Improving the accuracy of engineering models at shielded building facades: experimental analysis of turbulence scattering

Paper in proceeding, 2013

Noise mapping models are able to accurately predict directly exposed facade levels near busy roads on condition that sufficiently detailed traffic data is available. At the non-directly exposed side of the building, however, common practice application of standard methods strongly underpredicts sound pressure levels, potentially leading to an incorrect assessment of noise annoyance and sleep disturbance. The concept of background noise mapping was proposed before, which has the important advantage that it can increase the accuracy of existing noise maps at a limited computational cost. In this study, long-term meteorological and noise data showed that turbulence scattering contributes significantly to the noise level at shielded facades, already at sound frequencies below 1 kHz. Periods with strong atmospheric turbulence are dominant for long-term equivalent noise levels as typically used in strategic noise maps. A comparison between predictions and measurements show that rather high turbulence strengths should be used when producing noise maps.