D.W.R. Green

Doctor of Philosopy, Oxford University, November 1983

A Wind Tunnel Study of Flow Over Hills


This research is concerned with the wind tunnel modelling of atmospheric boundary-layer flow over three-dimensional hill. The first set of experiments was performed using hills of "cosine-squared" cross-section, and results for fractional speed-up were compared with predictions of an analytical model MS3DJH/3.1 described by Taylor et al (1983) - good agreement was found between them. These hills were also used for further tests to isolate the effect of changing model datum level. The hills were raised onto platforms in the wind tunnel and the speedup over their crest re-measured. A linear superposition of two effects was found, so that the fractional speed-up of the raised hill was equal to the sum of those caused by hill topography and change in elevation. This agrees well with the assertions of Jensen & Peterson (1978) following their analysis of the experiment at Riso.

The main work involved modelling the flow over two "real" hills, Kettles Hill and Askervein. Topographical models of each were constructed at a scale of 1/800 and measurements of mean flow speed-up were then compared with existing data. For Kettles Hill the agreement with full-scale results is good except very near the surface where the wind tunnel results are considerably lower than full- scale.

The Askervein mean flow speed-up results are compared with preliminary field data at 10m above the ground and good agreement is obtained, particularly along the main tower line A.

Extensive turbulence measurements were also made, and a careful examination revealed that above the hill-crest the flow could be divided into four separate regions in which the behaviour of (sigma-u)squared was quite different. These findings are discussed in relation to the rapid-distortion theory of Britter, Hunt & Richards (1981) and some doubt is cast over whether this theoretical work is a valid model of the hill-top turbulence structure.

(No thesis available)