Prediction and abatement of noise in long enclosures

Abstract

The study of noise reduction in long enclosures, such as tunnels, underground stations and long corridors, begins with the examination of sound characteristics in these spaces. In the first part of my study, the 'coherent' model, or the complex image model, is extended to predict the reverberation time (RT30 and EDT) and the speech transmission index (STI) in a long enclosure. The approach is different from previous energy-based methods. The interference effects between the direct and reflected waves are included, and the coherent model takes into account the phase information of the sound waves. The sound field is computed in the frequency domain, and the impulse response is generated by applying the inverse Fourier transform. Subsequent calculations are performed on the impulse response to deduce the reverberation time and STI accordingly. The numerical model is validated by comparing the predictions with measured data. The numerical model is modified to consider the existence of impedance discontinuity on the boundary surfaces in the second part of the thesis. A single change of impedance in a two-dimensional duct is focused as the fundamental study of the problem. The diffraction effect at the impedance discontinuity is proved to be insignificant, and it is ignored in the formulation. With the assumption that the diffraction effect is not important, investigation is moved on to a rectangular long enclosure. A set of equations are developed on the basis of the coherent model to predict the noise reduction and acoustic indices (EDT and STI) in a long enclosure with impedance discontinuities. Experiments are conducted in two scale models, and the predictions are in excellent agreement with the data collected. Finally, the verified coherent model is used as a tool to investigate the optimal positioning of sound absorption material in a long enclosure. Several cases in an imaginary long enclosure are presented as examples to show how to determine the location with the numerical model. It is perhaps not surprising to find that an increase in the amount of absorption material does not always result in a remarkably higher degree of noise reduction. Moreover, when the absorption material is meant to be installed on two surfaces, perpendicular boundaries are preferred to parallel planes. The prediction scheme can be used to evaluate the optimal arrangement of sound absorption material. It facilitates the goals of noise reduction and improvement of speech intelligibility in a long enclosure.