three decades. For higher frequencies, however, the LES data fell due to the coarser resolution, etc. This indicates that the inevitable spatial averaging imposed by the mesh means that LES is not capable of directly predicting the larger, very short time scale concentration peaks observed in the wind tunnel data. The exponent p in the power law function is smaller for the sensors in the central position of the building array (behind the buildings) than for those located in short streets or at intersections and they also decreased more slowly with distance from the source. WT data were used to investigate the influence of sensor height on the peak-to-mean ratio as a function of the averaging time and p-values were found much higher for sensors located above the building height.

More information

Identifiers

ORCID for Zheng-Tong Xie: orcid.org/0000-0002-8119-7532

Catalogue record

Export record

Altmetrics