Outdoor Ground Impedance Models
Invited paper
The Open University
Tuesday 2 june, 2015, 16:20 - 16:40
0.8 Rome (118)
Abstract:
Many models are available for representing the frequency dependence of outdoor ground impedance. The increasing popularity of time domain calculations has lead to several investigations of the physical admissibility of these models. Recently, modifications of a single-parameter (effective flow resistivity) semi-empirical equivalent fluid power law model, first introduced in 1990 [Y. Miki "Acoustical properties of porous materials - modifications of Delany-Bazley models" J. Acoust. Soc. Japan (E) 11 19-24 (1990)], have been used to represent frequency-dependent soil impedance. However, it is shown that, as is the case with the Delany and Bazley model on which they are based, these forms of the Miki model lead to non-physical predictions of the real part of complex effective density and the surface impedance of a hard-backed layer at low frequency. Short-range predictions of the level difference spectra between vertically-separated microphones obtained by using the two Miki-based impedance models in classical expressions for the field due to a point source over an impedance plane are compared with data and those obtained when using other ground impedance models. Models that require use of porosity and tortuosity as well as flow resistivity can be made into two-parameter forms by assuming a relationship between tortuosity and porosity. It is found that, although several models enable acceptably accurate predictions for grassland surfaces, a two parameter variable porosity impedance model enables the best fits to the short range data. The single-parameter semi-empirical models give rise to significantly poorer predictions of short-range propagation over relatively low flow resistivity ground surfaces including railway ballast, gravel and forest floors than obtained after using other impedance models. Estimates of the potential discrepancies in predictions of outdoor sound propagation at longer ranges are made. The best-fit parameters are used to explore the potential of replacing acoustically-hard ground by soft ground for reducing traffic noise.