Experimental investigation of

fluid mechanical transport processes in street canyons

Petra Kastner-Klein

Doctor thesis defended at the faculty of Civil engineering, University of Karlsruhe, Germany, July 1999.

Abstract

Dispersion of atmospheric pollutants is an important aspect of urban air-quality studies. Vehicle emissions represent the main group of pollutants in cities. It turned out that the increase of traffic has compensated emission reductions, which were expected from the implementation of catalytic converters in cars. Concentrations measured in cities are often close to threshold values. Improvement of air quality in residential areas is necessary in order to avoid risk to human health. Development of a general approach to the investigation of pollutant dispersion in the urban canopy is a very hard task. The climate in urban areas is dominated by microscale influences. The arrangement of sources and targets there is extremely variable. To enable verification of numerical model predictions, a group of basic urban-canopy elements, which represent typical flow configurations, should be defined. Street canyons represent one group of basic urban-canopy elements.

Wind-tunnel studies of dispersion on the urban-canopy scale can provide data sets of high resolution, which are obtained under fixed and repeatable conditions. In the present study, flow fields and concentration patterns in idealised street-canyon configurations were investigated in an atmospheric boundary-layer wind tunnel. An isolated two-dimensional street canyon was used as a reference case for a systematic study with parameter variations. The idea was to cover a wide range of street canyon situations and to quantify effects of changing flow configurations. The corresponding data sets include results of concentration measurements for different building dimensions, building roof shapes, wind directions and source positions with respect to the axis of the canyon. In addition, situations where one or two additional rows of buildings were located upwind of the street canyon were also considered.

Flow field measurements with the laser Doppler velocimetry and laser light sheet visualisations were used for comprehensive analysis of the flow dynamics in the canyons. It turned out that a vortex rotating in the canyon, which is usually considered to be the dominating element of the flow pattern in the case with the approach flow oriented perpendicular to the canyon, occurs only in particular regions and under specific boundary conditions. Vorticity zones near the lateral building edges have a significant influence on the flow dynamics and on the concentration field in the canyon. They usually extend up to a distance of 2-3 times the building height and cause a flow along the canyon which transports pollutants to the canyon centre. Additionally the roof-shape geometry has a strong influence on the in-canyon flow regime. For certain roof configurations, a rotating vortex was not observed in the case of perpendicular approach flow. For oblique wind directions the flow component along the canyon axis is a dominant transport mechanism already when the approach flow deviates by 15 degrees from the perpendicular direction. It is concluded that small-scale features of the building configuration, and approach flow conditions are important parameters determining the flow and concentration fields in street canyons.

The study was extended by including an experimental technique for modelling the effects of vehicle induced turbulence, which can be another important factor of pollutant dispersion in urban areas, especially under conditions of low wind speeds which are typical for street canyons. The concept of this study is based on an idea proposed by E.J. Plate in 1982. The movement of vehicles was simulated in a boundary - layer wind tunnel by small metal plates mounted on two belts moving in a modelled street canyon. The scaling factor is based on the ratio of turbulence production by cars to that by wind flow. The variation of traffic conditions are represented by variations of the vehicle velocity and traffic density.

The influence of traffic on concentration patterns at the canyon walls was studied. It was found that the concentration decreases with an increasing ratio of vehicle to wind velocity and with an increase of traffic density. For situations with two-way traffic, mean flow conditions are not affected by the traffic movement and a dimensionless product of vehicle to wind velocity ratio and one-third power of the density factor was proved to be a similarity parameter describing the dependence of concentration on the vehicle-induced turbulence. A comparison of the experimental results with numerical calculations has shown good overall agreement. In the case of one-way traffic the moving vehicles caused a strong change of the mean flow dynamics: a flow component along the canyon axis was observed which destroys the rotating vortex. The verification of this result should be an aim of further investigations in real urban conditions, as well as in the laboratory.