Prediction of airborne noise inside railway vehicles

Abstract

Railway vehicles are important means of transportation and vital to the public. However, the noise within railway vehicles is an important aspect that can affect the passengers’ and train crew’s comfort. High interior noise can be attributed to strong external noise sources and insufficient sound insulation by the train wall structures. To predict the interior noise is challenging. Existing approaches such as the finite and boundary element (FE/BE) method or statistic energy analysis (SEA) are not suitable for a wide frequency spectrum solution and are either of high computational cost or of low accuracy. The aim of this thesis is to develop a comprehensive modelling approach to predict the interior airborne noise of modern railway vehicles, taking into account different noise sources and transmission paths as well as the complexity of the car body structure. An approach is presented for modelling the noise propagation beneath the train floor and this is applied to rolling noise sources by assuming that the sound incident on the train floor is made up of a direct and a reverberant component. An equivalent source model is used to represent the direct component, and SEA model is used for the reverberant part. The sound power of the rolling noise is obtained by using the TWINS model. A wavenumber-domain boundary element (2.5D BE) approach is adopted to predict the propagation of rolling noise to the train external surfaces. Comparisons are made with measurements showing good agreement. Noise propagation from the pantograph to the train external surfaces is studied considering the influence of flow on the sound propagation. The total sound power from the pantograph is calculated based on the component-based approach and a database of factors of influence created by previous researchers. The 2.5D BE method is again employed to calculate the relevant sound propagation. The influence of flow on the sound propagation is modelled either by a uniform mean flow or by allowing for the variation of velocity through the turbulent boundary layer. Laboratory experiments and the ray tracing approach verified the 2.5D models for predicting the pantograph noise propagation. To calculate the noise transmission through extruded train wall structures, use is made of a 2.5D FE/BE model and an SEA model. The 2.5D FE model is used to study the bending waves in the extruded panel to calibrate the input parameters for the SEA model. With the aid of such a calibration using the 2.5D FE model, the SEA model can give good quality predictions of the sound transmission loss and radiation efficiency of the extruded panel in comparison with the measurements. Finally, an overall framework is provided to predict the airborne noise inside railway vehicles, in which another SEA model is created for the interior space of the train. The power input to the interior SEA model is determined from the incident sound power on the train external surfaces. The framework of interior noise prediction is verified against measurements on a metro train. It is found that the predictions agree reasonably well with the measurements in terms of sound spectra and overall sound pressure levels.

More information

Identifiers

ORCID for David Thompson: orcid.org/0000-0002-7964-5906

Catalogue record

Export record