Management of a ground-borne noise prediction methodology for railway traffic in tunnels

Abstract

The purpose of this thesis was to understand how a new prediction methodology for ground-borne noise should be maintained and continuously updated. The scope included ground-borne noise from trains in tunnels and focused on the predictions to fulfill the requirements for the post-construction phase. Research questions were formulated and they were as follows: What new data should be added to the methodology? What are the risks when managing the methodology? What strategy should be used to ensure a successful management? The method was composed of a literature review, case studies, interviews and a numerical analysis. In total, 22 respondents were interviewed. The results from the case studies demonstrated that under-ballast mats are common and that assumptions related to the foundation differ between projects. However, assumptions used when obtaining the ground-borne noise level in a room are similar. In Sweden, the same noise criteria have been utilized in practice. The results from the interview sessions demonstrated that more information regarding the coupling between the tunnel and the building is needed. Also, the contribution from railway components can be added to the implemented model. Measures should not be added to the implemented model since it could alter the creativity during development. However, a default value for an under-ballast mat can be added. Advantages with the methodology is that it contains a great deal of data, a detailed literature study, includes the uncertainties and that it is transparent. The source strength in the model was developed using data from two specific tunnels. One identified risk is that the model might not make sufficient predictions for cases with significantly different conditions. Another risk is that it stagnates when it is no longer maintained. To ensure a successful management, a reference group is needed that can discuss it. The process of using the methodology and the implementation of the model should be further explained in a step-by-step user guide. New measurements are required and results from other methods, such as specific transfer functions can be used. The numerical part demonstrated for a specific case that no measures on the track are needed for dwellings situated further than 100 m away from the tunnel when passenger trains are investigated. For freight trains, the same distance must exceed 200 m. A difference of 5 dBA exists between results from the implemented model and results from the EIA (Environmental Impact Assessment) in the Lärketorpet project. Important for fulfilling the criteria is that an upgraded rail structure is used. This will mitigate the ground-borne noise levels. Future studies could focus on comparing numerical results from other projects and investigate ground-borne noise in the construction phase.