Applicable directivity description of railway noise sources

Doctoral thesis, 2010

For a sound source, directivity is an important parameter to specify. This parameter also reflects the physical feature of the sound generation mechanism. For example, turbulence sound is of quadrupole directivity whilst fluid-structure interaction often induces a sound of dipole characteristic. Therefore, to reach a proper directivity description is in fact a process of understanding the sound source in a better way. However, in practice, this is often not a simple procedure. As for railway noise engineering, several noise types of different directivity characteristics are often mixed together, such as wheel and rail radiation, engine and cooling fan noise, scattered fluid sound around the bogie areas and turbulent boundary layer noise along the train side surfaces. Moreover, it is a question if the horizontal directivity of a line source can be measured directly. All these factors increase difficulties in reaching a proper directivity description that may explain why modelling directivities of railway noise sources is so far behind modelling their sound powers. This thesis work aims at working out an applicable directivity description of railway noise sources. The study focuses on the two most important railway noise types, i.e. rolling noise and aerodynamic noise. Directivities of these two noise types are studied based on measurement investigation, theoretical problem solution and model analysis. As for wheel/rail radiation, a model of a perpendicular dipole pair (PDP) was proposed to interpret those measurement specified directivity characteristics. This model naturally explains why rail radiation is of different horizontal and vertical directivities and why a vibrating railway wheel does not present dipole directivity. As for aerodynamic noise, it has been found that pantograph noise is also of perpendicular dipole components. Moreover, for aerodynamic noise around bogie areas, scattering of the air flow was proposed to be the dominant mechanism of the noise generation. This understanding leads to a different directivity description for this noise component. And, once again, scattered fluid sound around the bogie areas and turbulent boundary layer noise along the train side surfaces can be treated as a PDP source. Finally, to complete directivity description of railway noise, directivities of other important noise types have been studied as well; the directivity characteristics of these noise types become understood although lack the relevant directivity data. With all these outputs integrated, a survey of the directivities of all important railway noise sources has been achieved and applicable directivity functions have been worked out. Hopefully, this directivity study provides values not only for railway noise engineering but also for a better understanding of railway noise.