D. Schröder¹, U. P. Svensson², M. Vorländer¹
¹
Institute of Technical Acoustics, RWTH Aachen University, Aachen, Germany
²
Acoustics Group, Norwegian University of Science and Technology, Trondheim,
Norway
Diffraction is a wave phenomenon that is
experienced in mostly any daily life situation. Just imagine a walk through a
corridor of an office building where sound sources are located in each room.
The sound field does not change abruptly when an open door is passed. The
transition is rather smooth which comes from sound energy that is bent around
the door entrance's edges, i.e. sound diffraction. When a sound wave hits an
obstacle, a frequency-dependent shadow zone occurs behind the object (related to
the direction of sound propagation). If the object is small in comparison to
the wave length, the incident wave remains unaffected. However, a shadow zone
appears and grows clearer and sharper with decreasing wave length and
increasing frequency, respectively. This shadow zone results from a total
cancellation of the incident wave by the diffracted wave which is radiated from
the edges or perimeter of the respective object.
Thus, sound diffraction must not be
neglected in computer simulations, especially in indoor and large-scale urban
scenarios. The development of noise barriers is a typical application where
computer simulation methods of edge diffraction are taken in account in order
to get detailed information about the barrier's efficiency in the respective
scenario. Unfortunately, all known simulation methods, either based on
Geometrical Acoustics or the numerical solving of the wave equation, are just
approximations. They work fine for distinct simple test cases within a certain
frequency range but none of them covers the effect of diffraction in its whole
complexity. Consequently, huge effort is currently put into further improving
these simulation models to enable a realistic prediction of edge diffraction even
in complicated real-world situations. This demands for
taking into account the influence of sound scattering and multiple wave
diffraction, too. Therefore, appropriate test scenarios are required in order
to validate these extended prediction models.
In this contribution various measurement series of a scaled noise barrier model are presented which aim to give developers and acoustic consultants the possibility to test their prediction methods. The model features three types of ground layers: an absorbing surface, a rigid surface and a highly scattering surface. The latter was realized by a self-constructed Skyline Diffuser so that the occurring sound scattering can be simulated in two ways, either deterministic or stochastic. The measurements were carried out in a full anechoic chamber and a turntable was used to rotate the scale-model during the measurements with a resolution of one degree. All measurements together with geometrical models of the scale-model (with/without diffuser), detailed information on sources and receivers, material data (absorption- and scattering coefficients) and helpful Matlab tools are freely available for download (www.openmeasurements.net).