Wind-tunnel page of Petra Kastner-Klein

email: pkklein@ou.edu

It is an old tradition that the colleagues create a kind of vehicle and a hat for the Ph.D. candidate with some features relevant to the work, behavior.... of the candidate. After you passed the exam the next one starts: You have to survive being transported around the campus!!

I am sitting in a model of the Atmospheric Boundary Layer Wind Tunnel which I used for my studies and also the hat is decorated with a small really working wind tunnel. The topic of my thesis was Experimental investigation of fluid mechanical transport processes in street canyons. That is the reason why my chair corresponds to a model of street-canyon buildings.

If you would like to know how it was looking like in reality move on to the next topic.

The photo to the right gives you an impression of the experimental set up during my street-canyon studies. You see the wind-tunnel model of an idealized street-canyon configuration mounted in the Atmospheric Boundary Layer Wind Tunnel. The two buildings are surrounded by homogeneously distributed roughness elements which are important for the simulation of a boundary - layer flow similar to the atmospheric boundary layer. Additionally vortex generators are installed at at the entrance of the wind-tunnel test section. The ceiling of the tunnel can be adjusted to avoid pressure gradients.

The idea of the experiments was to study the influence of small-scale parameters describing the building configuration, like e.g. building dimensions or roof geometry, on flow field and dispersion inside the canyon. A tracer gas was released by two line sources. The basic configuration was characterized by two 12 cm high buildings with flat roof and perpendicular approach flow. The buildings were 180 cm long, which means their horizontal extension was up to the tunnel walls. Additionally an inflow at the lateral street edges was blocked by plates. In this case two dimensional flow patterns could be observed in the central region. The width of the street was 12 cm, which means the aspect ratio, defined as height of buildings to width of street , was one. The length and width of buildings, number of buildings, roof geometry and wind direction were varied.

I found that small-scale features of the building configuration have a strong influence on flow field and dispersion inside the street canyon. The photos to the right give an impression on the changes of the vortex dynamics. They result from laser light sheet visualizations which I made for different roof configurations. The smoke was released near the ground and the light sheet was placed in the central plane. The wind is blowing from the right to the left. For situations with perpendicular approach flow a vortex rotating in the street can be usually observed (upper photo). This type of flow occurred for situations with flat roofs at the upstream building (right building on the photo) and no step-down. The lower photo shows a significant change if a slanted roof is placed on the upstream building but similar effects were also found for several other step-down situations. No clear rotating vortex forms and the region polluted with smoke is much higher.

Data sets of concentration distributions at the building walls and mean and turbulent flow field characteristics are available for inter-comparison and model-evaluation studies. You should read the detailed description (pdf-file) of the experimental setup (inflow conditions, location of sampling points...) and contact me pkklein@ou.edu if you are interested in any of them.

Vehicle-induced turbulence can be an important factor of pollutant dispersion in urban areas. But there is only little information on the quantitative effect and methods to include this parameter in dispersion models. This was the motivation for a wind-tunnel study with emphasis on influence of vehicle motions on the dispersion characteristics inside street canyons. Based on a similarity concept proposed by E.J. Plate (1982) a modeling technique was developed. The experimental setup is shown on the photo to the right. The motions of vehicles are simulated by small metal obstacles which are mounted on belts. They generate turbulence on a characteristic scale related to the obstacle geometry.

We modeled two traffic lanes by two belts, which can move with different velocities and and in different directions. Situations with two-way traffic as well as one-way traffic could be studied.

The traffic density (number of obstacles per unit length), the wind velocity and vehicle velocity were varied. A tracer gas was released by two line sources placed next to the belts. The concentration distribution at the building walls was measured.

Concentration distributions at the leeward canyon wall (diagram below) demonstrate the influence of vehicle motions.

The results for the reference case (without traffic), are shown in the upper plot. In this case the flow characteristics are determined by the interaction of a canyon vortex and vorticity zones near the building edges. The pollutants are transported to the leeward canyon wall where a three-dimensional, symmetric concentration pattern was observed. The maximum concentration was found in the canyon center near the ground. For the two-way traffic situation similar results were found. The moving traffic did not essentially affect the concentration pattern, but the maximum concentration was lower than in the reference case. In the case of a one-way traffic situation the concentration field changes fundamentally. The moving traffic leads to a pronounced transport of pollutants along the canyon axis. The concentration pattern at the leeward canyon wall is no more symmetric and the point of maximum concentration is shifted to the canyon end.

Systematic studies for situations with two-way traffic showed that the maximum concentration decreases with an increasing ratio of vehicle to wind velocity and with an increase of traffic density in the case of perpendicular approach flow. A dimensionless combination of vehicle to wind velocity ratio and density factor was proved to be a universal parameter describing the dependence of concentration on vehicle-induced turbulence.