1 of 1

Spires

Roughness 1

Roughness 2

Roughness 3

Test section

Modelling the Atmospheric Boundary Layer in the Recirculating Water Tunnel

Benjamin Clare

Supervisor: Dr Christina Vanderwel

Background

The atmospheric boundary layer (ABL) describes the flow of wind along the surface of the earth. The structure of the ABL is similar to a high Reynolds number, flat plate, turbulent boundary layer over a rough wall. Therefore, it can be modelled in wind and water tunnels.

Urban aerodynamics studies, including pollutant dispersion are becoming much more necessary, especially as urban development continues at its current rate. However the incoming conditions must be correctly modelling the ABL in order for the data to be useful.

​

Aims

The main aim of the project was to use spires and roughness elements to develop an incoming boundary layer in the RWT that accurately models the atmospheric boundary layer.

This was checked on 3 conditions:

  • Logarithmic region in its mean velocity profile
  • A constant Reynolds stress region
  • Consistent turbulence intensity measurements

Methodology

Irwin spires and a false floor with roughness elements were installed into the RWT. The channel height was set at 0.6m and the velocity of the flow at 0.5 m/s.

Particle image velocimetry (PIV) measurements were taken in the vertical streamwise plane. 1000 image pairs were taken, then post processed in DaVis10 and Matlab. Time and streamwise averaging was performed and profiles were plotted.

Friction velocity was estimated using the peak of the Reynolds shear stress profile and the comprehensive shear stress method developed by Womack et al. It was then iteratively increased until the correct velocity profile was plotted. Uτ varied from 0.022 m/s to 0.028 m/s depending on the setup. The friction velocity Reynolds number, Reτ, was between 11,000 and 15,000, depending on the setup.

​

Results

The addition of roughness elements and spires drastically changed the characteristics of the flow compared to the smooth floor. The rough floor caused an initial decrease in velocity and increase in stresses and turbulence. When the spires were introduced the effects are amplified.

​

​

​

U∞

​

Recommendations

The roughness elements should be extended until the edge of the test section, removing the rough → smooth transition. Uniform roughness could also be used in future studies to keep the IBL consistent.

Spires should also be used to create the correct amount of streamwise turbulence, in combination with a full rough floor.

Conclusions

Due to the roughness setup and the transition to smooth panels upstream of the test section multiple internal boundary layers grew along the fetch. This lead to multiple logarithmic regions in the velocity profile, and Uτ. The addition of spires increased this effect.

The existence of internal boundary layers is also proven by plotting velocity against y1/2. The intersection of the fit lines indicate the edge of an internal boundary layer.

Overall, the setup that most accurately models the ABL is the rough wall with no mounting plate (orange), it has largest log region. This type of setup could be used to model an urban → rural ABL.

​

(Budapest University of Technology and Economics, 2017)