Berechnung komplexer Beugungsfelder in Kanal mit vollständig absorbierender Wandung

Abstract

To date, the acoustic diffraction field behind reflectors, such as noise barriers, buildings, etc., has been calculated using a very simple but sometimes inaccurate approximation. The purpose of this work is to develop a more precise calculation method. This work presents a new mathematical model for calculating two-dimensional diffraction fields. The model is based on a two-dimensional duct consisting of two horizontal and parallel walls. The numerical results indicate that, beyond a certain duct width and at small distances from the source, the upper wall of the duct has no impact on the sound pressure. This makes it possible to simulate a semi-free field using the two-dimensional duct. In order to further minimize the influence of the upper wall, it can be decomposed in such a way that the impedance depends on the angle of incidence of the source. The two-dimensional duct allows the calculation of noise barrier diffraction. Since the sound field can be synthesised as a series of modes, the analytical solution for the pressure describes the sound field in the entire space between the two boundaries of the duct. Noise barriers with attachments parallel or perpendicular to the floor could, in principle, be calculated using this method. This work examines the effect of T and μ profiles on sound barriers. For the T profile, a flap length optimization and the impedance of the T profile were estimated. In doing so, it was determined that the effect of the T profile was heavily dependent on the source position. In spite of this, improvements were made in every source position considered. Both profiles were examined using measurements, and improvements of up to 6 dB were confirmed compared with a simple sound barrier. The two-dimensional duct was also used to calculate the diffraction of geometric shapes on the ground. This shape is decomposed into a series of small rectangular partial pieces. The effects of the discretization remain small enough if the length of a partial piece is smaller than smaller than onetwelfth of the wavelength. Due to the large number of modes and partial pieces, the significant amount of processing power and memory required for the calculation could cause problems.