Prediction, measurement and reduction of traffic noise

Abstract

This thesis is concerned with several efficient prediction models for devising effective noise reduction measures and investigating their interactions when they are incorporated into realistic urban environments. More specifically, retrofitting noise barriers near high-rise building facades and deliberating artificially-roughened hard surfaces in traffic tunnels are investigated. The first proposed model is an extension of the ray-based coherent model, originally developed by Panneton for predicting the shielding performance of a pair of noise barriers erected on an outdoor ground. Different sets of multiple imaging sources and receivers and their validity conditions are established in order to elucidate the total sound field at various receiver locations. Compared with the computationally intensive boundary element method (BEM) and the scale model measurement results, the model predicts reasonably well. It is shown that multiple reflections between the pair of parallel barriers can degrade the barrier performance expected from each barrier which may not be adequately quantified using the existing engineering scheme and guidelines. Parametric analyses show several geometry-related factors for controlling the degradation. With the more complicated sets of ideal ray paths and image validity conditions, the proposed model is subsequently developed and validated for assessing the acoustic performance of the parallel barriers placed in an urban environment either in front of a row of tall buildings or in a street canyon. The effectiveness of using parallel absorbent barriers is also explored. The Twersky boss formulation is applied to the image source model for the calculation of the sound pressure levels in a rigid tunnel with the presence of a 2D hard rough surface on one of its vertical walls. The proposed model was validated by comparing with experimental measurements conducted in two model tunnels. Comparing with experimental data, it was demonstrated that the presence of a hard rough surface on one of the vertical walls in the tunnel introduced effective impedance on the boundary. The presence of apparent impedance on the boundary surfaces led to the reduction in the average noise levels in an otherwise hard tunnel. In the models studies, it is found that the introduction of a hard rough surface on one of the vertical walls of the model tunnels can lead to an average reduction in the noise level of about 3 dB over the frequency range from 500 to 5000 Hz. By considering relatively high frequencies, the third proposed model is developed for estimating the acoustic performance of a barrier placed on an outdoor ground near either geometrically or diffusely reflecting tall buildings. The formulation used combines a radiosity equation with an energetic diffraction term. It is shown that the barrier performance is consistently predicted against the numerical results obtained from the BEM computations. At last, the actual noise exposure of residents who frequently complain about air, rail and road traffic noises in three typical urban areas is explored.