Flow over regular roughness

Objectives

To measure vertical and horizontal structure of turbulence statistics for various roughness densities in a wind tunnel.

 

Methodology

The thermally-stratified open-circuit blower boundary-layer wind tunnel at the Research Institute for Applied Mechanics (RIAM) at Kyushu University, Japan was used to measure turbulent flow over a rough model surface (neutral conditions only). The dimensions of the wind tunnel are 1.5 x 1.2 x 13.5 m (W x H x L). Boundary-layers were developed over 4 different arrays of cube-roughness (30 x 30 x 30 mm) with linearly increasing roughness density. Boundary-layer height varied between 180 - 300 mm.

The photograph on the right is a view towards the wind tunnel intake of the roughness configuration termed E-type which is the highest roughness density used in this experiment.

  

 

  Mean U, W and fluctuations about the mean, u and w, were measured with a 2-D Laser Doppler Anemometer (LDA) mounted on a high-precision positioning device (photograph on the left). Seeding was done at the intake of the wind tunnel using water-glycol particles. The advantage of the LDA over conventional x-wires is increased directional response and the ability to identify positive and negative flow directions. Both are necessary to measure the turbulence statistics with confidence in regions of high flow inhomogeneities such as between roughness elements. Measurements were taken at x = 5 m and free flow velocity was U = 5 m/s. Turbulence properties were measured in the transverse (y-direction) and vertical (z-direction) at various positions and levels.

 

Preliminary Results

Analysis of observations is still incomplete - please write to the NUS administration and suggest to reduce my administrative workload and teaching duties if you think this work should be expedited. The results below show the strong variability of the flow over a rough surface. Here C- and E-type cases are both from staggered (diamond-shaped) roughness arrays with E-type having the larger roughness density.

 

The panel on the right shows mean vertical wind velocity (in cm/s) from traverses (y-direction) measured at 5 heights (h = height of roughness element) at various locations along the normalized transverse axis (d = distance between cubes). The thick blue bars show the location of the cubes relative to the measurement: solid (dashed) line refer to measurements behind (ahead) of a cube). Below h, W is generally negative behind and between cubes but strongly positive just ahead of an obstacle.

  

 

 

The panel on the left shows standard deviations of u normalized by the constant flux layer friction velocity as a function of non-dimensional height. Also indicated are data from 2 urban studies (Z86 and S92) and measurements from a wind tunnel study performed by M. Raupach and co-workers (Reference).