The effect of large roughness on the aerodynamic noise emitted by bluff bodies was investigated in a series of experiments on rough circular cylinder flow as well as rough wall turbulent boundary layer. The far field acoustic tests were performed in an open-jet rig at the anechoic chamber at the University of Southampton. A set of phased microphone array measurements were performed to obtain information regarding the source locations. A broadband peak at high frequencies was detected and associated with roughness noise, in both the flat wall and the circular cylinder.
The measured far field roughness noise spectra were compared with two prediction models, which were successful in the case of the flat wall. A simple extension of the flat wall model to a circular cylinder was able to predict the peak frequency, but the peak level was slightly over-predicted.
During approach to landing airframe noise is a dominant source of emitted noise, and in particular landing gear noise. Landing gears of commercial aircraft have a large amount of hoses, protrusions, joints, etc. These elements are typically much smaller than the size of the wheels and the struts. Based on current noise prediction models, the small elements are responsible for most of the high-frequency noise range. However the prediction schemes are almost entirely empirical and don't include geometrical information such as the shape and size of the elements, or their location. The modelling of small elements needs to include this information in order to be able to include acoustic considerations in landing gear design and optimization.
With the goal of improving the acoustic modelling of the high frequency components of aircraft landing gears, the effect of large roughness on the radiated noise by a circular cylinder has been studied experimentally. Existing prediction models of noise generated by interaction of a TBL with a rough flat wall have been validated, and one of them has been simply extended to the case of a circular cylinder. The noise peak is reasonably well predicted, but there is over-prediction for higher frequencies. The cylindrical elements are less efficient in generating roughness noise than the hemispheres, especially in the case of a circular cylinder, and they cause a significant increase in noise at high frequencies.